ProFound: A moderate-sized vision foundation model for multi-task prostate imaging

TL;DR

ProFound is a multi-task vision foundation model trained on diverse prostate MRI data, demonstrating superior performance across various clinical tasks and offering a scalable, generalizable tool for prostate cancer imaging analysis.

Contribution

The paper introduces ProFound, a domain-specific vision foundation model for prostate MRI that leverages self-supervised pre-training on multi-institutional data, enabling effective multi-task clinical applications.

Findings

ProFound outperforms state-of-the-art models in multiple prostate MRI tasks.

Pre-training on diverse data improves generalization across clinical tasks.

The model is effective across a broad spectrum of diagnostic and segmentation tasks.

Abstract

Many diagnostic and therapeutic clinical tasks for prostate cancer increasingly rely on multi-parametric MRI. Automating these tasks is challenging because they necessitate expert interpretations, which are difficult to scale to capitalise on modern deep learning. Although modern automated systems achieve expert-level performance in isolated tasks, their general clinical utility remains limited by the requirement of large task-specific labelled datasets. In this paper, we present ProFound, a domain-specialised vision foundation model for volumetric prostate mpMRI. ProFound is pre-trained using several variants of self-supervised approaches on a diverse, multi-institutional collection of 5,000 patients, with a total of over 22,000 unique 3D MRI volumes (over 1,800,000 2D image slices). We conducted a systematic evaluation of ProFound across a broad spectrum of $11$ downstream clinical tasks on over 3,000 independent patients, including prostate cancer detection, Gleason grading, lesion localisation, gland volume estimation, zonal and surrounding structure segmentation. Experimental results demonstrate that finetuned ProFound consistently outperforms or remains competitive with state-of-the-art specialised models and existing medical vision foundation models trained/finetuned on the same data.