Background Abbreviated MRI offers faster interpretation and shorter acquisition times than full-protocol MRI. Purpose To assess the diagnostic performance of abbreviated breast MRI compared with full-protocol MRI in determining pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) in patients with breast cancer. Materials and Methods This retrospective, single-center study included patients with breast cancer who underwent full-protocol MRI before and after NAC from January 2016 to May 2020, followed by surgical resection. Three readers independently interpreted abbreviated MRI data, which included an axial T2-weighted sequence and precontrast and early postcontrast T1-weighted sequences. After a 4-week washout period, the readers interpreted the full-protocol MRI data. Response was classified as complete, partial, stable, or progressive disease. Interpretation times were recorded, and residual enhancement was measured (longest and shortest orthogonal lesion diameters). For statistical analysis, abbreviated MRI and full-protocol MRI were compared using two pCR definitions: the absence of invasive cancer and the absence of both invasive cancer and ductal carcinoma in situ. Statistical tests included generalized estimating equation models, Wilcoxon rank sum tests, and Pearson correlation. Results The study sample included 478 patients (mean age, 51 years ± 12 SD; all female). Full-protocol MRI yielded higher specificity than abbreviated MRI did (82.1% vs 85.4% using pCR definition 1; 85.4% vs 88.6% using pCR definition 2; P = .047 for both). There was no evidence of a difference in sensitivity, negative predictive value, positive predictive value, or overall accuracy between abbreviated MRI and full-protocol MRI across pCR definitions and reader (all P > .22). The median interpretation times for abbreviated MRI and full-protocol MRI were 39-44 seconds and 42-64 seconds, respectively (all P Supplemental material is available for this article. See also the editorial by Slanetz and Yeh in this issue.
Gullo et al. (Wed,) studied this question.