Clinical Performance of Contrast-Enhanced Mammography versus Breast MRI in Women at Increased Breast Cancer Risk

This dataset contains the multireader, multicase (MRMC) raw data from a retrospective study comparing the diagnostic performance of contrast-enhanced mammography (CEM), low-energy CEM (LE-CEM), and breast MRI in women at intermediate or high risk for breast cancer. The study included 461 women (mean age 50 ± 9 years) enrolled who underwent both CEM and breast MRI between March 2019 and October 2022. The provided CSV file contains the individual reader assessments used to evaluate sensitivity, specificity, and Area Under the ROC Curve (AUC) for three imaging modalities. Column Description: readerID: Identifier for the eight breast radiologists involved in the study. R1, R2, R3, R4: Group 1 readers, who are experienced in CEM (2–5 years) and interpreted both LE-CEM and CEM datasets R5, R6, R7, R8: Group 2 readers, who are experts in breast MRI (8–15 years) modality: The imaging technique used for the assessment. LE-CEM: Low-energy images considered equivalent to standard digital mammography CEM: Contrast-enhanced images generated through dual-energy subtraction (DES) MRI: Standardized dynamic contrast-enhanced breast MRI caseID: Unique anonymized identifier for each of the 461 participants. truth: The established ground truth for each case. For positive cases, truth was determined by histopathology from biopsy or surgery For negative or benign cases, truth was verified by at least 24 months of stable imaging follow-up score: The BI-RADS assessment score assigned by the reader for each case and modality. In the study’s analysis, a score >3 (BI-RADS 4 or 5) was considered positive for malignancy, while scores ≤3 were considered negative