Early prediction of response to neoadjuvant chemotherapy in breast cancer patients: comparison of single-voxel (1)H-magnetic resonance spectroscopy and (18)F-fluorodeoxyglucose positron emission tomography.

Abstract

To prospectively compare performances of single-voxel proton magnetic resonance spectroscopy ((1)H-MRS) and (18)F-fluorodeoxyglucose positron emission tomography (FDG-PET) in predicting pathologic response to neoadjuvant chemotherapy (NAC) in breast cancer patients. Thirty-five breast cancer patients who received NAC and subsequent surgery were prospectively enrolled. MRS and FDG-PET were performed before and after the 1st NAC cycle. Percentage changes of total choline-containing compounds (tCho) via MRS, and maximum and peak standardized uptake values (SUVmax, SUVpeak) and total lesion glycolysis (TLG) via FDG-PET were measured, and their performances in predicting pathologic complete response (pCR) were compared. Of the 35 patients, 6 showed pCR and 29 showed non-pCR. Mean % reductions of tCho, SUVmax, SUVpeak, and TLG of the pCR group were larger than those of the non-pCR group (-80.3 ± 13.9% vs. -32.1 ± 49.4%, P = 0.025; -54.7 ± 22.1% vs. -26.3 ± 33.7%, P = 0.058; -60.7 ± 18.3% vs. -32.3 ± 23.3%, P = 0.009; -89.5 ± 8.5% vs. -52.6 ± 36.2%, P = 0.020). Diagnostic accuracy (area under ROC curve; Az, 0.911) of the % reduction of tCho was comparable to those of %SUVmax (0.822), SUVpeak (0.862), and TLG (0.879) in distinguishing pCR from non-pCR (all P > 0.05). MRS showed comparable performance to FDG-PET in early prediction of pCR in breast cancer patients. • MRS can predict response to NAC in breast cancer post-1 (st) cycle. • Changes in tCho and SUV after NAC reflect tumour cellularity changes. • MRS can be an alternative to FDG-PET in predicting response to NAC.