LinGuinE: Longitudinal Guidance Estimation for Volumetric Tumour Segmentation

TL;DR

LinGuinE is a flexible PyTorch framework that combines image registration and guided segmentation to improve longitudinal tumour segmentation and tracking without requiring training on longitudinal data.

Contribution

It introduces a novel, versatile framework that enables lesion tracking and segmentation across multiple scans using a single prompt, without the need for longitudinal training data.

Findings

Achieves state-of-the-art segmentation and tracking performance.

Minimal degradation in segmentation with increasing temporal separation.

Framework is temporally direction agnostic and adaptable.

Abstract

Longitudinal volumetric tumour segmentation is critical for radiotherapy planning and response assessment, yet this problem is underexplored and most methods produce single-timepoint semantic masks, lack lesion correspondence, and offer limited radiologist control. We introduce LinGuinE (Longitudinal Guidance Estimation), a PyTorch framework that combines image registration and guided segmentation to deliver lesion-level tracking and volumetric masks across all scans in a longitudinal study from a single radiologist prompt. LinGuinE is temporally direction agnostic, requires no training on longitudinal data, and allows any registration and semi-automatic segmentation algorithm to be repurposed for the task. We evaluate various combinations of registration and segmentation algorithms within the framework. LinGuinE achieves state-of-the-art segmentation and tracking performance across four datasets with a total of 456 longitudinal studies. Tumour segmentation performance shows minimal degradation with increasing temporal separation. We conduct ablation studies to determine the impact of autoregression, pathology specific finetuning, and the use of real radiologist prompts. We release our code and substantial public benchmarking for longitudinal segmentation, facilitating future research.