Diffusion-Weighted MRI in Predicting the Efficacy of Neoadjuvant Chemotherapy of Breast Cancer.

Abstract

Abstract Objective: To approach the apparent diffusion coefficient(ADC) obtained on diffusion weighted imaging (DWI) in predicting the response of breast cancer to neoadjuvant chemotherapy.Methods: 53 locally advanced breast cancer patients participated in this prospective study on neoadjuvant chemotherapy. All patients were treated with weekly neoadjuvant paclitaxel plus carboplatin for 4 cycles. DWI and ADC value were examined before chemotherapy(the first time point), after the first and the forth cycle of chemotherapy(the second and the third time point). Patients were categorized as high responders if their ADC at the second time point were higher by 2 times the standard deviation than the mean pre-chemotherapy ADC, and otherwise the patients were considered as low responders. Logistic regression was conducted to examine the independent factors predictive of clinical complete response(CR) and pathologic complete response(pCR).Results: CR was documented in 13 patients (24.5%); partial response in 36 patients (67.9%); stable disease in 4 patients (7.5%) and no patient had progressive disease. 11 patients(20.8%) have achieved pCR. There was a strong positive correlation between the percentage changes in ADC value at the third time points and the degree of the tumor reduction. A significant increase in ADC value was observed at the second time point compared with the first time point (1.104±0.117×10-3mm2/s, 1.220±0.158×10-3mm2/s, P<0.001), while the decrease of the longest diameter of the tumors measured by physical examination, ultrasound or MRI or the tumor volumn measured by MRI were not (P>0.05). Logistic regression revealed that percentage changes of ADC value at the second time point were significantly associated with CR(P=0.022, hazard ratio[HR] =1.08, 95% confidence interval[CI]: 1.01∼1.16) and pCR(P=0.022, HR =1.11, 95%CI: 1.02∼1.21). The AUC value of the receiver operating characteristic curve were 0.721(95%CI: 0.512∼0.931, P=0.017) and 0.775(95%CI: 0.598∼0.952, P=0.005) in the two diagnostic tests for CR and pCR, and the sensitivity and specificity were 76.9%, 63.6% and 77.5%, 90.5%, respectively. According to categorizing system described above, 10 patients(18.9%) were considered as high responders and 43 patients(81.1%) low responders. 7 patients(70%) of the high responders have achieved clinical CR and pCR while for the low responders there were only 6 patients(14.0%) and 4 patients (9.3%) proven to have CR and pCR. The difference have both reached statistical significance(P=0.001, P<0.001). Logistic regression confirmed that the categorizing system was independently predictive of CR(P=0.006, HR=42.26, 95%CI: 2.86∼624.08) and pCR(P=0.013, HR=25.19, 95%CI: 2.00∼318.91). Taking this categorizing system as a dignostic test, the sensitivity and specificity were 53.8% and 92.5% for CR and 63.6% and 92.9% for pCR, respectively.Conclusions: The change of ADC value obtained on DWI after the first cycle of neoadjuvant chemotherapy occured prior to the morphologic changes of the tumor and at this time ADC value could to some extent predict the efficacy of neoadjuvant chemotherapy. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 4025.