The role of [18F]16α-fluoro-17β-fluoroestradiol (FES) PET in predicting response to endocrine therapy (ET) in advanced breast cancer.

Abstract

e13092 Background: Seventy percent of breast cancers (BC) are estrogen receptor-positive (ER+). ET improves clinical outcomes in ER+ BC; however, ET effectiveness relies on functional ER. While immunohistochemistry (IHC) is routinely used to quantify ER presence in sampled tissue, its clinical utility has inherent limitations. ER+ by IHC does not confirm function of ER, as illustrated by up to 50% of patients experiencing disease progression while on ET. FES is a radiolabeled form of estrogen and quantification of functional ER binding via FES PET may predict potential response to ETs. FES PET quantification, as measured by standardized uptake value (SUV), provides evidence of functional ER and presence of ER heterogeneity. Methods: A systematic literature review identified advanced ER+ BC patients with comparable FES SUV and progression-free survival (PFS) measures. A lesion with FES SUVmax > 1.8 was initially defined as FES-positive (FES+); other FES SUVmax thresholds were explored based on the lesion type. PFS was defined as time from initiating ET until disease progression / death from any cause and was censored at last follow-up. PFS was estimated via the Kaplan-Meier method. Cox proportional hazards regression was used to evaluate associations between FES+ and PFS, summarized via hazard ratios (HRs). HRs were estimated using different thresholds to define FES+ lesions (SUVmax thresholds 1.8-6, every 0.1). Results: The individual-level data was merged, resulting in 101 distinct patients with 878 quantified lesions (median lesions/patient [inter-quartile range]: 7 [4 - 12]). By FES PET, 53 patients had soft-tissue only lesions; 30 with mixed lesions; and 18 with bone-only lesions. Lesions were homogeneous (100% FES+) in 75% of patients and heterogeneous in 25%. Overall, median PFS was less in the heterogeneous vs. homogeneous group (5.5 vs. 21.6 months), with a corresponding HR of 5.4 (95% confidence interval [CI]: 3.2-9.4, p < 0.001). When the SUVmax threshold for FES+ was progressively raised beyond 1.8 for bone lesions, the HR numerically decreased (HR range: 1.8 [95% CI: 0.9-2.5] to 5.2 [95% CI: 3.0-8.9]), particularly starting at SUVmax ≥ 2.2 (HR: 4.4, 95% CI: 2.6-7.6). The HR also decreased when the threshold was raised for non-bone lesions instead (HR range: 1.5 [95% CI: 0.9-2.5] to 3.6 [95% CI: 2.2-6.0]), even at SUVmax ≥ 1.9 (HR: 3.4 [95% CI: 2.0-5.7]). No combinations of thresholds resulted in a higher HR than the original threshold of > 1.8 for all lesion types. Conclusions: FES was strongly predictive of PFS using a threshold of SUVmax > 1.8 for FES+. Raising the threshold for SUVmax did not improve risk stratification. Any amount of heterogeneity among lesions was associated with shorter PFS with ET compared to patients with homogeneous FES+ expression. The results support the utility of FES in predicting clinical benefit to ET.