PET for predicting chemoradiotherapy survival in cervical cancer: the search continues.

Abstract

Objective: Cervical c ancer is still a global health concern. The push for personalized care has been the mainstay and the use of noninvasive approaches to stratify patients is a corollary of this management strategy. 18 F-fluorodeoxyglucose PET/computed tomography is frequently used for cervical c ancer staging. Identifying predictive factors PET-based imaging scores could mean increased benefit for these patients, at no increased cost.

Methods: An observational study was conducted with stage IIB-IIIB cervical cancer patients receiving erlotinib concomitantly to standard therapy. Survival analyses employed Kaplan-Meier and Cox proportional hazards models, waterfall, and heatmap plots to investigate survival associations. Maximum and average PET measures were obtained before and 3 months after treatment.

Results: Thirty patients were enrolled in this phase I/II study. IIB Staging was significantly associated with longer survival when compared with IIIB. Simultaneous lower 2SD algorithm mean standard uptake value (SUV 2SD ) (<4.1) and greater BMI reduction (< -0.17 kg/m 2 ), or lower PERCIST SUV (SUV per ) (<7.6) and greater BMI reduction, predicted worse overall and progression-free survival ( P < 0.05). Unifying pretreatment maximum SUV (SUV max ), SUV per , and SUV 2SD demonstrated a trend toward significant overall survival stratification ( P = 0.057). In the heatmap analysis, there was a lack of high percentile pretreatment tumor-to-liver ratio, SUV max , SUV per , and SUV 2SD features in patients with worse endpoints. The waterfall plots suggested pretreatment lesion volume may predict volume reduction.

Conclusion: SUV 2SD and SUV per , paired with BMI variation, were associated with survival prediction. PET might play an important role in individualizing patients before treatment decisions for optimal management of locally advanced disease.