Abstract: A multi-model medical scan analysis system is operable to generate a generic model by performing a training step on image data of a plurality of medical scans and corresponding labeling data. A plurality of fine-tuned models are generated by performing a fine-tuning step on the generic model. Abnormality detection data is generated for a new medical scan by utilizing the generic model. A first one of the plurality of abnormality types that is detected in the new medical scan is determined based on a corresponding one of the plurality of probability values. Additional abnormality data is generated by performing a fine-tuned inference function on the image data of the new medical scan that utilizes one of the plurality of fine-tuned models that corresponds to the first one of the plurality of abnormality types. The additional abnormality data is transmitted for display.

Abstract: A model-assisted annotating system is operable to receive a first set of annotation data for a first set of medical scans from a set of client devices. A computer vision model is trained by utilizing first set of medical scans and the first set of annotation data. A second set of annotation data for a second set of medical scans is generated by utilizing the computer vision model. The second set of medical scans and the second set of annotation data is transmitted to the set of client devices, and a set of additional annotation data is received in response. An updated computer vision model is generated by utilizing the set of additional annotation data. A third set of annotation data is generated for a third set of medical scans by utilizing the updated computer vision model for transmission to the set of client devices for display.

Abstract: A report generating system is operable to generate inference data for a medical scan indicating a first subset of a plurality of anatomical features of the medical scan are normal. A set of default natural language text corresponding to the first subset of the plurality of anatomical features are identified based on report template data. Preliminary report data is generated to include the set of default natural language text corresponding to the first subset of the plurality of anatomical features based on the inference data. The preliminary report data is displayed an interactive user interface, and review data is received based on user input in response to at least one prompt displayed via the interactive user interface. Final report data that includes natural language text data for each of the plurality of report sections is generated based on the review data.

Abstract: A peer-review flagging system is operable to train a computer vision model and to generate automated assessment data by performing an inference function on a first medical scan by utilizing the computer vision model. Human assessment data is generated based on a first medical report written by a medical professional in conjunction with review of the first medical scan. First consensus data is generated based on the automated assessment data, the human assessment data, and a first threshold, and the first medical scan is determined to be flagged based on the first consensus data. A second threshold is selected use in generating second consensus data for a second medical scan and a second medical report written by the medical professional in conjunction with review of the second medical scan, and is selected to be stricter than the first threshold based on determining to flag the first medical scan.

Abstract: A medical scan artifact detection system is operable to receive a medical scan of a patient. Artifact detection data is generated by executing an artifact detection function on the medical scan, where the artifact detection data indicates at least one artifact detected in the medical scan that includes a motion artifact or a nipple shadow. A notification is generated for display via a display device, where the notification indicates the at least one artifact.

Abstract: A medical scan triaging system is operable to generate a global abnormality probability for each of a plurality of medical scans by utilizing a computer vision model trained on a training set of medical scans. A triage probability threshold is determined based on user input to a client device. A first subset of the plurality of medical scans, designated for human review, is determined by identifying medical scans with a corresponding global abnormality probability that compares favorably to the triage probability threshold. A second subset of the plurality of medical scans, designated as normal, is determined by identifying ones of the plurality of medical scans with a corresponding global abnormality probability that compares unfavorably to the triage probability threshold. Transmission of the first subset of the plurality of medical scans to a plurality of client devices associated with a plurality of users is facilitated.

Abstract: An automatic patient recruitment system is operable generate abnormality data for medical scans by performing at least one inference function on image data of each medical scans by utilizing a computer vision model trained on a training set of medical scans. A subset of a plurality of patients is identified to be eligible for a pharmaceutical study by identifying medical scans having abnormality data that compares favorably to abnormality criteria of the pharmaceutical study. A size of the subset is compared to a minimum participant count requirement. A notification indicating the subset of the plurality of patients is transmitted based on the size of the subset comparing favorably to the minimum participant count requirement.

Abstract: A method comprises displaying, via an interactive interface, a medical scan and a plurality of prompts of each prompt decision tree of a plurality of prompt decision trees in succession, beginning with automatically determined starting prompts of each prompt decision tree, in accordance with corresponding nodes of each prompt decision tree until a leaf node of each prompt decision tree is ultimately selected. Labeling data indicating the ultimately selected leaf node of each prompt decision tree is determined for the medical scan.

Abstract: A medical scan interface feature evaluator system is operable to generate an ordered image-to-prompt mapping by selecting a set of user interface features to be displayed with each of an ordered set of medical scans. The set of medical scans and the ordered image-to-prompt mapping are transmitted to a set of client devices. A set of responses are generated by each client device in response to sequentially displaying each of the set of medical scans in conjunction with a mapped user interface feature indicated in the ordered image-to-prompt mapping via a user interface. Response score data is generated by comparing each response to truth annotation data of the corresponding medical scan. Interface feature score data corresponding to each user interface feature is generated based on aggregating the response score data, and is used to generate a ranking of the set of user interface features.

Abstract: A medical scan labeling quality assurance system is operable to transmit a selected set of medical scans to a set of client devices associated with an expert user and a selected set of users. The client devices display medical scans are displayed to the expert user and the set of users, and a set of labeling data generated via user input to each client device is received from each client device. A set of performance score data is generated based on comparing each set of labeling data to a set of golden labeling data that was received from the client device of the expert user. The set of performance score data is used to update user profiles of the set of users, and is transmitted to the set of client devices for display to the set of users.

Abstract: A peer-review flagging system is operable to receive a medical scan and a medical report written by a medical professional in conjunction with review of the medical scan. Automated assessment data is generated by performing an inference function on the medical scan by utilizing a computer vision model trained on a plurality of medical scans. Human assessment data is generated by performing an extraction function on the medical report. Consensus data is generated by comparing the automated assessment data to the first human assessment data. A peer-review notification is transmitted to a client device for display. The peer-review notification indicates the medical scan is flagged for peer-review in response to determining the consensus data indicates the automated assessment data compares unfavorably to the human assessment data.

Abstract: An automatic patient recruitment system is operable to determine a set of eligibility criteria, which includes abnormality criteria and other patient criteria, for each of a plurality of pharmaceutical studies. Abnormality data is generated for received medical scans by performing at least one inference function on image data of each medical scans by utilizing a computer vision model trained on a training set of medical scans. One of a plurality of patients is identified to be eligible for a pharmaceutical study by determining a medical scan of the patient has abnormality data that compares favorably to the abnormality criteria of the pharmaceutical study and by determining that the patient has patient data that compares favorably to the other patient criteria of the pharmaceutical study. A notification indicating the identified patient is eligible for the pharmaceutical study is generated for transmission to a client device.

Abstract: A medical scan hierarchical labeling system stores labeling application data that includes application operational instructions and a plurality of prompt decision trees. A medical scan and the labeling application data are sent to a client device for storage. The client device executes the application operational instructions of the labeling application data, causing the client device to display, via an interactive interface, the medical scan and a plurality of prompts of each prompt decision tree in succession, beginning with automatically determined starting prompts of each prompt decision tree, in accordance with corresponding nodes of each prompt decision tree until a leaf node of each prompt decision tree is ultimately selected. The client device transmits labeling data indicating the ultimately selected leaf node of each prompt decision tree. A medical scan entry of the medical scan in a medical scan database is populated based on the set of labels.

Abstract: An intensity transform augmentation system is operable to generate a plurality of sets of augmented images by performing a set of intensity transformation functions on each of a training set of medical scans. Each of the set of intensity transformation functions are based on density properties of corresponding anatomy feature present in the training set of medical scans. A computer vision model is generated by performing a training step on the plurality of sets of augmented images, where each augmented image of a set of augmented images is assigned same output label data based on a corresponding one of the training set of medical scans. Inference data is generated by performing an inference function on a new medical scan by utilizing the computer vision model on the new medical scan. The inference data is transmitted to a client device for display via a display device.

Abstract: A multi-model medical scan analysis system is operable to generate a generic model by performing a training step on image data of a plurality of medical scans and corresponding labeling data. A plurality of fine-tuned models corresponding to one of a plurality of abnormality types can be generated by performing a fine-tuning step on the generic model. Abnormality detection data can be generated for a new medical scan by performing utilizing the generic model. One of the plurality of abnormality types is determined to be detected in the new medical scan based on the abnormality detection data, and a fine-tuned model that corresponds to the abnormality type is selected. Additional abnormality data is generated for the new medical scan by utilizing the selected fine-tuned model. The additional abnormality data can be transmitted to a client device for display via a display device.

Abstract: A medical scan annotator system is operable to select a medical scan for transmission via a network to a first client device and a second client device for display via an interactive interface, and annotation data is received from the first client device and the second client device in response. Annotation similarity data is generated by comparing the first annotation data to the second annotation data, and consensus annotation data is generated based on the first annotation data and the second annotation data in response to the annotation similarity data indicating that the difference between the first annotation data and the second annotation data compares favorably to an annotation discrepancy threshold. The consensus annotation data is mapped to the medical scan in a medical scan database.

Abstract: A multi-label heat map display system is operable to receive a medical scan and a set of heat maps set of heat maps that each correspond to probability matrix data generated for each of a set of abnormality classes. An interactive interface that displays image data of the medical scan and at least one of the set of heat maps is generated for display on a display device associated with the multi-label heat map display system. User input to a client device is received, and an updated interactive interface that includes a change to the display of the at least one of the set of heat maps by the second portion of the interactive interface in response to the user input is displayed.

Abstract: An intensity transform augmentation system is operable to generate a plurality of sets of augmented images by performing a set of intensity transformation functions on each of a training set of medical scans. Each of the set of intensity transformation functions are based on density properties of corresponding anatomy feature present in the training set of medical scans. A computer vision model is generated by performing a training step on the plurality of sets of augmented images, where each augmented image of a set of augmented images is assigned same output label data based on a corresponding one of the training set of medical scans. Inference data is generated by performing an inference function on a new medical scan by utilizing the computer vision model on the new medical scan. The inference data is transmitted to a client device for display via a display device.

Abstract: A multi-label heat map display system is operable to receive a medical scan and a set of heat maps set of heat maps that each correspond to probability matrix data generated for each of a set of abnormality classes. An interactive interface that displays image data of the medical scan and at least one of the set of heat maps is generated for display on a display device associated with the multi-label heat map display system. User input to a client device is received, and an updated interactive interface that includes a change to the display of the at least one of the set of heat maps by the second portion of the interactive interface in response to the user input is displayed.

Abstract: A triage routing system is operable to receive a medical scan via a receiver. Inference data for the medical scan is generated by performing an inference function, where the inference function utilizes a computer-vision model trained on a plurality of medical scans. One of a plurality of medical professionals is selected to review the medical scan based on the inference data. Triage routing data that indicates the medical scan and the one of the plurality of medical professionals is generated. The medical scan is transmitted to a client device associated with the one of the plurality of medical professionals for display via a display device in accordance with the triage routing data.