Risk stratification for febrile neutropenia in patients with testicular germ cell tumors

Abstract The aim of this study was to detect risk factors for febrile neutropenia (FN) in patients with testicular germ cell tumors (TGCT). In this retrospective cohort study at the Medical University of Graz, we included 413 consecutive TGCT patients who received adjuvant or curative treatment with cisplatin‐based chemotherapy. FN occurred in 70 (16.9%) of 413 patients. In univariable logistic regression, higher age (odds ratio (OR) per 5 years = 1.17, 95% CI: 1.02–1.35, P = 0.022), reduced performance status (PS) (OR = 2.73, 1.47–5.06, P = 0.001), seminomatous histology (OR = 2.19, 1.26–3.78, P = 0.005), poor IGCCCG risk class (OR = 4.20, 1.71–10.33, P = 0.002), and prior radiotherapy (pRTX) (OR = 8.98, 2.09–38.61, P = 0.003) were associated with a higher risk of FN. In multivariable analysis adjusting for age and risk classification, only poor PS (OR = 2.06, 1.05–4.03, P = 0.035), seminomatous histology (OR = 2.08, 1.01–4.26, P = 0.047), and pRTX (OR = 7.31, 1.61–33.17, P = 0.010) prevailed. In the subgroup of seminoma patients (n = 104), only pRTX predicted for FN risk (OR = 5.60, 1.24–25.34, P = 0.025). Five of eight seminoma patients with pRTX developed FN (63%), as compared to 22 FN cases (23%) in the 96 seminoma patients without pRTX (P = 0.027). The eight seminoma patients who received pRTX had significantly lower pre‐chemo white blood counts (4.7 vs. 6.5 G/L), neutrophil counts (3.2 vs. 4.3 G/L), and platelet counts (185 vs. 272 G/L) than patients without pRTX (all P < 0.0001). TGCT patients with a reduced performance status or who had been previously treated with radiotherapy have an increased risk for neutropenic fever during chemotherapy.


Introduction
Testicular germ cell tumors (TGCT) represent the most common solid malignancy among men aged 15-40 years and can be divided into two subgroups: seminomatous (SGCT) and nonseminomatous germ cell tumors (NSGCT) [1]. Long-term prognosis of testicular germ cell tumors (TGCT) is excellent. The 5-year survival rate of patients with TGCT in stage I approximates 99%. Treatment options for stage I seminoma are active surveillance, adjuvant treatment with one cycle of carboplatin or adjuvant radiotherapy [2,3]. NSGCT patients with stage I disease can be managed with active surveillance or may be offered one adjuvant chemotherapy cycle with bleomycin, etoposide, and cisplatin (BEP) [4]. For metastatic cases, the International Germ Cell Cancer Collaborative Group (IGCCCG) has identified three prognostic groups: good, intermediate, and poor risk. With optimal management the 5-year survival rate approximates 91%, 79% and 48% in metastatic disease with good, intermediate and poor
However, treatment is associated with non-negligible complications and a negative impact on quality of life [7][8][9][10][11]. Therefore, optimizing the complication rate of TCGT treatment holds promise to further increase the already favorable prognosis of these patients.
Febrile neutropenia (FN) is a serious complication of myelosuppressive chemotherapy. FN rates during curative BEP chemotherapy are between 10 and 20%. Routine prophylactic use of granulocyte colony-stimulating factors (G-CSF) is not recommended [12]. A personalized FN risk stratification approach to the patient with TCGT in need of systemic chemotherapy may optimize the indication and thus risk-benefit ratio of prophylactic G-CSF support in this population. Thus, the aim of this study was to identify prognostic factors for FN in order to delineate subgroups with the highest risk of FN.

Subjects
All consecutive patients (n = 960) with histologically confirmed TGCT, presenting to the Division of Oncology at the Medical University of Graz between January 1994 and September 2013, were retrospectively reviewed. Patients were initially staged using computed tomographic (CT) scans of the abdomen, CT scan or X-ray of the chest, and postoperative tumor markers αfetoprotein (AFP), human chorionic gonadotropin (HCG), and lactate dehydrogenase (LDH). Patients with disseminated disease were stratified according to the IGCCCG risk classification [13,14].
Follow-up data were retrieved until January 2015. Follow-up investigations at our center were performed according to a local protocol and were adapted in 2007 and 2012 according to recent publications [15][16][17]. Because the primary endpoint of this study was the risk of febrile neutropenia in patients undergoing cisplatin-based chemotherapy, patients who were managed with active surveillance, adjuvant or curative radiotherapy, adjuvant carboplatin and did not experience relapse were excluded from further analysis (n = 547). A total of 413 patients who received adjuvant or curative cisplatin-based chemotherapy during the course of disease were selected for this study (Table 1). A total of 377 (91%) of 413 patients had not received any treatment (except orchiectomy) prior to cisplatin-based chemotherapy, eight (2%) patients had undergone adjuvant or curative radiotherapy to paraaortic/ iliac lymph nodes prior to chemotherapy and one (0%) patient had received one cycle of adjuvant carboplatin. Because 27 patients received several cycles of cisplatinbased chemotherapy, all cycles were counted until an episode of FN occurred. In patients without any FN episode, all cycles of cisplatin-based chemotherapy were counted. Electronic and paper medical records of all 413 consecutive TGCT patients were retrospectively reviewed, and febrile neutropenia was documented in our in-house administrative system. FN was defined as an oral temperature of >38.3°C or two consecutive readings of >38.0°C for 2 h and an absolute neutrophil count (ANC) of <0.5 × 10 9 /L, or expected to fall below 0.5 × 10 9 /L [18].
The study was approved by the Institutional Review Board of the Medical University of Graz (No. 26-196 ex 13/1).

Statistical methods
All statistical analyses were performed using Stata (Version 14.0, Stata Corp., Houston, TX) and IBM SPSS Statistics (Release 23.0.0. 2015. Chicago (IL), USA: SPSS Inc., an IBM Company). Continuous variables were reported as medians [25th-75th percentile] and count data as absolute frequencies (%). Means between two or more groups were compared with t-tests, rank-sum tests, and Kruskal-Wallis tests, respectively. Spearman's rank-based correlation coefficient was used for examining the correlation between radiotherapy parameters and blood counts. The association between FN and clinical covariables was quantified with uni-and multivariable logistic regression [19].

Discussion
The present study demonstrated and confirmed that FN is a frequent complication in TGCT patients undergoing curative or adjuvant cisplatin-based chemotherapy. In detail, the FN risk in our cohort was 17%. However, treatment delay was negligible and the case fatality rate of FN in TGCT appears to be very low. Nevertheless, 80% of patients with FN had to be admitted to hospital for a median duration of 1 week, which obviously comprises quality of life during the course of chemotherapy and increases healthcare costs. With establishment of current guidelines and identification of "low-risk" patients, the admission rate nowadays might be lower. However, primary G-CSF support may have a role in reducing the burden of FN in this population. Guidelines of the major societies in the field, such as the EORTC algorithm, recommend primary G-CSF support when the predicted FN risk is above 20%. In case the FN risk is assumed to be between 10% and 20%, the decision on whether or not to prescribe primary G-CSF support should be supported by an individual assessment of FN risk factors for each patient [12]. Our data show that with a FN risk of 17%, TCGT patients fall into this "individual risk assessment" group. However, the risk factors that should be considered for this individual risk assessment according to the EORTC algorithm, such as age > 65 years poor performance status, poor nutrition status, female sex, anemia, and comorbidity, are not very prevalent in the often young and minimally comorbid TGCT population. Therefore, our study aimed to identify FN risk factors specifically for TGCT patients in order to facilitate FN risk assessment.
In detail, we could confirm a previous report by Feldman et al. that even age above 50 years is an important risk factor for FN in TGCT patients. These authors observed a 44% FN risk in TCGT patients above 50 years of age, which prompted them to recommended primary G-CSF administration for all patients in this age group [20]. Two other strong FN risk factors identified by our analysis were poor performance status and poor IGCCCG risk classification.
Moreover, seminomatous histology and prior radiotherapy emerged as risk factors for FN in the overall study cohort. Because radiotherapy is only used as adjuvant or curative treatment in SGCT and not NSGCT patients, this might explain why seminomatous histology appeared to be a risk factor for developing FN ("confounding by radiotherapy"). Clinically, a link between prior radiotherapy and a higher FN risk is highly plausible [21]. The acute depletion of bone marrow components following irradiation has been ascribed to the direct effect of radiation depleting the stem cell compartment [22]. The ability of the bone marrow compartment to recover and regenerate is dependent on the volume of bone marrow within the irradiated field. Radiotherapy to paraaortal lymph nodes involves around 25% of bone marrow [22,23]. In our analysis, the eight seminoma patients who received pRTX had significantly lower pre-chemo white blood counts, neutrophil counts, and platelet counts than patients without pRTX (all P < 0.0001). The mean time from radiotherapy to relapse and subsequent chemotherapy was 1 year. This supports the hypothesis that irradiated bone marrow might not have the ability to recover or that a time interval of 1 year is too short for recovery which makes the risk of FN for relapsed seminoma patients that high.
We did not observe that the magnitude of reduction in mean blood count was associated with certain radiation parameters like radiation dose or radiation field size among seminoma patients even though there were differences in radiation dose and volume. The differences in dose and volume prescriptions are attributable to the long observation period and the change in radiation dose from 30 Gy to 20 Gy after the results from the EORTC trial 30942 in 2005 [24,25].
Our study is limited by its retrospective character and the small sample size of the subgroup of relapsed irradiated seminoma patients. The most used regimen in our study was PEB chemotherapy (in 96% of patients) and only 4% received multiple chemotherapy regimens. Our risk stratification therefore mainly applies for TGCT patients who are treated with PEB chemotherapy as firstline treatment.
In summary, we observed a considerable risk of FN in men with TGCT undergoing cisplatin-based adjuvant or curative chemotherapy. Our study identified (1) higher age, (2) poor performance status, (3) poor IGCCCG risk classification, and (4) prior radiotherapy in the seminoma subpopulation as risk factors for FN in patients with testicular cancer.

Conflict of Interest
None of the contributing authors have any conflict of interests, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript.