# Preprint reports faster, better-aligned AI evaluation of radiology reports

_Published Friday, October 16, 2026 at 3:59 AM EDT · AI, Science · Latest · Tier 2 — Notable_

Researchers report that lightweight fine-tuning of Qwen3-30B improved agreement with radiologists by up to 25% and reduced latency by 37.2x in a preprint evaluating AI systems that judge radiology reports. The study used two expert-annotated datasets spanning multiple imaging modalities and anatomical regions.

Their proposed evaluation metric, VERT, improved Kendall's correlation with radiologist judgment by up to 26% over GREEN, the strongest existing metric tested. Although the automated metrics correlated with expert ratings overall, they tended to miscount errors such as omissions.

## Sources

- [cs.AI updates on arXiv.org](https://arxiv.org/abs/2604.03376)

---
Canonical: https://techandbusiness.org/newswire/KP8EAtamR4BbKKMJriLKUv
Published: 2026-10-16T07:59:41.331Z
Story chronology: 2026-10-16T04:00:00.000Z
Retrieved: 2026-10-16T10:21:42.343Z
Publisher: Tech & Business (techandbusiness.org)