CT-$Δ$Bench: A Benchmark for Longitudinal 3D Medical Imaging Difference Reporting with Vision-Language Models

2026-08-11

AI Summary

Existing medical foundation models struggle to capture temporal dynamics in longitudinal CT imaging, limiting their clinical utility for lesion evolution analysis. To address this gap, this work introduces CT-ΔBench—the first benchmark specifically designed for generating radiology reports that describe differences between paired time-point CT scans. The benchmark features patient-level data splits to prevent information leakage, change-aware evaluation metrics, and expert clinician validation. Building upon a vision–language architecture, we develop DeltaMed, a baseline model that integrates 3D image registration, change semantics extraction, and generative report synthesis. Experimental results demonstrate that direct two-timepoint inference significantly outperforms conventional two-stage approaches, with generated reports exhibiting clinically reliable descriptions of pathological events and accurate capture of temporal changes as assessed by radiologists.

Abstract

In medical imaging, the clinical value of Computed Tomography (CT) lies not only in depicting current disease status, but crucially in enabling longitudinal comparison of serial scans to determine disease evolution, a process that underpins response assessment, recurrence detection, and ongoing patient management. Yet, despite this central role of temporal comparison in clinical decision-making, existing medical foundation models remain largely confined to single-study understanding, leaving temporally grounded cross-examination insufficiently addressed. To address this gap, we study longitudinal imaging difference reporting, a task in which a model takes two temporally separated scans from the same patient and generates a clinically meaningful report describing interval changes between them. We introduce CT-$Δ$Bench, a dedicated benchmark for this task with patient-level splitting to prevent information leakage. To better evaluate this task beyond surface-level text similarity, we further develop change-aware metrics specifically designed to capture clinically meaningful longitudinal changes, and conduct an independent physician validation to assess the reliability of the synthesized references and event extraction pipeline. We also compare direct paired-CT reasoning with an indirect two-stage pipeline that first generates single-timepoint reports and then performs textual differencing. Finally, we propose DeltaMed, a baseline model for direct paired-CT difference reporting, and train it on the benchmark training set. Together, these contributions lay the groundwork for temporally aware medical foundation models that better reflect real-world longitudinal clinical reasoning.

Problem

• longitudinal medical imaging
• difference reporting
• temporal comparison
• CT scans
• vision-language models

Innovation

• longitudinal medical imaging
• vision-language models
• difference reporting
• change-aware metrics
• DeltaMed