Diagnostic efficacy in screening mammograms does not improve with peer reading strategy : a Sino-Australian study

Unlike Australia, China has no population-based early detection screening program with radiological expertise being a barrier to implementation. This study explores observer performance between breast radiologists from China and Australia, and the role of peer-Assisted reading in Chinese radiologists performance. A test set of 60 high density screening mammograms (40 normal, 20 cancer cases) was constructed with 8 Chinese and 17 Australian radiologists reading the test set independently, while another 10 Chinese radiologists read the test set as a peer-duo, where discussion was encouraged but lesion marking was done separately. For independent readings by radiologists who read >20 cases per week, Chinese reader had lower performance in sensitivity, lesion sensitivity, AUC and JAFROC. There was no significant difference in performance between independent reading and peer-Assisted reading Chinese readers and this strategy may have limited valued in improving diagnostic efficacy. Breast screening programs for the early detection of small lesions is important to reduce cancer deaths. Unlike Australia, China has no population-based early detection screening program, with radiological expertise one barrier for the successful implementation. Currently, little is known about the diagnostic accuracy of Chinese breast radiologists and comparative studies with other international radiology cohorts could allow for benchmarking. This study explores difference in performance between breast radiologists from China and Australia, and the role of peer-Assisted reading in Chinese radiologists performance. A test set containing 60 high density screening mammograms (40 normal and 20 cancer cases) was constructed. Eight Chinese and seventeen Australian radiologists read the test set independently while another ten Chinese radiologists read the test set with peer assistance as a duo, but lesion marking was done separately. All radiologists were asked to localise any suspicious lesion, giving each marked lesion a grade of 1 5 on both mammographic views and scores generated for case sensitivity, specificity, lesion sensitivity, AUC and JAFROC. For independent readings, Chinese radiologists who read >20 cases per week demonstrated significant lower scores in sensitivity (p=0.026), lesion sensitivity (p=0.003), AUC (p=0.003) and JAFROC (p=0.008) than Australian radiologists. There was no significant difference in performance between Chinese radiologists independent reading and peer-Assisted reading although a better performance in specificity with the strategy of peer-Assisted reading (p=0.004) was observed for radiologists that read >20 cases per week. Overall, peer-Assisted reading may have limited valued in improving diagnostic efficacy of Chinese radiologists. 2022 SPIE.