Comparing the prognostic and predictive utility of serum thymidine kinase 1 and CA 15-3 in patients with hormone receptor positive metastatic breast cancer starting first-line endocrine therapy in SWOG S0226.

Abstract

1076 Background: Serum levels of thymidine kinase 1 activity (TKa), a fundamental enzyme in DNA synthesis and cellular proliferation, are prognostic of benefit from endocrine therapy (ET) in patients (pts) with hormone receptor positive (HR+) metastatic breast cancer (MBC) who participated in SWOG S0226, a prospective randomized trial comparing anastrozole vs. anastrozole and fulvestrant given in 4 week cycles. The assay for TKa (DiviTum TKa) was FDA approved in July 2022 for monitoring of postmenopausal HR+ MBC pts. The assay for circulating MUC1, CA 15-3, is routinely utilized in monitoring of pts with MBC. We compared the prognostic and predictive utilities of CA 15-3 and TKa in pts from S0226. Methods: TKa was measured in 1725 sera from 432 pts while CA 15-3 was measured in 1298 sera from 326 pts at baseline (BL), cycles 2, 3, 4 and 7. Prespecified cutoff levels of ≥ 250 DuA and ≥ 30 U/mL were considered high for TKa and CA 15-3 respectively. Progression-free survival (PFS) and overall survival (OS) were analyzed by Kaplan-Meier and Cox regression stratified by prior adjuvant tamoxifen use. To examine markers over time, a landmarked Cox regression analysis was done starting at initiation of Cycle 4. BL and recent (Cycle 3 if available, otherwise Cycle 2) markers were assessed for simultaneous prediction of subsequent PFS and OS. Results: BL TKa was elevated in 190/432 (44%) and CA 15-3 was elevated in 254/326 (78%). Agreement between assays was 53%. Pts with high versus low BL TKa had significantly worse PFS (median 11.6 vs. 17.2 months, HR = 1.68, 95% confidence interval (CI) 1.37-2.06) and OS (median 34 vs. 58 months, HR = 2.16, 95% CI 1.73-2.70) whereas pts with high versus normal BL CA 15-3 had no significant difference in PFS (median 13.6 vs. 16.1 months, HR = 1.23, 95% CI 0.93-1.62) but worse OS (median 45 vs. 66 months, HR = 1.82, 95% CI 1.30-2.53). A multivariable Cox model with both markers shows only TKa as being prognostic for PFS (TKa HR = 1.61, 95% CI 1.28-2.03; CA 15-3 HR = 1.21, 95% CI 0.91-1.59), but both prognostic for OS (TKa HR = 2.09, 95% CI 1.61-2.71; CA 15-3 HR = 1.70, 95% CI 1.22-2.38). In the landmarked multivariable analysis, positive TKa at BL was a strong predictor of PFS (HR = 1.65, 95% CI 1.28-2.14), but recent TKa was not (HR = 1.25, 95% CI 0.93-1.68). In contrast, positive CA15-3 at BL was not a predictor of PFS (HR = 0.73, 95% CI 0.46-1.17), but recent CA15-3 was (HR = 1.97, 95% CI 1.26-3.06). Conclusions: BL TKa is highly prognostic in pts with HR+ MBC initiating first-line systemic ET, with low BL TKa conferring superior prognosis. In contrast, CA 15-3 is only modestly prognostic at BL, but is prognostic after 3 cycles of treatment. These hypothesis-generating data suggest further study is needed to determine how these biomarkers should be employed in a complementary manner to monitor response to systemic therapy in HR+ MBC. Clinical trial information: NCT00075764 .