WO2005004726A2

WO2005004726A2 - Method, apparatus and system for diagnosing tumors using velocity spectrum

Info

Publication number: WO2005004726A2
Application number: PCT/IL2004/000617
Authority: WO
Inventors: Solange Akselrod; Ron Tepper; Yodfat Shaharabany
Original assignee: Ramot at Tel Aviv University Ltd
Current assignee: Ramot at Tel Aviv University Ltd
Priority date: 2003-07-09
Filing date: 2004-07-08
Publication date: 2005-01-20
Anticipated expiration: 2006-01-09
Also published as: WO2005004726A3

Abstract

A method of analyzing a Doppler flow image of a region containing a tumor, the method comprising: the steps of representing the Doppler flow image as a three-dimensional flow representation (12) and calculating at least one parameter characterizing a velocity spectrum of the three-dimensional flow representation (14), so as to determine malignancy likelihood of the tumor (16); thereby analyzing the Doppler flow image.

Description

METHOD, APPARATUS AND SYSTEM FOR DIAGNOSING TUMORS USING VELOCITY SPECTRUM

FIELD AND BACKGROUND OF THE INVENTION The present invention relates to tumor diagnosis and, more particularly, to a method, apparatus and system employing velocity spectrum analysis of Doppler flow images for characterizing and diagnosing tumors. Cancer is a major cause of death in the modern world. Effective treatment of cancer is most readily accomplished following early detection of malignant tumors. Female gynecological cancers include cervical cancer of the uterus, endometrial cancer of the uterus, ovarian cancer, choriocarcinoma, etc. Among them, generation of cervical cancer of the uterus has been constantly reduced year by year, and the choriocarcinoma can be now expected to have high therapeutic effect. However, endometrial cancer of the uterus and ovarian cancer have many problems in early diagnosis, therapy, prognosis, monitoring, etc. Ovarian cancer is generated generally in post-menopausal subjects, and is increasing year by year from the age of about 50 with a peak in the eighths decade. Ovarian cancer is the second most common cancer of the female reproductive organs and the fourth most common cause of cancer deaths among women, because its diagnosis is typically possible when the disease has progressed to a late stage of development. Approximately 75 % of women diagnosed with such cancers are already at the high-stage (III and IV) of the disease at their initial diagnosis. Due to the high percentage of high-stage initial detections of the disease, neither prognosis nor five year survival have greatly improved for these patients over the past two decades. Therefore, the challenge remains to develop new methods to improve early diagnosis and reduce the percentage of high-stage initial diagnoses. In order for malignant cells to grow, spread or metastasize, they must have the capacity to invade local host tissue, dissociate or shed from the primary tumor, enter and survive in the bloodstream, implant by invasion into the surface of the target organ and establish an environment conducive for new colony growth (including the induction of angiogenic and growth factors). A variety of imaging techniques have been used to detect and diagnose diseases. Representative examples include X-ray imaging, magnetic resonance imaging (MRI), nuclear imaging, ultrasonic imaging and the like. Images obtained through X-rays technique reflect the attenuation coefficient of the object being imaged. Contrast agents such as barium or iodine are used to attenuate or block X-rays such that the contrast between various structures is increased. X-ray imaging, however, is less favored over other techniques because of the cumulative nature of various deleterious effects caused by ionization. MRI utilizes a quantum mechanical phenomenon named nuclear magnetic resonance. Although MRI is capable of providing high quality images of internal organs, this technique has various drawbacks such as expense and the fact that it typically cannot be conducted as a portable examination. In addition, MRI is not available at many medical centers. In nuclear imaging, radionuclides such as technetium labeled compounds are injected into the patient and images thereof are obtained by gamma cameras. Nuclear imaging, however, suffers from poor spatial resolution and exposes the patient to the deleterious effects of radiation. Furthermore, the handling and disposal of radionuclides is problematic. Ultrasonic imaging is frequently used for a variety of diagnostic procedures because it is non-invasive, low cost, and has a fast response time. These qualities are especially valuable in medical fields where an added benefit is reducing or eliminating a patient's exposure to radiation. Hence, unlike nuclear and X-rays imaging techniques, ultrasound does not expose the patient to the harmful effects of ionizing radiation. Moreover, unlike MRI, ultrasound is relatively inexpensive and can be conducted as a portable examination. The first recorded use of ultrasound as an imaging technique was by Dr. Karl Dussik, an Austrian Psychiatrist who tried to locate brain tumors using ultrasound. Dr. Dussik used two opposed probes, a first probe for transmitting ultrasound waves and a second probe for receiving them. The received signal was used to visualize the cerebral structure by measuring the ultrasound beam attenuation. Generally, ultrasonic imaging is accomplished by first generating and directing an ultrasonic wave into a media under investigation, then observing any resulting waves that are reflected back from dissimilar tissues and tissue boundaries within the media under investigation. The resulting waves are received as signals. These received signals are then post-processed and imaged (e.g., on a display device) by plotting a spot whose intensity is proportional to the amplitude of a reflected wave from a given location. The location of a particular spot in the image is based upon a known transmission and re-radiation rate after an ultrasonic wave is pulsed into the media under investigation. Doppler ultrasound is based upon the well-known Doppler shift phenomenon, occurring when waves, such as ultrasound waves, are generated by, or reflected off, a moving object. In Doppler ultrasonic imaging, when the object reflecting the ultrasound waves is moving, it changes the frequency of the echoes, creating a higher frequency if it is moving toward the ultrasound probe and a lower frequency if the object is moving away from the ultrasound probe. The shift in frequency is directly proportional to the object velocity component parallel to the direction of the ultrasound beam. The frequency shift is the same for any object moving at a given velocity, whereas the amplitude of the detected signal is a function of the acoustic reflectivity of the moving object reflecting the ultrasound. Pulse Doppler ultrasound systems commonly produce a spectrogram of the detected return signal frequency as a function of time in a particular sample volume. Being representative of the object velocity, the spectrogram provides the physician with velocity information supplementing the ultrasonic image. Typically, Doppler ultrasound is used to measure the flow rate in blood vessels or in the heart. Blood flow modifications occur in the female reproductive organs in many specific conditions, normal or pathological. Initial studies in primates showed that a normal blood flow profile of the ovary was correlated with the maturation and selection of the dominant follicle. The introduction of Doppler imaging techniques has made it possible to study uterine and ovarian perfusion also in humans. Angiogenesis, the process of forming new blood vessels from pre-existing blood vessels, is believed to play an important role in tumorigenesis. Doppler imaging has demonstrated that increased angiogenesis plays an important role in the tumorigenesis of ovarian, endometrial and cervical malignancies. Doppler imaging offers a noninvasive, in vivo method for assessing tumor angiogenesis that can be repeated when necessary. The development of an adequate vascular network and blood nourishment is crucial to the growth and development of a cancer as well as its metastasis. Compared to the normal vessel architecture, blood vessels located at the advancing cancer front lack muscular coating and they are consist mostly of endothelial lining and may contain tumor cells. Because of this characteristic architecture of the tumor blood vessels, tumor vascularization can be analyzed in terms of vessel location, vessel arrangement and Doppler waveform signal features, such as shape and resistance to blood flow. Vascular resistance to blood flow had been one of the major features in the assessment of tumor vascular characteristics. Blood vessels in malignant adnexal lesions show lower resistance to blood flows that in benign adnexal masses. Two major clinical parameters, derived from Doppler flow waveform, are commonly used for differentiation between benign and malignant adnexal tumors: the resistance index (RI) and pulsatility index (PI). An additional measure of tumor malignancy is the decrease of the Doppler waveform from systole to diastole, which is typically characterized by an exponential decay constant. For example, with respect to RI, it has been found that when RI values obtained from a tumor are within a certain discriminatory zone (typically below 0.4) there is a greater likelihood that the tumor is malignant. Yet, there is still a considerable range in which RI readings are inconclusive. Similar evidence where found for the aforementioned PI and exponential decay constant. Thus, prior art Doppler flow parameters have limited differentiation capability, in particular between benign and malignant ovarian masses. There is thus a widely recognized need for, and it would be highly advantageous to have a method, apparatus and system for characterizing and diagnosing tumors, devoid of the above limitations.

SUMMARY OF THE INVENTION According to one aspect of the present invention there is provided a method of analyzing a Doppler flow image of a region containing a tumor, the method comprising: (a) representing the Doppler flow image as a three-dimensional flow representation; and (b) calculating at least one parameter characterizing a velocity spectrum of the three-dimensional flow representation, so as to determine malignancy likelihood of the tumor; thereby analyzing the Doppler flow image. According to another aspect of the present invention there is provided a method of determining malignancy likelihood of a tumor, the method comprising: (a) imaging a region surrounding the tumor using a Doppler ultrasonic device, thereby providing a Doppler flow image; (b) representing the Doppler flow image as a three- dimensional flow representation; (c) calculating at least one parameter characterizing a velocity spectrum of the three-dimensional flow representation; and (d) using the at least one parameter for determining the malignancy likelihood of the tumor. According to further features in preferred embodiments of the invention described below, the method further comprises smoothing the Doppler flow image. According to still further features in the described preferred embodiments the smoothing of the Doppler flow image, is by a median filter. According to still further features in the described preferred embodiments the method further comprises defining a plurality of velocity ranges of the velocity spectrum. According to yet another aspect of the present invention there is provided a software apparatus for analyzing a Doppler flow image of a region containing a tumor, the apparatus comprising: (a) a representation unit, for representing the Doppler flow image as a three-dimensional flow representation; and (b) a parameters calculator, for calculating at least one parameter characterizing a velocity spectrum of the three- dimensional flow representation; and (c) a determinator for determining malignancy likelihood of the tumor, using the at least one parameter. According to further features in preferred embodiments of the invention described below, the apparatus further comprises a smoother, for smoothing the Doppler flow image. According to still further features in the described preferred embodiments the smoother comprises a median filter. According to still further features in the described preferred embodiments the median filter has a fixed window size. According to still further features in the described preferred embodiments the median filter has a variable window size. According to still further features in the described preferred embodiments the apparatus further comprises a velocity range definer, for defining a plurality of velocity ranges of the velocity spectrum. According to still another aspect of the present invention there is provided a system for determining malignancy likelihood of a tumor, the system comprising: (a) a Doppler ultrasonic device, for imaging a region surrounding the tumor, thereby to provide a Doppler flow image; and (b) a software apparatus, communicating with the

Doppler ultrasonic device and capable of analyzing the Doppler flow image, the software apparatus having (i) a representation unit, for representing the Doppler flow image as a three-dimensional flow representation; (ii) a parameters calculator, for calculating at least one parameter characterizing a velocity spectrum of the three- dimensional flow representation; and (iii) a determinator, for determining the malignancy likelihood of the tumor, using the at least one parameter. According to further features in preferred embodiments of the invention described below, the three-dimensional flow representation comprises at least one of: velocity information, time information and intensity information. According to still further features in the described preferred embodiments the intensity information represents flowing volume. According to still further features in the described preferred embodiments the intensity information represents number of flowing cells. According to still further features in the described preferred embodiments the intensity information is characterized by a gray-level measure. According to still further features in the described preferred embodiments the at least one parameter comprises velocity distribution slope. According to still further features in the described preferred embodiments the velocity distribution slope is selected from the group consisting of an end-diastolic velocity distribution slope, DVD_S, and a systolic velocity distribution slope, SVDJS. According to still further features in the described preferred embodiments the end-diastolic and the systolic velocity distribution slopes, are each independently calculated using a linear fit. According to still further features in the described preferred embodiments the at least one parameter comprises a minimal average velocity. According to still further features in the described preferred embodiments the calculation of the minimal average velocity comprises: (i) averaging the velocity spectrum, so as to obtain an average velocity function; and (ii) locating a minimal value of the average velocity function. According to still further features in the described preferred embodiments the at least one parameter comprises an average velocity difference. According to still further features in the described preferred embodiments the calculation of the average velocity difference comprises: (i) selecting a plurality of end-diastolic instants of the three-dimensional flow representation; (ii) for each end- diastolic instant of the end-diastolic instants calculating a difference between a maximal velocity and an average velocity, thereby providing a plurality of velocity differences; and (iii) averaging the plurality of velocity differences, thereby providing the average velocity difference. According to still further features in the described preferred embodiments the plurality of the velocity ranges comprises a high velocity range and a low velocity range, respectively defined above and below a velocity threshold. According to still further features in the described preferred embodiments the velocity threshold is predetermined. According to still further features in the described preferred embodiments the velocity threshold substantially equals a minimal end-diastolic velocity of the velocity spectrum. According to still further features in the described preferred embodiments the at least one parameter comprises a high-to-low velocity ratio. According to still further features in the described preferred embodiments the calculation of the high-to-low velocity ratio comprises: (i) obtaining a velocity distribution; and (ii) calculating a ratio between a first area under the velocity distribution and a second area under the velocity distribution, the first and the second areas respectively corresponding to the high and the low velocity ranges; thereby calculating the high-to-low velocity ratio. According to still further features in the described preferred embodiments the velocity distribution is obtained by averaging the velocity spectrum. According to still further features in the described preferred embodiments the at least one parameter comprises an angle parameter. According to still further features in the described preferred embodiments the calculation of the angle parameter comprises: (i) obtaining a velocity distribution; and (ii) calculating an angle between a first line representing the low velocity range and a second line representing the high velocity range; thereby calculating the angle parameter. According to still further features in the described preferred embodiments the velocity distribution is characterized by maximal intensity over the three-dimensional flow representation. According to still further features in the described preferred embodiments the first and the second lines, are each independently obtained by a linear fit. According to still further features in the described preferred embodiments the at least one parameter comprises a resistance index. According to still further features in the described preferred embodiments the at least one parameter comprises pulsatility index. According to still further features in the described preferred embodiments the at least one parameter comprises a decay constant. According to still further features in the described preferred embodiments the at least one parameter comprises a weighted decay constant. According to still further features in the described preferred embodiments the region comprises a pelvis, adnexa uteri, a uterus, an ovary, a breast, a prostate, a hepatic artery, a liver and the like.

The present invention successfully addresses the shortcomings of the presently known configurations by providing a method, apparatus and system employing velocity spectrum analysis of Doppler flow images for characterizing and diagnosing tumors.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. In case of conflict, the patent specification, including definitions, will control. In addition, the materials, methods and examples are illustrative only and not intended to be limiting. Implementation of the method and system of the present invention involves performing or completing selected tasks or steps manually, automatically, or a combination thereof. Moreover, according to actual instrumentation and equipment of preferred embodiments of the method and system of the present invention, several selected steps could be implemented by hardware or by software on any operating system of any firmware or a combination thereof. For example, as hardware, selected steps of the invention could be implemented as a chip or a circuit. As software, selected steps of the invention could be implemented as a plurality of software instructions being executed by a computer using any suitable operating system. In any case, selected steps of the method and system of the invention could be described as being performed by a data processor, such as a computing platform for executing a plurality of instructions.

BRIEF DESCRIPTION OF THE DRAWINGS The invention is herein described, by way of example only, with reference to the accompanying drawings. With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of the preferred embodiments of the present invention only, and are presented in the cause of providing what is believed to be the most useful and readily understood description of the principles and conceptual aspects of the invention. In this regard, no attempt is made to show structural details of the invention in more detail than is necessary for a fundamental understanding of the invention, the description taken with the drawings making apparent to those skilled in the art how the several forms of the invention may be embodied in practice. In the drawings: FIG. 1 is a flowchart diagram illustrating a method of analyzing a Doppler flow image, according to preferred embodiments of the present invention; FIG. 2 is a is a schematic illustration of an apparatus for analyzing a Doppler flow image, according to preferred embodiments of the present invention; FIG. 3 is a schematic illustration of system for determining malignancy likelihood of a tumor, according to a preferred embodiment of the present invention; FIGs. 4a-b show contours of a single wave which were automatically extracted before (Figure 4a) and after (Figure 4b) a smoothing procedure, according to a preferred embodiment of the present invention; FIG. 5 shows a Doppler flow image, in which two velocity ranges are defined, according to a preferred embodiment of the present invention; FIGs. 6a-b show probability as a function of a velocity measured in arbitrary units, according to a preferred embodiment of the present invention; FIG. 7a shows a velocity spectrum corresponding to a selection of high velocities, according to a preferred embodiment of the present invention; FIG. 7b shows a velocity distribution which is characterized by a maximal probability along the envelope of the velocity spectrum shown in Figure 7a, according to a preferred embodiment of the present invention; FIG. 8 is a demonstration of a calculation of minimal average velocity and average velocity difference at the end-diastolic time instants, according to a preferred embodiment of the present invention; FIG. 9 shows a spectrum of velocities as a function of time, in which the low velocity range is highlighted at the low part of the spectrum, according to a preferred embodiment of the present invention; FIGs. lOa-b is a demonstration of a calculation of an end-diastolic velocity distribution slope, DVDJS, for a subject having benign ovarian tumor; FIGs. lla-b is a demonstration of a calculation of an end-diastolic velocity distribution slope, DVDJS, for a subject having malignant ovarian tumors; FIGs. 12a-b show Receiver Operating Characteristic (ROC) curves of resistance index (Figure 12a) and a weighted decay constant, τ, (Figure 12b), calculated in accordance to a preferred embodiment of the present invention; FIG. 12c shows values of conventional resistance index (RI) versus the weighted decay constant, τ, calculated in accordance with preferred embodiments of the presents invention, for benign masses (green circles) and malignant tumors (red circles); FIGs. 13a-d show ROC curves of average velocity difference (Figure 13a), minimal average velocity (Figure 13b), β (Figure 13 c) and high- to-low velocity ratio (Figure 13d); FIGs. 14a-d show relations between several parameters: Figure 14a shows resistance index versus high-to-low velocity ratio, Figure 14b shows the average velocity difference versus the high-to-low velocity ratio, Figure 14c shows the minimal average velocity versus the high-to-low velocity ratio, and Figure 14a shows resistance index versus β; FIGs. 15a-b show the ROC curves of DVD_S (Figure 15a) and SVDJS (Figure

15b), according to a preferred embodiment of the present invention; FIG. 16a shows relations between RI and DVDJS, according to a preferred embodiment of the present invention; FIG. 16b shows relations between the average velocity difference and DVDJS, according to a preferred embodiment of the present invention; FIG. 16c shows on a three-dimensional plot, relations between the weighted decay constant, RI and DVDJS, according to a preferred embodiment of the present invention; and FIGs. 17a-b show malignancy classification using conventional resistance index (Figure 17a) and DVDJS (Figure 17b), according to a preferred embodiment of the present invention.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention is of a method, apparatus and system employing velocity spectrum analysis of Doppler flow images, which can be used for characterizing and diagnosing tumors. Specifically, the present invention can be used to determine malignancy likelihood of a tumor present in, for example, a uterus, an ovary, a breast, a prostate, a liver and the like. The principles and operation of a method, apparatus and system for velocity spectrum analysis of Doppler flow images according to the present invention may be better understood with reference to the drawings and accompanying descriptions. Before explaining at least one embodiment of the invention in detail, it is to be understood that the invention is not limited in its application to the details of construction and the arrangement of the components set forth in the following description or illustrated in the drawings. The invention is capable of other embodiments or of being practiced or carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein is for the purpose of description and should not be regarded as limiting. According to one aspect of the present invention there is provided a method of analyzing a Doppler flow image of a region containing a tumor. The Doppler flow image can be obtained by imaging the region using any Doppler ultrasonic device known in the art. The type of device and imaging procedure depends on the type of tumor and the imaged region. Representative examples for ultrasound imaging include, without limitation, transvaginal (endovaginal) ultrasound imaging, transrectal ultrasound imaging, endoscopic ultrasound imaging and non-invasive ultrasound imaging. Representative examples of regions which can be imaged include, without limitation a pelvis, adnexa uteri, a uterus, an ovary, a breast, a prostate, a hepatic artery and a liver. Referring now to the drawings, Figure 1 is a flowchart diagram illustrating the method according to preferred embodiments of the present invention. Hence, in a first and optional step, shown in Block 10 of Figure 1, the region surrounding the tumor is imaged, as stated, using a Doppler ultrasonic device so as to provide the Doppler flow image. In a second step of the method, shown in Block 12, the Doppler flow image is represented as a three-dimensional flow representation. As used herein, "three-dimensional flow representation" refers to any flow image containing three-dimensional information characterizing the flow. The term "three-dimensional flow representation" should not be confused with the known term "three-dimensional image," which is generally used in the context of an image having three spatial dimensions (e.g., length, width and height, commonly referred to as x-y-z dimensions). It is to be understood, however, that spatial images are not excluded from the scope of the present invention, provided that such images are supplemented by the three-dimensional flow representation. For example, the three-dimensional flow representation can be supplemented by a two- or a three-dimensional image generated by any conventional imaging device, such as, but not limited to, an ultrasonic or MRI device. The implementation of Block 12 is not limited. For example, in one embodiment, the three-dimensional information which characterizes the flow can comprise velocity information, time information and intensity information. Thus, in this embodiment, "three-dimensional flow representation" refers to a "velocity-time- intensity" representation. As can be understood by one of ordinary skill in the art, the velocity information typically represents the flow rate of the blood, the time information typically represents the instant in which the velocity or flow rate was measured, and the intensity information typically represents the amount of blood flowing through the blood vessel (e.g., volume or number of cells). All the above information can be measured using any Doppler ultrasonic device, whereby the velocity information is typically represented numerically in units of centimeters per second, the time information is typically represented numerically in units of milliseconds and the intensity information is typically characterized by a gray- level measure, which can be represented numerically (e.g. , using 256 gray-level units). The three-dimensional flow representation can have many graphical realizations. Representative examples include, without limitation: (i) a surface defined by a velocity axis, a time axis and an intensity axis; (ii) a plurality of velocity distributions in the velocity-intensity plane, each corresponding to a different time instant; (iii) a plurality of velocity distributions in the velocity-time plane, each corresponding to a different intensity (e.g., a different gray-level value); and (iv) a plurality of intensity distributions in the intensity-time plane, each corresponding to a different velocity. In any event, each point on the graphical realization (in the above examples the surface or the plurality of distributions) corresponds to a particular amount of blood, flowing in a particular velocity at a particular time instant. In a third step of the method, shown in Block 14, at least one parameter is calculated. The parameter preferably characterizes a velocity spectrum of the three- dimensional flow representation. According to a preferred embodiment of the present invention the method further comprises an additional step, shown in Block 16, in which the calculated parameter(s) are used for determining the malignancy likelihood of the tumor. In other words, the parameter(s) are used for determining to what certainty the tumor in question is malignant. The malignancy likelihood can be expressed in any way known in the art. For example, in one embodiment, the malignancy likelihood can be expressed as a Boolean function which can have one of two values, say, "true" for a malignant tumor and "false" for a benign tumor. In another embodiment, malignancy likelihood can be expressed by way of a statistical function, e.g., a likelihood function, a probability function and the like. In this embodiment, the statistical function can have more than two values, e.g., large values for greater certainty and small values for lesser certainty. Such function can also be normalized such that its value is bounded to a predetermined range (say, from 0 to 1, or from 0 to 100), whereby the low bound can represent a benign tumor and the high bound can represent a malignant tumor. Typically, but not obligatory, one or more of the calculated parameters) is associated with a predetermined threshold facilitating the determination of the malignancy likelihood. For example, when the malignancy likelihood is expressed as a Boolean function, then, a value of a particular parameter which is above the predetermined threshold corresponds to "true", and a value of the particular parameter which is below the predetermined threshold corresponds to a "false." In a more complicated case in which the malignancy likelihood is expressed in terms of a statistical function having more than two values, the calculated parameters can be used as arguments of the statistical function, whereby different values of the parameters correspond to different values of the statistical function, hence to different malignancy likelihood. Additionally, the malignancy likelihood can be expressed in terms of regions on a multidimensional graph. For simplicity, suppose that there are two parameters each being associated with a single threshold, so that when a value of the parameter is above its respective threshold, it is said to be "high" and if it is below the threshold it is said to be "low." Plotting the values of the two parameters on a two-dimensional graph, results in four regions ("low-low," "low-high," etc.), which can be used to better specify the malignancy likelihood. When more parameters (or more thresholds) are used, the number of dimensions and/or regions is increased to further improve the specificity. According to a preferred embodiment of the present invention the method comprises an additional step, shown in Block 11, in which the Doppler flow image is smoothed, for example, by a median filter. The smoothing procedure facilitates better waveform delimitation. The median filter can have either a fixed or a variable window size, as desired. A typical window size for the median filter is, without limitation, about 11x11 pixels. As used herein the term "about" refers to + 10 %. Many parameters which characterize the velocity spectrum are contemplated. For example, one such parameter is a velocity distribution slope, which can be defined as the slope of a predetermined portion of a velocity distribution. As stated, the three- dimensional flow representation can be realized as a plurality of velocity or intensity distributions. According to a preferred embodiment of the present invention, the velocity distribution slope is calculated by averaging over several distributions (e.g., several velocity distributions) so as to obtain an average distribution, and calculating the slope of a predetermined (preferably central) portion of the average distribution.

Any type of averaging can be implemented, using either equal or different weights for each distribution. Representative examples include, without limitation, arithmetic averaging, geometric averaging and harmonic averaging. For a given portion of a given distribution or average distribution, the calculation of the slope can be performed, in any way known in the art, such as, but not limited to, using a linear fit. The value of the velocity distribution slope depends on the distributions over which the averaging procedure is implemented. There is therefore more than one way for calculating the velocity distribution slope. Hence, in one embodiment, the slope is calculated from an average of the velocity distributions during end-diastole of the blood flow. In another embodiment the slope is calculated from an average of the velocity distributions during systole of the blood flow. Slopes calculated during end-diastole, and systole are referred to hereinafter as end-Diastolic Velocity Distribution Slopes (DVDJS) and Systolic Velocity Distribution Slopes (SVD_S), respectively. The average velocity distribution can be used for calculating many other parameters, characterizing the velocity spectrum of the three-dimensional flow representation. Hence, in one embodiment, the parameter is a minimal average velocity. This parameter is preferably calculated by locating a minimal value of velocity distribution. In another embodiment, the parameter is an average velocity difference. This parameter is preferably calculated as follows. First, a plurality of end-diastolic instants is selected. Second, the difference between a maximal velocity of each such end-diastolic instant and an average velocity is calculated such that a plurality of velocity differences, each representing the value of maximal velocity with respect to the average velocity. Third, the velocity differences are averages to provide the desired average velocity difference. Referring to Block 13 of Figure 1 , according to a preferred embodiment of the present invention the method further comprises an additional step in which a plurality of velocity ranges of the velocity spectrum are defined. The ranges can be defined by selecting one or more velocity thresholds separating between the ranges. In the simplest embodiment in which Block 13 is employed, a single velocity threshold is defined, separating between a high velocity range and a low velocity range, respectively defined above and below the velocity threshold. The velocity threshold can be either a predetermined threshold in which case its value is selected a priori, or, more preferably, an adapted threshold in which case its value is adapted to the Doppler flow image of relevance. It has been found by the Inventors of the present invention that it is useful to define the velocity threshold to be equal or close to the minimal end- diastolic velocity of the velocity spectrum. Typically, the minimal end-diastolic velocity of the velocity spectrum is, without limitation, from about 10 cm/s to about 70 cm/s. The velocity ranges can be exploited for the purpose of calculating other parameters, characterizing the velocity spectrum. Hence, according to a preferred embodiment of the present invention the parameter is a high-to-low velocity ratio, which is related to areas under the velocity distributions, such as, but not limited to, the averaged velocity distribution. For example, the high-to-low velocity can be defined as a ratio between two areas under the averaged velocity distribution: a first area, corresponding to the high velocity range and a second area corresponding to the low velocity range. In another embodiment, the parameter is an angle parameter, which is preferably defined as an angle between two lines (e.g., straight lines), each representing a different portion of a particular velocity distribution. For example, the angle parameter can be defined as the angle between straight lines representing the low and high velocity ranges of a velocity distribution which is characterized by a maximal intensity over the three-dimensional flow representation. The straight lines can be obtained, for example, by calculating linear fits to the respective portions of the selected velocity distribution. According to another aspect of the present invention there is provided a software apparatus 20 for analyzing a Doppler flow image of a region containing a tumor. Apparatus can be used for implementing one or more of the method steps described hereinabove, with reference to Figure 1. Reference is now made to Figure 2, which is a schematic illustration of apparatus 20. Hence, apparatus 20 preferably comprises a representation unit 22, a parameters calculator 24 and a determinator 26. Unit 22 serves for representing the Doppler flow image as a three-dimensional flow representation, for example, the aforementioned "velocity-time-intensity" representation. Calculator 24 serves for calculating at least one parameter, such as, but not limited to, the velocity distribution slope(s), the minimal average velocity, the average velocity difference, the high-to-low velocity ratio and the angle parameter, as further detailed hereinabove. Determinator 26 serves for determining the malignancy likelihood of the tumor, using any of the parameters calculated by calculator 24, as further detailed hereinabove. According to a preferred embodiment of the present invention apparatus 20 further comprises a smoother 28, for smoothing the Doppler flow image, e.g., using a median filter, as detailed hereinabove. According to an additional aspect of the present invention there is provided a system for determining malignancy likelihood of a tumor, generally referred to herein as system 30. Reference is now made to Figure 3, which is a schematic illustration of system 30. System 30 comprises a Doppler ultrasonic device 32, for imaging a region surrounding the tumor. Device 32 can be any Doppler ultrasonic device known in the art, including, without limitation, non-invasive, invasive and minimal invasive (e.g., endoscopic) Doppler ultrasonic devices. System 30 further comprises software apparatus 20 communication with device 32 and capable of analyzing the Doppler flow image as further detailed hereinabove.

It is expected that during the life of this patent many relevant Doppler technologies will be developed and the scope of the terms "Doppler ultrasonic device," and "Doppler flow image" is intended to include all such new technologies a priori. Additional objects, advantages and novel features of the present invention will become apparent to one ordinarily skilled in the art upon examination of the following example, which is not intended to be limiting. Additionally, each of the various embodiments and aspects of the present invention as delineated hereinabove and as claimed in the claims section below finds experimental support in the following example. EXAMPLE Reference is now made to the following example, which together with the above descriptions illustrates the invention in a non limiting fashion. Quantitative Analysis of Doppler Flow Images

Methods Records of 53 post-menopausal subjects with ovarian mass, who were under medical evaluation at the gynecological department of the ultrasound unit of Sapir Medical Center, Israel, were examined. The subjects were selected according to the following criteria: (i) histologically confirmed ovarian tumor; and (ii) ovarian mass evaluated by Doppler ultrasound in a transvaginal sonography (TVS) examination. All TVS and recorded Doppler flow examinations, followed by histo- pathological analysis of the same patients, were included in the study. TVS and Doppler were performed using an ultrasound system equipped 128XP10 (Acuson, Mountain- View, CA) with high resolution endovaginal transducer of 5-7 mHz. The TVS and recorded Doppler flow examinations resulted in 53 Doppler flow images of ovarian mass (33 malignant and 20 benign). Each record was digitized into a 512 x 512 pixels image with 256 grey levels, using a DT-2853 frame grabber (Data Translation, Marlboro, MA, USA). The tumors were examined histologically according to the World Health Organization classifications. Each Doppler flow image was smoothed by a median filter with a window size of 1 lxl 1 pixels, so as to improve waveform delimitation. The effect of the smoothing procedure is demonstrated in Figures 4a-b showing contours of a single wave which were automatically extracted before (Figure 4a) and after (Figure 4b) smoothing. For each Doppler flow image, a conventional resistance index (RI) and a conventional decay constant were extracted. The values of the conventional decay constants were determined by fitting the waveforms to exponential functions, and extracting the time coefficients therefrom. In addition, a weighted decay constant, denoted hereinafter by τ, was calculated by averaging the conventional decay constants over several waveforms, using the quality of each individual fit as a relative weight. Once smoothed, each Doppler flow image was represented as a three- dimensional flow representation having time information, velocity information and intensity information expressed in gray-level measure. The gray-level of each pixel is proportional to the number of the red blood cells, passing through the blood vessel at a specific time and velocity. Reference is now made to Figure 5 showing a typical Doppler image. Each pixel on the time axis corresponds to about 11.4 ms. The instantaneous flow spectrum was divided into two velocity ranges: a high velocity range which included velocities with values higher than the minimal end-diastolic velocity of the image, and a low velocity range which included all other velocities. The minimal end-diastolic velocity of the image is shown as a horizontal line in Figure 5. Figures 6a-b show the gray-level values normalized to the region of 0 to 1, as a function of the velocity in arbitrary units. Each curve in Figures 6a-b corresponds to one instant of time, and referred to herein as velocity distribution. Being scaled to the range o to 1, a normalized value of the gray-level reflects the probability of finding a specific velocity at a specific time instant. Thus, the normalized values of the gray- level are interchangeably referred to herein as "probabilities." Figure 6a shows two velocity distributions corresponding to the systole time and the end-diastolic time. Figure 6b shows a plurality of successive velocity distributions, with a time scale of a few cardiac beats between successive curves. Also shown in Figure 6b, is the minimal end-diastolic velocity, presented in Figure 6b by a vertical dashed line, and the average velocity distribution presented by an emphasized solid curve. The average velocity distribution can be regarded as the average occurrence for each velocity at all times. The vertical line divides the area under the emphasized curve into two areas. Figure 6b thus demonstrates the calculation of the high-to-low velocity ratio, which is the ratio between the two areas. Figure 7a shows a portion of the spectrum shown in Figure 6b, which corresponds to the aforementioned high velocity range. Figure 7b shows a velocity distribution which is characterized by a maximal probability along the envelope of the velocity spectrum shown in Figure 7a. Also shown in Figure 7b, two linear fits applied to the right portion and the left portion of the maximal probability curve. A portion of about 40 % was selected for each linear fit. Figure 7b thus demonstrates the calculation of the angle parameter, β, which is the angle between the two linear fits. Figure 8 demonstrate the calculations of the minimal average velocity and the average velocity difference at the end-diastolic time instants, over a sequence of several beats. Shown in Figure 8 are two adjacent average velocity curves, calculated before (blue curve) and after (green curve) the smoothing procedure. Also shown is a black curve representing maximal velocities. The systolic and the end-diastolic time instants are shown as vertical lines in Figure 8 and are designated by "S" and "ED," respectively. Hence, for every end-diastolic time instant, a weighted average velocity was calculated, using the gray levels as weights. Once the average velocity was calculated, it was subtracted from several end-diastolic maximal velocities, to thereby provide several velocity differences, shown in Figure 8 as thick vertical red streaks. The average velocity difference was defined as the average value of the velocity difference over several beats (four in the example shown in Figure 8). The minimal average velocity was calculated by locating a minimal value of velocity distribution, and is represented in Figure 8 by a black pentagram. The velocity spectrum was further analyzed by calculating velocity distribution slopes, in accordance with preferred embodiments of the present invention. The velocity distribution slopes were calculated immediately after the end-systolic peak and before end-diastole. It is recognized that immediately after the end-systolic peak the influence of the heart on the blood flow is maximal, whereas before end-diastole the effect of the heart is minimal but the expression of the blood vessel characteristics is maximal. Figure 9 shows a spectrum of velocities as a function of time, in which a horizontal white line separates the low velocity range from the high velocity range. Also shown in Figure 9, are windows representing the time intervals in which the velocity distribution slopes were calculated. Time interval including the systole time instant are designated by "S" and time intervals including the end-diastolic time instants are designated "ED." The width of each window was about 6 pixels, which is about 10 % of duration of one wave. Figures lOa-b and l la-b demonstrate the calculation of the end-diastolic velocity distribution slope, DVD_S, for subjects having benign (Figures lOa-b) and malignant (Figures l la-b) ovarian tumors. Hence, each of Figure 10a and Figure 11a shows a plurality of velocity distributions at the low velocity range and during the end- diastolic time intervals. The distributions were averaged, and the corresponding average distributions are shown in Figure 10b and Figure l ib, respectively. Also shown are linear fits of a central portion of the average distributions from which the values of the DVDJS were extracted. For average distribution, only a portion of the curve was used for the linear fit.

Specifically the left and, in some cases, the right portion of each curve was discarded.

The probability values at the left portion of the curve were considered as false values resulting from a high-pass filtering procedure employed by the ultrasonic Doppler device. This filtering is for the purpose of eliminating extrinsic low frequency (velocity) components, which arise predominantly from the vessel walls or other adjacent slow-moving structures. In some cases, the right portion of the distributions contained small probability values that result from setting a shifted maximum waveform curve. This shift was due to a corresponding shift at the automatically extracted contour, used for deriving the maximum waveform curves. In case the latter shift is upwards, the automatic contour is drawn above the boundary layer between the actual wave and the background. Since this layer represents velocities higher than the maximal velocities of the wave, these values were considered as false values hence were discarded. As shown, for the subject with the benign tumor the value of DVDJS was 13.5 (see Figure 10b) while for the subject with the malignant tumor the value of DVDJS was 0.66 (see Figure 1 lb). Results The values obtained for the conventional RI parameter were: for malignant tumors RI = 0.40 ± 0.14 and for benign tumors RI = 0.56 ± 0.10. The values of τ were 108 ± 81 and 57±18 for malignant and benign tumors, respectively. A Receiver Operating Characteristic (ROC) curve was drawn for each parameter. The area under the ROC curve was 0.816 (PO.0001) for RI and 0.723 (P=0.0008) for τ. Figures 12a-b show the ROC curves of RI (Figure 12a) and τ (Figure 12b). Figure 12c shows RI versus τ, for all subjects, where benign masses are shown as green circles and malignant tumors are shown in red circles. As can be understood from inspecting Figure 12c, these parameters can discriminate between malignant and benign tumors for small values of RI and large values of τ. When the RI becomes significantly large (above about 0.4) and τ significantly low (below about 100), no separation can be made benign and malignant tumors, as the points overlap. Table 1 below summarizes the values of the parameters which characterize the velocity spectrum of the three-dimensional flow representation. Table 1

Figures 13a-d show the ROC curves of the average velocity difference (Figure 13 a), minimal average velocity (Figure 13b), β (Figure 13 c) and high-to-low velocity ratio (Figure 13d). Figures 14a-d, show relations between several parameters: Figure 14a shows RI versus the high-to-low velocity ratio, Figure 14b shows the average velocity difference versus the high-to-low velocity ratio, Figure 14c shows the minimal average velocity versus the high-to-low velocity ratio, and Figure 14a shows RI versus β. As shown, particularly in Figures 14c-d, the new parameters improve the discrimination between benign and malignant tumors. Yet, a certain degree of overlap is still present. Figures 15a-b show the ROC curves of DVDJS (Figure 15a) and SVDJS (Figure 15b). Table 2 below present the classification of all flow images according to the value of DVDJS. The discrimination between malignant and benign tumors (the malignancy likelihood) was obtained by comparing the value of the respective DVDJS, to a threshold of 3.5, 4, 4.5 and 5. Table 2

PPV =Positive Predictive Value. ΝPV =Negative Predictive Value. Figures 16a-c show relations between RI, the average velocity difference, the weighted decay constant and DVDJS. Figure 16a shows RI versus DVDJS, Figure 16b shows the average velocity difference versus DVDJS and Figure 16c shows, on a three-dimensional plot, τ versus RI and DVDJS. In Figure 16a-c, benign and malignant tumors are shown in green and red circles, respectively. As shown, the benign and malignant tumors are separated into well defined clusters. Thus, DVDJS can serve for determining the malignancy likelihood of the tumors. Discussion Generally, the parameters which characterized the three-dimensional representation of the Doppler flow images achieved better results than the Doppler parameters calculated by conventional method, i.e., devoid of the three-dimensional representation. The separation capacity between benign and malignant tumors was significantly influenced by the selection of the time intervals in which the parameters were calculated. Parameters obtained mainly from the end diastole were found to be of particular advantage because end-diastolic measurement minimizes the effect of cardiac contractility and emphasizes the vascular effect on blood flow, thus stressing the difference in the flow characteristics related to ovarian malignancy. Thus, comparing the separation capacity of the end-systolic parameter (SJ^/D_S), which is strongly influenced by cardiac function, to the end-diastolic parameter (DVDJS), revealed that the latter offers extremely sensitive discrimination ability between malignant and benign tumors. In fact, DVDJS provides much better results than the other parameters, including conventional RI and decay constant. Figures 17a-b show malignancy classification using conventional RI (Figure

17a) and DVDJS (Figure 17b), where green and red circles respectively correspond to benign and malignant tumors. A horizontal arrow in Figure 17a marks the position of a conventional RI threshold, and a vertical arrow in Figure 17b marks the position of a DVDJS threshold, according to a preferred embodiment of the present invention. Points lying below the RI threshold in Figure 17a and points lying to the left of a DVDJS threshold in Figure 17b are classified as having high malignancy likelihood. As shown in Figure 17a, there is a considerable number of malignant points which cannot be classified using RI values. Note that this inability of RI to classify cannot be improved by changing the threshold level in Figure 17a, because the region of RI > 0.4 is occupied by both green (benign) and red (malignant) points. On the other hand, Figure 17b demonstrates that DVDJS, can be used for determining malignancy likelihood of most of the images, both for RI < 0.4 and RI > 0.4, with a greater sensitivity. The main reason for the somewhat lesser discrimination ability of the high-to- low velocity ratio and β, is that both parameters were obtained by analyzing the entire velocity spectrum at all times along the heartbeats. A better discrimination can be achieved, for example, when every wave from systole to end-diastole is divided into a number of time intervals, such that the calculation of the parameters is performed on each interval separately, resulting in a set of values for each parameter. In addition, the discrimination abilities of the parameters were affected by the universal selection of the two velocity ranges. It is expected that different choices for the velocity ranges can improve the discrimination abilities.

It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination.

Although the invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims. All publications, patents and patent applications mentioned in this specification are herein incorporated in their entirety by reference into the specification, to the same extent as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated herein by reference. In addition, citation or identification of any reference in this application shall not be construed as an admission that such reference is available as prior art to the present invention.

Claims

WHAT IS CLAIMED IS:

1. A method of analyzing a Doppler flow image of a region containing a tumor, the method comprising: (a) representing the Doppler flow image as a three-dimensional flow representation; and (b) calculating at least one parameter characterizing a velocity spectrum of said three-dimensional flow representation, so as to determine malignancy likelihood of the tumor; thereby analyzing the Doppler flow image.

2. The method of claim 1, wherein said three-dimensional flow representation comprises at least one of velocity information, time information and intensity information.

3. The method of claim 2, wherein said intensity information represents flowing volume.

4. The method of claim 2, wherein said intensity information represents number of flowing cells.

5. The method of claim 2, wherein said intensity information is characterized by a gray-level measure.

6. The method of claim 1, wherein said at least one parameter comprises velocity distribution slope.

7. The method of claim 1, wherein said velocity distribution slope is selected from the group consisting of end-diastolic velocity distribution slope and systolic velocity distribution slope.

8. The method of claim 7, wherein said end-diastolic and said systolic velocity distribution slopes, are each independently calculated using a linear fit.

9. The method of claim 1 , wherein said at least one parameter comprises a minimal average velocity.

10. The method of claim 9, wherein said calculation of said minimal average velocity comprises: (i) averaging said velocity spectrum, so as to obtain an average velocity function; and (ii) locating a minimal value of said average velocity function.

11. The method of claim 1, wherein said at least one parameter comprises an average velocity difference.

12. The method of claim 11, wherein said calculation of said average velocity difference comprises: (i) selecting a plurality of end-diastolic instants of said three-dimensional flow representation; (ii) for each end-diastolic instant of said end-diastolic instants calculating a difference between a maximal velocity and an average velocity, thereby providing a plurality of velocity differences; and (iii) averaging said plurality of velocity differences, thereby providing said average velocity difference.

13. The method of claim 1, further comprising smoothing the Doppler flow image.

14. The method of claim 13, wherein said smoothing of the Doppler flow image, is by a median filter.

15. The method of claim 13, wherein said median filter has a fixed window size.

16. The method of claim 13, wherein said median filter has a variable window size.

17. The method of claim 1, further comprising defining a plurality of velocity ranges of said velocity spectrum.

18. The method of claim 17, wherein said plurality of said velocity ranges comprises a high velocity range and a low velocity range, respectively defined above and below a velocity threshold.

19. The method of claim 18, wherein said velocity threshold is predetermined.

20. The method of claim 18, wherein said velocity threshold substantially equals a minimal end-diastolic velocity of said velocity spectrum.

21. The method of claim 18, wherein said at least one parameter comprises a high-to-low velocity ratio.

22. The method of claim 21, wherein said calculation of said high-to-low velocity ratio comprises: (i) obtaining a velocity distribution; and (ii) calculating a ratio between a first area under said velocity distribution and a second area under said velocity distribution, said first and said second areas respectively corresponding to said high and said low velocity ranges; thereby calculating said high-to-low velocity ratio.

23. The method of claim 22, wherein said velocity distribution is obtained by averaging said velocity spectrum.

24. The method of claim 18, wherein said at least one parameter comprises an angle parameter.

25. The method of claim 24, wherein said calculation of said angle parameter comprises: (i) obtaining a velocity distribution; and (ii) calculating an angle between a first line representing said low velocity range and a second line representing said high velocity range; thereby calculating said angle parameter.

26. The method of claim 25, wherein said velocity distribution is characterized by maximal intensity over said three-dimensional flow representation.

27. The method of claim 25, wherein said first and said second lines are each independently obtained by a linear fit.

28. The method of claim 1 , wherein said at least one parameter comprises a resistance index.

29. The method of claim 1, wherein said at least one parameter comprises pulsatility index.

30. The method of claim 1, wherein said at least one parameter comprises a decay constant.

31. The method of claim 1 , wherein said at least one parameter comprises a weighted decay constant.

32. The method of claim 1 , wherein the region comprises an ovary.

33. The method of claim 1 , wherein the region comprises a pelvis.

34. The method of claim 1, wherein the region comprises adnexa uteri.

35. The method of claim 1 , wherein the region comprises a uterus.

36. The method of claim 1 , wherein the region comprises a breast.

37. The method of claim 1 , wherein the region comprises a prostate.

38. The method of claim 1 , wherein the region comprises a hepatic artery.

39. The method of claim 1 , wherein the region comprises a liver.

40. A method of determining malignancy likelihood of a tumor, the method comprising: (a) imaging a region surrounding the tumor using a Doppler ultrasonic device, thereby providing a Doppler flow image; (b) representing the Doppler flow image as a three-dimensional flow representation; (c) calculating at least one parameter characterizing a velocity spectrum of said three-dimensional flow representation; and (d) using said at least one parameter for determining the malignancy likelihood of the tumor.

41. The method of claim 40, wherein said three-dimensional flow representation comprises at least one of: velocity information, time information and intensity information.

42. The method of claim 41, wherein said intensity information represents flowing volume.

43. The method of claim 41, wherein said intensity information represents number of flowing cells.

44. The method of claim 41, wherein said intensity information is characterized by a gray-level measure.

45. The method of claim 40, wherein said at least one parameter comprises velocity distribution slope.

46. The method of claim 40, wherein said velocity distribution slope is selected from the group consisting of end-diastolic velocity distribution slope and systolic velocity distribution slope.

47. The method of claim 46, wherein said end-diastolic and said systolic velocity distribution slopes, are each independently calculated using a linear fit.

48. The method of claim 40, wherein said at least one parameter comprises a minimal average velocity.

49. The method of claim 48, wherein said calculation of said minimal average velocity comprises: (i) averaging said velocity spectrum, so as to obtain an average velocity function; and (ii) locating a minimal value of said average velocity function.

50. The method of claim 40, wherein said at least one parameter comprises an average velocity difference.

51. The method of claim 50, wherein said calculation of said average velocity difference comprises: (i) selecting a plurality of end-diastolic instants of said three-dimensional flow representation; (ii) for each end-diastolic instant of said end-diastolic instants calculating a difference between a maximal velocity and an average velocity, thereby providing a plurality of velocity differences; and (iii) averaging said plurality of velocity differences, thereby providing said average velocity difference.

52. The method of claim 40, further comprising smoothing the Doppler flow image.

53. The method of claim 52, wherein said smoothing of the Doppler flow image, is by a median filter.

54. The method of claim 52, wherein said median filter has a fixed window size.

55. The method of claim 52, wherein said median filter has a variable window size.

56. The method of claim 40, further comprising defining a plurality of velocity ranges of said velocity spectrum.

57. The method of claim 56, wherein said plurality of said velocity ranges comprises a high velocity range and a low velocity range, respectively defined above and below a velocity threshold.

58. The method of claim 57, wherein said velocity threshold is predetermined.

59. The method of claim 57, wherein said velocity threshold substantially equals a minimal end-diastolic velocity of said velocity spectrum.

60. The method of claim 57, wherein said at least one parameter comprises a high-to-low velocity ratio.

61. The method of claim 60, wherein said calculation of said high-to-low velocity ratio comprises: (i) obtaining a velocity distribution; and (ii) calculating a ratio between a first area under said velocity distribution and a second area under said velocity distribution, said first and said second areas respectively corresponding to said high and said low velocity ranges; thereby calculating said high-to-low velocity ratio.

62. The method of claim 61, wherein said velocity distribution is obtained by averaging said velocity spectrum.

63. The method of claim 57, wherein said at least one parameter comprises an angle parameter.

64. The method of claim 63, wherein said calculation of said angle parameter comprises: (i) obtaining a velocity distribution; and (ii) calculating an angle between a first line representing said low velocity range and a second line representing said high velocity range; thereby calculating said angle parameter.

65. The method of claim 64, wherein said velocity distribution is characterized by maximal intensity over said three-dimensional flow representation.

66. The method of claim 64, wherein said first and said second lines are each independently obtained by a linear fit.

67. The method of claim 40, wherein said at least one parameter comprises a resistance index.

68. The method of claim 40, wherein said at least one parameter comprises pulsatility index.

69. The method of claim 40, wherein said at least one parameter comprises a decay constant.

70. The method of claim 40, wherein said at least one parameter comprises a weighted decay constant.

71. The method of claim 40, wherein the region comprises a pelvis.

72. The method of claim 40, wherein the region comprises adnexa uteri.

73. The method of claim 40, wherein the region comprises a uterus.

74. The method of claim 40, wherein said region comprises an ovary.

75. The method of claim 40, wherein said region comprises a breast.

76. The method of claim 40, wherein said region comprises a prostate.

77. The method of claim 40, wherein said region comprises a hepatic artery.

78. The method of claim 40, wherein said region comprises a liver.

79. A software apparatus for analyzing a Doppler flow image of a region containing a tumor, the apparatus comprising: (a) a representation unit, for representing the Doppler flow image as a three-dimensional flow representation; and (b) a parameters calculator, for calculating at least one parameter characterizing a velocity spectrum of said three-dimensional flow representation; and (c) a determinator for determining malignancy likelihood of the tumor, using said at least one parameter.

80. The apparatus of claim 79, wherein said three-dimensional flow representation comprises at least one of velocity information, time information and intensity information.

81. The apparatus of claim 80, wherein said intensity information represents flowing volume.

82. The apparatus of claim 80, wherein said intensity information represents number of flowing cells.

83. The apparatus of claim 80, wherein said intensity information is characterized by a gray-level measure.

84. The apparatus of claim 79, wherein said at least one parameter comprises velocity distribution slope.

85. The apparatus of claim 79, wherein said velocity distribution slope is selected from the group consisting of end-diastolic velocity distribution slope and systolic velocity distribution slope.

86. The apparatus of claim 85, wherein said end-diastolic and said systolic velocity distribution slopes, are each independently calculated using a linear fit.

87. The apparatus of claim 79, wherein said at least one parameter comprises a minimal average velocity.

88. The apparatus of claim 79, wherein said at least one parameter comprises an average velocity difference.

89. The apparatus of claim 79, further comprising a smoother, for smoothing the Doppler flow image.

90. The apparatus of claim 89, wherein said smoother comprises a median filter.

91. The apparatus of claim 89, wherein said median filter has a fixed window size.

92. The apparatus of claim 89, wherein said median filter has a variable window size.

93. The apparatus of claim 79, further comprising a velocity range definer, for defining a plurality of velocity ranges of said velocity spectrum.

94. The apparatus of claim 93, wherein said plurality of said velocity ranges comprises a high velocity range and a low velocity range, respectively defined above and below a velocity threshold.

95. The apparatus of claim 94, wherein said velocity threshold is predetermined.

96. The apparatus of claim 94, wherein said velocity threshold substantially equals a minimal end-diastolic velocity of said velocity spectrum.

97. The apparatus of claim 94, wherein said at least one parameter comprises a high-to-low velocity ratio.

98. The apparatus of claim 94, wherein said at least one parameter comprises an angle parameter.

99. The apparatus of claim 79, wherein said at least one parameter comprises a resistance index.

100. The apparatus of claim 79, wherein said at least one parameter comprises pulsatility index.

101. The apparatus of claim 79, wherein said at least one parameter comprises a decay constant.

102. The apparatus of claim 79, wherein said at least one parameter comprises a weighted decay constant.

103. A system for determining malignancy likelihood of a tumor, the system comprising: (a) a Doppler ultrasonic device, for imaging a region surrounding the tumor, thereby to provide a Doppler flow image; and (b) a software apparatus, communicating with said Doppler ultrasonic device and capable of analyzing said Doppler flow image, said software apparatus having a representation unit, for representing said Doppler flow image as a three- dimensional flow representation; a parameters calculator, for calculating at least one parameter characterizing a velocity spectrum of said three-dimensional flow representation; and a determinator, for determining the malignancy likelihood of the tumor, using said at least one parameter.

104. The system of claim 103, wherein said three-dimensional flow representation comprises at least one of velocity information, time information and intensity information.

105. The system of claim 104, wherein said intensity information represents flowing volume.

106. The system of claim 104, wherein said intensity information represents number of flowing cells.

107. The system of claim 104, wherein said intensity information is characterized by a gray-level measure.

108. The system of claim 103, wherein said at least one parameter comprises velocity distribution slope.

109. The system of claim 103, wherein said velocity distribution slope is selected from the group consisting of end-diastolic velocity distribution slope and systolic velocity distribution slope.

110. The system of claim 109, wherein said end-diastolic and said systolic velocity distribution slopes, are each independently calculated using a linear fit.

111. The system of claim 103, wherein said at least one parameter comprises a minimal average velocity.

112. The system of claim 103, wherein said at least one parameter comprises an average velocity difference.

113. The system of claim 103, wherein said software apparatus comprises a smoother, for smoothing the Doppler flow image.

114. The system of claim 113, wherein said smoother comprises a median filter.

115. The system of claim 113, wherein said median filter has a fixed window size.

116. The system of claim 113, wherein said median filter has a variable window size.

117. The system of claim 103, wherein said software apparatus comprises a velocity range definer, for defining a plurality of velocity ranges of said velocity spectrum.

118. The system of claim 117, wherein said plurality of said velocity ranges comprises a high velocity range and a low velocity range, respectively defined above and below a velocity threshold.

119. The system of claim 118, wherein said velocity threshold is predetermined.

120. The system of claim 118, wherein said velocity threshold substantially equals a minimal end-diastolic velocity of said velocity spectrum.

121. The system of claim 118, wherein said at least one parameter comprises a high-to-low velocity ratio.

122. The system of claim 118, wherein said at least one parameter comprises an angle parameter.

123. The system of claim 103, wherein said at least one parameter comprises a resistance index.

124. The system of claim 103, wherein said at least one parameter comprises pulsatility index.

125. The system of claim 103, wherein said at least one parameter comprises a decay constant.

126. The system of claim 103, wherein said at least one parameter comprises a weighted decay constant.