arXiv cs.AI· Hrishikesh Singh, Aarti Sharma, Divyanshu Kumar, Millie Pant·· 13 小时前AI 评分29
从分割掩码读取 BI-RADS 描述符：乳腺 MRI 分割的可解释性理由生成
Reading the Mask: Descriptor-Grounded BI-RADS Rationale Generation for Breast MRI Segmentation
AI 导读
研究团队基于 BI-RADS 指南设计含 8 个节点的决策树，用影像组学等 8 项定量测量结合病灶大小与成分数量推导描述符，再由大语言模型生成自然语言理由。在 MAMA-MIA 数据集全部 1,506 例的专家掩码与自动掩码上共生成 3,012 条 BI-RADS 理由，忠实度达 99.98%，反事实准确率 99.3%。
正文 · 原文
Abstract:Deep learning based models accurately delineate tumours on dynamic contrast-enhanced MRI (DCE-MRI), but segmentation masks alone do not provide the rationales or reasoning used by radiologists to assess the lesion suspicion. This limits the explainability as well the clinical integration. We investigate whether clinically relevant descriptors can instead be derived from the segmentation masks and their dynamic contrast-enhancement series. We have used Breast Imaging Reporting and Data System (BI-RADS) guidelines as a structured framework. Using these guidelines we designed a BI-RADS decision tree with eight quantitative measurements using radiomics, together with the lesion size and component count, that determine the descriptors at eight nodes. Using these descriptors we generate natural-language rationale using a large language model. The model was constrained to express only the derived records from the decision tree. After applying this framework to the expert and automatic masks of each of the 1,506 cases in the MAMA-MIA dataset, we generated a reproducible corpus of 3,012 BI-RADS grounded rationales. We obtained a faithfulness score of 99.98% and a counterfactual accuracy of 99.3%, indicating that the generated rationales faithfully verbalises the measured descriptors and correctly reflects changes in the underlying record. Agreement between the descriptors from expert and automatic masks falls from 0.832 to 0.549 as mask quality falls from Good to Missed. The generated rationales are intended as draft representations for radiologist review and final sign-off rather than autonomous clinical reports.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.29358 [cs.CV] |
| (or arXiv:2610.29358v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.29358 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Hrishikesh Singh [view email]
[v1]
Wed, 21 Oct 2026 07:50:20 UTC (3,283 KB)
来源：arXiv cs.AI · arxiv.org