# Beyond Benchmark Scores: Auditing Medical Vision-Language Models for Chest X-Ray Tuberculosis Screening

arXiv: https://arxiv.org/abs/2609.21763 · HTML: https://mlsanity.com/paper/2609.21763/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Medical vision-language models, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 16 of its 31 references that are in the map.
- First posted on arXiv 18 Sep 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Medical vision-language models](https://mlsanity.com/mllm/medical-vision-language-models/)

## Builds on (most cited first, 16 of 16)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Reproducible scaling laws for contrastive language-image learning](https://mlsanity.com/paper/2212.07143/) (2022)
- [Selective Classification for Deep Neural Networks](https://mlsanity.com/paper/1705.08500/) (2017)
- [BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image-text pairs](https://mlsanity.com/paper/2303.00915/) (2023)
- [MedGemma Technical Report](https://mlsanity.com/paper/2507.05201/) (2025)
- [Confounding variables can degrade generalization performance of radiological deep learning models](https://mlsanity.com/paper/1807.00431/) (2018)
- [VinDr-CXR: An open dataset of chest X-rays with radiologist's annotations](https://mlsanity.com/paper/2012.15029/) (2020)
- [Hidden Stratification Causes Clinically Meaningful Failures in Machine Learning for Medical Imaging](https://mlsanity.com/paper/1909.12475/) (2019)
- [FUTURE-AI: International consensus guideline for trustworthy and deployable artificial intelligence in healthcare](https://mlsanity.com/paper/2309.12325/) (2023)
- [FairMedFM: Fairness Benchmarking for Medical Imaging Foundation Models](https://mlsanity.com/paper/2407.00983/) (2024)
- [MediConfusion: Can you trust your AI radiologist? Probing the reliability of multimodal medical foundation models](https://mlsanity.com/paper/2409.15477/) (2024)
- [Robustness tests for biomedical foundation models should tailor to specifications](https://mlsanity.com/paper/2502.10374/) (2025)
- [How Does Diverse Interpretability of Textual Prompts Impact Medical Vision-Language Zero-Shot Tasks?](https://mlsanity.com/paper/2409.00543/) (2024)
- [A data- and compute-efficient chest X-ray foundation model beyond aggressive scaling](https://mlsanity.com/paper/2602.22843/) (2026)

## Abstract (arXiv)

A medical model's benchmark score does not establish that the same conclusion holds under a different evaluation. This study tests whether claims about model ranking, score reliability and screening performance survive changes in cohort, prompt, negative spectrum, specified prevalence and operating threshold. We audit three medical vision-language models (BioMedCLIP, CheXficient, and MedSigLIP) and a general-domain OpenCLIP comparator on 12,200 chest radiograph records from four datasets (Montgomery, Shenzhen, TBX11K, and VinDr-CXR). Five fixed prompt families yield 244,000 model--image--prompt scores. No model leads every cohort and reliability criterion. Prompt-family changes alter AUROC in 21 of 48 multiplicity-controlled comparisons. Replacing healthy controls with sick non-tuberculosis controls reduces AUROC by 0.075--0.306 across all four models. On VinDr-CXR, the three medical models distinguish tuberculosis from no-finding controls substantially better than from pneumonia or lung tumor; their AUROC point estimates for both named diseases fall below 0.5. CheXficient has documented VinDr-CXR pretraining exposure, which limits the interpretation of its results. Thresholds chosen for 95% sensitivity on TBX11K training retain that constraint by point estimate in only four of sixteen target evaluations. A five-seed supervised source model reaches 0.999 AUROC on TBX11K validation but 0.629 on each of two external cohorts. Conservative exclusion of perceptual-overlap candidates narrows this gap without closing it. These retrospective, single-task results show that discrimination, score reliability and threshold retention support different portability claims. Evidence for chest X-ray tuberculosis screening should identify the complete evaluation specification rather than attribute clinical portability to a checkpoint alone.