What question did this study set out to answer?

The study aims to investigate how weekly reading hours impact radiologist performance in detecting breast cancer and whether multiple training test sets improve accuracy.

July 21, 2026Journal of Cancer Education0 citationsOpen Access

Optimising Radiologist Performance in Cancer Detection on Mammograms: The Impact of Weekly Reading Hours and Test Set Experience on BreastScreen and Non-BreastScreen Readers

JWJayden WellsThe University of Sydney PTPhuong Dung TrieuThe University of Sydney MBMelissa L. BarronUniversity of East Anglia

Key Points

The study aims to investigate how weekly reading hours impact radiologist performance in detecting breast cancer and whether multiple training test sets improve accuracy.
Retrospective analysis of 871 digital mammography and digital breast tomosynthesis readings from three radiologist groups across 13 BREAST test sets (2014–2024)
Performance metrics measured included sensitivity, specificity, and ROC AUC; associations were statistically analyzed with p-values less than 0.05 considered significant.
Determined performance plateau thresholds for different radiologist cohorts based on their weekly reading hours.
Performance improved consistently with increased reading hours for all groups; significant positive correlations noted for sensitivity and specificity (p<0.05).
JAFROC plateau thresholds at 20-30 HPW for BreastScreen, 10-20 HPW for non-BreastScreen, and 5-10 HPW for registrars.
No significant performance improvements were seen for completing ten or more BREAST test sets.

Abstract

Abstract Early detection through mammographic screening plays a central role in improving breast cancer outcomes, with diagnostic accuracy relying heavily on radiologist performance. While experience and case volume are recognised contributors to accuracy, the specific effect of weekly mammogram reading hours (HPW) remains unclear. This study examined the relationship between HPW and diagnostic performance among BreastScreen (BS) radiologists, non-BreastScreen radiologists, and radiology registrars, and explored whether completing multiple BREAST training test sets enhanced diagnostic accuracy. This retrospective observer-performance study analysed digital mammography (DM) and digital breast tomosynthesis (DBT) readings from 871 test set completions by BreastScreen radiologists, non-BreastScreen radiologists and radiology registrars across 13 BREAST test sets completed between 2014–2024. Performance metrics included sensitivity, specificity, lesion sensitivity, ROC AUC, and JAFROC. Demographic and performance data of participants were collected through the BREAST platform. Statistical analyses were performed to assess associations between HPW and diagnostic performance, and to identify performance plateau thresholds across HPW categories. Performance of readers improved consistently with increasing HPW for all cohorts, with significant positive correlations observed for most performance metrics (p<0.05). JAFROC plateau values occurred at different thresholds: 20–30 HPW for BreastScreen radiologists, 10–20 HPW for non-BS radiologists, and 5–10 HPW for registrars. In DBT analyses, BreastScreen radiologists achieved significantly higher JAFROC scores than non-BreastScreen radiologists only within the lowest HPW category (p<0.05). Completing ten or more BREAST test sets was not associated with significant improvements in diagnostic performance. Increasing HPW was associated with improved diagnostic performance up to cohort-specific thresholds, beyond which additional reading yielded limited benefit. Establishing optimal HPW levels may help screening services maintain high diagnostic standards while minimising fatigue and supporting workforce sustainability.

问 AI

Bookmark

View Full Paper