A SAM2 and VLM-Based Automatic Radiological Reporting System for Pneumothorax Detection
- 1Faculty of Technology, Department of Software Engineering Firat University, Elazig 23000, Türkiye;
- 2Faculty of Medicine Hospital, Department of Radiology, Firat University, Elazig 23000, Türkiye;
Abstract
Background/Objectives: Pneumothorax, caused by air accumulation in the pleural space, can lead to lung collapse and, if diagnosis is delayed, respiratory or cardiovascular deterioration. Most deep-learning studies on pneumothorax address classification or segmentation alone. This study proposes a holistic automatic radiological reporting system combining U-Net-based segmentation, Segment Anything Model 2 (SAM2)-based boundary refinement, and a Vision-Language Model (VLM)-based reporting module for pneumothorax detection on computed tomography (CT) images. Methods: A MobileNetV2-encoder U-Net was trained on 224×224 CT patches from 171 patients (15,555 patches). A frozen, zero-shot SAM2 module, guided by a box prompt derived from the U-Net mask, was combined with U-Net via an Intersection over Union (IoU)-agreement rule. The refined mask and rule-based geometric features were passed to a frozen VLM (Qwen2-VL-2B-Instruct) to generate structured reports with a combined confidence coefficient. Performance was evaluated with Accuracy, Dice, IoU, and Loss, over all patches and over pneumothorax-positive patches only. Results: U-Net achieved 93.09/92.09/99.71% (Dice/IoU/Accuracy) on validation and 93.68/92.79/99.79% on the test set; these largely reflect correct empty-mask prediction on negative patches. On positive patches only, Dice was 0.6295 (validation) and 0.6215 (test), with patch-level sensitivity of 82.2% (validation) and 79.9% (test) and specificity above 97%. SAM2 refinement achieved 92.44% Dice and 91.29% IoU on a random test subsample (n ≈ 300), close to but not exceeding U-Net alone; an ablation analysis confirmed no consistent Dice/IoU or 95th-percentile Hausdorff distance (HD95) gain from SAM2. BLEU/ROUGE and Likert evaluation frameworks for the VLM reports were defined but not yet scored against reference reports. Conclusions: U-Net-based segmentation reliably identifies pneumothorax-containing patches with moderate boundary accuracy, while frozen SAM2 refinement adds limited value. The rule-based reporting layer is reproducible and shows triage-support potential, whereas the VLM free-text layer carries a documented hallucination risk and requires radiologist verification before clinical use.
Related
- Recommended Future Directions for the ICRP Related to Improving the System of Radiological Protection
- Clinical-Radiological Discordance in the Detection of Chronic Constipation in Children Newly Diagnosed with Autism Spectrum Disorder: A Retrospective Study.
- Right Coronary Artery Thrombosis Causing ST-Elevation Myocardial Infarction After a Coronary Artery Disease Reporting and Data System Score of 0, With Disseminated Intravascular Coagulation During Idiopathic Acute Pancreatitis: A Case Report
- Age-stratified signals and early onset of adverse events with pregabalin: a disproportionality analysis of the FDA adverse event reporting system
- Derivation of a clinical-radiological scoring system to predict foreign body aspiration in children
BibTeX
@article{tuncer2026sam2,
title = {{A SAM2 and VLM-Based Automatic Radiological Reporting System for Pneumothorax Detection}},
author = {Savaş Tunçer and Mustafa Yıldırım and Özal Yıldırım},
journal = {Biomedicines},
year = {2026},
month = oct,
doi = {10.3390/biomedicines14102312},
url = {https://europepmc.org/article/PMC/PMC13701554},
note = {CC BY},
}
RIS
TY - JOUR
TI - A SAM2 and VLM-Based Automatic Radiological Reporting System for Pneumothorax Detection
AU - Savaş Tunçer
AU - Mustafa Yıldırım
AU - Özal Yıldırım
PY - 2026
DA - 2026/10/01
JO - Biomedicines
DO - 10.3390/biomedicines14102312
UR - https://europepmc.org/article/PMC/PMC13701554
AB - Background/Objectives: Pneumothorax, caused by air accumulation in the pleural space, can lead to lung collapse and, if diagnosis is delayed, respiratory or cardiovascular deterioration. Most deep-learning studies on pneumothorax address classification or segmentation alone. This study proposes a holistic automatic radiological reporting system combining U-Net-based segmentation, Segment Anything Model 2 (SAM2)-based boundary refinement, and a Vision-Language Model (VLM)-based reporting module for pneumothorax detection on computed tomography (CT) images. Methods: A MobileNetV2-encoder U-Net was trained on 224×224 CT patches from 171 patients (15,555 patches). A frozen, zero-shot SAM2 module, guided by a box prompt derived from the U-Net mask, was combined with U-Net via an Intersection over Union (IoU)-agreement rule. The refined mask and rule-based geometric features were passed to a frozen VLM (Qwen2-VL-2B-Instruct) to generate structured reports with a combined confidence coefficient. Performance was evaluated with Accuracy, Dice, IoU, and Loss, over all patches and over pneumothorax-positive patches only. Results: U-Net achieved 93.09/92.09/99.71% (Dice/IoU/Accuracy) on validation and 93.68/92.79/99.79% on the test set; these largely reflect correct empty-mask prediction on negative patches. On positive patches only, Dice was 0.6295 (validation) and 0.6215 (test), with patch-level sensitivity of 82.2% (validation) and 79.9% (test) and specificity above 97%. SAM2 refinement achieved 92.44% Dice and 91.29% IoU on a random test subsample (n ≈ 300), close to but not exceeding U-Net alone; an ablation analysis confirmed no consistent Dice/IoU or 95th-percentile Hausdorff distance (HD95) gain from SAM2. BLEU/ROUGE and Likert evaluation frameworks for the VLM reports were defined but not yet scored against reference reports. Conclusions: U-Net-based segmentation reliably identifies pneumothorax-containing patches with moderate boundary accuracy, while frozen SAM2 refinement adds limited value. The rule-based reporting layer is reproducible and shows triage-support potential, whereas the VLM free-text layer carries a documented hallucination risk and requires radiologist verification before clinical use.
SN - 2227-9059
ER -