LoGSAM: Parameter-Efficient Cross-Modal Grounding for MRI Segmentation

TL;DR

LoGSAM introduces a parameter-efficient, speech-guided MRI tumor segmentation framework that leverages pretrained foundation models with minimal updates, achieving state-of-the-art accuracy and high case-level reliability.

Contribution

It presents a novel modular pipeline combining speech transcription, NLP, vision-language detection, and segmentation models with minimal parameter updates for MRI tumor segmentation.

Findings

Achieved 80.32% dice score on BRISC 2025 dataset.

Attained 91.7% case-level accuracy on unseen MRI scans.

Enabled efficient domain adaptation with only 5% of model parameters updated.

Abstract

Precise localization and delineation of brain tumors using Magnetic Resonance Imaging (MRI) are essential for planning therapy and guiding surgical decisions. However, most existing approaches rely on task-specific supervised models and are constrained by the limited availability of annotated data. To address this, we propose LoGSAM, a parameter-efficient, detection-driven framework that transforms radiologist dictation into text prompts for foundation-model-based localization and segmentation. Radiologist speech is first transcribed and translated using a pretrained Whisper ASR model, followed by negation-aware clinical NLP to extract tumor-specific textual prompts. These prompts guide text-conditioned tumor localization via a LoRA-adapted vision-language detection model, Grounding DINO (GDINO). The LoRA adaptation updates using 5% of the model parameters, thereby enabling computationally efficient domain adaptation while preserving pretrained cross-modal knowledge. The predicted bounding boxes are used as prompts for MedSAM to generate pixel-level tumor masks without any additional fine-tuning. Conditioning the frozen MedSAM on LoGSAM-derived priors yields a state-of-the-art dice score of 80.32% on BRISC 2025. In addition, we evaluate the full pipeline using German dictations from a board-certified radiologist on 12 unseen MRI scans, achieving 91.7% case-level accuracy. These results highlight the feasibility of constructing a modular, speech-to-segmentation pipeline by intelligently leveraging pretrained foundation models with minimal parameter updates.