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 automatically processes 3D anatomical image and 3D functional image to identify, and measure uptake of radiopharmaceutical in, cancerous lesions within a subject having or at risk for a cancer. The method receives a 3D anatomical image obtained using an anatomical imaging modality, automatically identifies within the 3D anatomical image a first skeletal volume, a first aorta volume, and a first liver volume using one or more machine learning modules, and determines a 3D segmentation map representing a plurality of 3D segmentation masks including a skeletal mask, an aorta mask, and a liver mask.

The method receives a 3D functional image of the subject obtained using a functional imaging modality, and automatically identifies within the 3D functional image, using the 3D segmentation map, corresponding second skeletal, second aorta, and second liver volumes. The method then automatically detects within the second skeletal volume one or more hotspots determined to represent lesions based on intensities of voxels within the second skeletal volume.

For each detected hotspot, the method determines an individual hotspot index value by determining an aorta reference intensity level from voxels in the second aorta volume and a liver reference intensity level from voxels in the second liver volume. The corresponding individual hotspot intensity level is determined from voxels of the detected hotspot, and the corresponding individual hotspot index level is determined from the hotspot intensity level, the aorta reference intensity level, and the liver reference intensity level.

The invention is also recited as a system including a processor and memory instructions that cause the processor to perform the same automatic identification, segmentation map determination, hotspot detection, and reference-based hotspot indexing steps.

Claims Coverage

The provided material includes two independent claims, a method claim and a system claim. Both independent claims include four inventive features covering machine-learning based region identification, corresponding functional-image region identification, hotspot detection, and reference-based hotspot indexing.

Across both independent claims, the coverage centers on machine-learning based identification of skeletal, aorta, and liver regions in a 3D anatomical image to create a segmentation map, applying that segmentation map to identify corresponding regions in a 3D functional image, detecting lesion-representing hotspots within the skeletal region using voxel intensities, and computing an individual hotspot index level using hotspot intensity relative to aorta and liver reference intensity levels.

Stated Advantages

Documented Applications