WO2008061296A1 - Testing device and method for use on soft tissue - Google Patents

Abstract

A device and method processing the health and/or characteristics of a soft tissue are described. The device and method are particularly well suited to cartilage. A testing device is provided with an indenter (50), a monitor (51, 52) for monitoring the recovery of the surface and underlying matrix. The indenter is connected to a processor (53, 54, 58) for receiving the information and assessing the health of cartilage based on a hybrid recovery index or other characteristics. The result may be shown in a visible indicator (55, 56, 57). The device may further incorporate monitors for assessing stretch characteristics of the surface. The monitors may also be an ultrasound monitor and/or a mere infrared monitor. The ultrasound may detect echoes for establishing a coefficient of reflectants between the surface layer and osteochondral junctions. The ultrasound may also be used for providing a profile of the soft tissue. Diffuse reflectants near infrared spectroscopy may also be used to provide information on health of the tissue.

Description

TITLE

TESTING DEVICE AND METHOD FOR USE ON SOFT TISSUE

FIELD OF THE INVENTION

The present invention is directed to a testing device for use on soft tissue, particularly, but not exclusively, cartilage, to assist in characterising the health of that soft tissue. The invention extends to a method for characterising the health of soft tissue, particularly, but not exclusively, cartilage.

BACKGROUND OF THE INVENTION Currently, a surgeon has no accurate way of judging the integrity of cartilage encountered at surgery and must rely on his or her visual impression to judge the health status of the tissue. Although this visual impression is important, it is now widely accepted that visually normal cartilage is not necessarily healthy. Although grading systems can be used to classify the stage of osteoarthritis, they cannot on their own facilitate treatment [1].

With the development of localised treatment technologies such as tissue engineered cartilage replacement and unicompartmental knee replacement, it has become important to develop precise and objective measures which are capable of detecting the extent of articular cartilage degeneration at an early stage. The ability to extend the current arthritic grading systems to optimise and leave the maximum amount of native host cartilage in a joint would facilitate optimal treatment. Associated benefits may include a reduction in recovery time, rehabilitation and costs. Due to the high degree of structure-function coupling in articular cartilage, any change in the integrity of the constituents of the matrix, or in their interactions with each other, manifests as a change in biomechanical performance. Recent publications [2,3] have suggested that alterations in mechanical properties may be notable before any gross morphological change is apparent.

Previous mechanical testing of both normal and osteoarthr^'rtic cartilage has been extensive but broad in its approach. A general understanding of the bulk properties of cartilage and the subsequent changes due to osteoarthritis have been reported. These include changes such as softening [2,4,5], different stress relaxation behaviour [5,6], and increased permeability [2]. In-depth investigation of the outward progression of a focal region of disease and its relation to osteoarthritic severity, however, has not been attempted. There has been increased interest in developing arthroscopic systems that can quantify and qualify the mechanical properties of articular cartilage at surgery [7-1 1]. Such systems concentrate on the stiffness alone and are incapable of distinguishing normal from degraded articular cartilage [24]. Buschmann et al. [12], in contrast, have used the streaming potential under indentation to detect matrix degradation by mechano- electrical analysis. Other researchers are developing ultrasound techniques to measure stiffness, although these face similar problems to basic indentation [5,9,13-15].

Published results from such techniques show wide variations in stiffness for normal, healthy tissue, in some cases as much as ±52% [9]. These measures can be used accurately and effectively to track changes in the health of a particular specimen. However, they cannot detect a small deviation from normal tissue properties with any confidence due to these inherent variations. When using the currently available mechanical techniques, a tissue must therefore be severely softened for a diagnosis to be made with a high degree of confidence. As the surgeon is already able to determine moderate to severe softening without the aid of a separate diagnostic instrument, a more sensitive indicator is required to aid or influence decision making. It has been shown that while the biophysical properties of isolated general matrix may change significantly with degeneration, a measurement of in situ stiffness by indentation may be an unsatisfactory indicator of the tissue's health, provided the articular surface is intact. In their study on the physical indicators of cartilage health, Broom and Flaschmann [16] found that on-boπe compliance tests alone were insensitive to degenerative changes in the structural integrity of the underlying matrix that could be readily detected by microscopic and swelling analyses. The authors hypothesised that in situ, the articular surface layer, the subchondral attachment and the wider continuum of articular cartilage tend to restrict fibrillar network expansion, even when degenerative influences have led to a substantial dismantling of much of the crucial interconnectivity in the matrix.

It should also be noted that, irrespective of health, the basic indentation properties of a tissue will vary considerably over the surface of a normal joint, between joints and between people. It is important to distinguish between these natural variations and variations due to disease or degradation. This important distinction will need to be made for any in vivo test to be satisfactory and unequivocally relevant. It is generally accepted that in the earliest stages of osteoarthritis, there is a disruption of the collagen structure in the superficial layer leading to a loss of proteoglycan [17,18], Broom et al [19] studied the structural changes in the matrix surrounding focal arthritic regions and found significant changes in the collagen integrity of macroscopically normal, intact tissue, particularly in the superficial zones. The ease of accessibility of the articular surface, coupled with Its compromised integrity early in disease progression, make it a prime target for diagnosis.

The collagen meshwork in the superficial layer is aligned tangentially to the articular surface, providing a strain-limiting function [20]. Published work shows that significant stretching is seen in this meshwork during crack propagation [21].

There has been increased interest in developing arthroscopic systems that can quantify and qualify the physical properties of articular cartilage. Such systems tend to concentrate on the stiffness of the cartilage matrix measured from cartilage surface echoes or indentation. OBJECT OF THE INVENTION

It is an object of the invention to provide a testing device and method to assist in characterising the health of soft tissue, in particular cartilage.

SUMMARY OF THE INVENTION In a first broad aspect the invention resides in a testing device for use in characterising the health of soft tissue, the testing device comprising: an indenter for indenting a surface of the soft tissue; a monitor for monitoring a change or changes in the surface during and after indentation; and a display for displaying information on the change or changes.

The indenter may include a piston. The piston may be any suitable shape but semi-spherical or cylindrical is preferred, A diameter of around 0.1 to 2 mm or a spherical radius of 0.1 to .5 mm. may be particularly suitable. Other configurations may also be acceptable and a diameter of 1-5 mm may be adopted.

The indenter preferably further includes a driver for propelling the indenter. The driver may be manually activated by an operator. Alternatively the driver may be activated by drive means such as mechanical, gas, electric, spring, magnetic or other suitable drive arrangement. The driver may be selectively variable for speed of advance and/or applied load.

The soft tissue is preferably cartilage. The monitor may comprise one or more of an ultrasonic monitoring component; fibreoptic monitoring component and NIR spectroscopic monitoring component. The monitor may include a displacement transducer such as a linear variable displacement transducer. The preferred change for monitoring is recovery of the surface after indentation. It is to be understood that recovery of the surface in this represents recovery of the tissue. The change or changes may include or, alternatively comprise, stretch of the surface. The changes may also include initial displacement of the surface. The monitor may also provide information on structure and chemistry of the soft tissue such as the integrity of the collagen meshwork, proteoglycan content and/or chemical changes associated with osteoarthritis progression. The monitor may comprise a single monitoring component or two or more monitoring components In a preferred embodiment of the monitor, the monitoring components may comprise a linear variable displacement transducer to monitor initial deformation, one or more fibreoptic and/or ultrasound transducers to monitor stretch of the surface and one or more fibreoptic and or ultrasound transducers to monitor recovery of the matrix, The fibreoptic and/or ultrasound transducers to monitor stretch are preferably offset to provide information at two or more different distances from the point of indentation. The ultrasound monitoring component may also scan cartilage and/or subchondral bone.

The display may include a processor for receiving, storing and/or analysing information on the changes. The processor may be a computer. The computer may determine a Recovery Index for the soft tissue. The display may include a display screen. The display may provide a visual, auditory and/or tactile indicium of soft tissue health. The visual indicium may be displayed on the display screen. The processor may determine and display on a display screen a map or fingerprint of soft tissue health.

In a further aspect, the invention may reside in a testing device for use in characterising the health of soft tissue, the testing device comprising an indenter for indenting the surface of the soft tissue, a monitor for monitoring a change or changes in the soft tissue, a processor for receiving, storing and analysing data on the change or changes wherein the change or changes comprises or includes rebound of the surface after withdrawal of the indenter.

The soft tissue is preferably cartilage and the processor may be programmed to identify the rebound strain calculated as the distance the surface recovers from loading at a given time after withdrawal of the indenter. The processor is preferably programmed to further provide a recovery index calculated with respect to relative deformations and/or quotient of dividing the rebound strain by the maximum

indentation stress,

where SR is the recovery index. ει the "short- term" time-dependent elastic rebound strain, and εo is original deformation, or indentation strain σ is maximum indentation reaction stress.

The indenter may have a diameter in the range of around 0.1mm to around 5mm and is preferably adapted for indentation for over a period up to 5 seconds. The device may include one or more of a fibre optic monitoring component, an ultrasonic monitoring component, a near infrared monitoring component and a linear variable displacement transducer for determining the travel of the indenter and/or the rebound of the surface.

The indenter may be advanced until it reaches either a preset stress or reaches a selected strain preferably between 5 to around

30% or even a pre-set distance. The testing device may further comprise an ultrasonic apparatus for emitting ultrasonic waves and detecting ultrasonic echoes.

The processor may be programmed to determine a ratio of reflection coefficients from the surface and a related osteochondral junction for use in characterising the health of the tissues. The processor may be programmed to analyse and display the frequency profile of ultrasound echoes from the soft tissue, preferably up to 25 MHz.

The testing device may further comprise a near infrared spectroscope for conducting diffuse reflectance near infrared spectroscopy of the soft tissue and the processor may be programmed to analyse results of the spectroscopy to provide an indicator of the health of the soft tissue. Most preferably the testing device is mounted to or mountable to an arthroscope.

In a still further aspect, the invention may reside in a method of characterising the health of soft tissue, preferably cartilage, the method comprising the steps of: applying a linear load to a surface of the soft tissue; withdrawing the load; and monitoring the recovery of the surface.

Applying a linear load to a surface of the soft tissue may comprise operating an indenter to propel a piston against the surface. Preferably the method includes the step of applying the linear load substantially perpendicularly to the surface.

Monitoring the recovery of the surface may include the steps of: assessing the amount of initial deformation of the surface; assessing the degree of recovery of the surface towards its original conformation; and characterising the health of the soft tissue.

Assessing the degree of recovery of the surface towards its original conformation may include the step of assessing the amount of recovery over one or more time periods. Assessing the amount of recovery of the surface towards its original conformation preferably includes the step of determining a recovery index according to the algorithm implementing either of the recovery indices calculated with respect to relative deformations 3i_D = ε_!/ε₀ and/or quotient of dividing the rebound strain by the maximum indentation stress,

where 91 is the recovery index. εj the "short- term" time-dependent elastic rebound strain, and εo is original deformation, or indentation strain σ is maximum indentation reaction stress. The method may further include or, alternatively, may comprise the step of monitoring stretch of the surface around a point of indentation, deriving information on the stretch, analysing the information and characterising the health of the soft tissue based on the analysis. The method may further include monitoring the stiffness at the point of indentation, deriving information on the stiffness, analysing the information and characterising the health of the soft tissue based on the analysis. The preferred method utilises ultrasound.

The method may further include chemical/molecular analysis of the soft tissue utilising NIR spectroscopy,

The method may further include the steps of ultrasonically scanning the cartilage and/or subchondral bone and analysing the ultrasound scans to provide information on the health of the cartilage. The method and device may be suitable for use on intravertebral discs. In a still further aspect, the invention may reside in a method of assessing the health of soft tissue, the method comprising indenting a surface of the soft tissue, monitoring rebound of the surface and analysing data on the rebound to provide an indicator of the health of the soft tissue. Indenting a surface of the soft tissue preferably comprises indenting a surface of cartilage, and analysing the data includes determining rebound strain calculated as the distance the surface recovered from loading at a given time after indenting.

Analysing the data may further comprise the step of determining a recovery index calculated either with respect to relative deformations yi_D = εJε₀ and/or quotient of dividing the rebound strain by the maximum indentation stress,

: where SR is the recovery index. ει the "short- term" time-dependent elastic . rebound strain, and εo is original deformation, or indentation strain σ is maximum indentation reaction stress.

The method may further comprise the step of directing ultrasonic waves into the soft tissue, detecting ultrasound echoes from the soft tissue, and characterising the health of the soft tissue. The step of characterising the health of the soft tissue may include determining a ratio of reflection coefficients from the surface and a related osteochondral junction.

The method may include the step of determining a frequency profile of ultrasound echoes from the soft tissue and analysing the frequency profile to provide an indicator of the health of the soft tissue.

Preferably when determining a frequency profile the orientation of an ultrasonic probe does not vary more than +/- 1.2 degrees.

The method may further comprise conducting diffuse reflectance near infrared spectroscopic examination of the soft tissue and analysing the result of the spectroscopic examination to assist in characterising the health of the soft tissue.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a plan schematic view of cartilage of a synovial joint with a defect;

FIG. 2 comprises FIGS. 2A - 2C showing a schematic side view of the application of an indenter of the present invention;

FiG. 3 is a schematic bottom view of an indenter of the present invention showing one arrangement of transducers;

FIG 4. is a schematic side sectional view of the operation of a piston in an indenter;

FIG. 5 is a schematic side sectional view of the use of an indenter on normal and degraded surfaces; FIG. 6 is a schematic side sectional view of the use of an indenter on normal and degraded surfaces for measurement of the stretch parameter;

FIG. 7 is a flow chart of a map and embodiment of a method for characterising the health of cartilaginous soft tissue; FlG. 8 is a schematic view of a "fingerprint" provided by an indenter of the present invention along with optional indicia of the health of soft issue;

FIG. 9A and 9B are a graphical representations of recovery profiles for normal cartilage and degraded cartilage; FIGS. 10-13 show recovery Vs time characteristics of normal and degraded samples under different conditions;

FIG. 14 shows a recovery profile around an osteoarthritic defect;

FIG. 15A and 15B shows results of axial strain and effective surface stretch characteristics of cartilage; FIG. 16A, 16B and16C show representative data from normal and degraded joints for axial strain, effective surface stretch and their ratio;

FlG. 17 shows a comparison of a frequency profile for normal cartilage compared to a proteoglycan depleted sample;

FIG. 18 shows a schematic representation of an ultrasound apparatus for use in the present invention;

FlG. 19 shows a series of ultrasound reflection patterns; and FlG. 20 shows reflection coefficients for surface and bone. FIG. 21 shows the effect of orientation on the reflected signal from the articular surface;

FIGS 22 and 23 show the frequency profiles of the surface and osteochondral junction;

FIG. 24 shows frequency profiles; FlG. 25 shows baselined ultrasound reflection frequencies of highest overlays and significance;

FIG. 26 shows spectral profiles of articular cartilage-on-bone using DR-NIR spectroscopy;

FIGS 27 and 28 show plots of eigenvectors; FIG. 29 shows box and whisker plots showing the spread of results; and

FIG 30 is a schematic view of a further testing device. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Referring to FIG. 1 there is seen a representation of a clinical situation in which the present invention may be of particular benefit. A cartilaginous layer 15 of a synovial joint is shown having a focal defect 16 which is a lesion in the joint severe enough to require intervention by a surgeon. The more problematic consideration is the area of interest 17 which may not be visibly damaged but which may have pathological changes sufficient to require surgical intervention.

FIG. 2 is a schematic view of the operation of the device and method of the present invention. Articular cartilage 18 overlies bone 19. A surface 20 of the articular cartilage is subject to indentation by the application of a linear force which is preferably substantially perpendicular to the surface. The force is represented by arrow 21 which is schematically representative of an indenter which is in turn connected to a processor and a display.

FIG. 2B shows application of the force 21 at a point of the surface with subsequent indentation 22 of the surface 20 as well as surface stretch as shown by arrow 23. The term "point" refers to the surface area of cartilage contacted by the indenter.

Recovery represents the ability of the cartilage to "spring back" towards its original conformation. FIG. 2C shows the recovery as represented by arrow 24. Therefore, FIG.2A represents the cartilage on bone before indentation. As the load is applied, the shaded area 25 represents the volume of the indentation, which is believed to equal the amount of fluid removed. This volume will depend on how the load is applied and also resistance to the flow of the fluid which, in turn, may be related to the structure. The load is also resisted by the surface stretch as shown by arrow 23. FIG. 2C shows the tissue after the load has been withdrawn. It is thought that the process involves a solid structure- dominated instantaneous recovery coupled with a fluid recovery. This recovery provides a measure of structural integrity without being influenced by the load. The inventors do not wish to be bound to any one or more theories expressed in this specification and directed to the possible reasons for the effectiveness and physiological mechanism or operation of the invention. All comments on any theory of operation of the present invention are offered only as non-binding suggestions.

FIG. 3 shows the bottom view of one embodiment of an indenter 26. Initial deformation caused by the load can be measured by using a linear variable displacement transducer ("LVDT", not shown) attached to the piston 31 of the indenter and placed within the cylinder. The surface profile around the indenter is measured by fibreoptic and/or ultrasound transducers to analyse the stretch. These transducers 27,28,29,30 are shown in this embodiment as offset to provide measurements at two distances from the indenter. They can also be used to ensure that the indenter is perpendicular to the surface on operation. The piston 31 is centrally located for extrusion outwardly and into contact with the cartilaginous surface. Once the piston 31 is retracted, recovery is measured by ultrasound or fibreoptical/NIR transducers embedded in the indenter. In the present embodiment, such a transducer 32 is shown located centrally in the device. In some preferred embodiments, secondary chemical or physical information may also be gathered by this central transducer. This may include information regarding the integrity of the collagen meshwork and the proteoglycan content of the tissue using ultrasound or NIR. The present testing device may be used as a stand alone instrument for assessment of soft tissue health, in one particularly useful embodiment, the device may be incorporated into an existing arthroscope using simple rigid snap-on/off components. This arrangement may provide great utility during procedures such as arthroscopic joint surgery. The present invention is particularly well suited for an assessment of cartilage. It may however be useful in other soft tissues and materials with soft matrices and cores such as brain tissue, muscles, tendons, ligaments, flesh and cancerous tissue, eyes, fruits, vegetables, foams, fabrics, environmental/ground water probing, soil and clay evaluation and grading, tissue engineering scaffolds and cell-scaffold interfaces, and outcomes of soft tissue treatments.

FIG. 4 shows a schematic representation of the operation of a piston 33 of an indenter. An air supply 34 is provided and operated by switch 35. Air is passed through a pressure reduction arrangement 36 before delivery to a cylinder 37 and activation of the piston 33. A coiled spring 38 biases the piston into a retracted position so that release of the pressurized air from the cylinder 37 causes automatic retraction of the piston 33. Retraction is preferably accomplished in under 1 second. It is essential that the design provides a safe repeatable and simple way of applying the initial load. While a mains voltage electric arrangement may provide a suitable motive force for the device, it is generally not considered safe at surgery. A battery powered arrangement may be used to power the driving mechanism. Most devices used in surgery are powered by pressurised air. It is envisaged the present invention may be operated from hospital's pneumatic supply which allows safe repetitive load application. Indentation can be provided by a single cylinder and piston returned by a spring. The piston may then act as the indenting object using pneumatic pressure to apply the load and the spring to retract it. A similar effect may be achieved using a double acting cylinder.

FIG. 5 shows a schematic operation of a testing device comprising an indenter with piston 39 applied to a cartilaginous surface 40 of articular cartilage 41 on top of subchondral bone 42. The piston is urged downwardly by a driving force represented by arrow 43. The left- hand side 44 of the figure shows the reaction of normal or healthy cartilage to the piston 39. The right-hand side 45 shows the reaction of a degraded surface. A normal surface layer will deform less than a degraded one forcing the matrix to deform around the indenter and creating the appearance of a classic elastic deformation. A degraded surface may simply stretch around the matrix adjacent to the indenter as shown. This test highlights the opportunity provided by the present invention for a surgeon to determine the integrity of the surface by measuring an area outside the indentation, unlike a simple stretch determination under an indentation which is both unreliable and very difficult to measure at surgery due to the need for a larger and impractical indenter. The offset transducers of FIG. 3 may be used to measure at two distances from the piston 39 as shown in FIG. 6. The first distance measurement shown by arrow 46 and a more remote measurement of arrow 47. These may be compared with the minimal stretch and displacement of arrow 48 for degraded tissue. Referring to FIG. 7, there is shown a flow chart for one embodiment of the method of the present invention. A cartilaginous surface is indented 50 after which monitors in the indenting device monitor recovery 51 and stretch 52. Information from the monitors is passed on for data processing 53 using the application of algorithms and preferably in calculating a recovery index. Data processing provides information which is then measured against a referenced database to provide a decision 54, The decision may be communicated on a display. The display may be as simple as a green light 55 for healthy tissue, red light 56 for definite removal and an intermediate light for the surgeon's decision 57. The information may be stored in a storage device 58. Additional information, such as three dimensional location, may be inputted into the storage device 59. The information may be displayed as a graphic representation of the joint, such as an osteoarthritis/defect "fingerprint" 60. This can also be seen as a schematic representation of a testing device of the present invention with an indenter 50, a monitor (51, 52), a processor/computer 53, 54, 58 and an indicator 55, 56, 57. The device may also include positioning monitors for displaying a fingerprint 59, 60. At the processing stage of the method, the parameters are compared to experimental results for normal joints, collagen-specific and proteoglycan-specific degradations and arthritic joints. This may be performed in a simple regression algorithm. This regression is used to decide whether the tissue is healthy and should be kept, or degraded and should be removed. Results may be included into the reference database to thereby broaden the basis of comparison. It is preferred that new results are added to the database to increase its contents and accuracy.

The present invention may be combined with existing three dimensional positional technologies to thereby provide an arthritic/defect "fingerprint" for the surgeon.

FIG. 8 shows one schematic view of such a display wherein the three indicator lights 55,56,57 as shown above an outline 61 of the cartilaginous layer with a defect 62 apparent plus an area of degraded tissue 63 and an intermediate or unknown area 64. A surgeon may therefore be informed by the present invention that he/she needs to exercise more widely than just the obvious defect and can also exercise clinical judgement in deciding where to terminate the excision of the tissue. The fingerprint provides the surgeon with information on structural degradation in the joint on which the surgeon can act with far more confidence than with any other available system. This extends visual grading systems to provide a decision making too! that will optimize transplantation procedures and have significant benefits for patients.

The difference between normal and degraded tissue is highlighted in FIGS 9A and 9B in which the recovery index of normal cartilaginous tissue is plotted against proteoglycan depleted tissue. Both are after a one hour treatment with trypsin. Fig 9A is with load applied at

0.1 s^'1 and FIG 9B at 0.025 s^"1. The plot is recovery index against time.

The inventors have found that the use of recovery of the surface and therefore matrix and surrounding cartilage is a reliable guide to the health of the tissue. It is preferred to determine a recovery index to give an objective indicator of tissue health. In an extension of the invention, an assessment of surface stretch may be incorporated to provide specific information on the integrity of the surface, which is often the starting point of a degradation process. In yet a further addition to the invention, the recovery and stretch assessments may be usefully supported by techniques to provide secondary information on the structure and function of the tissue. Ultrasound can provide information about underlying bone changes that occur with arthritis and other joint defects, and also physical information about the cartilage matrix. Spectroscopy particularly, using a near infra-red technique, can provide information about the nature of the tissue including its chemical components.

The recovery index is measured by the amount the tissue

"bounces back" after an indentation. This varies with the integrity of the structure and makes the recovery method more reflective of degeneration than the existing methods which depend more on the nature of the applied load. The profile of the recovery curve over time is indicative of aspects of soft tissue health. A preferred period of time is one to five seconds which will usually be sufficient for analysis. The patterns of the time-dependent recovery index for normal cartilage is consistently and significantly higher than that for degraded cartilage. In this case, the cartilage is degraded by proteoglycan removal.

Deformation can also be measured with light by using a fibreoptic displacement transducer. The absorbed and reflected light at different wave lengths is useful as a secondary parameter to gather chemical information about the cartilage matrix. A preferred approach uses a Near infra-Red ("NIR") probe or apparatus to analyse the chemical constituents of the tissue and identify changes with degeneration. Chemical profiles of normal and arthritic joints may be mapped to establish a reference system for use in analysing the health of tissue at surgery. Example 1 :

This example highlights structural response of the tissue, independent of the manner in which a load is applied. A measure of the recovery, which is governed by the integrity of the collagen fibres and their interactions, in addition to the proteoglycan content and configuration, provides important information about the health of the tissue. The compliance functional tftc presented here is the recovery index or, also known as the hybrid functional index. This is calculated based on the original deformation, or strain εo, and the short-term time dependent reduction in strain εj.

"3{ -ε_Rjσ where ε_R was the rebound strain and σ was the measured maximum reaction stress to mechanical indentation of a sample. Samples of normal intact, artificially degraded and osteoarthritic cartilage were subjected to quasi-static compressive loading at 0.1 and 0.025 s^"\ to the same level of deformation (30% strain) and then unloaded. The rebound was measured over the first 60 seconds after the removal of the load and compared to indentation stiffness.

A significant reduction in instantaneous rebound was observed for artificially and naturally degraded samples, particularly during the first 5 seconds of rebound. The hybrid index provided a 300% improvement over indentation alone in the ability to discriminate normal and artificially degraded samples in a pairwise test, with significant changes observed in osteoarthritic samples (p<0.005) in the early stages of rebound.

Samples were tested from normal, artificially degraded and arthritic patellae. Intact joints were taken from oxen within 24 hours of slaughter, wrapped in a 0.15M saline soaked cloth and stored at -2O⁰C. Prior to testing, the joints were thawed in saline, sectioned into 14 x 14 mm squares and labelled according to their position on the patella. The cartilage samples were set perpendicular to the indenter in Palapress (Heraeus Kulzer GmbH & Co, Germany) dental acrylic and then returned to saline to recover for 2 hours.

18 normal samples were tested under compressive loading, and artificially degenerated to either disrupt the collagen network or deplete the proteoglycan content of the tissue. After testing under compressive loading, 12 samples were treated for one hour in O.img.ml^'1 of Trypsin (from bovine pancreas T4665, Sigma-Aldrich) in phosphate buffered saline. These were then blotted dry and returned to 0.15M saline before retesting. A second group of 4 samples was placed in 0.15M saline containing 3OU. ml^"1 collagenase (Sigma C0773 protease free, Sydney, Australia) for 16 hours at 37^CC and retested.

To test the application of the technique to the characterisation of naturally degraded tissue, a second group of 10 normal and 10 osteoarthritic patellae was obtained and tested under the same conditions as the previous normal samples. For the osteoarthritic patellae, samples were subjected to load-unload tests in the areas of focal fibrillation, and adjacent regions of visually normal tissue at distances of 6 and 12 mm from the edge of the defect, After testing, these samples were subjected to histological analysis to quantify the proteoglycan content and the presence of an orientated collagen meshwork.

The histological techniques used here are fully explained in a previous paper [25]. For the analysis of proteoglycan content, samples were sectioned at 7 μm, fixed and stained with safranin-0 using standard histological procedures. Optical absorbance profiles were taken using a microscope and a monochromatic light source, and processed with IMAGEJ software (1.33u, National Institutes of Health, Bethesda, USA) to determine the proteoglycan concentration as a function of distance from the articular surface. The presence of an orientated collagen meshwork was measured using polarised light microscopy. Birefringence profiles were taken from 20 μm, fixed sections and similarly analysed in IMAGEJ.

The normal and degraded samples were subjected to two loading regimes on a Hounsfield Testing Machine (Hounsfield Testing Equipment, Salsford, England), using a φ4 mm plane ended indeπter. These regimes included 0.1 s^"1 and 0.025 s^"1 quasi-static loading to 30% strain, at which point the load was immediately removed and the deformation measured for 60 seconds: The normal/osteoarthritic group were tested at 0.1 s^'\ representing the load rate that can be expected to be applied by a surgeon at arthroscopy. The rebound strain was then calculated as the distance that the sample recovered from loading at a given time. The hybrid functional index was calculated as ϋl = ε_n/σ where ε_R was the rebound strain and σ was the measured stress at 30% strain. After each load-unload cycle, the samples were allowed to recover for 2 hours before retesting. Axial force and displacement vs. time data were taken using a Yokogowa DL-708E digital oscilloscope (Yokogawa, Tokyo, Japan).

The level of variation within the normal samples was calculated as a relative difference, in order to take into account magnitude differences between measurement types. The relative difference was

calculated as The Student's t-test and the overlap

between the paired normal and degraded samples, was applied to determine the ability of the rebound to distinguish normal from artificially degraded samples. This analysis was then applied to the unpaired second group of normal and osteoarthritic samples, with results correlated against position on the joint, proteoglycan content and collagen birefringence.

The rebound vs. time characteristics of normal and artificially degraded samples are presented in FIGS 9A and 9B. These figures show the characteristic unloading behaviour under 0.1 s^~1 0.025 s^'1 driven load scenarios respectively, comparing normal behaviour with that after 1 hour of trypsin treatment.

Table 1

Table 1 presents the variation in the rebound and stiffness for norma! samples as a relative difference. It was observed that the variation in stiffness across the joint was approximately double that of the rebound for the same sites and the same indentation. The overlap scores for each load rate are presented in Table 2, below:

This table shows the probability of distinguishing a randomly obtained sample as being either normal or degraded. The statistical results for the 2, 5 and 10 second rebound of normal and osteoarthritic samples are presented in Table 3.

This shows statistically significant differences between normal and osteoarthritic samples for the 2 and 5 second rebounds, and the hybrid parameter at 2 and 5 seconds. It should be noted that the variation in proteoglycan content, measured by optical absorbance between normal (mean 216, SD 13.5) and osteoarthritic (mean 214, SD 13.8) samples was visible but very small. A similarly small variation was observed for collagen birefringence results (normal 19.2, SD 1.9; osteoarthritic 18.9, SD 1.8).

The results of this study show that the rebound of articular cartilage after compressive loading, particularly in the first few seconds, is sensitive to the integrity of the matrix structure with respect to the discrimination between normal and degraded tissue, and can be combined with a classical indentation measurement to create a hybrid index of functional integrity. FIGS 9A and 9B show that the artificial degeneration of normal cartilage by proteoglycan depletion resulted in a consistent 20 to 30% reduction in rebound. At the loading rate of 0.025 s^"1 the proteoglycan depleted samples showed a significant change, particularly for the first 15 seconds of rebound (p<0.001), with p<0.005 and p<0.01 at 30 and 60 seconds respectively. Similar results were observed under the higher loading rate, with the exception of a reduced statistical significance for proteoglycan depleted samples at 2 and 5 seconds (p<0.005).

The results in Table 1 show a wide variation in stiffness across the normal joint (64% relative difference). The variation in the rebound under the same compressive load was considerably smaller (24 to 41). This advantage is reflected in the overlap scores, presented in Table 2, which show that the ability to distinguish samples as being either normal or degraded is considerably improved by using the rebound in comparison to indentation stiffness for the same samples, with the best results at 2, 5 and 10 seconds of rebound from the 0.1 s^"1 indentation. This potential is further demonstrated in Figures 9A and 9B, indicating that the possibly varying load rates applied by the surgeon(s) will have less effect on the results using rebound (p>0.2) than indentation stiffness (p<0.05). It is preferred that the load be applied for a short (preferably t<5 s) time to produce the most consistent and practically obtainable results.

Applying the rebound and indentation response of the samples as a hybrid index showed an improvement in overlap scores, particularly at the loading rate of 0.1 s^"1, with significant changes at 2, 5 and 10 (all p<0.001) seconds rebound. It can be observed from Table 2 that the hybrid index considerably improves the ability to distinguish normal from degraded samples.

Although the histological changes in the osteoarthritic samples were very small, significant differences were observed between the normal and osteoarthritic samples at 2 (p<0.01) and 5 (p<0.05) second rebound. These two rebound times also showed the best correlations with position relative to the defect and proteoglycan content, while collagen birefringence correlated equally with 2, 5 and 10 second rebound (see Table 3). These correlations were generally considered to be weak when compared to the large changes from the artificial degradation processes, probably due to the small histological differences between the normal tissue and the tissue at ≥6 mm from the osteoarthritic defects. The inclusion of the load-dependent parameter in the hybrid index lead to stronger correlations with position and collagen birefringence, but weaker with proteoglycan content compared to the rebound on its own, though each correlations were more consistent with rebound time. A significant difference could be observed between normal and osteoarthritic tissue for 2 (p<0.005) and 5 (p<0.01) seconds rebound using the hybrid index.

The measurement of rebound can be easily incorporated into current indentation probes such as ARTSCAN [8] and ACTAEON [22] which use ultrasound and fibreoptic displacement measurements respectively. While the stress measured at indentation provides a useful parameter for a quasi-static indentation with controlled displacement, this parameter will preferably be an indentation displacement parameter for statically applied loads or to calculate a rebound strain when the sample thickness is unknown.

The load rate dependence of the stiffness and recovery parameters in normal samples is presented in Table 4. The difference in the stiffness parameter is an order of magnitude larger than that of the recovery parameter over the two load rates,

Table 4: Load rate dependent variation in a representative normal sample. The recovery v's time characteristics of normal and artificially degraded samples are presented in FIGS. 10-13. FIGS. 10 to 12 show the characteristic unloading behaviour after 1 hr of trypsin treatment under 4.5 mm/min, 1.5 mm/min and static loading scenarios. FIG. 13 discloses the unloading behaviour before and after collagenase treatment. FIG. 14 presents the recovery as a function of the distance from a focal osteoarthritic defect. It can be seen that the characteristic recovery increases with distance until approximately.8mm, after which it plateaus.

These results show the recovery index is very sensitive to degenerative changes in the arthritic joint, and superior to the currently used stiffness parameter in detecting arthritic change. This technique has potential application in a wide range of mechanical diagnoses of soft tissue, particularly in the diagnosis of osteoarthritis at arthroscopy.

It is important that the normal variation be as small as possible to detect small degenerative changes with a high degree of confidence. Stiffness values demonstrate a wider variation across the normal joint than the recovery measured under the same compressive load. It is apparent from Table I that there is a large natural variation (64% relative difference) in stiffness across a normal joint. The variation in the recovery index under the same level of loading is considerably smaller (24 to 30%), and is therefore more sensitive to degenerative change.

It is impossible for a surgeon to apply a constant loading rate at arthroscopy. The measurement system is preferably sufficiently robust to allow for the different abilities of the surgeon. From the information presented in Table 1 it is apparent that the recovery index is considerably less dependant on the applied loading rate. It should be noted, however, that after static loading for 30s, the recovery was reduced in all of the tested tissues due to the gradual outflow of fluid from the matrix over this time. In general, therefore, it is preferred that the loading is applied for a short (<5s) time. As the load rate dependence is an order of magnitude smaller than for stiffness, the recovery index is. far superior in this respect.

The artificial degeneration of normal cartilage has resulted in a consistent 20 to 30% reduction in recovery index, with the exception of the static loading of trypsin-treated samples, which showed a smaller change due to the fluid flow during the 30s creep period. This response was not seen in the static loading of collagen-disrupted cartilage. The artificial degeneration times and solution strength in this study were smaller than those generally used by researchers in this field producing a very mild, but easily detectable degenerative change. The arthritic changes measured in this example, however, were considerably larger than the artificially induced changes and fell well outside of the natural variation in normal samples. The recovery index can be expected to detect naturally degenerative changes with a high level of confidence.

In the osteoarthritic joint, the method was able to successfully track the transition from degenerated to normal tissue, as shown in FIG. 14. The results show a gradual increase in recovery to a distance of approximately 8mm, after which the response plateaus. The value at the plateau is smaller than the recovery index of normal tissue. This shows that there is both a focal degeneration which extends into the visually normal areas of the joint, as well as a general degenerative change across the entire joint. The general change in the joint is considered reversible and can be returned to homeostasis. The focal change, however, is irreversible and the focal area described by this test will need to be removed in a localised treatment procedure such as mosaicplasty or tissue engineered cartilage replacement,

The recovery method can be implemented in a hand held testing device for assessing soft tissue and in particular, used to determine the extent of arthritic damage in a joint and optimise the amount of tissue removed at surgery. The recovery methodology is significant in improving decisions relating to the treatment of osteoarthritis. There is a benefit for both knee replacement surgery and articular cartilage transplantation of all kinds. The recovery method is useful at arthroscopy in any case that potentially involves a focal tissue replacement, such as mosaicplasty, artificial biomaterial implants, artificial biomaterial or tissue engineered replacement and similar. Unicompartmental knee replacement will also benefit from this technology. This provides procedural benefits in surgery where more precise decisions can lead to the optimisation of the amount of cartilage removed in the treatment of osteoarthritis. This also aids developments in micro-surgery of the joint and enables significant improvements in tissue replacement outcomes for patients. The ability to optimise and leave the maximum amount of native host cartilage in a joint provides a significant benefit to the issues of most relevance to the patient relating to the reduction in rehabilitation and costs.

Example 2:

The following example considers the surface stretch characteristics as an aid in assessing health of cartilage. Current in vivo indentation techniques, which primarily rely on stiffness measurements based on axial data, are unable to adequately distinguish between healthy and degraded tissue. The present in vitro study investigates the effects of controlled artificial degradation on the effective surface stretch, comparing the results with those obtained from the peripheral cartilage surrounding focal osteoarthritis. Results suggest that this technique is highly sensitive, showing a maximum range of 14% effective surface stretch in a normal joint compared to 42% for axial strain measurements. It is further shown that the technique can discriminate between degenerative changes and the intrinsic variations in cartilage properties across the normal joint. The relationship between indentation and the in-plane strain field under the indenter can better distinguish degraded tissue than the currently used stiffness techniques.

This example is directed to the superficial zone in its functional role. By loading the tissue with a spherical or cylindrical indenter, the collagen network in the superficial layer produces a strain- limiting behaviour, which plays a significant role in maintaining cartilage thickness in the consolidation process. The superficial zone will stretch under the indenter, and will also act in the region adjacent to the indentation. A measure of the surface stretch, which is governed by the integrity of the collagen fibres and their interactions, for a given stress- strain characteristic will provide important information about the integrity of the collagen network in this zone. A strongly bonded network will resist this stretch, but a weakly bonded network will be able to deform more. By combining this data with the classical axial indentation data, a more thorough picture of the biomechanical integrity of the joint is possible.

26 samples were tested from normal, artificially degraded and arthritic patellae. Intact joints were taken from oxen within 24 hours of slaughter, wrapped in a 0.15M saline soaked cloth and stored at -2O⁰C.

Prior to testing, the joints were thawed in saline, sectioned into 14 x 14 mm squares and labelled according to their position on the patella. The cartilage samples were taken off the bone and immediately glued to stainless steel disks in order to eliminate the influence of the bone on the mechanical response of the tissue. A fine grid was imprinted using a waterproof marker and the samples returned to saline to recover for 2 hours at 4°C.

8 Normal samples from 2 patellae were tested under static compressive loading, and artificially degraded to either disrupt the collagen network or deplete the proteoglycan content of the tissue. After testing under compressive loading, four samples were placed in 0.15M phosphate-buffered saline (PBS) containing 3OU. ml^"1 collagenase (Sigma C0773 protease free, Sydney, Australia) for four hours at 37°C. These were then blotted dry and returned to 0.15M saline before retesting. The samples were then subjected to a further four hour collagenase treatment and retested. A second group of four samples was treated for one hour in O.img.ml^'1 of Trypsin (from bovine pancreas T4665, Sigma-AIdrich) in PBS and retested. It should be noted that although collagenase treatment in PBS has minimal effect on proteoglycans, Trypsin can further attack collagen which has previously been cleaved.

To test the application of the technique in naturally arthritic tissue, 6 visually normal samples from different patellae were harvested adjacent to focal arthritic defects and were tested under the same conditions as the normal samples. The focal defects in the 6 arthritic patellae were all characterised by exposed bone and occurred in either the lateral-distal area of the joint or the central-distal area. Samples were harvested from the nearest visually normal tissue to the defect on the lateral side. A custom-built indenter was used to facilitate the photographing of the indentation imprint and stretching of the area around the indenter. This indenter comprises a 6 mm diameter glass disk glued onto a 20 mm thick glass plate of 100 mm diameter. The plate was set in a stainless steel frame with an angled mirror to allow direct viewing through the indenter.

Samples were subjected to a static indentation load of 85N for 13s, corresponding to approximately 0.33MPa₁ on a Hounsfield Testing Machine (Hounsfield Testing Equipment, Salsford, England). This figure was chosen to represent a pressure that can be expected to be applied by a surgeon at arthroscopy. Axial force and displacement vs. time data were taken using a Yokogowa DL-708E digital oscilloscope (Yokogawa, Japan). Engineering strain was calculated by ε= (h_o - h)/h_o where h = instantaneous thickness and ho = original thickness.

Lateral stretch at the surface both under and surrounding the indentation was captured optically at different stages of the indentation using a Canon EOS 20-D digital camera equipped with a Tamron 90mm f2.8 macro lens. ImageJ software (National Institutes of Health, USA) was used to track the x-y coordinates of the centroid of each grid point throughout the indentation, giving deformation data. A superimposed plot of the lateral deformation, or stretch, in the articular surface under indentation was developed. Under a plane-ended indenter, the articular surface undergoes deformation in a radial direction, increasing in magnitude with increased distance from the centre of indentation. In the region adjacent to the indenter, the visible articular surface strain continued in an outward direction for up to 2 mm. The inside end of each ray showed the initial position of a point, while the outer end shows its position after 20s at 0.33 MPa. The effective surface stretch measurement was calculated by normalising the length of the rays against the indenter size. Evaluation based on effective stretch as defined in this study was considered appropriate. Use in surgery is based on a known indenter geometry rather than a parameter that requires calculation based on the initial positions of grid points relative to the centre of indentation. The effective surface stretch can therefore be derived as ESS= JfD where s is the average stretch, or ray length, and D is the diameter of the indenter.

Each specimen was removed after testing and a 2mm by 4mm biopsy taken. Each biopsy sample was placed on a metal mount and embedded in optimal cutting temperature (OCT) medium (IA018, ProSciTech) and rapidly frozen in liquid nitrogen to limit damage to the tissue samples. The sample was then placed within a cryostat and sectioned at 7μm in the transverse direction. The sections were immediately picked up by a microscope slide and left to dry within a sealed container, at 4°C overnight. The presence and concentration of proteoglycans was determined by Safraniπ-O staining. Safranin O is a cationic dye that binds stoichiometrically to mucopolysaccharides, that is, one positively charged dye molecule binds to one negatively charged carboxyi or sulphate group [25-28]. Slides were fixed in 95% alcohol for 30 seconds and left to dry in air. Proteoglycans are hydrophitic and easily extracted during conventional histological aldehyde fixation, hence the use of a less disruptive fixative regime. The slides were then hydrated in distilled water, rinsed in 1% Acetic Acid for 8 dips and then stained in 0.1% Safranin-0 for 5 minutes. Finally, the slides were dehydrated in 95% alcohol for 6 dips, followed by 8 dips in 100% alcohol. Colour photographs were captured by light microscope.

Polarised light microscopy (PLM) was used to detect the presence of an oriented collagen network. PLM measurements were performed using a Nikon Labo-Phot PLM. To ensure reproducibility, the camera was calibrated to the light intensity of a λ/4 wave plate, with a known birefringence. The exposure time was then adjusted to the thin section of cartilage and kept constant for the remaining samples. Slides were placed on the moveable stage, with the articular surface facing toward the body of the microscope. The stage was then turned 45" to ensure maximum light emittance through the polarisers and therefore maximum brightness. The image was captured, and converted to ImageJ software for analysis of the retardance profiles.

To test the statistical significance of deviations from the normal, the Student's t-distribution was used to obtain a P-value. P-values of greater than 0.1 accepted the null hypothesis of no change (equivalent to 90% confidence interval). P-values of less then 0.1 were labelled for comparison.

FIG. 15 shows the variations across a normal joint as measured by each technique. The axial strains at 0.33 MPa at different positions on the same normal patella are illustrated in FIG. 15A. FiG. 15B shows the variation in the effective surface stretch for the same indentation. These graphs illustrate the considerably larger variation in axial strain than effective surface stretch across the joint. The mean axial strain and effective surface stretch were 0.24 and 0.027 respectively. The range as a percentage of the mean was 42 and 14% respectively.

Representative data from specimens subjected to collagenase and trypsin treatments, and visually normal specimens harvested adjacent to arthritic defects are illustrated in FIG. 16. These graphs show the axial strain, effective surface stretch and their ratio (ESS : έ) respectively. Four and eight hour degradation in collagenase resulted in a considerably higher effective surface stretch, p<0.01 and p<0.001 respectively, with a visible but a smaller increase in axial strain. The axial strains for the samples that underwent collagenase degradation fell within the normal range. One hour degradation in trypsin produced a large increase in axial strain p<0.01. This degradation also resulted in an increased effective surface stretch, but this was lower than for the 8hr collagenase treatment. Osteoarthritic samples, labelled OA-1 to OA-6, were characterised by increased effective surface stretch (all p<0.001), but only OA-4 showed an increased axial strain p<0.01. The ratio of effective surface stretch to axial strain increased for collagenase treatment and decreased for trypsin treatment. The arthritic samples show similar effective surface stretch to axial strain ratios, all higher than the normal (p<0.1).

The results show that the effective surface stretch measurement is very sensitive to degenerative changes originating at the surface of the cartilage layer, and appears to be superior in detecting this mode of arthritic change. For degenerative changes originating in the deeper zones of the matrix, the compliance measurement is likely to be more effective. By considering the relationship between the axial strain and effective surface stretch, a more accurate picture of health may therefore be determined.

Axial strain values demonstrate a wider variation across the normal joint than the effective surface stretch measured under the same compressive load. There is a large natural range (42% of mean) in axial strain, and therefore stiffness and compliance, across a normal joint. The variation in effective surface stretch under the same level of loading is considerably smaller (14%), and is therefore more sensitive to the changes in the integrity of the collagen network induced by collagenase treatment. The increase in effective surface stretch for the artificially degraded samples was 34 (p<0.01), 119 (p<0.001) and 89% (pθ.001) for 4hr collagenase, 8hr collagenase and 1hr trypsin treatment respectively. The measured effective surface stretch for the osteoarthritic samples was at least 37% higher than the largest stretch measured in the normal samples. As each of these values falls well outside of the natural range it appears that the effective surface stretch measurement can detect degenerative changes with a high degree of confidence.

Although a measurement of compliance is also sensitive to the degenerative changes induced in this experimental program, it is limited by its inability to distinguish small collagen-disruptive or arthritic changes from the natural variations across the joint. The increase in axial strain presented in FIG. 16A for 4hr collagenase, 8hr collagenase and 1hr trypsin treatment were 14, 34 and 110% (p<0.01) respectively. Each of the strain values for samples that underwent collagen degradation fall within the natural variation in the normal samples, suggesting that these results alone cannot be used to evaluate small changes. From FIG. 16A₁ we can conclude that the axial strain appears sensitive to artificially induced changes in the proteoglycan content. The arthritic samples however, which were characterised by reduced proteoglycan content, were generally not detected using axial strain values from compressive loading.

Because collagenase acts to loosen up the collagen network, it is reasonable to argue that its effect will be that of increasing the surface stretch, thereby providing an assessment protocol that is considerably more sensitive than axial strain for this type of degradation. Axial strain, however, is more influenced by the depletion of proteoglycans through trypsin treatment. It therefore appears that useful information about the nature of the degeneration in a given tissue may also be obtained by considering the relationship between the axial strain and effective surface stretch in addition to the effective surface stretch measurement alone. FIG. 16C shows that a reduction in the integrity of the collagen network increases the ratio of effective surface stretch to axial strain, while the depletion of proteoglycan decreases the ratio. Applying this relationship to the visually normal samples adjacent to the arthritic defects generally showed an increased ratio, or increased effective surface stretch behaviour accompanied by a small or no increase in axial strain, indicating that a disruption of the collagen network has occurred. Arthritic sample OA-4 shows very large axial strain behaviour and a large effective surface stretch, indicating a more substantial deep proteoglycan loss in addition to considerable collagen disruption.

It is known that collagen disruption at the surface is apparent in the early stages of osteoarthritis. The PLM technique used in these investigations, however, was not sufficiently sensitive to show significant collagen degradation in the samples, with the exception of sample OA-1. Safranin-0 staining, however, revealed a substantial proteoglycan depletion associated with the disease process.

It can be concluded that the development of effective surface stretch in articular cartilage is more sensitive to anatomical changes in articular cartilage, including small changes in the collagen network, and can be employed for distinguishing structural and biomechanical alterations due to disease and biomechanical/biochemical degradation. By considering the relationship between the basic indentation data and the concomitant surface stretch, a more sensitive indication of the condition of the entire cartilage matrix may be possible. Such a technique may allow a more meaningful analysis of health to be made at arthroscopy. This is especially important in determining the extent of the disease for tissue replacement and tissue removal at surgery.

In a preferred embodiment, ultrasound is incorporated into the testing device as a way of measuring the deformation and recovery. It may also provide information on the physical characteristics of underlying bone and matrix by measuring reflection ratio or using frequency analysis.

Example 3;

A total of 160 ultrasound echoes were taken from both the saline-cartilage and cartilage-bone interfaces across a normal joint, a joint with an artificially compromised surface, a joint with an artificially compromised matrix and a naturally arthritic joint to determine the effect of different modes of degeneration on the ultrasonic patterns of the cartilage laminate. Intact joints were taken from oxen within 24 hours of slaughter, wrapped in a 0.15M saline soaked cloth and stored at -20C.

Prior to testing, the joints were thawed overnight in saline, according to the method of Broom and Flaschmann [16].

A-mode echograms were obtained randomly over the surface of normal patellae immersed in saline, which were then artificially degraded to either compromise the surface by removing surface active phospholipids (SAPLs), or compromise the general matrix by depleting the proteoglycan content or disrupting the collagen meshwork of the tissue.

Both of these phenomena are apparent in natural degradation. SAPLs were removed by wiping the surface with a methanol soaked cloth, and immediately returning the joint to saline. Following treatment and retesting, the surface was further degraded by light abrasion with a coarse cloth.

In order to removed proteoglycans, joints were treated for four hours in O.img.ml^"1 of trypsin (from bovine pancreas T4665, Sigma- Aldrich, Australia) in 0.15M phosphate buffered saline (PBS) at 37°C. To disrupt the collagen meshwork, the joints were placed in 0.15M PBS containing 30U.ml-1 collagenase (Sigma C0773 protease free, Sydney, Australia) for 18 hours at 37°C. Trypsin generally acts only to remove proteoglycans, but may attack collagen molecules which have been previously cleaved. This effect is likely to be minimal in the 4 hour degradation of a normal joint but the lack of specificity of the action of trypsin on the general matrix of articular cartilage should be noted as a possible limitation in this study. To test the application of the technique in naturally arthritic tissue, three patellae with 6- 8mm diameter focal defects characterised by subchondral bone exposure were tested under the same conditions as the normal samples. These defects occurred in the lateral distal quadrant and echograms (n=12) were taken at random from macroscopically normal tissue between 5 and 20mm from the edge of the defect.

The ultrasonic examinations were made at an approximate distance of 3mm using a 10MHz transducer (Panametrics Inc., Massachusetts USA) connected to a pulser/receiver. This was in turn connected to an oscilloscope and a PC through an analogue to digital converter (Pico Technology Limited, Cambridgeshire, UK). To calculate the reflection coefficient, the reflected amplitudes were compared to that of an acoustic mirror, which was made from a highly polished stainless steel disk and tested in saline under the same conditions as the cartilage samples. The reflection coefficient of the mirror was calculated to be approximately 0.941 Rayl from the known acoustic impedance values of the two materials. The amplitude reflection coefficient of the ultrasound echoes from the cartilage-on-bone was determined from R = (/Wiiage x 0.941 )/A_mir_ror, where A is the amplitude of the reflected signal. The purpose of this investigation was to observe and relate patterns of ultrasonic reflection to degradation, rather than use absolute quantities as the basis for comparison. Where statistics are mentioned, however, p-values were obtained from a double-sided Student's t-test using a "no difference" null hypothesis. FIG. 17 shows a comparison of a frequency profile for normal cartilage compared to that of a proteoglycan depleted sample. The top two printouts are obtained from normal tissue while the bottom two printouts are from proteoglycan depleted tissue.

FlG. 18 shows a schematic of one suitable ultrasound apparatus for assembly with the other component or components of the testing device. The ultrasound apparatus 80 comprises an oscilloscope 81 , a pulser/receiver 82 and a transducer 83. An analogue/digital converter 84 is in communication with the other components and the processor 85 which may be a laptop or other computer. FIG. 19 presents the typical ultrasonic reflections from normal (A), proteoglycan depleted or collagen disrupted (B), surface delipidized (C), abraded (D) and arthritic (E) tissues. These patterns are similar in approximately 80% of the tests, with deviations from these patterns generally occurring when a weaker signal was obtained for either the surface or bone reflections. Bone reflections showed more variation than surface reflections.

The normal sample (FIG. 19A) gives strong signals from both the surface and bone, with the bone signal having slightly higher amplitude. Proteoglycan depletion by exposure to trypsin or collagen depletion (FlG.19B) reduced the surface amplitude (p < 0.01) and slightly increased the bone reflection, though this increase was not statistically significant. Surface disruption by the removal of SAPLs (FIG. 19C) had a non-significant, but visible effect of reducing the surface echo and increasing the bone echo. Further disruption of the surface by abrasion significantly reduced the surface echo (p < 0.1), with no further change in the bone echo.

In the arthritic sample, a considerably larger bone echo could generally be seen, as presented in FIG. 19D. A large variance was seen in this measurement and the change was consequently non- significant. The surface and bone reflections of all samples are presented as reflection coefficients in FIG. 20.

On closer examination of the reflections, a recurring pattern was found in the ratio of surface to bone reflection coefficients. This ratio changed significantly for each degeneration protocol tested in this investigation, particularly for arthritic degradation. The significance of each change is summarized in Table 5.

The most apparent feature of the results was the large variation within each sample group, particularly for the bone reflections. This is most likely due to the considerable natural variation in properties both across and between joints, combined with error due to the alignment of the transducer with the surface. Although a great deal of caution was taken to minimize this misalignment, it is well accepted that a small departure from the correct angle will produce a large change in reflection. Despite these variations, it was found that surface reflection is more influenced by the nature of the matrix than the integrity of the surface. The surface abrasion produced a visible roughening of the surface, but had a less significant effect (p < 0.1) on the surface reflection than the trypsin treatment (p < 0.01). Delipidization of the surface had the smallest effect on surface reflection of all the degradations. In each case, it was found that the bone reflection was too varied to have a significant change with degradation, perhaps explaining why it is seldom measured. In this specification it is to be understood that the term "soft tissue" may include the osteochondral junction and calcified cartilage sublayer in general.

The results show that the ratio of the reflection coefficients from the cartilage surface and osteochondral junction form a coefficient that describes degenerative changes to the articular surface, the superficial and deep zones of the cartilage matrix, and the osteochondral junction. It provides a more consistent indicator of artificial degeneration than the reflection coefficients from the individual surface or osteochondral junction echoes, and was found to be particularly effective in distinguishing naturally degraded from normal tissue. Due to the complementary effect of combined degradations on the value of the ratio, it may provide an effective coefficient for distinguishing degraded from normal tissue in the osteoarthritic joint, independent of the site of initiation of the osteoarthritic process.

It should be noted, however, that the bone reflections provide important information for characterising degradation. The patterns of ultrasonic response remained similar across the samples within each group, and when quantified by the ratio of surface to bone reflection coefficients, gave more significant information about the change than either the bone or surface individually, even though the individual averages often changed by a large amount. The complementary relationship between the reflections in degraded samples, and the results presented in Table 5 show that the application of this ratio may allow a better discrimination between normal and degraded tissues (all p < 0.001), particularly in arthritic joints (p < 0.001). Combining this with the matrix sensitive surface reflection may further provide information about the nature of degeneration in macroscopically normal tissues. The inventors found that information from specific bands in the frequency domain of reflected ultrasound signals has the potential to differentiate normal from degenerate cartilage. In this respect an incidence angle-independent parameter for characterising the structure and structural changes in articular cartilage that are due to degeneration such as osteoarthritis can be established. Furthermore, the inventors found that the frequency-dependent absorption, relaxation and resonance characteristics of articular cartilage-on-bone depends on the structural integrity of the matrix and the cohesion between the proteoglycans and the collagen meshwork. For the same site on the cartilage surface, a change in the incident direction of an ultrasonic pulse relative to the material interface will affect the amplitude of the reflected wave, as can be seen in FlG, 21. The frequency profile remained consistent up to +/- 1.2°after which an increased peak in the 3-5 MH_Z band was observed, as well as a left shifted overall peak. This effect may indicates a difference between the specular and diffuse reflection profiles. With a change in orientation of less than +/- 1.2°, the value of the amplitude at any single frequency changed, but the overall pattern/profile of the reflected signal's frequency response remained the same. This suggests there may be advantage if the angle of incidence can be controlled within a +/- limit.

EXAMPLE 4

A total of 240 ultrasound echoes (20 per joint in a 4x5 grid) were taken from the saline-cartilage and cartilage-bone (osteochondral junction) interfaces of 12 normal bovine patellae. 6 of these patellae were then depleted of their proteoglycan content and 6 were treated to disrupt the collagen network, before retesting. Both of these types of modification are commonly observed in naturally degraded articular cartilage, and can therefore be argued to represent discrete parts of the degradation process [2,3,13]. Finally, a joint with focalised osteoarthritis was tested to examine the application of the reflection coefficient to the natural degradation process. Intact joints were taken from oxen within 24 hours of slaughter, wrapped in a 0.15 M saline soaked cloth and stored at -20⁰C. Prior to testing, the joints were thawed overnight in saline, according to the method of Broom and Flaschmann [16]. Ultrasound scans were obtained over the surface of normal cartilage-on-bone samples immersed in saline. All ultrasonic examinations were made at a fixed distance of approximately 3 mm using a φ3 mm, plane 10 MHz contact transducer (V129 Panametrics Inc., Massachusetts USA), sampling at 50 MHz and connected to a pulser/receiver as illustrated in Figure 2. This was in turn connected to an oscilloscope and a personal computer through an ADC200 analogue to digital converter (Pico Technology Limited, Cambridgeshire, UK). Surface and bone-end echoes were captured in each test and recorded using PicoScope software (Pico Technology Limited, Cambridgeshire, UK). After testing, the normal cartilage-on-bone samples were artificially degraded by either disrupting the collagen meshwork or removing a large portion of the proteoglycan content. To disrupt the collagen meshwork the intact patellae were immersed in 0.15 M phosphate-buffered saline containing 30 U. ml^"1 collagenase (Sigma C0773 protease free, Sydney, Australia) for 18 hours at 37°C. These were then blotted dry and returned to 0.15 M saline before re-testing.

In order to remove the proteoglycans, the normal intact patellae were treated for four hours in 0,1 mg.mf¹ of trypsin (from bovine pancreas T4665, Sigma-Aldrich, Australia) in phosphate buffered saline. Trypsin generally acts only to remove proteoglycans, but may attack collagen molecules which have been previously cleaved. This effect is likely to be minimal in the 4 hour degradation of a normal articular cartilage matrix. However, the lack of specificity of the action of trypsin on the general matrix of articular cartilage should be noted as a possible limitation in this study.

The osteoarthritic patellae contained a 7 mm diameter International Cartilage Repair Society grade 4 (exposed bone) [23] focal defect at the distal extremity of the patellar ridge. The remaining joint surface was visually normal. Ultrasound scans were taken at the edge of the defect and at intervals along the patellar ridge, measured by distance from the edge of the defect. After completion of the degradation programme, the joints were sectioned and frozen for histology, arresting the enzyme action. The histological quantification techniques are fully explained above. In summary, the proteoglycan content and distribution was measured by staining with Safranin-O, followed by absorbance profiling under monochromatic light source using a Nikon Labo-Phot polarized light microscope (PLM). PLM was also used without staining to detect and quantify the presence of an oriented collagen network, quantified by birefringence. Standard histological procedures were employed.

The captured ultrasound signals from the cartilage surface and osteochondral junction were processed and converted to the frequency domain using a Fast Fourier Transform routine in MATLAB (The MathWorks Inc, USA, version 7.0.4.352). Zero padding was employed to increase the apparent spectral resolution. Two parameters were used to determine the most appropriate frequencies for distinguishing normal from degraded articular cartilage, statistical significance of the change in reflection derived from the student's t-test, and the level of absolute overlap.

The t-test, while powerful, assumes a normal distribution in the samples and as such may be limited in its ability to elucidate the changes if the distribution of samples is non-normal. The overlap parameter was therefore used in addition to the statistical significance parameter to describe the probability of distinguishing a randomly selected sample as being either normal or degraded, without assuming a particular sample distribution. An overlap value closest to 1 is preferred as it represents a better discrimination between the sample groups.

In order to determine the frequency ranges that best describe changes to the proteoglycan and collagen content, the results from the student's t-test and overlapping test were obtained and the values were plotted in their respective graphical formats. The resulting frequency bands were then compared for the two parameters. In an attempt to account for amplitude changes due to the alignment of the ultrasound transducer with the sample, results were baselined against a set frequency. For clarity, we have used the notation Fo for the frequencies from the original data, and F₁ for the frequencies adjusted against the baseline frequency FBL- The student's t-test and the overlapping test were also coded in the MATLAB environment, using a band criterion of >0.5 MHz to reduce the influence of noise. The frequency profile of- the signal reflected from the acoustic mirror (highly pofished stainless steel) in saline is used as a reference. The peak reflection occurred at a frequency of 8.4 MHz, with a gradual decay with increasing and decreasing frequency. The bandwidth of interest for this study is limited to the 0-10 MHz range. The collagen meshwork was found to be reduced in the superficial area of the cartilage following the 18 hour treatment but the deeper zones did not appear to be affected. It was apparent that the proteoglycan content is almost completely depleted after the four hour treatment. FIGS 22 and 23 show the typical frequency profiles of the ultrasound echoes obtained from the surface and osteochondral junction respectively. Both sets of curves were characterised by a band of low reflection between 1 and 2.2 MHz, and a reflection peak between 7 and 8.4 MHz.

Qualitatively, the average frequency profile of the surface echo for proteoglycan-depleted samples, shown in FIG 22, showed distinct differences when compared to the normal samples. In this respect, it was characterised by a right-shifted and lower amplitude band of low reflection, and a more exaggerated 3.6 MHz peak. Further differences included a lower amplitude and left-shifted reflection peak at approximately 7.8 MHz and a more distinct trough at 4.4 MHz. The surface echoes from the collagen-disrupted samples showed a generally lower amplitude compared to the normal and proteoglycan-depleted samples. The profile was similarly proportioned to that of the normal samples, with an extended and more complex band in the 1-5 MHz range.

The frequency profiles of the reflected signals from the osteochondral junction, shown in FIG 23, were more tightly grouped than the surface echoes. When compared to the normal samples, the proteoglycan-depleted samples were characterised by a right-shifted band of low reflection, and a consistently lower amplitude signal in the 1 to 5.6

MHz range. A right-shifted and slightly higher amplitude reflection peak was further observed at 7.6 MHz. The collagen-disrupted samples were characterised by a distinct 3.2 MHz trough which was not observed in the normal samples. Also observed were a deeper 5 MHz trough and a lower and rϊght-shrfted reflection peak amplitude at approximately 7.4 MHz.

FIG. 24 shows the frequency profiles, baselined at the frequency corresponding to the reflection peak, for the data presented in

Figure 21. The frequency profiles at angles smaller than ±1.2° were tightly grouped, with a distinctly different profile at angles greater than ±1.2° at the same site.

Using results from the student's t-test and overlap test, the frequency reference values that showed the most change with degradation for both the surface and bone reflections were obtained and presented in Table 6. The alternate test scores for the corresponding "best" frequencies are also provided, ie, where the statistical significance test is used to find the best frequency reference value, F₀, the results of the overlap test for this frequency are also shown for comparison.

The baselined frequency profiles were analysed in the same way as the original profiles and are presented in Table 7 below TABLE 7

for the statistical significance test (P). For each case, the baselining frequency and alternate test scores are provided.

The results demonstrate that changes to the frequency profile characteristics of articular cartilage-on-bone, due to scattering, absorption and reflection correlate with the structural integrity of the matrix and the cohesion between the proteoglycans and the collagen meshwork. The inventors have been able to specify the frequency changes between the reflected ultrasound signals from normal intact cartilage-on-bone, and those from degraded and diseased samples.

The penetrating nature of ultrasound allows it to be more readily applied for detecting changes in deeper zones of articular cartilage than the more commonly used surface echo or mechanical indentation techniques. The surface echoes, shown in FIG.22, are most likely to give information and insight into the general bulk modulus and density of the matrix, as well as superficial damage such as fibrillation and microcrackiπg. The osteochondral junction echoes, for which the signal passes through the full depth of the cartilage matrix, is expected to allow more insight into the deeper zone changes to the collagen meshwork and proteoglycan macromolecules. As can be seen in FIG.23, however, the reflected signals from the osteochondral junction do not necessarily provide a more complex profile when compared to the cartilage surface echoes in FIG.22.

The amplitude of the ultrasound signal from articular cartilage is highly dependent on the orientation of the transducer with respect to the reflective surface, the strongest signal occurring when the transducer is perpendicular to the surface, with signal strength rapidly decreasing with increasing deviation. Although a guide was used to hold the ultrasound transducer in place and minimise this misalignment in our experiments, it can be assumed that this type of effect was not completely eliminated, thereby introducing some degree of error, it would be expected, however, that in each case the error was less than ±1° from the perpendicular.

An interesting finding was that the frequency profile remained consistent up to ±1.2°, after which an increased peak in the 3-5 MHz band was observed, as well as a left shifted overall peak. This effect may indicate a difference between the specular and diffuse reflection profiles, though further work will be required in this area to delineate the responses and relate them to structural characteristics. With a change in orientation of less than ±1.2°, the value . of the amplitude at any single frequency changed, but the overall pattern/profile of the reflected signal's frequency response remained the same. This observation suggests that certain frequency characteristics of the proteoglycan-depleted or collagen- disrupted samples could contribute to the differentiation of normal from degraded joint tissues only when the angle of incidence can be controlled within a +1.2° limit, The specularity of the reflected signals suggests that scattering within the matrix may play a major role in the profile of reflected signals.

Most qualitative changes in the frequency profile of the osteoarthritic sample's surface echo occurred at lower frequencies. It was seen that although an amplitude change occurred up to approximately 15 mm from the defect at high frequencies (6-10 MHz), the overall profile changed very little. The clear difference between the frequency of the major reflection peaks for normal and degraded samples in Figures 5 and 6 were not observed in either of the osteoarthritic sample profiles, indicating that the entire sample could be degraded, despite the fact that only a focal defect was visible.

In the frequency band between 1 and 5 MHz, the surface reflection profile changed considerably, resembling the more complex patterns of the trypsin- and collagenase-treated samples. The higher frequency band above 6 MHz changes very little in shape, despite an apparent amplitude change.

The frequency profiles from the osteochondral junction echoes show a more distinct qualitative change, particularly in the region within 5 mm of the defect, than the profile from the surface echo, most likely due to changes on the surface of the subchondral bone coupled with structural changes in the overlying cartilage matrix. These profiles appear similar to the collagenase-treated samples shown in FIG. 23, with more exaggerated reflection bands in the lower frequency range at approximately 3.2 and 5 MHz. The more subtle changes in the frequency profiles of the reflections from the osteochondral junctions of the trypsin- treated samples were not clearly observed in the profile for the osteoarthritic osteochondral junctions, relative to the normal samples. Of additional interest was the ultrasound reflection taken from within the defect (0 mm), which shows a complex pattern with multiple, sharp peaks and troughs due to the highly degraded state of the reflecting surface and the absence of the overlying cartilage matrix.

The ratios provided a generally consistent profile around an osteoarthritic defect. The surface reflection parameters based on collagen meshwork disruption peaked at 6-8 mm from the defect, before decreasing slightly and levelling out, The reflection parameters from the osteochondral junction, based on proteoglycan depletion dropped considerably in the first 5 mm before levelling out. Each of these parameters reached a plateau at approximately 10 mm from the edge of the defect.

A non-destructive, diffuse reflectance near infrared spectroscopic (DR-NIRS) approach is of use for determining the low-level (molecular) structural properties of the full depth articular cartilage matrix. DR-NIRS may be used to detect changes in proteoglycan quantity, using principal components analysis as the statistical basis of characterisation. A useful comparison is of the results from a directly applied probe to those obtained from a probe that was offset by 1.5 mm. The results show that this technique, particularly using the offset probe, can reliably (R²>0.9) distinguish normal intact cartilage from cartilage that has lost some of its proteoglycan content using vector normalising and second derivative pre- treatments. The ability of this technique is to gather information over the full depth of the matrix and its extension to the determination of collagen meshwork and subchondral bone integrity recommended, The use of DR- NIRS enables the probing of the deeper layers of cartilage, thus allowing the assessment of both the osteochondral junction and the advancing calcification front associated with osteoarthritis.

EXAMPLE 5

Due to the novelty of the application of DR-NIRS to articular cartilage, a number of preliminary experiments were required to establish the most suitable sampling technique and the areas of interest within a DR-NIRS spectrum relative to the proteoglycan content. Because of the high absorption of NIR signals by water, the most appropriate sampling geometry with potential for in vivo application is either with the probe positioned against the articular surface, or recessed with an air gap. To determine the optimum air gap distance between the probe tip and the articular surface, the probe was positioned on a reflectance standard (Spectralon, Labsphere USA) and adjusted to the position with the highest signal response (approximately 1.5 mm). This gave an estimate for the position of most efficient light collection from a diffuse reflecting sample. Eight normal patellae were tested using a 5 x 5 grid pattern

(n=200 samples) and a fibre optic probe with a specially manufactured adjustable gap setting. After testing the visually normal intact samples, a quantity of proteoglycans was removed form each patella by immersion in 0.1 mg.ml^'1 of trypsin (from bovine pancreas T4665, Sigma-Aldrich, Australia) dissolved in 0.15M phosphate buffered saline (PBS) at 37°C for four hours.

To investigate the level of penetration of NIR signals in cartilage and bone, a cartilage-on-bone specimen was tested and compared to the same cartilage removed from the bone and placed on the reflectance standard, which is an ideal diffuse reflector in the NIR region. As the cartilage was 1.6 mm thick, the test was repeated for two layers of cartilage (3.1 mm total thickness) to test the penetration into depths more representative of human tissue. Experimental data were collected using a Nicolet Nexus FT-

IR system with OMNIC v5.1 software (both ThermoNicolet, UK). The φ4 mm fibre optic probe was coupled to the FT-IR system via a Grasby SPECAC NIR fibre port accessory. The spectrum was acquired from the 4000 to 12500 cm^"1 region: The calibration of NIR spectra becomes quite complex when dealing with biological samples. This often results in detection limits as high as 0.5-1% absolute, even for major constituents. Further, Beer's law and Hooke's law break down for diffuse reflectance at high overtones/harmonics, leading to the need for multivariate analysis. An in-house multivariate analysis package combined with a commercial package (The Unscrambler v7.5, 1999, CAMINO ASA, Norway) was used to condition and analyse the spectral data. The data were smoothed using a low pass filter, Fourier differentiated and mean centred. Due to the noise resulting from the high absorbance in the regions of the spectrum below 7000 cm^"1 for the directly applied probe, only the 7400-10000 cm^'1 region was analysed. The fuli DR-NIRS spectrum was analysed for the air-gap probe.

Principal component analysis (PCA), principal component regression (PCR) and partial least squares regression (PLS) techniques were then used to model the difference between normal and proteoglycan depleted samples. Principal component analysis (PCA) is a matrix method that was used to manipulate the data set to probe for relationships between variables. This method reduces the dimensionality of the original data without loss of information and with an added benefit of reduction in noise. PCR was employed to determine the number of useful principal component, preventing the over-fitting of the data by PLS. The PLS then used these principal components to create a more accurate model. This investigation used categorical regression assignments, using a value of 1 for the normal samples and -1 for the proteoglycan depleted samples.

Figure 26 shows the spectral reflectance profiles of articular cartilage-on-bone using directly applied and offset probes respectively. Bands of very low reflection, and therefore low signal to noise ratio, are apparent in the profiles of the directly applied probe at 4000-5300 and 6500-7150 cm^"1. Consequently, the data from the directly applied probe was only analysed in the region above 7300 cm^"1.

The eigenvectors for the PCA are presented in Figures 27 and 28. PCR and PLS R² values for reflectance and its 2nd Derivative are shown for raw and vector normalised data in Figures 5a and b respectively. These results show that, although the PLS correlations appear strong, the PCR results were relatively weak, indicating that the PLS results may be including noise in its prediction. The R² results take 'longer' to reach a good value using the directly applied probe, compared to the air-gap probe. The inventors have identified the ability of the DR-NIRS method to detect proteoglycan loss in articular cartilage-on-bone using both a directly applied probe and an offset probe, separated from the articular surface by an air-gap.

Based on the geometry of the sample, the probe measured a combination of the effects of scattering by the cartilage ultrastructure, absorbance, and transmittance through to the bone.

The eigenvectors for the PCA showed similar trends for both the directly applied and offset arrangements. The first few components appear to consist of variation relating to the absolute offset in the spectra, while the middle (PC's 3-9) shows more interesting peaks, indicating frequency bands of interest. For the directly applied probe these bands occur in the 7500-9000 cm^'1 region. Higher components generally appear noisy and may be of .little value to the analysis. For the offset probe, the principal components show a number of peaks throughout the spectrum, particularly at approximately 5100 cm^'1. This peak is the most apparent feature of PC 5 which, as shown in Figure 29, is capable of separating normal from proteoglycan depleted tissue on its own.

The inventors have shown that normalised and second derivative DR-NIRS reflectance can be used to reliably discriminate between normal and proteoglycan-depleted cartilage.

While the preferred testing device combines all the described apparatus and methods, the inventors believe the rebound method and device, the ultrasound method and device, both using ratios and frequency profiles, the surface stretch method an device and the DR- NIR method and device are all separately novel, non-obvious and industrially applicable. They expressly reserve the right to select any one or any combination of two or more separate methods and devices as patentable inventions in their own right. As an example FIG 30 shows a schematic view of a testing device with a single ultrasound transducer 81 and an array of NIR fibreoptics having sending outlet s82 and receiving inlet 83s scattered aorunfd the base.

Although research has been conducted principally in oxen, it is expected inter-species variation will be minimal. However, even if that variation is significant, the present method and device may be simply calibrated and applied to a new species including man, by straight forward processes readily apparent to a skilled addressee. It is intended that "characterising the health" includes assessing the general condition of a soft tissue.

The benefits of the present invention are numerous and significant. A surgeon may reliably assess the health status of a soft tissue, especially cartilage. This will lead to better, more focussed surgical interventions and beneficial outcomes. However the advantages are not so limited. The invention may be used to assess the quality of tissue engineered cartilage for example, prior to its recruitment in a therapeutic environment.

Reference to any prior art documentation in this specification is not an acknowledgement that such documentation forms part of the common general knowledge in Australia or any other country.

References:

\ 1. Brismar, B. H., Wredmark, T., Movin, T., Leandβrsson, J., and Svensson, O. (2002) Observer reliability in the arthroscopic classification of osteoarthritis of the knee. J Bone Joint Surg Br ^■: j 84(1), 42-47

2. Appleyαrd, R. C, Ghosh, P., αndSwαin, M. V. (1999)

Biomechαnicαl, histological and immunohistological studies of patellar cartilage in an ovine model of osteoarthritis induced by lateral meniscectomy. Osteoarthr Cartilage 7, 281-294 3. Arokosfci, J₁ Jurvelin, J., Kiviranta, L, Tammi, M., and Helminen, H. J. (1994) Softening of the lateral condyle articular cartilage in the canine knee joint after long distance (up to 40 km/day) running training lasting one year. Int J Sports Med 15, 254-260 4. Lyyra, T., Arokoski, J. P. A., Oksala, N, Vihko, A., Hyttinen, M., Jurvelin, J. S., and Kiviranta, I. (1999) Experimental validation of arthroscopic cartilage stiffness measurement using enzymatically degraded cartilage samples. Phys Med Biol 44, 525-535 5. Tδyrάs, J., Rieppo, J., Nieminen, M. T., Helminen, H. J., and

Jurvelin, J. S. (1999) Characterisation of enzymatically induced degradation of articular cartilage using high frequency ultrasound. Phys Med Biol 44, 2723-2733

6. Zeytin, S., Konduk, B. A., Ipek, M., Bindal, C, and Ucisik, A. H. (2002) An evaluation of human articular cartilage on femoral head. Mat Sci Eng 20, 219-222

7. Niederauer, M. Q., Cristante, S., Niederauer, G. M., Wilkes, R. P., Singh, S. M., Messina, D. F., Walter, M. A., Boyan, B. D., DeLee, J. C, and Niederauer, G. (1998) A novel instrument for quantitatively measuring the stiffness of articular cartilage.

Transactions of the Orthopaedic Research Society 23, 905

8. Lyyra, T., Jurvelin, J, Pitlcanen, P., Vάάtάinen, U., and Kiviranta, I. (1995) Indentation instrument for the measurement of cartilage stiffness under arthroscopic control, Phys Med Biol 17(5), 395-399

9. Laasanen, M,, Tδyrάs, J., Hirvonen, J., Saarakkala, S., Korhonen, R. K., Nieminen, M. T., Kiviranta, I, andJurvelin, J. S. (2002) Novel mechano-acoustic technique and instrument for diagnosis of cartilage degeneration. Physiol Meas 23(3), 491- 503

10. Frank, E. K₁ Salant, E. P., and Grodzinsky, A. (1993) Probe, system and method for detecting cartilage degeneration. In. US Patent, MT, USA

11. Athanasiou, K., Constantinides, G., and Lanctot, D. R. (1996) Arthroscopic cartilage evaluator and method for using the same. In. US Patent, USA

12. Buschmann, M. D., Guardo, R., Garo, M., Guyader, P. L., and Savard, P. (1998) Medical Apparatus for the diagnosis of cartilage degeneration via a spatial mapping or compression- induced electrical potentials. In. US Patent, USA

13. Saarakkala, S., Laasanen, M. S., Jurvelin, J., Tόrrόnen, K., Lammi, M. J., Lappalainen, R., and Tδyrάs, J. (2003) Ultrasound indentation of normal and spontaneously degenerated bovine articular cartilage. Osteoarthr Cartilage 11, 697-705

14. Nieminen, H. J., Tδyrάs, J., Rieppo, J., Nieminen, M. T., Hirvonen, J., Korhonen, R., and Jurvelin, J. S. (2002) Real-time ultrasound analysis of articular cartilage degradation in vitro.

Ultrasound Med Biol 28(4), 519-525

15. Suh, J., Youn, I, and Fu, F. H. (2001) An in situ calibration of an ultrasound transducer: a potential application for an ultrasonic indentation test of articular cartilage. J Biomech 34, 1347-1353

16. Broom, N. D., and Flachsmann, R. (2003) Physical indicators of cartilage health: the relevance of compliance, thickness, swelling and fibrillar texture. J Anat 202(6), 481-494 17. Kempson, G. E. (1982) Relationship between the tensile properties of articular cartilage from the human knee and age. Ann Rheum Dis 41, 508-5 Jl

18. Montella, A., Manunta, A., Espa, E., Gasparini, G., De Santis, E., and Gulisano, M. (1992) Human articular cartilage in osteoarthritis. I. The matrix. Transmission electron microscope study. Int J Anat Embryo 97, 1-12

19. Broom, N.D., Ngo. T., and Tham, E. (2006) Tranversing the intact/fibnllated joint surface: a biomechanics! interpretation. J Anat 2006, 55-67.

20. Benninghoff, A. (1925) Form und bau der Gelenknorpel in ihren Bezeihungβn zurFunktion. Z Zellforsch Mikrosk Anat 2, 783-825

21. Stok, K., and Oioyede, A. (2003) A qualitative analysis of crack propagation in articular cartilage at varyting rates of tensile loading. Connect Tissue Res 44, 109-120

22. . Dudα G., et αl. (2004), A new device to detect early cartilage degradation. American Journal of Sports Medicine, 32, 693- 698.

23. International Cartilage Repair Society (2000). ICRS cartilage injury evaluation package in ICRS 2000 standards workshop,

Switzerland.

24. Brown, C P., Crawford, R. W., Oioyede, A. (2007). Indentation stiffness does not discriminate between normal and degraded articular cartilage. Clinical Biomechanics, 22, 843-848. 25. Moody, H.R., et al (2006). In vitro degradation of articular cartilage - Does trypsin treatment produce consistent results ? Journal of Anatomy, 209, 259-267.

Claims

1. A testing device for use in characterising the health of soft tissue, the testing device comprising: an indenter for indenting a surface of the soft tissue; a monitor for monitoring a change or changes in the soft tissue; a processor for receiving, storing and analysing data on the change or changes; wherein; the change or changes comprises or includes rebound of the surface after withdrawal of the indenter,

2. The testing device of claim 1 wherein the soft tissue is cartilage and the processor is programmed to identify the rebound strain calculated as the distance the surface recovers from loading at a given time after withdrawal of the indenter.

3. The testing device of claim 2 wherein the processor is programmed to further provide a recovery index calculated with respect to relative deformations nd/or quotient of dividing the rebound

strain by the maximum indentation stress,

where Jl is the recovery index. ε,- the "short- term" time-dependent elastic rebound strain, and εo is original deformation, or indentation strain σ is maximum indentation reaction stress.

4. The testing device of Claim 1 wherein the indenter has a diameter in the range of around 0.5mm to around 5mm and is adapted for indentation for over a period up to 5 seconds; and further including one or more of a fibre optic monitoring component, an ultrasonic monitoring component, an NIR monitoring component and a linear variable displacement transducer for determining the travel of the indenter and/or monitoring the rebound of the surface.

5. The testing device of Claim 2 where the indenter is advanced until it reaches either a preset stress or reaches a selected percentage of strain, preferably in the range of 5% to 30%.

6. The testing device of Claim 2 further comprising an ultrasonic apparatus for emitting ultrasonic waves into the soft tissue and detecting ultrasonic echoes.

7. The testing device of Claim 6 wherein the processor is further programmed to determine a ratio of reflection coefficients from the surface and a related osteochondral junction for use in characterising the health of the tissue.

8. The testing device of Claim 6 wherein the processor is programmed to analyse the frequency profile of ultrasound echoes for use in characterising the health of the tissue.

9. The testing device of Claim 6 wherein the processor analyses ultrasound echoes up to 25 MHz.

10. The testing device of Claim 2 further comprising a near infrared spectroscope for conducting diffuse reflectance near infrared spectroscopy of the soft tissue and the processor is programmed to analyse results of the spectroscopy for use in characterising the health of the soft tissue.

11. The testing device of Claim 1 mounted on or mountable to an arthroscope.

12. The testing device of Claim 2 further including positional hardware and/or software to produce a map of the soft tissue highlighting unhealthy areas.

13. A method of assessing the health of soft tissue, the method comprising:

Indenting a surface of the soft tissue; monitoring rebound of the surface; and analysing data on the rebound to provide an indicator of the health of the soft tissue.

14. The method of Claim 13 wherein indenting a surface of the soft tissue comprises: indenting a surface of cartilage; and analysing the data includes determining rebound strain calculated as the distance the surface recovered from loading at a given time after indenting.

15. The method of Claim 13 further comprising the step of determining a recovery index calculated with respect to relative deformations yi_υ

and/or quotient of dividing the rebound strain by the maximum indentation stress, M_c

: where 91 is the recovery index. E_\ the "short- term" time-dependent elastic rebound strain, and εo is original deformation, or indentation strain σ is maximum indentation reaction stress.

16. The method of Claim 13 further comprising the step of: directing ultrasonic waves into the soft tissue; detecting ultrasound echoes from the soft tissue; and analysing data on the echoes to provide an indicator, at least in part, of the health of the soft tissue.

17. The method of Claim 16 wherein analysing data on the echoes includes: determining a ratio of reflection coefficients from the surface and a related osteochondral junction.

18. The method of Claim 16 wherein analysing data on the echoes includes: determining a frequency profile of ultrasound echoes from the soft tissue; and analysing the frequency profile to provide an indicator, at least in part, of the health of the soft tissue.

19. The method of Claim 16 further comprising the step of : orientating an ultrasonic probe within a range of less than ±1.2 degrees of perpendicular.

20. The method of Claim 13 further comprising the steps of: conducting diffuse reflectance near infrared spectroscopic examination of the soft tissue; and analysing data from the spectroscopic examination to provide an indicator, at least in part, of the health of the soft tissue.

21. A method of assessing the health of cartilage, the method comprising :

Indenting a surface of the soft tissue; monitoring rebound of the surface; and analysing data on the rebound to provide an indicator of the health of the soft tissue; and one or more of; directing ultrasonic waves into the soft tissue; detecting ultrasound echoes from the soft tissue; and analysing data on the echoes to provide an indicator, at least in part, of the health of the soft tissue; or conducting diffuse reflectance near infrared spectroscopic examination of the soft tissue; and analysing data from the spectroscopic examination to provide an indicator, at least in part, of the health of the soft tissue; and assessing stretch characteristics of the surface when indented.

22 A testing device for use in characterising the health of soft tissue, the testing device comprising; an indenter for indenting a surface of the soft tissue; a monitor for monitoring a change or changes in the soft tissue; wherein: the change or changes comprises or includes rebound of the surface after withdrawal of the indenter; and one or more of; an ultrasonic apparatus for emitting ultrasonic waves and detecting ultrasonic echoes; or a near infrared spectroscope for conducting diffuse reflectance near infrared spectroscopy of the soft tissue and the processor is programmed to analyse results of the spectroscopy to provide an indicator of the health of the soft tissue; or a device for assessing the surface stretch of the surface during indentation ; and a processor for receiving, storing and analysing data on the change or changes and data received from the ultrasonic apparatus and/or the spectroscope and/or the device for assessing surface stretch.