# CT-Flow: Orchestrating CT Interpretation Workflow with Model Context Protocol Servers

Source: [CT-Flow: Orchestrating CT Interpretation Workflow with Model Context Protocol Servers](<https://www.sota2.com/research/paper/69a66bdd2d603e6de3e3ccb2>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-04T05:00:05.078Z

Authors: Yannian Gu, Xizhuo Zhang, Linjie Mu, Yongrui Yu, Zhongzhen Huang, Shaoting Zhang, Xiaofan Zhang

Published: 2026-02-23T07:19:30.000Z

## Abstract

Recent advances in Large Vision-Language Models (LVLMs) have shown strong potential for multi-modal radiological reasoning, particularly in tasks like diagnostic visual question answering (VQA) and radiology report generation. However, most existing approaches for 3D CT analysis largely rely on static, single-pass inference. In practice, clinical interpretation is a dynamic, tool-mediated workflow where radiologists iteratively review slices and use measurement, radiomics, and segmentation tools to refine findings. To bridge this gap, we propose CT-Flow, an agentic framework designed for interoperable volumetric interpretation. By leveraging the Model Context Protocol (MCP), CT-Flow shifts from closed-box inference to an open, tool-aware paradigm. We curate CT-FlowBench, the first large-scale instruction-tuning benchmark tailored for 3D CT tool-use and multi-step reasoning. Built upon this, CT-Flow functions as a clinical orchestrator capable of decomposing complex natural language queries into automated tool-use sequences. Experimental evaluations on CT-FlowBench and standard 3D VQA datasets demonstrate that CT-Flow achieves state-of-the-art performance, surpassing baseline models by 41% in diagnostic accuracy and achieving a 95% success rate in autonomous tool invocation. This work provides a scalable foundation for integrating autonomous, agentic intelligence into real-world clinical radiology.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Medical Vision-Language Modeling on 3D-RAD](<https://www.sota2.com/research/sota/medical-vision-language-modeling-on-3d-rad.md>): Medical Vision-Language Modeling; 3D-RAD; BLEU-4: 36.96.

- [Medical Vision-Language Modeling on CT-FlowBench](<https://www.sota2.com/research/sota/medical-vision-language-modeling-on-ct-flowbench.md>): Medical Vision-Language Modeling; CT-FlowBench; QA Score: 43.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.00123>)

[PDF](<https://arxiv.org/pdf/2603.00123.pdf>)