Abstract PD3-04: Which measure of the interim changes in breast tumoral volume at breast MRI in response to neoadjuvant chemotherapy best predicts final pathological response?

Abstract BACKGROUND: Interim changes in breast tumour volume at magnetic resonance imaging (MRI) can predict ultimate response to neoadjuvant chemotherapy (NAC), but there is little data on the best measure of volumetric change. PURPOSE: To assess whether changes in measurements of semi-automated enhancing tumour volume (ETV) or fully automated functional tumour volume (FTV) between baseline and interim contrast-enhanced MRI are equivalent in predicting ultimate pathological response to neoadjuvant chemotherapy (NAC) for primary breast cancer, assessed using the residual cancer burden (RCB) score. MATERIALS & METHODS: 78 patients undergoing treatment with NAC for primary breast cancer underwent contrast-enhanced MRI on a 1.5T or 3.0T MRI scanner using a dedicated bilateral breast coil before and after two or three cycles of NAC. Image analysis was performed using either semi-automated, user-defined thresholding (ITK-Snap; ETV) or fully-automated (Siemens SyngoVia BreVis; FTV) approaches. For ETV, the two-minute post-contrast subtracted volumes were analysed, with enhancing pixels thresholded to define tumour volume. FTV was measured using a manufacturer default setting of 50% enhancement threshold, relative to pre-contrast signal intensity, to define tumour volume. ETV intra-observer reproducibility was assessed by repeat analysis one month after initial analysis and a second observer also repeated the measure. Coefficient of reproducibility (CoR) and intraclass correlation coefficents (ICC) were calculated for intra- and inter-observer repeatability. ETV and FTV percentage reduction between baseline and interim examinations was compared with final pathological response, as assessed using the residual cancer burden (RCB) score on resected cancer specimens. Correlation of the two volumetric measures was performed using a Pearson Intra-class Correlation Coefficient (ICC) and pair-wise comparisons of ETV and FTV changes between RCB groups carried out using a Mann-Whitney U test. All statistical assessment was performed using SPSS, v21, with p<0.05 considered significant. RESULTS: There was significant correlation between ETV and FTV (ICC= 0.744, p<0.05). Intra and inter observer reproducibility for ETV was excellent, with ICC 0.940 and 0.861 respectively and corresponding CoRs of 11.6% and 14.8%. Average percentage reductions in ETV for each pathological response category were: pCR 96.4% (n=12), RCB-I 66.6% (n=10), RCB-II 62.9% (n=39) and RCB-III 27.3% (n=17). Corresponding values for FTV were 88.8%, 70.6%, 54.6% and 20.8%. Significant differences in percentage ETV changes were found for pCR vs. RCB-I (p<0.008), II (p<0.001) & III (p<0.001) and RCB-II vs. RCB-III (p<0.001). For FTV, significant differences were measured only for pCR vs. RCB-II & III (p<0.001). CONCLUSION: changes in the semi-automated ETV measuement between baseline and interim MRI may provide more useful predictive information on final pathological response to NAC than FTV, as the changes are better able to discriminate between pCR and minimal residual disease (RCB-I). The ability to confidently predict pCR versus all other residual disease categories could facilitate planning of enhanced approaches to surgical management. Citation Format: Thompson AM, Vinnicombe SJ, Waugh SA, Purdie CA, Evans AJ, Brunton T, Fuller-Pace FV. Which measure of the interim changes in breast tumoral volume at breast MRI in response to neoadjuvant chemotherapy best predicts final pathological response? [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD3-04.