Incorporating Clinical Guidelines through Adapting Multi-modal Large Language Model for Prostate Cancer PI-RADS Scoring

TL;DR

This paper presents a novel multi-modal large language model approach that incorporates clinical guidelines into prostate cancer MRI scoring, improving accuracy without additional annotations or network complexity.

Contribution

It introduces a two-stage fine-tuning process to adapt a multi-modal large language model for PI-RADS scoring by integrating clinical guidelines directly into the model.

Findings

Improved scoring accuracy on public and in-house datasets.

Effective integration of clinical guidelines without extra annotations.

Enhanced model performance through feature distillation.

Abstract

The Prostate Imaging Reporting and Data System (PI-RADS) is pivotal in the diagnosis of clinically significant prostate cancer through MRI imaging. Current deep learning-based PI-RADS scoring methods often lack the incorporation of common PI-RADS clinical guideline~(PICG) utilized by radiologists, potentially compromising scoring accuracy. This paper introduces a novel approach that adapts a multi-modal large language model (MLLM) to incorporate PICG into PI-RADS scoring model without additional annotations and network parameters. We present a designed two-stage fine-tuning process aiming at adapting a MLLM originally trained on natural images to the MRI images while effectively integrating the PICG. Specifically, in the first stage, we develop a domain adapter layer tailored for processing 3D MRI inputs and instruct the MLLM to differentiate MRI sequences. In the second stage, we translate PICG for guiding instructions from the model to generate PICG-guided image features. Through such a feature distillation step, we align the scoring network's features with the PICG-guided image features, which enables the model to effectively incorporate the PICG information. We develop our model on a public dataset and evaluate it on an in-house dataset. Experimental results demonstrate that our approach effectively improves the performance of current scoring networks. Code is available at: https://github.com/med-air/PICG2scoring