Textural features of 18 F-FDG PET after two cycles of neoadjuvant chemotherapy can predict pCR in patients with locally advanced breast cancer

Abstract

Objective

This study was designed to evaluate the utility of textural features for predicting pathological complete response (pCR) to neoadjuvant chemotherapy (NAC).

Methods

Sixty-one consecutive patients with locally advanced breast cancer underwent ¹⁸F-FDG PET/CT scanning at baseline and after the second course of NAC. Changes to imaging parameters [maximum standardized uptake value (SUV_max), metabolic tumor volume (MTV), total lesion glycolysis (TLG)] and textural features (entropy, coarseness, skewness) between the 2 scans were measured by two independent radiologists. Pathological responses were reviewed by one pathologist, and the significance of the predictive value of each parameter was analyzed using a Chi-squared test. Receiver operating characteristic curve analysis was used to compare the area under the curve (AUC) for each parameter.

Results

pCR was observed more often in patients with HER2-positive tumors (22 patients) than in patients with HER2-negative tumors (5 patients) (71.0 vs. 16.7%, p < 0.001). ∆ %SUV_max, ∆ %entropy and ∆ %coarseness were significantly useful for differentiating pCR from non-pCR in the HER2-negative group, and the AUCs for these parameters were 0.928, 0.808 and 0.800, respectively (p = 0.003, 0.032 and 0.037). In the HER2-positive group, ∆ %SUV_max and ∆ %skewness were moderately useful for predicting pCR, and the respective AUCs were 0.747 and 0.758 (p = 0.033 and 0.026). Although there was no significant difference in the AUCs between groups for these parameters, an additional 3/22 patients in the HER2-positive group with pCR were identified when ∆ %skewness and ∆ %SUV_max were considered together (p = 0.031). The absolute values for each parameter before NAC and after 2 cycles cannot predict pCR in our patients. Neither ∆ %MTV nor ∆ %TLG was efficiently predictive of pCR in any group.

Conclusions

The early changes in the textural features of ¹⁸F-FDG PET images after two cycles of NAC are predictive of pCR in both HER2-negative and HER2-positive patients; this evidence warrants confirmation by further research.

http://ift.tt/2sxEpo6