Assessmentof saudi radiologist performance in reading screening mammography

To evaluate the diagnostic performance of Saudi radiologists in interpreting screening mammograms. A total of 72 readers participated, including 33 Saudis and 39 non-Saudis. Participants interpreted 30 mammograms (15 normal and 15 abnormal) using the DetectedX online platform. For each participant, key performance metrics were collected, including sensitivity, specificity, and lesion sensitivity. In addition, lesion detection performance was assessed with the Jackknife Free-Response Receiver Operating Characteristics (JAFROC) Figure of Merit (FOM). The analysis explored the relationship between performance metrics and key reader characteristics. Non-Saudi radiologists demonstrated significantly higher JAFROC scores (median = 0.53 vs. 0.43, p = 0.043) and specificity (median = 0.80 vs. 0.55, p < 0.001) compared to Saudi radiologists. However, Saudi radiologists showed higher sensitivity (median = 0.80 vs. 0.60, p = 0.024). In the JAFROC model, nationality was a significant predictor, with non-Saudis scoring higher than Saudis (R 2 = 10.2 %, p = 0.04), whereas in the sensitivity model, Saudis outperformed non-Saudis (R 2 = 7.4 %, p = 0.02). Lesion sensitivity was predicted by position (R 2 = 15.2 %), while the specificity model explained the highest variance (R 2 = 44.7 %) based on both position and nationality. While Saudi radiologists showed strong abilities in detecting breast cancer, their lower specificity highlights the need for improving diagnostic accuracy to reduce false positives. To enhance overall diagnostic accuracy in breast cancer screening, it is essential to implement targeted professional development programs focusing on balancing sensitivity and specificity.