Yan-Ling Li1, Xiao-Peng Zhang1, Jie Li1, Kun Cao1, Yong Cui1, Xiao-Ting Li1, Ying-Shi Sun2. 1. Department of Radiology, Peking University Cancer Hospital & Institute, Key laboratory of Carcinogenesis and Translational Research (Ministry of Education),Beijing, China. 2. Department of Radiology, Peking University Cancer Hospital & Institute, Key laboratory of Carcinogenesis and Translational Research (Ministry of Education),Beijing, China. Electronic address: sys@bjcancer.org.
Abstract
OBJECTIVE: To select effective indicators for diagnosis of pathological complete response (pCR) by MRI and to establish an appropriate diagnostic program to maximize the accuracy of pCR detection by MRI. MATERIALS AND METHODS: Twenty-one pCR patients and 22 non-pCR randomly selected patients receiving neoadjuvant chemotherapy (NAC) and subsequent surgery were recruited for the study. All patients underwent breast MRIs both before and after chemotherapy. Changes in diameter, area and dynamic variables between the first and final MRI were compared between the two groups. Logistic and ROC analysis were performed to select effective indicators for predicting pCR on MRI. RESULTS: Eleven out of 43 patients had no residual enhanced areas on MRI, and the sensitivity and specificity for predicting pCR on MRI under the current criterion was 52.38% and 100%, respectively. Logistic regression analysis revealed that changes in diameter, SIpeak and area were effective in predicting pCR by MRI. The latter two parameters had a greater impact on diagnosis than the diameter change. Two new independent criteria were established to predict pCR on MRI: (1) a reduction of ≥78% in area; and (2) a combination of a reduction of ≥27% in SIpeak and of ≥78% in area on MRI. Both had diagnostic accuracy of 88.37% and criterion 1 had higher sensitivity of 90.48%. However, criterion 2 had perfect specificity of 100%. CONCLUSION:MRI is an effective means for detecting pCR from non-pCR patients. Changes in area and SIpeak can be used to establish two new independent criteria which perform better in diagnosing pCR on MRI than the current criterion.
RCT Entities:
OBJECTIVE: To select effective indicators for diagnosis of pathological complete response (pCR) by MRI and to establish an appropriate diagnostic program to maximize the accuracy of pCR detection by MRI. MATERIALS AND METHODS: Twenty-one pCR patients and 22 non-pCR randomly selected patients receiving neoadjuvant chemotherapy (NAC) and subsequent surgery were recruited for the study. All patients underwent breast MRIs both before and after chemotherapy. Changes in diameter, area and dynamic variables between the first and final MRI were compared between the two groups. Logistic and ROC analysis were performed to select effective indicators for predicting pCR on MRI. RESULTS: Eleven out of 43 patients had no residual enhanced areas on MRI, and the sensitivity and specificity for predicting pCR on MRI under the current criterion was 52.38% and 100%, respectively. Logistic regression analysis revealed that changes in diameter, SIpeak and area were effective in predicting pCR by MRI. The latter two parameters had a greater impact on diagnosis than the diameter change. Two new independent criteria were established to predict pCR on MRI: (1) a reduction of ≥78% in area; and (2) a combination of a reduction of ≥27% in SIpeak and of ≥78% in area on MRI. Both had diagnostic accuracy of 88.37% and criterion 1 had higher sensitivity of 90.48%. However, criterion 2 had perfect specificity of 100%. CONCLUSION: MRI is an effective means for detecting pCR from non-pCR patients. Changes in area and SIpeak can be used to establish two new independent criteria which perform better in diagnosing pCR on MRI than the current criterion.