Scholarlink

<p>Aligning radiological features with clinical text descriptions remains a key challenge for zero-shot disease recognition in chest radiography. We propose DVLM (Dual-Head Vision-Language Model with Neural Memory), a framework combining Vision Transformer visual encoding with ClinicalBERT-based text processing through parallel contrastive and supervised learning branches. A neural memory module stores disease-relevant patterns during training for improved generalization to unseen pathologies. We evaluated DVLM on CheXpert, MIMIC-CXR, and PadChest using multi-seed validation (five seeds <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\times \)</EquationSource> <EquationSource Format="MATHML"><math> <mo>×</mo> </math></EquationSource> </InlineEquation> fivefold cross-validation), controlled ablation studies, and statistical significance testing. DVLM achieved 90.0% ± 0.28% macro-averaged AUROC on CheXpert (95% CI, 89.5–90.6%), with the neural memory module contributing +3.3% improvement (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(p&lt;0.001\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>p</mi> <mo>&lt;</mo> <mn>0.001</mn> </mrow> </math></EquationSource> </InlineEquation>, Cohen’s <InlineEquation ID="IEq3"> <EquationSource Format="TEX">\(d=0.89\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>d</mi> <mo>=</mo> <mn>0.89</mn> </mrow> </math></EquationSource> </InlineEquation>). For zero-shot classification (25% held-out diseases), DVLM achieved 73.5% AUROC, outperforming MedKLIP by 2.3%. Temperature scaling reduced calibration error by 72%, and Grad-CAM localization achieved an IoU of 0.642 against radiologist annotations. Subgroup analysis confirmed equitable performance across demographic groups (maximum disparity, 1.3%). While DVLM demonstrates strong ranking capability suitable for triage applications, threshold-based classification for rare diseases remains limited (F1, 24.8–30.1%), indicating the need for radiologist confirmation in clinical deployment.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Zero-Shot Lung Disease Detection Using Radiological Symptomatic Descriptors and Pretrained Neural Networks

  • 期刊论文

  • 发表日期：

    2026年3月30日

查看全文
  • Sabbir Ahmed,
  • Md. Abdul Hamid,
  • Muhammad Mostafa Monowar,
  • Mohammad Abu Yousuf

摘要

Aligning radiological features with clinical text descriptions remains a key challenge for zero-shot disease recognition in chest radiography. We propose DVLM (Dual-Head Vision-Language Model with Neural Memory), a framework combining Vision Transformer visual encoding with ClinicalBERT-based text processing through parallel contrastive and supervised learning branches. A neural memory module stores disease-relevant patterns during training for improved generalization to unseen pathologies. We evaluated DVLM on CheXpert, MIMIC-CXR, and PadChest using multi-seed validation (five seeds \(\times \) × fivefold cross-validation), controlled ablation studies, and statistical significance testing. DVLM achieved 90.0% ± 0.28% macro-averaged AUROC on CheXpert (95% CI, 89.5–90.6%), with the neural memory module contributing +3.3% improvement ( \(p<0.001\) p < 0.001 , Cohen’s \(d=0.89\) d = 0.89 ). For zero-shot classification (25% held-out diseases), DVLM achieved 73.5% AUROC, outperforming MedKLIP by 2.3%. Temperature scaling reduced calibration error by 72%, and Grad-CAM localization achieved an IoU of 0.642 against radiologist annotations. Subgroup analysis confirmed equitable performance across demographic groups (maximum disparity, 1.3%). While DVLM demonstrates strong ranking capability suitable for triage applications, threshold-based classification for rare diseases remains limited (F1, 24.8–30.1%), indicating the need for radiologist confirmation in clinical deployment.