Chemotherapy-induced neutropenia as a prognostic factor in patients with extensive-stage small cell lung cancer

Chemotherapy-induced neutropenia (CIN) is a dose-limiting factor for cytotoxic chemotherapy, but recently, it was suggested that CIN contributes to prolonged survival. In this study, we examined the association between severe CIN and survival and determined whether CIN affected survival in patients with extensive-stage small cell lung cancer (ES-SCLC). The medical records from 214 patients with ES-SCLC treated with etoposide or irinotecan in combination with cisplatin (EP/IP) between 2012 and 2016 were collected and retrospectively analyzed. Landmark analysis was performed at the end of cycle 4, and the relationship between severe CIN and survival was determined by a log-rank test. In addition, a multivariate analysis using the COX proportional hazard model was performed to identify independent predictive factors. The Landmark analysis included 102 patients in the IP group and 47 patients in the EP group. No significant difference was found between grades 0–3 and grade 4 neutropenia and overall survival (OS) in the EP group (P = 0.57). Contrariwise, for the IP patients, the median OS was 444 days for grades 0–3 and 633 days for grade 4 neutropenia, which was significantly longer for patients who developed grade 4 neutropenia (P = 0.03). Multivariate analysis adjusted for potential factors revealed that the development of grade 4 CIN was identified as a significant predictor of longer OS (hazard ratio [HR], 0.50; 95% confidence interval (CI), 0.28–0.87, P = 0.015). The results indicated that the development of severe CIN with IP therapy is associated with prolonged OS.


Introduction
Lung cancer remains the leading cause of cancer deaths worldwide including Japan. Small cell lung cancer (SCLC) accounts for 10-15% of all lung cancers and is biologically, molecularly, and clinically very different from other lung cancers [1]. Although SCLC is more sensitive to chemotherapy and radiotherapy compared with other lung cancers, it recurs in most cases. SCLC may be classified into two types: limited stage and extensive stage (ES). ES-SCLC is a refractory disease with a poor prognosis and a 5-year survival rate of less than 5% [2].
Cisplatin combined with etoposide (EP) or irinotecan (IP) are two of the most commonly used chemotherapeutic regimens for first-line treatment of ES-SCLC. In particular, IP is widely used in Japan as first-line treatment based on the results of a Japanese phase III trial (JCOG 9511) and a meta-analysis including this clinical trial [3][4][5].
Chemotherapy-induced neutropenia (CIN) is a significant dose-limiting toxicity for many cytotoxic anticancer drugs, but an interesting association between the occurrence of CIN and a favorable clinical outcome has been reported in several solid tumors [6][7][8][9][10]. In lung cancer, post hoc analyses of clinical trials for non-small cell lung cancer have also reported that the occurrence of CIN contributes significantly to prolonged survival [11,12]. However, because of the limited number of SCLC cases, few studies have been conducted thus far to evaluate the prognostic role of CIN in patients with ES-SCLC. Furthermore, the JCOG 9511 study conducted in Japan for ES-SCLC revealed a high incidence of CIN [3]. Intriguingly, the incidence of severe CIN was higher compared with that reported in the non-Asians studies [13,14]. The uridine diphosphate glucuronosyltransferase (UGT) 1A1, which transforms SN-38 into SN-38 glucuronide, is a key metabolic enzyme involved in irinotecan. Since there are genetic polymorphisms in UGT1A1 and Japanese are known to be frequent carriers of *6 in addition to *28, such patients may lead to severe toxicity with irinotecan including neutropenia. Thus, the relationship between the development of severe CIN and survival is an important issue to consider. Therefore, we conducted a multicenter study in Japanese patients to determine the association between severe CIN development and survival in patients with ES-SCLC treated with IP or EP therapy.

Data source and data collection
Patients diagnosed with ES-SCLC during the 5-year period from January 1, 2012, to December 31, 2016, at 14 study sites in Japan were included. Patients aged 20 years or older who received cisplatin-based chemotherapy as first-line treatment were eligible. The regimens included in this study were IP (intravenous cisplatin 60 mg/m 2 on day 1 and irinotecan 60 mg/m 2 on days 1, 8, and 15, every 28 days) and EP (intravenous cisplatin 80 mg/m 2 on day 1 and etoposide 100 mg/m 2 on days 1 through 3, every 21 days), which are standard therapies for ES-SCLC.
The exclusion criteria were as follows: (1) patients without distant metastasis; (2) patients who received a chemotherapy regimen other than IP or PE as initial therapy; and (3) patients who received primary prophylactic administration of granulocyte colony-stimulating factor (G-CSF). Clinical records of patients with ES-SCLC receiving IP or PE therapy were extracted from the electronic medical records system maintained by each institution and retrospectively reviewed. Clinical parameters included age, sex, BMI, number of metastasis sites, baseline biochemical and hematological data (within 28 days of initiation of first-line chemotherapy), secondary prophylactic use of G-CSF, and the date of confirmed recurrence or death from any cause.

Patients and treatments
Patients who received standard doses of IP or EP as initial therapy for advanced and previously untreated SCLC were selected from the electronic medical record system of each participating institution. A total of 214 patients were selected and divided into two groups: those who received IP (n = 138) and those who received EP (n = 76).

Clinical data and assessment of CIN
Clinical data including patient characteristics and prognostic factors were collected from each hospital's electronic medical records. Hematologic toxicity was assessed based on the Common Terminology Criteria for Adverse Events, version 5.0 (CTCAE v5.0). CIN grade was determined based on the lowest neutrophil count for each patient recorded during first-line treatment. The endpoint of the study was severe CIN survival. This was defined as the time from the start date of cisplatin-based chemotherapy to the date of death or the last day of the follow-up period.

Statistical analysis
Neutropenia occurring during the treatment period was graded (grade 0-4) according to CTCAE v5.0. Survival curves for overall survival (OS) were generated for each grade of neutropenia using the Kaplan-Meier method and compared using the log-rank test. To analyze the effect of severe CIN development on clinical outcomes in patients with ES-SCLC, a multivariate analysis was performed using the COX proportional hazards model. Hazard ratios and 95% confidence intervals were calculated. A multivariate analysis with a P value < 0.05 was considered statistically significant. In this study, Landmark analysis [15] was performed, and patients who died within 112 days (equivalent to 4 cycles of IP) or 84 days (equivalent to 4 cycles of EP) of starting IP or EP and patients who received less than 4 cycles of IP or EP were excluded from the analysis. Patients with distant metastatic cancer may die early during chemotherapy, and these patients may receive less chemotherapy, which reduces the likelihood of developing CIN. In other words, patients with longer survival who receive more chemotherapy are more likely to develop CIN, which may lead to a large bias in the association of CIN with prolonged survival. To avoid this bias, we performed a Landmark analysis. Statistical analysis was performed using IBM SPSS Statistics version 22 (IBM Japan Ltd. Tokyo, Japan) and EZR version 1.55 [16].

Patient characteristics
A total of 214 patients with a diagnosis of ES-SCLC were selected from the electronic medical record system at each participating hospital. Of 138 patients who received IP, 35 patients received less than the planned four cycles of chemotherapy (112 days) or received four cycles, but died within 112 days, whereas 1 patient with missing neutrophil count data was excluded. Of the 76 patients who received EP, 28 patients received less than four cycles (84 days) or received four cycles but died within 84 days and were excluded. One patient with missing neutrophil count data was also excluded. A flow diagram of the patients who participated in the study is shown in Fig. 1, and the baseline and clinical characteristics are listed in Table 1. In the Landmark analysis, 84 of the 102 patients who received IP therapy died. Of the 47 patients who received EP, 43 died.

Relationship between OS and CIN
Of the 102 patients who received IP therapy, 84 were confirmed dead by the time of the analysis, and 18 were censored or lost to follow-up. Two groups were compared: the severe CIN group (grade 4) and the absent/mild CIN group (grade ≤ 3). Figure 2 shows the cumulative OS curves of the severe CIN group and the absent/mild CIN group. The  Of the 47 patients who received EP therapy, 43 were confirmed dead by the time of the analysis and 4 were censored or lost to follow-up. Two groups were compared: the severe CIN group and the absent/mild CIN group. Figure 3 shows the cumulative OS curves of the severe CIN group and the absent/mild CIN group. There was no statistically significant difference in median survival time between the two groups (443 days [95% CI, 281-533] vs. 411 days [95%

Discussion
In this study, we examined the relationship between OS and the occurrence of severe CIN to identify prognostic factors for patients with ES-SCLC that were treated with standard IP or EP as first-line therapy. A multivariate analysis revealed that pretreatment LDH levels ≥ 245 IU/L and grade 4 CIN were two significant independent factors that influenced the OS of patients treated with IP. High level of serum LDH has long been reported to be a poor prognostic factor in solid tumors, and it has also been reported to be a prognostic factor in SCLC [17,19]. This finding was consistent with the results of the current study and confirmed serum LDH as an important prognostic factor. We also showed that in addition to high serum level of LDH at baseline, the absence of severe CIN was an independent factor affecting the poor prognosis of patients with ES-SCLC treated with IP. In recent years, an integrated analysis of 6 randomized clinical trials in non-SCLC revealed that the occurrence of severe CIN has a positive impact on prognosis [20]. CIN  is not just an adverse event, but may serve as a marker of improved patient outcome. The development of grade 4 neutropenia in patients treated with IP was also associated with longer OS in our study. Although the detailed mechanism of the association between the development of severe CIN and favorable prognosis is unclear, it is believed that the hematological toxicity associated with the administration of cytotoxic drugs may be a surrogate marker for the biological activity of anticancer drugs [21].
Since hematologic toxicity, such as neutropenia, is a dosedependent adverse event, patients who do not develop CIN may have received inadequate doses of the anticancer drugs or may be resistant to treatment [22]. Currently, the dose of anticancer drugs is administered to the patient based on body surface area (BSA) or creatinine clearance [23]; however, it has been suggested that BSA-based dosing calculations may be inadequate [24]. In two randomized trials of gastric and breast cancer, a trend toward improved progression-free and recurrence-free survival were observed in the latter group compared with the BSA-based dosing chemotherapy group and the group in which the dose of anticancer drugs was adjusted based on the incidence of adverse reactions, such as CIN [25]. Although no studies have been conducted in SCLC, this retrospective study revealed that grade 4 CIN may be an indicator of IP response. In other words, lowgrade CIN results from an underdose of IP therapy, and prognosis may be improved by increasing the dose of the anticancer drugs. However, there is currently insufficient evidence for dose escalation for IP therapy in patients with SCLC; thus, it is necessary to conduct prospective clinical trials to evaluate dose escalation and safety. In contrast, there was no association between the development of CIN and prolonged survival in patients who received EP therapy as opposed to IP therapy. A high incidence of grade 3 or higher CIN was associated with EP therapy (92.2%) in a phase III trial of Japanese patients with ES-SCLC [3]. In the present study, we found that grade 3 or higher CIN occurred more frequently in the EP therapy (87.2%) group; in particular, 61.7% of the patients had grade 4 CIN. The incidence of FN in the group of patients who received EP therapy was 19.1% (9/47), which was higher compared with that of patients who received IP therapy (4.9%, 5/102). Thus, the occurrence of FN may have had a significant effect as a poor prognostic factor in patients treated with EP. In fact, the MST was 473 days (95% CI, 371-599) for the group without FN and 367 days (95% CI, 219-443) for the group with FN (log-rank, P = 0.040).
To our knowledge, there have been no studies on the relationship between the occurrence of CIN and ES-SCLC patient survival. This study showed for the first time that the occurrence of CIN is not associated with prognosis following EP therapy and the development of severe CIN is associated with a favorable prognosis following IP therapy. Irinotecan is converted to its active metabolite, SN-38, by carboxylesterases. SN-38 is inactivated and excreted into the bile by glucuronidation, which is catalyzed by UDP-glucuronosyltransferases [26]. Since irinotecan and SN-38 are both substrates for the ABC transporters, polymorphisms in these transporter genes may affect the pharmacokinetics of irinotecan and irinotecanassociated toxicity. The association between the development of CIN and polymorphisms in the UGT1A1 and ABCB1 genes is becoming clearer [27][28][29]. Unfortunately, data on UGT1A1 gene polymorphisms were only available for 21 patients in the IP group in this study, so we were unable to examine on this point in detail. It is expected that the prediction of response to IP therapy will be possible by examining these genetic polymorphisms prior to chemotherapy.
Recently, a three-drug combination chemotherapy regimen consisting of platinum and etoposide plus an immune checkpoint inhibitor (ICI) has been recommended as a standard regimen for ES-SCLC. The results of two phase III trials indicated the efficacy of triplet chemotherapy with atezolizumab or durvalumab, showing a median OS time of 12.3 and 12.9 months, respectively, for patients treated with ICIs [30,31]. The median OS of all patients treated with IP in this study was 423 days (approx.14.1 months), whereas the median OS of patients who developed grade 4 CIN was 555 days (approx. 18.5 months). Although a simple comparison cannot be made, the median survival time of patients treated with IP therapy is comparable to that of combined therapy with platinum, etoposide, and ICI. Moreover, there appears to be a clear trend toward improved median survival time, particularly in patients who developed grade 4 CIN. Combination chemotherapy regimens with ICI and anticancer drugs have been implemented worldwide for multiple cancer types; however, combination therapy with atezolizumab or durvalumab has been estimated to be less costeffective compared with conventional chemotherapeutic regimens for first-line treatment of patients with ES-SCLC [32,33]. In addition, biomarkers for predicting the effect of ICIs, such as PD-L1 expression and tumor mutational burden, have not been established in ES-SCLC [34]. Therefore, the selection of responders may be an important therapeutic strategy in conventional chemotherapy, in which the cost of treatment is relatively low.
There are several limitations to the current study. First, this was a retrospective study with small sample size. Second, the sample size was not determined a priori using hypothesis testing. Thus, even if some strong associations are shown, no causal inferences can be drawn. Third, although pretreatment serum LDH level was shown to be an influencing factor in this study, this study used a collective serum LDH values commonly used in clinical practice, not the isozymes of LDH. Although the serum LDH level may be influenced by other diseases such as infection, anemia, and heart failure, we were unable to examine all relevant factors. Also, the patients were Japanese, and it has been suggested that there are differences in the development of CIN based on race [35][36][37]. Thus, the findings of this study may not be applicable to patients in other countries or regions of Japan. Despite these limitations, the results of this study are important because they are the first to show that the occurrence of severe CIN with IP therapy is a prognostic factor for ES-SCLC and revealed that median OS is long in patients exhibiting severe CIN.

Conclusion
Severe CIN is a useful prognostic factor for ES-SCLC patients treated with IP therapy, but not EP therapy. These findings suggest that severe CIN may be used as a marker to predict a good clinical response to IP therapy in patients with ES-SCLC. Furthermore, if CIN can be predicted prior to the initiation of chemotherapy, physicians may be able to stratify patients receiving IP chemotherapy and monitor neutrophil counts and patient status. It may be possible to identify the ES-SCLC patients who are expected to have a longer survival. It is also anticipated that the use of prediction models for CIN will lead to individualized treatment for ES-SCLC.