Going from double to single reading for screening exams labelled as likely normal by AI: what is the impact?

Balta C, Rodriguez-Ruiz A, Mieskes C, Karssemeijer N, Heywang-Köbrunner S
Proc. SPIE 11513, 15th International Workshop on Breast Imaging (IWBI2020), 115130D (22 May 2020); doi: 10.1117/12.2564179

Aim and Objective
The goal of this study was to investigate whether Transpara®, a deep learning-based artificial intelligence (AI) decision support system, can be used to improve breast cancer screening workflow efficiency by making a pre-selection of likely normal screening mammograms where double-reading could be safely replaced with single-reading.

Materials and Method
18,015 consecutively acquired screening exams were collected, including the independent reading assessments by each radiologist of the double reading process, and the information about whether the case was recalled and if so the recall outcome. Transpara (version 1.6.0) assigned a 1-10 score to each screening exam denoting the likelihood of cancer. We simulated the impact on recall rate, cancer detection rate, and workload if single reading would have been performed for the mammograms with the lowest AI scores.

Results
– Doing single instead of double reading for exams with Transpara Exam Score 1 to 7 was found to be optimal
– Recall rate would have reduced significantly (p<0.0001) by 11.8% (from 5.35% to 4.79%), without a significant impact in cancer detection rate
– Workload would have been reduced by 32.6%
– The PPV of screening would have increased significantly (p<0.0001) by 10.54% (11.90% to 13.30%)
– The number of cases going through consensus would have decreased significantly (p<0.0001) by 20.79% (2400 to 1987)
– 14 cancers were detected by Transpara (Exam Score 10) that were only detected by a single reader

Conclusion
Transpara Exam Score has the potential to be used as an effective triaging tool to determine which cases benefit from a second reader and which do not, in order to reduce workload and potentially increase the quality of screening by safely reducing the recall rate when avoiding the recall of cases with low Transpara Exam Score.