Research

Paper

AI LLM March 12, 2026

Prototype-Based Knowledge Guidance for Fine-Grained Structured Radiology Reporting

Authors

Chantal Pellegrini, Adrian Delchev, Ege Özsoy, Nassir Navab, Matthias Keicher

Abstract

Structured radiology reporting promises faster, more consistent communication than free text, but automation remains difficult as models must make many fine-grained, discrete decisions about rare findings and attributes from limited structured supervision. In contrast, free-text reports are produced at scale in routine care and implicitly encode fine-grained, image-linked information through detailed descriptions. To leverage this unstructured knowledge, we propose ProtoSR, an approach for injecting free-text information into structured report population. First, we introduce an automatic extraction pipeline that uses an instruction-tuned LLM to mine 80k+ MIMIC-CXR studies and build a multimodal knowledge base aligned with a structured reporting template, representing each answer option with a visual prototype. Using this knowledge base, ProtoSR is trained to retrieve prototypes relevant for the current image-question pair and augment the model predictions through a prototype-conditioned residual, providing a data-driven second opinion that selectively corrects predictions. On the Rad-ReStruct benchmark, ProtoSR achieves state-of-the-art results, with the largest improvements on detailed attribute questions, demonstrating the value of integrating free-text derived signal for fine-grained image understanding.

Metadata

arXiv ID: 2603.11938
Provider: ARXIV
Primary Category: cs.AI
Published: 2026-03-12
Fetched: 2026-03-14 05:03

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.11938v1</id>\n    <title>Prototype-Based Knowledge Guidance for Fine-Grained Structured Radiology Reporting</title>\n    <updated>2026-03-12T13:51:13Z</updated>\n    <link href='https://arxiv.org/abs/2603.11938v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.11938v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Structured radiology reporting promises faster, more consistent communication than free text, but automation remains difficult as models must make many fine-grained, discrete decisions about rare findings and attributes from limited structured supervision. In contrast, free-text reports are produced at scale in routine care and implicitly encode fine-grained, image-linked information through detailed descriptions. To leverage this unstructured knowledge, we propose ProtoSR, an approach for injecting free-text information into structured report population. First, we introduce an automatic extraction pipeline that uses an instruction-tuned LLM to mine 80k+ MIMIC-CXR studies and build a multimodal knowledge base aligned with a structured reporting template, representing each answer option with a visual prototype. Using this knowledge base, ProtoSR is trained to retrieve prototypes relevant for the current image-question pair and augment the model predictions through a prototype-conditioned residual, providing a data-driven second opinion that selectively corrects predictions. On the Rad-ReStruct benchmark, ProtoSR achieves state-of-the-art results, with the largest improvements on detailed attribute questions, demonstrating the value of integrating free-text derived signal for fine-grained image understanding.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <published>2026-03-12T13:51:13Z</published>\n    <arxiv:primary_category term='cs.AI'/>\n    <author>\n      <name>Chantal Pellegrini</name>\n    </author>\n    <author>\n      <name>Adrian Delchev</name>\n    </author>\n    <author>\n      <name>Ege Özsoy</name>\n    </author>\n    <author>\n      <name>Nassir Navab</name>\n    </author>\n    <author>\n      <name>Matthias Keicher</name>\n    </author>\n  </entry>"
}