AI Insight
Researchers developed and validated Inspectra CXR version 5, a deep learning system that identifies nine thoracic diseases on chest X-rays and localizes lesions within a single model. The system was trained on 874,858 chest radiographs from a Bangkok hospital and achieved a mean AUROC of 0.994 on internal testing and 0.970 across 13 independent Thai hospitals. Five radiologists rated the system highly usable with 93.6% classification concordance and 94.7% localization concordance.
Why it matters
This addresses the critical shortage of radiologists in Thailand and Southeast Asia by providing an accurate automated diagnostic tool that was specifically adapted to local populations. The system's ability to both classify diseases and show their locations on X-rays, combined with high clinician acceptance, suggests it could be practically deployed to improve diagnostic capacity in resource-limited healthcare settings.
Understand the Science
arXiv:2607.09305v1 Announce Type: cross
Abstract: Chest radiography (CXR) remains the most widely used thoracic imaging modality, yet expert interpretation is constrained by a severe shortage of radiologists in Thailand and across Southeast Asia. Local adaptation of deep learning models to Thai data has been shown to substantially improve accuracy on Thai populations. Here we present the development and comprehensive validation of the chest radiograph analysis model in Inspectra CXR version 5, a deep learning system that performs multi-label thoracic disease classification and weakly supervised lesion localization within a single model. The architecture couples a DenseNet-121 backbone with Attend-and-Compare Modules (ACM) and a Probabilistic Class Activation Map (PCAM) aggregation layer, producing a per-condition classification score and heatmap simultaneously. The model was developed on 874,858 frontal chest radiographs with paired radiologist reports from Siriraj Hospital, Bangkok. On a held-out, radiologist-verified in-domain test set of 19,871 cases, it achieved a mean AUROC of 0.994 (mean sensitivity 92.4%, specificity 98.6%) across nine clinically important conditions. On an independent generalization set of 5,992 cases from 13 hospitals across Thailand, the mean AUROC was 0.970, indicating robust transfer across sites. For localization, evaluated on 4,549 radiologist-annotated cases, the model attained a mean lesion-localization fraction (LLF) of 77.9% at 0.59 non-lesion localizations per image. In a usability evaluation with five thoracic radiologists, the system reached a classification concordance of 93.6%, a localization concordance of 94.7%, and a mean System Usability Scale (SUS) score of 89. These results indicate that a locally developed, localization-capable CXR system can deliver high accuracy, generalize across heterogeneous Thai hospitals, and earn the trust of practicing radiologists.