Paper Detail

DDX-TRACE: A Benchmark for Medical Diagnostic Trajectories in VLMs

Jiazhen Pan, Weixiang Shen, Jun Li, Julian Canisius, Felix Bitzer, Paula Roßmüller, Jiancheng Yang, Virginie Kreutzinger, Daniel Rueckert, Benedikt Wiestler

arxiv Score 19.6

Published 2026-05-22 · First seen 2026-05-25

General AI

Abstract

Medical diagnosis is not a single prediction from a fully specified vignette. It is a sequential workup: clinicians decide what evidence to obtain, revise a differential diagnosis, and stop when the diagnosis is sufficiently supported. Most medical AI benchmarks instead reveal the relevant context upfront and score only the final answer, making unsupported correct guesses, premature closure, inefficient workups, and poor uncertainty updating invisible. We introduce DDX-TRACE, a physician-adjudicated benchmark for multimodal neuroradiology that evaluates diagnostic trajectories under hidden evidence over 211 challenging cases. Each case begins with limited clinical history; models request imaging studies in free form, receive matched image bundles when available, update a probabilistic differential diagnosis after each turn, and stop with a localized final diagnosis. Evaluating state-of-the-art VLMs, we find that final diagnosis scores can substantially misrepresent workup quality: models may guess plausible diagnoses without essential evidence, request useful studies but misinterpret raw images, or acquire evidence inefficiently while updating uncertainty poorly. Controlled evidence variants isolate bottlenecks in planning, visual evidence extraction, and downstream differential reasoning. DDX-TRACE shifts medical AI evaluation from final answers to evidence-supported diagnostic trajectories.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{pan2026ddx,
  title = {DDX-TRACE: A Benchmark for Medical Diagnostic Trajectories in VLMs},
  author = {Jiazhen Pan and Weixiang Shen and Jun Li and Julian Canisius and Felix Bitzer and Paula Roßmüller and Jiancheng Yang and Virginie Kreutzinger and Daniel Rueckert and Benedikt Wiestler},
  year = {2026},
  abstract = {Medical diagnosis is not a single prediction from a fully specified vignette. It is a sequential workup: clinicians decide what evidence to obtain, revise a differential diagnosis, and stop when the diagnosis is sufficiently supported. Most medical AI benchmarks instead reveal the relevant context upfront and score only the final answer, making unsupported correct guesses, premature closure, inefficient workups, and poor uncertainty updating invisible. We introduce DDX-TRACE, a physician-adjudic},
  url = {https://arxiv.org/abs/2605.23629},
  keywords = {cs.CV},
  eprint = {2605.23629},
  archiveprefix = {arXiv},
}

Metadata

{}