Early predicting the pathologic complete response (pCR) after neoadjuvant chemotherapy in patients with locally advanced breast cancer (LABC) is closely associated with clinical outcomes. However, conventional metabolic parameters using baseline 18F-FDG PET/CT have failed to accurately predict the pCR. Breast cancer stem cells (CSCs) are known for their established role in chemoresistance. We designed a new PET parameter for CSC metabolism (MTVcsc) from pretreatment 18F-FDG PET/CT by using distinctive glucose metabolism between CSCs and differentiated cancer cells, and aimed to evaluate the prognostic value of the MTVcsc.
A total of 71 patients with LABC who underwent initial 18F-FDG PET/CT before NAC were included in this study. The SUV values of single voxels within the primary tumor were clustered by performing k-means clustering using R version 3.5.3 and MTVcsc was derived by calculating the volume of the most glycolytic cluster. The predictive values of the MTVcsc, as well as clinicopathologic and conventional metabolic parameters (SUVmax, MTV, TLG) for pCR, were analyzed by multivariable logistic regression.
Seventeen patients were excluded from the final analysis due to small tumor size (< 64 voxels). The lower MTVcsc and non-luminal subtypes were significantly associated with achieving pCR following NAC (Table). The MTVcsc outperformed the conventional PET parameters in predicting pCR. Table Univariable and multivariable logistic regression model of clinicopathologic and metabolic parameters for predicting pathologic complete response.
Parameters Univariable analysis Multivariable analysis OR 95% CI P value OR 95% CI P value Ki-67 Low, < 20% High, ≥ 20% 1.00 5.57 1.06-29.27 0.043 Molecular subtype Luminal A and B HER2 positive Triple negative 1.00 1.00 11.46 2.07-63.36 0.005 13.7 1.75-107.36 0.013 6.55 1.05-40.67 0.044 17.42 1.41-215.04 0.026 Metabolic parameters SUVmax Metabolic tumor volume (MTV) Total lesion glycolysis (TLG) MTVcsc 0.98 0.83-1.16 0.814 0.98 0.94-1.02 0.281 0.99 0.98-1.00 0.231 0.29 0.10-0.89 0.031 0.21 0.05-0.82 0.025
MTVcsc, a novel PET parameter for CSC metabolism, provides predictive value for pCR. By further stratifying LABC patients with a combination of MTVcsc and molecular subtype at initial staging workup, achieving pCR after NAC can be early predicted more accurately.
Has not received any funding.
All authors have declared no conflicts of interest.