[go: up one dir, main page]

WO2024102061A1 - Procédé de dépistage du corps entier pour la détection d'un mélanome - Google Patents

Procédé de dépistage du corps entier pour la détection d'un mélanome Download PDF

Info

Publication number
WO2024102061A1
WO2024102061A1 PCT/SE2023/051142 SE2023051142W WO2024102061A1 WO 2024102061 A1 WO2024102061 A1 WO 2024102061A1 SE 2023051142 W SE2023051142 W SE 2023051142W WO 2024102061 A1 WO2024102061 A1 WO 2024102061A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
mole
moles
macro
micro
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/SE2023/051142
Other languages
English (en)
Inventor
Gyorgy Marko-Varga
Peter Horvath
Istvan BALAZS NEMETH
Gábor HOLLANDI
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Malskin AB
Original Assignee
Malskin AB
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Malskin AB filed Critical Malskin AB
Publication of WO2024102061A1 publication Critical patent/WO2024102061A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G06T7/0014—Biomedical image inspection using an image reference approach
    • G06T7/0016—Biomedical image inspection using an image reference approach involving temporal comparison
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2200/00—Indexing scheme for image data processing or generation, in general
    • G06T2200/24—Indexing scheme for image data processing or generation, in general involving graphical user interfaces [GUIs]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30088—Skin; Dermal
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30196—Human being; Person

Definitions

  • This invention pertains in general to the field of patient screening for skin mole identification and characterization. More particularly the invention relates to a method for reading the accessible skin surface of a subject and identifying and characterizing skin defects, such as moles, for early detection of melanoma and melanoma burden factor.
  • Melanoma is a type of skin cancer that develops from the pigment-producing cells known as melanocytes. Melanomas typically occur in the skin, and in women, they most commonly occur on the legs, while in men, they most commonly occur on the back. As a sequential carcinogenesis, melanomas can develop from benign moles, but they may also display a de novo new malignant lesion without a premalignant lesion. Therefore, new skin moles, or changes in a mole that can indicate melanoma. Such changes include increase in mole size, irregular edges, change in color, itchiness, or skin breakdown.
  • Melanoma is the most dangerous type of skin cancer. Globally, in 2012, it newly occurred in 232,000 people. In 2015, 3.1 million people had active disease, which resulted in 59,800 deaths. One problem is that melanoma may lead to spread (metastasis) of the disease. Most people are cured if spread has not occurred, however, five-year survival rates in the United States were 65% when the disease has spread to lymph nodes, and 25% among those with distant spread. Thus, it is of utmost importance to capture melanoma early. Even so, in most countries, screening for melanoma is not mandatory.
  • Screening means testing people for early stages of a disease, before they see concrete symptoms of melanoma.
  • a skin cancer specialist or nurse examines the skin.
  • a specialist is trained to look out for moles that may be starting to become cancerous.
  • screening both has to pick up all suspicious moles without removing too many (non-melanoma) moles.
  • different individuals usually have different normal appearance of the skin. Thus, it may be hard for a specialist meeting a patient for the first time, to be able to spot what is normal or abnormal for that unique individual.
  • Another disadvantage of state of the art methods is that it usually focuses on single spots, and the treatment of single spots. If a skin defect is deemed potentially harmful, it will be removed and screened for malignant factors. However, except for the obvious patent characteristics (age, skin tone, number of spots) and possibly results from earlier removed spots, little data is available to the physician at time for a screening.
  • the present invention preferably seeks to mitigate, alleviate or eliminate one or more of the above-identified deficiencies in the art and disadvantages singly or in any combination and solves at least the above mentioned problems by providing A method of a multi-step whole body dermoscopic screening of skin defects and from skin images, comprising the steps of: a) performing a Counter Mapping Image Capture, wherein a pose estimation of a subject is created comprising at least 7 marker points, providing a coordinate system of the subject body image, b) performing a macro-analysis, wherein at least one macro-image, showing a larger portion of the subjects skin, is captured and the coordinates of the at least one macro-image are Linked to the coordinate system of the subject body image from step a), a detection algorithm determines the presence moles in the macro-image by an image processing method, and for each determined mole, a machine learning model, an image processing algorithm or a neural network selects moles more likely to show a malignant pattern, a unique mole ID is assigned to each selected mole, a
  • step a) during Counter Mapping Image Capture, the coordinate system of the subject body image is integrated with the existing coordinate system of the subject body image in the AVATAR model, in step b), during macro-analysis, determined moles are correlated with existing moles in the AVATAR model, such that moles with an existing mole ID will receive the same mole ID, and new moles will receive a new mole ID, in step c), during micro-analysis, the prediction of the malignancy for each existing mole ID is compared to the previous prediction of the malignancy, and in step d), adding the new information to the AVATAR model with a new time stamp.
  • a system for dermoscopic screening of a skin anomaly comprising a macro-camera, a micro-camera, a controller and computation unit, a graphical output device, and a database and/or storage device.
  • a method for dermoscopic screening of a skin anomaly wherein the macro-camera is used to capture a whole body image for Counter Mapping Image Capture, the macro-camera is used to capture at least one macro image of the skin surface of the patient, the controller and computation unit is used to perform the pose estimation based on the Counter Mapping Image Capture, the prediction and provides a list of selected moles deemed more likely to show a malignant pattern together with their coordinates on the graphical output device, the micro-camera is used to capture a micro image per mole of at least one, preferably all, of the selected moles, and classification of the malignity of each mole for which a micro image has been captured and provides an output on the graphical output device showing a list of the selected moles, their classification of malignancy and a mole identifier or mole coordinates, and all aquired or generated data is saved as an AVATAR model in a database on the controller or computation unit or on online.
  • a method for dermoscopic screening of skin anomalys from captured pose estimation and captured macro- and micro images comprises the steps of a) from a Counter Mapping Image, creating a pose estimation of the subject based on at least 7 marker points, that provides a coordinate system of the subject body image, b) from at least one macro-image showing a larger portion of the subjects skin, doing a macro-analysis, linking the coordinates of the at least one macro-image to the coordinate system of the subject body image from step a), using a detection algorithm to determine the presence moles in the macro-image by an image processing method, and for each determined mole, selects moles likely to show a malignant pattern using a machine learning model, an image processing algorithm or a neural network, assigning a a unique ID to each selected mole, and creating an output listing, for each mole ID, the mole position and preferably a cropped image of mole, preferably providing a visible output displaying the mole output listing, to facilitate micro-image capture of selected moles,
  • Fig- 1 shows an overview of the method of dermoscopic screening of a skin anomaly, with melanoma type of skin cancer classification from skin images, in accordance with the invention.
  • the overview is from image capture of the patient to final melanoma clinical status after excision of two spots.
  • Possible software algorithm analysis steps are shown, including identification/analysis of moles at various body locations, assessment and diagnostic read-out;
  • Fig- 2 shows an overview of a system for using the method a multi-step whole body dermoscopic screening of skin defects, with melanoma burden factor from skin images, in accordance with the invention.
  • the system comprises a macro-camera, optionally together to an instrument for Counter Mapping Image Capture (camera/3D scanner/laser scanner/LIDAR), a micro-camera (dermoscope), a controller and computation unit (computer), a graphical output device (monitor), and a database and storage device (cloud).
  • the graphical output device is as an example shown in (B) to output macro image with selected spots, micro images of selected spots and software assessment output;
  • Fig- 3 shows an overview of a multi-step whole body dermoscopic screening of skin defects, with melanoma burden factor from skin images, in accordance with the invention.
  • a Counter Mapping Image capture creates a coordinate system of the subjects body image, a preliminary prediction of at least one macro image of a subject's skin detects and accesses malignancy of moles and positions them on the body coordinate system, a micro analysis wherein a micro-image is captured for at least one, preferably for each, of the selected moles, followed by analysis and prediction for the malignant evolution of the mole and providing a visible output displaying the moles, and compiling an AVATAR model for the subject wherein all information from the analysis is saved together with a timestamp, to provide the coordinates and analysis result for a physician for further inspection or possible removal of the skin defects, and/or to be used as reference model for the next patient scan;
  • Fig- 4 shows and overview of a multi-step whole body dermoscopic screening of skin defects in accordance with Fig. 3, where an avatar model of the subject already exists, wherein the patient contour, spot occurrence, position and assessment is correlated with the existing contour, to enable a timeline evaluation to assist the determination of new spots, the assessment of malignity for assessed moles and any changes in the melanoma burden factor of the patient;
  • Fig- 5 shows a Counter image scan result with a colour coded display on the 25 measure points that connects and builds the counter part
  • Fig. 6 shows a macro mage of a patient and the outcome of the macro-image step is presented as blue designated spots
  • Fig. 7 shows an optional step, wherein during macro screening, at least one fingerprint comprising at least 3 strong spots are located in each macro image and used to correlate a new scan with an existing scan.
  • the new scan has detected new spots that have develop since the last scan (highlighted with light colored boxes).
  • pronounced spots that are existing in both svans are used to generate two “fingerprinf’-pattern.
  • the fingerprints are used to fine-adjust fine-adjust the coordinates of spots in the new screening with the existing AVATAR coordinate system. This ensures that the mole IDs are assigned correctly to the new scan;
  • Fig. 8 shows a micro-imaging follow up step, from the macro-imaging full body screening
  • Fig. 9 shows an example of micro-image captured moles
  • Fig. 10 shows the output of the method using the 8-channel mAIskin algorithm
  • Fig. 11 shows malignancy prediction for an individual mole taken at several different time points, wherein (c) represents a steady growth of a (here benign) mole, (b) represents a steady growth of a mole with a higher growth rate than in (c), and (a) shows a mole that suddenly changes behaviour and generates a higher malignancy prediction score;
  • Fig. 12 shows malignancy prediction for all moles of a subject at several different time points, wherein (B) shows a uniform melanoma burden, and (A) shows a subgroup of moles that generates a higher malignancy prediction score.
  • populations may also be correlated by other factors such as mole position, to show melanoma burden patterns;
  • Fig. 13 shows shows the summarized output of the method using the 8-channel mAIskin algorithm at two different time-points, a) (2021) and b) (2022) for the same moles.
  • a mole will obtain the same ID as during the first scan by a Counter image scan positon and spot fingerprinting.
  • the second prediction benefit from the first scan result during malignancy prediction.
  • the following description focuses on an embodiment of the present invention applicable to a multi-step whole body dermoscopic screening of skin defects, with individual mole malignancy prediction and patient melanoma burden factor from skin images.
  • the patient has a nude body allowing gross, naked-eye inspection of the existing moles on the whole body surface.
  • moles can have any colour change, but nearly skin-coloured or pink macules, papules or plaques are also noticed.
  • the topography of the moles are identified, also the macroscopically clearly different moles (in size, colour, or shape) from others are already initially marked (gross topography identification and initial pre-analysis, marking).
  • the order of the regions are the trunk, head and neck region, limbs (also palms and soles), and genital area.
  • the pattern analysis is made by the dermoscopic guidelines (ABCDE-rule, Assymetry, Border, Color, Diameter, Elevation and 7-points checklist: 3 points indicate suspicious for melanoma).
  • the regular mole check should preferably be 1 event/year, however, multiple atypical mole syndrome may indicate 3-6 monthly follow up. Similarly, certain dysplastic moles which have a moderately atypical pattern but do not step through the threshold criteria of excision, shorter follow up periods should be made.
  • the method of the invention operates as a whole body imaging system with image capture that covers a whole body skin surface in macro quality resolution, interfaced with a micro-imaging high quality resolution capture.
  • the subject is positioned within the Contour mapping unit, for a pose estimate.
  • the pose estimation builds on at least 7, such at least 15, but preferably at least 25 marker points, and are generated from the subject that provides a coordinate system of the subject body image.
  • the resolving power of the contour system may be as low as 1 mm.
  • An output of the contour-image system is provided in Figure 5.
  • the resulting contour will be correlated to the macro and micro images and selected spots in the subsequent steps. This makes it possible to give each selected spot a unique ID that is linked to a specific position on the patient contour map.
  • macro images showing larger skin regions of the patient's body are taken using a high resolution camera.
  • at least four macro images are taken to cover the larger skin regions of the patient's body such as wherein at least 40%, preferably at least 75%, more preferably at least 90%, of the skin surface of the patient.
  • a physician inspect these areas manually for best result, and input any micro scanned moles (see below) with a position relative to the subject body image.
  • the macro images should be captured with good light conditions with the camera at a known position, to ensure easier correlation between the position of the moles on the patient and in the macro image.
  • a preliminary prediction i.e. macro-prediction, is performed, during which spots are identified and highlighted.
  • a detection algorithm determines each mole by an image processing method. Since the image processing method only has to detect skin defects, several different algorithms work for this initial detection.
  • the mole detection may be based on a convolutional neural network to recognize and overlay the moles on the images.
  • the macro prediction is appended with a classification step, which further filters the samples to be examined.
  • a classification step which further filters the samples to be examined.
  • a melanoma probability score is assigned to each mole detected during processing, based on this the detected objects can be categorized medically. This may be done by for instance a deterministic or stochastic algorithm which generates a score for each mole, indicative of if the mole is likely to show a malignant pattern.
  • the moles that are deemed in need of follow up or visual inspection i.e. selects moles more likely to show a malignant pattern) are selected and highlighted on the macro overview image.
  • the melanoma probability score is generated using a neural network for skin anomaly classification of the malignity for the mole, or a neural network ensemble classification.
  • Macro images are correlated to the patient’s contour map. This makes it possible to give each selected spot a unique ID that is linked to the specific position on the patient contour map.
  • Selected moles are preferably visually displayed on a screen with a bounding box, and the position on the patient’s contour map is displayed. This helps the physician to locate the spot for further examination.
  • the high-resolution micro image preferably comprises at least 1 000 000 pixels per cm2 of imaged skin surface, preferably at least 3 000 000 pixels per cm2 of imaged skin surface.
  • Capturing a micro image of each mole/spot presents several advantages.
  • the high resolution image provides a lot more detail than the cropped macro image.
  • Another advantage is that the camera and lens can be selected to provide a favourable depth of field (DOF), which, under good light conditions, may help to capture mole details not just on the surface of the skin, but also features just under the skin surface, thereby possibly obtaining more characteristics of the mole.
  • DOF depth of field
  • the method of the invention performs a micro-prediction.
  • the malignancy of the moles are performed by prediction.
  • the micro-image algorithm is a neural network trained using ground truth data comprising an assessment/diagnosis of moles by a physician using the seven-point checklist scoring system, associated with the mole image data.
  • the final score can be an ensemble of the predictions of the multiple neural network models, each sensitive for distinct characteristics of moles via unique training approaches including number of epochs, learning rate, augmentation. These may also include metadata (which is the relevant previous medical information of the patient). This way, the decision of several independent "doctors" is simulated.
  • the algorithm weights the individual model predictions and combines them to a single final score which provides an overview of the mole's malignity.
  • metadata, non-image based characteristics such as age, skin cancer history, genetics (family history of skin cancer), skin tone, health status (such as weakened immune system), and reports of excessive UV-exposure) for the patient is collected, may be used to help the scoring algorithm to provide a better prediction
  • the prediction of the malignity may be weighted using patient metadata.
  • the metadata may also be saved in the AVATAR model (described below).
  • more than one macro image may be taken per mole, for instance one micro image (as described above) and one micro image using a polarizer.
  • the method may process the captured micro- or macro-images using a quality control module, to verify that the image(s) meet image requirements for reliable screening, rejecting images where the quality is too low for safe analysis. This way, sub-standard images will not be allowed to be used in the method, minimizing the risk for false positive of false negatives.
  • the data from the screening, together with possible metadata about the patient, is compiled and saved as an AVATAR model for the subject, together with a timestamp for the scan.
  • the AVATAR is a way to save different type of information in a correlated manner, such as patient metadata, the subject body image scan and positions, macroscan images, mole positions, mole predictions, micro scan images and mole malignancy predictions etc.
  • a database used for data storage in the method hosts a personal and individualized system which allows for repeated screenings, with comparative analysis of mole developments according to the ABCDE-classification.
  • the AVATAR model may be described as a (patient specific) database comprising the data from the multi-step whole body dermoscopic screenings, together with possible metadata about the patient, and a timestamp for each screening.
  • a method of a multi-step whole body dermoscopic screening of skin defects from skin images comprising the steps of:
  • a detection algorithm determines the presence moles in the macro-image by an image processing method, and for each determined mole, a machine learning model, an image processing algorithm or a neural network selects moles more likely to show a malignant pattern.
  • Each selected mole will be assigned a unique ID, and an output listing will be created, listing for each mole ID, the mole position and (preferably) a cropped image of the mole.
  • a visible output displaying the mole output listing is preferably provided, to facilitate micro-image capture of selected moles.
  • step b Performing a micro analysis, wherein a micro-image is captured for at least one, preferrably for each one of the selected moles of step b).
  • an algorithm analyzes each micro-image and provides a prediction of the malignity for the mole on the image, using classification to either of the two classes: benign and malignant phenotype on a continuous scale from 0 to 1.
  • the mole close-up image and prediction of malignity are then added to the output list for each mole ID.
  • step (d) Compiling and saving an AVATAR model for the subject wherein all information from step a) to c), including the mole output listing is saved together with a time stamp.
  • a visible output displaying the AVATAR model is preferably provided.
  • the method may also be used to provide a summarized melanoma burden factor for the patient.
  • the ground truth data used for training the neural network for micro-assessment and/or macro- assessment of malignancy may comprise assessment/diagnosis of moles and mole image data collected at several different time stamps during a period of several years, providing a predicted change of malignancy score for benign spots.
  • the algorithm/neural network has the advantage of being built from a longitudinal 5- year study, with follow up mole testing and clinical investigation from 500 Melanoma patients with developing mole progressions. These images as well as clinical data has been used as ground truth data for the Al-learning software development.
  • the ground truth data comprises assessment/diagnosis of moles and mole image data collected at several different time stamps during a period of several years, providing a predicted change of malignancy score for benign spots. This way, the prediction of the malignity may be weighted using the predicted change of malignancy score for benign spots.
  • the macro images of the AVATAR model and the current scan may be compared to indicate if the skin tone is different at the time of the scans. This may be if the subject is pale on one macro analysis and sun-burned on the other. To ensure that the image comparison is not affected by the skin tone difference, the current scan may be normalized to enable fair comparison of individual moles from the scan and from the AVATAR.
  • the AVATAR will not only register changes in individual moles, but also changes in population of moles in the patent. This becomes especially relevant when a removed mole has been classified as having mutations that are near-malignant or malignant. Then, the whole melanoma burden factor for the patent can be calculated, that is that changes in the AVATAR, such as clusters of new moles, or clusters of moles that suddenly get a higher melanoma probability score. Such a cluster can be seen in figure 11 (B).
  • the system can also track such systemic changes and use it to weigh the melanoma probability scores, or to alert the physician to such trends.
  • the subject is screened using method of a multi-step whole body dermoscopic screening above.
  • the coordinate system of the subject body image is integrated with the existing AVATAR coordinate system.
  • detected moles are overlapped with existing AVATAR moles. This way, moles with an existing mole ID will keep the same mole ID, while new moles will receive a new mole ID.
  • new mole IDs can be are highlighted (such as with an extra color), to help the physician see if the moles are regional, have a very high score (indicating fast growth rate), or if it is only benign moles in locations where they are expected (such as due to sun exposure).
  • the prediction of the malignity for each existing mole ID will be compared to the previous prediction of the malignity.
  • the mole growth rate can be estimated, and possibly the prediction can be enhanced.
  • All information will be saved in the AVATAR model with a new timestamp, preferably providing a visible output displaying the AVATAR model, highlighting; new mole IDs, representing now moles sine the last screening, mole IDs where the prediction of the malignity has increased, representing moles that has changed or worsened since the last screening.
  • the method may select a number of spots positions to create a patient “finger print” of spot positions. This allows for automatic adjustment/correlation of spot positions in relation to the AVATAR coordinate system, and allows for overlapping of macro images, or transformation of macro images, to ensure that existing mole IDs are assigned correctly to the same mole for each new scan.
  • An example of such a fingerprintig step can be seen in figure 7.
  • At least one fingerprint comprising at least 3 strong spots (spots that are the largest and/or the darkest spots in the macto image) are located in each macro image, and wherein overlapping of the fingerprint for the old screening and the new screening are used to fine adjust the coordinates of the new screening with the existing AVATAR coordinate system, to ensure that the mole IDs are correct.
  • the regular mole check should preferably be 1 event/year.
  • people come in for their first screening at very different ages, with very different amount of moles.
  • multiple atypical mole syndrome may indicate a shorter, such as 3-6 months interval, or even monthly follow up.
  • people over certain ages or having special skin conditions or atypical growth patterns should preferably have screenings with shorter intervals.
  • the AVATAR presents both a safety net and important tool for the physician.
  • patient metadata such as age and predisposition factors for melanoma
  • spot growth history and melanoma burden the AVATAR can ensure that the factor that presents highest risk (may for example be an individual mole or the whole patient melanoma burden factor) is used to set the recommended interval until next screening. Since it is especially favorable if early indications of melanoma can be detected, this helps ensure that patients do not wait too long before follow up screenings. Similarly, if all spots show benign growth patterns and metadata and melanoma burden is low, a longer interval can be suggested with low risk, in order to save health care resources and for patient convenience.
  • a physician may rely on the method of the invention not only for screening and scooring, but also for keeping the patient information between scans. Further, having been trained on a database showing natural spot progression from real spots, the method of the invention can predict normal spot development and warn if a spot shows signs of changing behaviour, which may be a sign of increaed risk for malignancy. This exanbles the physician to wait with surgical removal of medium risk spots (which always represents a risk and discomfort to the patient), and instead screen the patients regulgarly. Thus, the method of the invention provides a tool for screening individuals in order to capture melanoma early.
  • a system for dermoscopic screening of a skin anomaly comprises a camera for macro-scan (i.e. macro-camera), a camera for micro-scan (i.e. macro-camera), a controller and computation unit, a graphical output device, and a database and/or storage device.
  • a camera for macro-scan i.e. macro-camera
  • a camera for micro-scan i.e. macro-camera
  • a controller and computation unit i.e. macro-camera
  • a controller and computation unit i.e. macro-camera
  • the macro-camera is used to capture a whole body image for Counter Mapping Image Capture.
  • other scanning device may be used, such as a camera, a 3D scanner, a laser scanner or LIDAR system.
  • the macro-camera is used to capture at least one macro image of the skin surface of the patient. If different devices are used for counter mapping image capture and capturing macro images, known relative device positions may be used to correlate captured macro-images with the for Counter Mapping Image Capture.
  • the controller and computation unit is used to perform the pose estimation based on the Counter Mapping Image Capture, the prediction and provides a list of selected moles deemed more likely to show a malignant pattern together with their coordinates on the graphical output device.
  • the micro-camera is used to capture a micro image per mole of at least one, preferably all, of the selected moles.
  • the controller and computation unit is used to perform the classification of the malignity of each mole for which a micro image has been captured.
  • An output is provided on the graphical output device showing a list of the selected moles, their classification of malignancy and a mole identifier or mole coordinates. All aquired or generated data is saved as an AVATAR in a database on the controller or computation unit or on online.
  • pose estimation and captured macro- and micro images can be captured.
  • the method of the invention can also work from captured pose estimation and captured macro- and micro images, wherein the method comprises the steps of a) from a Counter Mapping Image, creating a pose estimation of the subject based on at least 7 marker points, that provides a coordinate system of the subject body image, b) from at least one macro-image showing a larger portion of the subjects skin, doing a macro-analysis, linking the coordinates of the at least one macro-image to the coordinate system of the subject body image from step a), using a detection algorithm to determine the presence moles in the macroimage by an image processing method, and for each determined mole, selects moles likely to show a malignant pattern using a machine learning model, an image processing algorithm or a neural network, assigning a a unique ID to each selected mole, and creating an output listing, for each mole ID, the mole position and preferably a cropped image of mole, preferably providing a visible output displaying the mole output listing, to facilitate micro-image capture of selected moles, c) from micro-images

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Measuring And Recording Apparatus For Diagnosis (AREA)

Abstract

L'invention se réfère à un procédé de caractérisation du cancer, dans lequel un procédé de dépistage dermoscopique en plusieurs étapes de défauts de la peau sur le corps entier à partir d'images de la peau, comprend les étapes suivantes : a) mettre en oeuvre une capture d'image de contre-mappage, b) mettre en oeuvre une macro-analyse, un algorithme de détection déterminant la présence de naevi dans la macro-image par un procédé de traitement d'image, et pour chaque naevus déterminé, un modèle d'apprentissage automatique, c) mettre en oeuvre une micro-analyse, une micro-image étant capturée pour au moins un, de préférence pour chacun des naevi sélectionnés à l'étape b), d) compiler et sauvegarder un modèle AVATAR pour le sujet, toutes les informations des étapes a) à c) comprenant la liste de sortie de naevi étant sauvegardées conjointement avec une estampille temporelle, produisant de préférence une sortie visible affichant le modèle AVATAR. L'invention concerne également un procédé de dépistage dermoscopique en plusieurs étapes de défauts cutanés sur le corps entier, un système de dépistage dermoscopique d'une anomalie cutanée et un procédé utilisant ledit système pour le dépistage dermoscopique d'une anomalie cutanée.
PCT/SE2023/051142 2022-11-11 2023-11-10 Procédé de dépistage du corps entier pour la détection d'un mélanome Ceased WO2024102061A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
SE2251328-7 2022-11-11
SE2251328 2022-11-11

Publications (1)

Publication Number Publication Date
WO2024102061A1 true WO2024102061A1 (fr) 2024-05-16

Family

ID=88863539

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/SE2023/051142 Ceased WO2024102061A1 (fr) 2022-11-11 2023-11-10 Procédé de dépistage du corps entier pour la détection d'un mélanome

Country Status (1)

Country Link
WO (1) WO2024102061A1 (fr)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170061621A1 (en) * 2015-08-31 2017-03-02 Massachusetts Institute Of Technology Systems and methods for tissue stiffness measurements
US20190188851A1 (en) * 2012-03-28 2019-06-20 University Of Houston System Methods for Screening and Diagnosing a Skin Condition
WO2022087132A1 (fr) * 2020-10-20 2022-04-28 Skinio, Llc Systèmes et procédés de surveillance d'anomalie de la peau

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190188851A1 (en) * 2012-03-28 2019-06-20 University Of Houston System Methods for Screening and Diagnosing a Skin Condition
US20170061621A1 (en) * 2015-08-31 2017-03-02 Massachusetts Institute Of Technology Systems and methods for tissue stiffness measurements
WO2022087132A1 (fr) * 2020-10-20 2022-04-28 Skinio, Llc Systèmes et procédés de surveillance d'anomalie de la peau

Similar Documents

Publication Publication Date Title
Mendonça et al. PH 2-A dermoscopic image database for research and benchmarking
EP1875863B1 (fr) Procedé et dispositif d'analyse de l'état de la peau, et moyen d'enregistrement sur lequel un programme d'analyse de l'état de la peau est enregistré
CN111602173B (zh) 脑部断层扫描数据分析方法
CN100483460C (zh) 皮肤疾患的自动检测
Yang et al. Development and validation of two artificial intelligence models for diagnosing benign, pigmented facial skin lesions
JP7638269B2 (ja) 脳卒中特徴取得のための画像処理
KR20190087272A (ko) 안저영상을 이용한 녹내장 진단 방법 및 이를 위한 장치
Veiga et al. Automatic microaneurysm detection using laws texture masks and support vector machines
US20230206435A1 (en) Artificial intelligence-based gastroscopy diagnosis supporting system and method for improving gastrointestinal disease detection rate
Primiero et al. A narrative review: opportunities and challenges in artificial intelligence skin image analyses using total body photography
Singh et al. Recent advancement in the early detection of melanoma using computerized tools: An image analysis perspective
Singh et al. Designing a retrieval-based diagnostic aid using effective features to classify skin lesion in dermoscopic images
Obukhova et al. Automated image analysis in multispectral system for cervical cancer diagnostic
Hatanaka et al. Improvement of automated detection method of hemorrhages in fundus images
US12067726B2 (en) Retina image annotation, and related training methods and image processing models
Gaber et al. Comprehensive assessment of facial paralysis based on facial animation units
Thamm et al. Artificial intelligence‐based PRO score assessment in actinic keratoses from LC‐OCT imaging using Convolutional Neural Networks
Pershin et al. Artificial intelligence for the analysis of workload-related changes in radiologists’ gaze patterns
Masood et al. Enhancing Skin Cancer Detection: A Study on Feature Selection Methods for Image Classification
US12087445B2 (en) Automatic cervical cancer diagnosis system
KR102809111B1 (ko) 합성곱 신경망의 계층 연관성 전파를 이용한 퇴행성 뇌질환 이미지 분류 장치 및 방법
Udrea et al. Real-time acquisition of quality verified nonstandardized color images for skin lesions risk assessment—A preliminary study
WO2024102061A1 (fr) Procédé de dépistage du corps entier pour la détection d'un mélanome
Basra et al. Exploring the utilization of imaging modalities in the diagnosis of basal cell carcinoma: A scoping review
Odstrcilik et al. Analysis of retinal nerve fiber layer via Markov random fields in color fundus images

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23809353

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 23809353

Country of ref document: EP

Kind code of ref document: A1