Systems and methods for platform agnostic whole body image segmentation

Inventors

Richter, Jens Filip Andreas • Johnsson, Kerstin Elsa Maria • Gjertsson, Erik Konrad • Anand, Aseem Undvall

Assignees

Exini Diagnostics AB

Abstract

Presented herein are systems and methods that provide for automated analysis of three-dimensional (3D) medical images of a subject in order to automatically identify specific 3D volumes within the 3D images that correspond to specific anatomical regions (e.g., organs and/or tissue). Notably, the image analysis approaches described herein are not limited to a single particular organ or portion of the body. Instead, they are robust and widely applicable, providing for consistent, efficient, and accurate detection of anatomical regions, including soft tissue organs, in the entire body. In certain embodiments, the accurate identification of one or more such volumes is used to automatically determine quantitative metrics that represent uptake of radiopharmaceuticals in particular organs and/or tissue regions. These uptake metrics can be used to assess disease state in a subject, determine a prognosis for a subject, and/or determine efficacy of a treatment modality.

Core Innovation

The invention provides a system for automatically segmenting a first group of tissue regions depicted within a 3D anatomical image of a subject. The system receives the 3D anatomical image and segments it using a selected trained machine learning module from a set of trained machine learning modules to produce a 3D segmentation map identifying corresponding 3D volumes.

The system receives a 3D functional image obtained following administration of a radiopharmaceutical comprising a prostate-specific membrane antigen (PSMA) binding agent. The system transfers the 3D segmentation map to the 3D functional image to identify a second plurality of 3D volumes corresponding to the first plurality of 3D volumes, and detects hotspots representing cancerous lesions within the 3D functional image using the identified second plurality of 3D volumes.

In embodiments directed to pelvic or abdomen regions, the system identifies a volume of interest (VOI) for the pelvic region or abdomen, segments within that VOI using a region-specific trained machine learning module, and performs hotspot detection within the identified functional volumes. The pelvic region module is trained to identify members consisting of left ilium, right ilium, prostate, urinary bladder, sacrum, and coccyx, and the abdomen module is trained to identify members consisting of liver, left kidney, right kidney, and gallbladder.

Claims Coverage

The independent-claim coverage includes three inventive features. Collectively, they cover automated 3D anatomical segmentation with trained machine learning modules, transfer of the segmentation to PSMA radiopharmaceutical functional images, and detection of hotspots representing cancerous lesions, with additional pelvic and abdomen region limitations.

The independent claims cover automated segmentation of 3D tissue regions using selected trained machine learning modules, alignment of those segmented volumes to PSMA radiopharmaceutical 3D functional images by transferring the segmentation map, and detection of cancerous-lesion hotspots within the functional volumes.

Stated Advantages

Documented Applications

No documented applications found