This retrospective bi-national multicenter diagnostic accuracy study evaluated the locked default-threshold performance of a commercial chest radiograph AI system for selected thoracic findings using a CT-anchored, radiographic-detectability reference framework. Consecutive eligible adult patients who underwent frontal chest radiography and chest CT at four tertiary-care centers in Turkey and Bulgaria between June and December 2024 were included. Reference labels were assigned using temporally paired CT, with adjudication of whether CT-confirmed abnormalities had a corresponding radiographic manifestation on the paired chest radiograph. For potentially dynamic findings, the allowable CT–CXR interval was restricted to ≤ 2 days. The AI system was evaluated at the manufacturer’s default threshold of 0.50. Diagnostic performance was summarized using prevalence, sensitivity, specificity, positive predictive value, negative predictive value, and false-positive burden with 95% confidence intervals. The study included 940 patients with paired CXR-CT examinations. Reference-positive prevalence ranged from 2.2% for pneumothorax to 21.0% for pleural effusion. Sensitivity was highest for fracture (91%) and pneumothorax (90%) and lowest for atelectasis (64%); specificity ranged from 82% for consolidation/opacity to 98% for fracture and pneumothorax. PPV ranged from 43% to 63%, indicating a non-trivial false-positive burden at the evaluated operating point. Because continuous probability scores, human-reader comparison, and workflow outcomes were unavailable, these findings should be interpreted as default-threshold technical validation and support further evaluation of the system as radiologist-supervised decision support with local performance monitoring, rather than standalone diagnosis.

Background/Objectives: Whether PCCT-derived virtual non-iodine (VNI) images can replace true non-contrast (TNC) for coronary artery calcium scoring (CACS) remains uncertain, particularly for small, low-density plaques. We aimed to evaluate agreement between VNI and TNC for CACS at the patient and lesion levels and to quantify risk-category reclassification. Methods: In this retrospective single-center sample (May 2024–May 2025), 211 patients without prior coronary intervention and with nonzero CAC on TNC underwent PCCT. VNI (55 keV, QIR 1; 60 keV, QIR 4; PureCalcium) and TNC were reconstructed with matched section thickness/increment and kernel. Agatston and total calcified volume were recorded. Paired comparisons used Wilcoxon tests; reclassification across CAC categories (0, 1–99, 100–399, ≥400) and lesion-level false negatives (FNs) were assessed with TNC as the reference. Results: Low-keV VNIs (55–60 keV) underestimated CAC versus TNC. The median Agatston score decreased from 35.9 (IQR, 10.3–121.2) on TNC to 23.6 at 55 keV (p = 0.0006) and 22.2 at 60 keV (p = 0.0003); the total volume declined from 37.8 mm³ to 20.2 mm³ (p = 0.001) and 18.3 mm³ (p < 0.0001), respectively. More than half of patients were reassigned to a lower CAC category; despite no patient being CAC = 0 on TNC, 46.9% (55 keV) and 47.4% (60 keV) were labeled CAC = 0 on VNI. Because this study deliberately included only patients with nonzero CAC on the TNC reference, these CAC = 0 rates on VNI represent misclassification within a CAC-positive sample and should not be interpreted as population-level prevalence. At the lesion level, 95% of patients had ≥1 FN plaques (430 FN plaques total), typically small (median 8 mm³) and of low density (median Agatston 6). Conclusions: In this single-center sample with relatively low-burden calcification, low-keV VNI (55–60 keV) significantly underestimates CAC and down-classifies patients, with frequent “false-zero” assignments (defined as CAC_VNI = 0 despite CAC_TNC > 0) driven predominantly by small, low-density plaques.