A Multicenter Study of Romiplostim to Treat Chemotherapy-Induced Thrombocytopenia in Solid Tumors and Lymphoid Malignancies

Introduction: Chemotherapy-induced thrombocytopenia (CIT) is a common complication in cancer patients causing chemotherapy treatment delays, dose reductions, and discontinuation. A major unmet clinical need, there is no FDA-approved agent available to manage CIT. Given this, we developed clinical pathways to treat CIT at 4 large academic hospitals. Data using romiplostim to manage CIT is limited to small single-center studies and case series of solid tumor patients only. We performed the first large multicenter study of romiplostim for CIT, including both solid tumor and non-myeloid hematologic malignancy patients.

Methods: We retrospectively collected demographics, cancer and chemotherapy data, romiplostim dosing, weekly platelet counts (Plt), and outcomes of romiplostim support for patients treated on institutional romiplostim CIT pathways from 2009-2019. For solid tumor patients, 2 institutions utilized intracycle dosing (romiplostim administered on average twice monthly on chemotherapy-off weeks) and 2 utilized standard weekly romiplostim dosing, facilitating comparison of dosing strategies. Romiplostim response was defined as a median on-romiplostim Plt of ≥75×10⁹/L and at least 30×10⁹/L higher than baseline. Baseline characteristics were analyzed for predictors of romiplostim non-response using multivariate logistic regression. Wilcoxon signed-rank and Mann-Whitney U tests compared paired and unpaired groups and Chi square and Fisher's exact tests compared categorical outcomes.

Results: 173 patients (153 solid tumor, 20 lymphoid malignancy) were treated for CIT with romiplostim to facilitate ongoing chemotherapy administration, with 170 (98%) receiving a median (range) of 4 (1-36) additional chemotherapy cycles over 10 (2-125) weeks of romiplostim support. A total of 61.3 patient-years of romiplostim treatment of CIT were analyzed for effectiveness and safety. Table 1 lists baseline patient characteristics and Table 2 details chemotherapy regimens precipitating CIT and supported on romiplostim. Median per-patient Plt on romiplostim was significantly higher than baseline (all patients, 112×10⁹/L vs. 54×10⁹/L, P<0.001; solid tumor patients, 116×10⁹/L vs. 60×10⁹/L, P<0.001), Figure 1. Median romiplostim starting dose was 3 µg/kg/week, which was also the median optimized romiplostim dose (dose achieving Plt >100×10⁹/L). While on romiplostim, bleeding occurred in 7.5% of patients and venous thromboembolism occurred in 4.6% (at Plt 65-306×10⁹/L). Compared with intracycle romiplostim dosing, weekly romiplostim dosing resulted in higher median Plt (Figure 2) and fewer chemotherapy delays/dose reductions with no difference in bleeding or thrombotic events (Table 3). Characteristics of solid tumor patients predicting non-response to romiplostim per multivariate regression included biopsy-proven bone marrow invasion by tumor (P<0.001), prior pelvic irradiation (P=0.029), and prior temozolomide (P=0.031) (Figure 3), with only 23%, 20%, and 46% of patients, respectively, achieving a romiplostim response compared with 82% of all other solid tumor patients. Plt improvement with romiplostim was modest in patients with myeloma (N=7, median per-patient Plt 39×10⁹/L with romiplostim vs. 19×10⁹/L at baseline, P=0.078) and aggressive lymphoma (N=13, 47×10⁹/L with romiplostim vs. 23×10⁹/L at baseline, P=0.033), Figure 3, with only 35% receiving more than 1 additional chemotherapy cycle on romiplostim. All myeloma and lymphoma patients were dosed weekly and had extensive bone marrow involvement by malignancy.

Conclusions: We present results of romiplostim treatment of CIT in a large cohort of 173 cancer patients, including lymphoid malignancies. Dosing strategies were compared and predictors of non-response were evaluated. Romiplostim was highly effective in solid tumor patients with CIT, allowing chemotherapy administration. Bleeding and thrombosis rates were similar to the general cancer patient population. Both weekly and intracycle dosing were effective and equally safe, but weekly dosing maintained higher Plt overall and higher Plt nadirs with fewer chemotherapy dose reductions or treatment delays. Bone marrow invasion, pelvic irradiation, and prior temozolomide predicted romiplostim non-response. Romiplostim had modest effectiveness to manage CIT in lymphoid malignancy patients with underlying bone marrow involvement.

View largeDownload slide

View largeDownload slide

Close modal