Summary:
UCLA Researchers in the Department of Radiology have developed a computer-aided diagnostic model that uses histology slides to interpret and detect prostate cancer diagnoses with high levels of accuracy.
Background:
In the United States, prostate cancer ranks as the most prevalent and second deadliest cancer among men. Diagnosing this disease typically involves manual histology reviews of biopsied tissue, which helps determine both the cancer's aggressiveness and the appropriate treatment. During this process, a pathologist must meticulously examine histology slides to identify regions of interest (ROIs) that are crucial for accurate diagnosis. However, this method can be time-consuming and is susceptible to human error and variability. As a result, there is a growing demand for computer-aided diagnosis (CAD) systems to expedite and enhance the diagnostic process. Existing CAD approaches often require labor-intensive labeling and pre-identified ROIs provided by pathologists. Consequently, there is an urgent need for advanced CAD techniques capable of leveraging easily accessible pathology reports with minimal human intervention.
Innovation:
Dr. Arnold and colleagues developed a two-stage, attention-based multiple instance learning (MIL) model for slide-level cancer grading and weakly-supervised ROI detection. In contrast to previous models that necessitate manual identification of ROIs by pathologists, this innovative method is trained solely using slide-level labels (weak labels) readily obtainable from pathology reports. The two-stage model first employs a lower-resolution approach to pinpoint potential regions of interest. Subsequently, it conducts a more in-depth analysis using a higher resolution, mimicking the process a pathologist would undertake during a manual histology review. Remarkably, this model achieved state-of-the-art performance, boasting an 85% accuracy rate in classifying benign, low-grade, and high-grade biopsy slides during an independent test.
Potential Applications:
• Cancer diagnosis
• Histology screening
• Computer-aided diagnostics
Advantages:
• High accuracy
• Minimal human input required
• Operates using weak labels
Development to Date:
Successful demonstration of invention reported in an independent test
Related Papers:
An attention-based multi-resolution model for prostate whole slide imageclassification and localization
A multi-resolution model for histopathology image classification and localization with multiple instance learning
Reference:
UCLA Case No. 2019-943
Lead Inventor:
Corey Arnold