AI-Assisted Chest X-Ray Reading Improves Sensitivity Without Reducing Specificity: A Crossover Study

Autoři

TRABÁLKOVÁ Zuzana ŠTEVÍK Martin ZELEŇÁK Kamil DANDÁR Jakub KVAK Daniel KVAKOVÁ Karolína OVESNÁ Petra

Rok publikování 2026
Druh Stať ve sborníku
Konference ICT for Intelligent Systems (ICTIS 2025)
Citace
Doi https://doi.org/10.1007/978-981-95-1353-6_38
Popis The growing demand for chest radiography, combined with radiologist shortages and increasing workloads, underscores the need for innovative diagnostic support tools. This crossover study evaluates the impact of a commercially available deep learning-based automatic detection software (DLAD) on radiologists' performance in interpreting chest X-rays (CXRs). Five radiologists independently assessed a dataset of 540 anonymized CXRs in two phases—without and with DLAD assistance—separated by a 30-day washout period. DLAD support significantly improved diagnostic performance: overall sensitivity increased from 0.762 (95% CI: 0.705–0.811) to 0.911 (0.870–0.941, p < 0.001), while specificity remained unchanged at 0.850 (0.805–0.887, p = 0.331). The positive predictive value (PPV) rose slightly from 0.810 (0.755–0.856) to 0.836 (0.788–0.876, p = 0.331), and the negative predictive value (NPV) improved from 0.810 (0.763–0.850) to 0.941 (0.882–0.947, p < 0.001). These improvements were consistent across all readers, with a marked reduction in false negatives. The findings demonstrate DLAD’s potential to enhance diagnostic accuracy, increase sensitivity, and support radiologists in chest X-ray interpretation without compromising specificity.

Používáte starou verzi internetového prohlížeče. Doporučujeme aktualizovat Váš prohlížeč na nejnovější verzi.