Additional value of 18F-FDG PET/CT response evaluation in axillary nodes during neoadjuvant therapy for triple-negative and HER2-positive breast cancer

Abstract

Background18F-FDG PET/CT can monitor metabolic activity in early breast cancer during neoadjuvant systemic therapy (NST), but it is unknown if the metabolic breast and axillary response differ. We evaluated the correlation between metabolic breast and axillary response at various time points during NST. Furthermore, we analysed if the combined metabolic response improves pathologic complete response (pCR) prediction compared to using the metabolic breast response alone.Methods18F-FDG PET/CT was performed at baseline (PET1), 2–3 weeks (PET2), and 6–8 weeks (PET3) of NST in patients with triple-negative (TN) and HER2-positive node-positive breast cancer. SUVmax and ∆SUVmax were determined separately for breast and axilla. Spearman’s correlation coefficients (r) between both localisations were calculated. The accuracy of pCR total (ypT0/is,ypN0) prediction using the metabolic response in breast, axilla or both was examined using logistic regression analysis.ResultsHundred-five patients were included: 45 TN and 60 HER2-positive tumours. The metabolic response in breast and axilla correlated moderately in TN tumours (r = 0.57) using ∆SUVmax between PET1-PET3 and poorly in HER2-positive tumours (r = 0.49) using SUVmax at PET2. In TN tumours, metabolic breast response predicted pCR well without improvement after adding axillary response (c-index 0.82 versus 0.85, p = 0.63). In HER2-positive tumours, metabolic breast response predicted pCR poorly with improvement after adding axillary response (c-index 0.64 versus 0.72, p = 0.06).Conclusions18F-FDG PET/CT response during NST differs between breast and axilla. In TN tumours, pCR total prediction can be made independent of metabolic axillary response. In HER2-positive tumours, axillary response may improve pCR total prediction. These findings may help guide PET/CT-response-based changes during NST.Trial registrationNTR NTR1797. Registered 29 May 2009, retrospectively registered.