TRV-2026-2087Certified recordPeer-reviewed

TurkRadLLM: A Compact MedGemma-Based Model for Structuring Turkish Lumbar and Cervical Spine MRI Reports-A Multi-reader Validation Study

To develop and validate a compact, locally deployable MedGemma-based model converting unstructured Turkish spine MRI reports into structured six-field output. In this retrospective single-institution study, 5210 lumbar and 5017 cervical Turkish spine MRI reports were used to fine-tune MedGemma 4B IT with QLoRA. Silver-standard labels from a proprietary large language model served as training targets and as the reference for automated fidelity analyses, but not as the clinical reference standard. Evaluation compr…

Health · The Trace — both readings · certified 2026-10-21 · v1 · article view · machine-readable

Current reading — gain

A fine-tuned compact MedGemma model produced valid structured output from Turkish spine MRI reports that readers accepted, and it ran on a single consumer GPU card.

Current reading — problem

Negative-finding completeness was reader-dependent, with one reader judging negative-finding lists incomplete in 57.8% of reports, and fidelity to the silver-label impression reached only 67.1%.

What this doesn’t fix

Labels came from a proprietary language model rather than a clinical reference standard, the study was single-institution, and negative-finding completeness varied by reader.

Evidence

Reader signal

How should this claim be treated?

Cite this record

Truvace Impact Record TRV-2026-2087, v1: “TurkRadLLM: A Compact MedGemma-Based Model for Structuring Turkish Lumbar and Cervical Spine MRI Reports-A Multi-reader Validation Study.” Truvace, 2026-10-21. /record/TRV-2026-2087 (accessed at citation time). sha256 017f522217eccd36…

Calibration history

Every change to this record since certification, in the open. None yet — the reading has held since it entered the record.

  1. Certifiedv1017f522217ec…

    Certified into the record

Verify this record
How to verify without trusting this page

Fetch the canonical text of any version from /api/record/TRV-2026-2087 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.