Abstract PO1-12-06: Is it appropriate to select patients for primary prophylactic use of pegfilgrastim based on the risk of febrile neutropenia?

Abstract Background: Febrile neutropenia (FN), a major hematologic adverse event in perioperative chemotherapy for breast cancer, is more prevalent among Asians than in Caucasian populations. Four main guidelines, i.e., ASCO, EORTC, NCCN, and ESMO, provide recommendations regarding the appropriate use of hematopoietic growth factors. These guidelines recommend primary prophylaxis with granulocyte colony-stimulating factor (G-CSF) based on the estimated incidence of FN, primarily defined by the chemotherapy regimen. G-CSF administration is recommended when the incidence of FN is ≥20% or 10–20% with risk(s). However, the incidence of FN varies depending on the trial and design, and patient-specific risk factors for FN remain unclear. We previously performed a multicenter prospective cohort study in Japan, the CSPOR-BC FN study, evaluating the incidence of FN during perioperative chemotherapy for breast cancer. Herein, we conducted an additional analysis to explore thresholds of patient-specific risk factors for FN that could be evaluated before chemotherapy. We analyzed the risk factors for FN and verified the appropriateness of patient selection for prophylactic G-CSF administration. Methods: The previously reported CSPOR-BC FN study (PubMed ID=33631458) consecutively enrolled 1005 patients with stage I–III breast cancer between August 2015 and July 2017. In the current additional analysis, we intended to have 477 patients in the visiting group that evaluated true-FN, defined as ≥37.5℃ and grade 4 neutropenia. Multivariate analysis of risk factors for FN was performed, followed by the determination of cutoff values for age and pretreatment absolute neutrophil count (ANC) as risk factors. Results: Chemotherapy regimens administered were FEC (fluorouracil, epirubicin, and cyclophosphamide) in 170 patients (36%), E(A)C (epirubicin (doxorubicin) + cyclophosphamide) in 165 patients (35%), and TC (docetaxel + cyclophosphamide) in 142 patients (30%). The incidence of FN was 28.7% (N = 137). To evaluate the risk factors for FN, we predefined regimens, age, performance status (PS), stage, chemotherapy or radiotherapy history, infectious wounds, open wounds, ANC, and renal and liver dysfunction as potential risk factors. Multivariate analysis of risk factors for FN identified regimen (TC), age (≥65), low pretreatment ANC, treatment setting, and no primary pegfilgrastim prophylaxis as significant risk factors. Two patient-specific risk factors (age and pretreatment ANC) were analyzed. Logistic regression analysis revealed that age≥65 years was a significant risk factor (odds ratio=2.24, 95% confidence interval: 1.34–3.75). Applying a cutoff of 65 years (age) to predict FN, the sensitivity and specificity were 28.4 and 83.8%, respectively, which were inaccurate. Despite applying a cutoff of 67 years optimized from the receiver operating characteristic (ROC) curve, the sensitivity (22.4%), specificity (90.7%), and area under the curve (0.519) indicated low discrimination for predicting FN. The incidence of FN by age group was as follows: < 45 years, 30.8%; 45–55 years, 26.4%; 55–64 years, 20.9%; ≥65 years, 39.6%. Additionally, applying a cutoff of 2436/µL (pretreatment ANC) optimized from the ROC curve, the observed sensitivity (27.6%) and specificity (83.8%), and area under the curve (0.558) also indicated low discrimination for predicting FN. The incidence of FN by stratum of pretreatment ANC was as follows: < 2000, 33.3%; 2000–3000, 30.7%; 3000–4000, 29.9%; and >4000, 24.6. Accordingly, setting a cutoff value to predict FN incidence using age and/or pretreatment ANC was deemed inappropriate. Conclusion: Selecting patients for primary prophylactic G-CSF based on the existing FN risk can be challenging. Primary prophylactic G-CSF would be considered for every patient undergoing perioperative chemotherapy for breast cancer, especially Asians. ROC curve for FN and age. The AUC was 0.519. ROC curve for FN and neutrophil count. The AUC was 0.558. Citation Format: Kazutaka Narui, Takashi Ishikawa, Ikumi Takashima, Kosuke Kashiwabara, Yukari Uemura, Yuichiro Kikawa, Naruto Taira, Hirofumi Mukai. Is it appropriate to select patients for primary prophylactic use of pegfilgrastim based on the risk of febrile neutropenia? [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-12-06.