Published in:
01-08-2013 | Breast Oncology
Analysis of Factors that Influence the Accuracy of Magnetic Resonance Imaging for Predicting Response after Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer
Authors:
Eun Sook Ko, MD, Boo-Kyung Han, MD, PhD, Rock Bum Kim, MD, PhD, Eun Young Ko, MD, PhD, Jung Hee Shin, MD, PhD, Soo Yeon Hahn, MD, Seok Jin Nam, MD, PhD, Jeong Eon Lee, MD, PhD, Se Kyung Lee, MD, PhD, Young-Hyuck Im, MD, PhD, Yeon Hee Park, MD, PhD
Published in:
Annals of Surgical Oncology
|
Issue 8/2013
Login to get access
Abstract
Purpose
The purpose of this study was to evaluate the accuracy of breast magnetic resonance imaging (MRI) to predict residual lesion size after neoadjuvant chemotherapy (NAC) and to determine the factors that influence the accuracy of response prediction.
Methods
This study comprised 166 patients who underwent MRI before and after NAC, but before surgery. The longest diameter of the residual cancer was measured using MRI and correlated with pathologic findings. Patients were further divided into subgroups according to various radiologic and histopathologic factors. Pathologic complete response (pCR) was defined as the absence of residual invasive cancer cells. The Pearson correlation was used to correlate tumor size as determined by MRI and pathology, and the Mann-Whitney U test and Kruskal-Wallis test were used to compare MRI-pathologic size discrepancies according to various clinical, histopathologic factors, and MRI findings.
Results
Of the 166 women, 40 achieved pCR. The overall sensitivity, specificity, and accuracy for diagnosing invasive residual disease by using MRI were 96, 65, and 89 %, respectively. The Pearson’s correlation coefficient between the tumor sizes measured using MRI and pathology was 0.749 (P < 0.001). The size discrepancy was significantly greater in patients with estrogen receptor-positive cancer (P = 0.037), in cancers with low nuclear grade (P = 0.007), and in cancers shown as diffuse non-mass–like enhancement on MRI (P = 0.001).
Conclusions
Size prediction is less accurate in cases with estrogen receptor-positive breast cancer, low nuclear grade, and diffuse non-mass–like enhancement on initial MRI.