The utility of risk factors proposed in a prospective clinical trial in the management of ER-positive, HER2-negative metastatic breast cancer patients: Feedback from the real world

Abstract

Background: Some risk factors (RFs) in the management of ER-positive, HER2-nega-tive metastatic breast cancer (ERθHER2-MBC), such as a shorter disease-free interval (DFI), visceral involvement or high tumor burden, have been identified in prospective clinical trials; however, the utility of those RFs in the real world has not been well discussed. Methods: We reviewed our medical records from 2002 to present to assess the utility of RFs (DFI<24 months [DFI<24M]; visceral metastases [VIS]; prior (neo)adjuvant anthracycline and/or taxane [A/T];or > 3 metastatic organs [>3 ORG]) defined in the TURANDOT risk factor analyses (Brodowicz T, Br J Cancer, 2014), a first-line bevaci-zumab trial of HR+HER2-MBC patients. According to the analysis, patients with >2 RFs were classified as “high-risk (HiR)”, and others were classified as “low-risk (LoR)”. Statistical analyses were performed using the Kaplan-Meyer method and a multivariate COX regression analysis. Results: We identified 311 ER+HER2-MBC (224 recurrent, 87 advanced) patients who underwent chemotherapy (CTx). The most common RF at the initiation of first-line CTxwas VIS (N = 186, 59.8%), followed by A/T, >3 ORGandDFI<24M. The distri-bution of RFs was as follows: 0 in 89 (28.6%), 1 in 93 (29.9%), 2 in 94 (30.2%), 3 in 30 (9.6%), and 4 in 5 (1.7%). The survival from the initiation ofCTx(OSCTx) was signifi-cantly poorer in HiR patients than LoR ones (median 815.0 vs. 1062.0 days, p < 0.001, log-rank) in all MBC patients, as well as in 87 advanced BC patients (median 825.0 vs. 1160.0 days, P < 0.05, log-rank). There was no significant difference in the OSCTx between patients with 0 and 1 RF (P = 0.90). In addition, in recurrent BC (rBC) patients, there was no significant difference in the OSCTx between patients with 2 and >3 RFs (P = 0.10). Multivariate analyses revealed >3 ORG and DFI<24M as significant RFs (P < 0.05) for all rBC patients (hazard ratios 1.61 and 1.49, respectively). Conclusions: Our review suggests that RFs such as high tumor burden and shorter DFI identified in prospective randomized studies are applicable to patients in the real world, even with heterogeneous backgrounds.