Cinepro: Robust Training of Foundation Models for Cancer Detection in Prostate Ultrasound Cineloops

TL;DR

Cinepro is a novel framework that enhances foundation models for prostate cancer detection in ultrasound cineloops by incorporating pathology-informed loss functions and leveraging temporal data, resulting in improved localization and accuracy.

Contribution

It introduces a robust training approach that integrates pathology data into the loss function and utilizes temporal information to improve cancer localization in ultrasound images.

Findings

Achieved AUROC of 77.1% on multi-center dataset.

Attained balanced accuracy of 83.8%, surpassing benchmarks.

Demonstrated improved robustness to label noise in ultrasound data.

Abstract

Prostate cancer (PCa) detection using deep learning (DL) models has shown potential for enhancing real-time guidance during biopsies. However, prostate ultrasound images lack pixel-level cancer annotations, introducing label noise. Current approaches often focus on limited regions of interest (ROIs), disregarding anatomical context necessary for accurate diagnosis. Foundation models can overcome this limitation by analyzing entire images to capture global spatial relationships; however, they still encounter challenges stemming from the weak labels associated with coarse pathology annotations in ultrasound data. We introduce Cinepro, a novel framework that strengthens foundation models' ability to localize PCa in ultrasound cineloops. Cinepro adapts robust training by integrating the proportion of cancer tissue reported by pathology in a biopsy core into its loss function to address label noise, providing a more nuanced supervision. Additionally, it leverages temporal data across multiple frames to apply robust augmentations, enhancing the model's ability to learn stable cancer-related features. Cinepro demonstrates superior performance on a multi-center prostate ultrasound dataset, achieving an AUROC of 77.1% and a balanced accuracy of 83.8%, surpassing current benchmarks. These findings underscore Cinepro's promise in advancing foundation models for weakly labeled ultrasound data.