Baseline predictors of hematological toxicity in patients with advanced cancer treated with ATR inhibitors in phase I/II clinical trials.

Abstract

3111 Background: Ongoing trials are exploring ATR inhibitors (ATRi) in genomically selected contexts. However, myelosuppression, particularly anemia, has limited the therapeutic window of this class of drugs. We sought to discover clinical biomarkers predicting severe hematological toxicity from ATRi. Methods: We retrospectively analyzed clinical parameters and peripheral blood cell indices retrieved from complete blood count (CBC) reports of patients (pts) pre- and during treatment with an oral ATRi on phase I/II trials at our center. Pts received ATRi monotherapy or in combination with a PD1 inhibitor (PD1i) or a PARP inhibitor (ATRi+PARPi) in dose-escalation and expansion cohorts, which included ATRi at potentially toxic doses. Results: 37553 indices from 2209 CBC reports of 141 pts treated with an ATRi from 10/2017 to 1/2022 were analyzed. 132 (93.6%) pts received ATRi +/- PD1i; 9 (6.4%) pts received ATRi+PARPi. The incidences of ≥ grade (G) 3 anemia, neutropenia and thrombocytopenia were 47.5%, 31.9% and 11.4%. 73/141 (51.8%) pts received red cell transfusion. Baseline risk factors predicting ≥G3 anemia on univariate analysis included: lower median (med) hematocrit (Hct) (hazard ratio (HR) (95% confidence interval) = 3.05 (1.82, 5.13) ≤med vs > med; p < 0.0001), hemoglobin (Hb) (HR = 2.74 (1.64,4.57) ≤med vs > med; p = 0.0001), mean corpuscular Hb concentration (HR = 1.85 (1.11,3.10) ≤med vs > med; p = 0.019); and higher median immature reticulocyte fraction (HR = 0.43 (0.25,0.71) ≤med vs > med; p = 0.0012), reticulocyte count (ct) (HR = 0.59 (0.35,0.97) ≤med vs > med; p = 0.037) and red cell distribution width (RDW) (HR = 0.54 (0.33,0.88) ≤med vs > med; p = 0.015). On multivariate analysis, lower median Hct (HR = 3.76 (2.15, 6.6) ≤med vs > med; p < 0.0001), higher immature granulocyte ct (HR = 1.71 (1.30, 2.25) per 1 fold increase; p = 0.0001), higher RDW (HR = 7.83 (1.70, 36.03) per 1 fold increase; p = 0.0082) and higher ATRi starting dose (HR = 1.40 (1.05, 1.86) per 1 fold increase; p = 0.022) significantly predicted ≥G3 anemia risk. Baseline risk factors for ≥G3 neutropenia on univariate analysis included: lower median absolute neutrophil ct (ANC) (HR = 2.26 (1.18, 4.33) ≤med vs > med; p = 0.015) or white blood cell ct (WBC) (HR = 2.73 (1.40, 5.33) ≤med vs > med; p = 0.0032). On multivariate analysis, lower median WBC (HR = 2.85 (1.45, 5.59) ≤med vs > med; p = 0.0024) was associated with higher risk of neutropenia, while ATRi+PARPi increased risk of neutropenia (ANC < 0.75) (HR = 4.15 (1.40, 12.3); p = 0.01) and thrombocytopenia (HR = 3.90 (1.47, 10.4); p = 0.0064). Conclusions: ≥G3 anemia was frequent in pts receiving ATRi. At baseline, lower median Hct and higher RDW predict severe anemia, while lower WBC predicts neutropenia from ATRi. ATRi+PARPi has increased risk of neutropenia and thrombocytopenia vs ATRi +/- PD1i. These indices may inform patient selection and CBC monitoring for future ATRi trials.