Abstract P2-08-38: Influence of prognostic factors on outcomes among metastatic breast cancer patients treated with CDK4&6 inhibitors in routine clinical practice

Abstract

Background: Evidence suggests that there are clinical features associated with a less favorable prognosis among patients with HR+/HER2- metastatic breast cancer (MBC) such as metastases to non-bone sites, including liver and lung, and negative progesterone receptor (PgR-) status. The objective of this study was to compare baseline characteristics and outcomes between those with and without these clinical factors among a cohort of HR+/HER2- MBC patients treated with a CDK4&6 inhibitor (CDK4&6i). Methods: This was a retrospective analysis of the Flatiron Health electronic health records-derived database for US patients diagnosed with MBC between 1/1/2011 and 9/30/2017. The study included a random sample of patients with HR+/HER2- MBC who were treated with a CDK4&6i on or after 6/30/2016. Baseline variables, including demographics, comorbidities, and sites of metastasis, were recorded at start of the first CDK4&6i containing line of therapy in the metastatic setting on or after this date. Dates of real-world progression were abstracted from patient charts. Descriptive statistics and appropriate statistical tests were used to compare baseline characteristics between patients with or without select clinical factors associated with unfavorable outcomes. In patients who received a CDK4&6i-based therapy, Kaplan–Meier methods and univariable Cox proportional hazards models were used to assess real-world progression free survival (rwPFS) by line from start of line to the date of first progression or death within line (unadjusted for treatment and other potential confounders). Results: 518 patients were included in this study. Median age at metastatic diagnosis was 66y (IQR; 59-73y); 99% female and 11.4% had PgR- status. At baseline, 20.5%, 46.3%% and 65.8% of patients had liver, visceral (defined as liver and/or lung), and non-bone only metastases, respectively. Among a total of 207 patients who received a CDK4&6i as initial therapy in the metastatic setting, 69.1% received it in combination with an aromatase inhibitor, 29.5% received it in combination with fulvestrant, and 1.4% as monotherapy. Within the same group, 58 had disease progression or died during first line (1L); median rwPFS measured from start of 1L was not reached (95% CI: 10.7 months, NA). Univariable analyses revealed the presence of liver metastases was associated with a higher risk of progression or death compared to no liver metastases (HR: 2.04, 95% CI: 1.13 - 3.68). Having non bone-only metastases was associated with a higher risk of progression or death compared to having bone-only metastases (HR: 2.23, 95% CI: 1.20 – 4.15). Univariable analyses did not reveal any statistically significant differences in first-line rwPFS by PgR status or presence of visceral metastases. Results from other lines of therapy are forthcoming. Conclusion: In a real world data set, and consistent with prior prospective data, presence of liver and non-bone only metastases were associated with a higher risk of progression among patients with HR+/HER2- MBC receiving initial therapy with a CDK4&6i. The heterogeneity of prognoses among this population reinforces the need to consider these clinical features in treatment decisions for optimal patient outcomes. Citation Format: Saverno KR, Carter GC, Li L, Battiato LA, Huang Y-J, Smyth EN, Price GL, Sheffield KM, Baxi SS, Kuk D, Seidman AD. Influence of prognostic factors on outcomes among metastatic breast cancer patients treated with CDK4&6 inhibitors in routine clinical practice [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-08-38.

Predicting optimal treatment regimens for patients with HR+/HER2- breast cancer using machine learning based on electronic health records

“…It is less likely that patients with liver metastasis received CDK4 & 6 inhibitor-based therapies based on current prescribing patterns (Supplementary Table 1). The presence of liver metastasis is a risk factor for poor survival [40,41]. As a result, it may be that the estimated optimal treatment regimen favored CDK4 & 6 inhibitor-based therapies simply because patients with liver metastasis were underrepresented in this class and so led to improved survival outcomes.…”

“…As the variable "Liver metastasis" was not present in the FHD, it was believed to have possibly been a potential unmeasured confounder in the main analyses. The SPD showed that presence/absence of liver metastasis was associated with survival [41,42]. Furthermore, liver metastasis status was unevenly distributed across regimen classes (with much lower prevalence among patients who received CDK4 & 6 inhibitor-based therapies; Supplementary Table 1).…”

“…The binary covariate in the SPD "Liver metastasis" was added to this data generation model. Liver metastasis status was simulated with strong association with the outcome and an uneven distribution over regimen classes and correlations with the rest of covariates, as observed in the SPD (Supplementary Table 1) [41,42]. By varying the hazard ratio associated with the "Liver metastasis" variable in the Weibull model and its distribution across treatment classes, the synthetic data were reanalyzed using the same RSF-based approach as for the FHD data.…”

“…Sensitivity analyses -evaluating the potential effect of an unmeasured confounder (liver metastasis) in the FHD data Liver metastasis was strongly associated with survival and unevenly distributed across regimen classes (with much lower baseline prevalence among patients who received CDK4 & 6 inhibitor-based therapy; Supplementary Table 1) [41,42]. The RSF method was repeatedly applied (N = 1000) to 80% of the synthetic training data.…”