Evaluation of factors associated with platinum-sensitivity status and survival in limited-stage small cell lung cancer patients treated with chemoradiotherapy

In this retrospective study, we analyzed the association of clinicopathological factors and therapeutic plans with platinum-sensitivity status and survival of limited-stage small cell lung cancer (LS-SCLC) patients. We enrolled 452 LS-SCLC patients with 279 platinum sensitive and 173 platinum refractory patients. The low serum neuro-specific enolase levels (NSE; p = 0.011), neutrophil-to-lymphocyte ratios (NLR; p = 0.013) and higher objective response rates (p = 0.003) were associated with sensitive group but not the refractory group. Multivariate analysis showed that treatment modality (HR = 0.267, p < 0.001), serum lactate dehydrogenase (LDH; HR = 1.894, p = 0.016), NLR (HR = 2.043, p = 0.043) and platinum-sensitivity status (HR = 0.561, p = 0.036) were independent prognostic factors for survival. We further showed that the numbers of chemotherapy cycles and response to first-line therapy were independent prognostic factors for refractory patients only. Our study demonstrates that platinum-sensitivity status is of prognostic importance, as it is strongly associated with survival in LS-SCLC patients.


INTRODUCTION
Nearly 13%-20% of lung cancer cases are small cell lung cancer (SCLC) [1], which have high incidence of widespread metastasis [2]. About 30% SCLC patients are diagnosed as limited disease (LD) and 70% as extensive disease (ED) [3]. Although patients with limited stage small-cell lung cancer (LS-SCLC) respond well to a combination of chemo-and radio-therapy, high recurrence rates result in low survival [4,5]. The recurrent cancer is generally refractory to therapy due to drug-resistant cancer cells and remains a problem for LS-SCLC survival [6].
LS-SCLC patients are denoted as platinum refractory or platinum sensitive based on shorter or longer relapse times (90 days from last platinum administered), respectively [7]. Patients with longer treatment-free interval generally demonstrate objective response (OR) to the same chemotherapy regimens used in initial treatment [8][9][10]. According to the National Comprehensive Cancer Network (NCCN 2016 edition) guidelines, the chosen second-line chemotherapy regimens [11] and the objective response rate (ORR) of second-line treatment were mainly dependent on the length of time from the end of chemotherapy to disease progression [12]. Although platinum-sensitivity status is important in clinical treatments, the risk factors in LS-SCLC are not known.
The circulating tumor cells (CTC) [13] and molecular factors such as vascular endothelial growth factor (VEGF) are widely considered as independent indicators for LS-SCLC survival [14]. However, the association between sensitivity to platinum-based treatment and overall survival is unclear [15]. Therefore, in this retrospective study, we evaluated the correlation of various clinicopathological factors with the platinumsensitivity status and overall survival of LS-SCLC patients.

Patient characteristics
The characteristics of patients were summarized in Table 1. We enrolled 452 LS-SCLC patients with a Clinical Research Paper median follow-up of 35.1 months. Among these, 279 patients were platinum sensitive and 173 patients were platinum refractory. The median age was 56 years (range: 27-82 years) and 75.2% of patients were male. Eastern Cooperative Oncology Group performance status (ECOG PS) score was 0-1 for 357 patients (79%) and 2-3 for 95 patients (21%). 71.2% patients had smoking history and 86.3% of these patients smoked more than 30 pack-years.

Univariate analysis of clinical factors associated with platinum-sensitivity status
The comparison of clinical variables in the platinum sensitive and refractory patient groups is shown in Table  2. Univariate analysis showed that ECOG PS scores were better in the sensitive patients compared to refractory patients (0-1 in sensitive vs. 0-1 in refractory, 81.7% vs. 74.6%, p = 0.070). The objective response (complete plus partial responses) was also higher in the sensitive patients compared to refractory patients (75.3% in sensitive vs. 60.7% in refractory, p = 0.001). Furthermore, more number of sensitive patients received chemoradiotherapy compared to the refractory group (85.3% vs. 78.0% p = 0.048). The tumor size (p = 0.149) and white blood counts (WBC; p = 0.306) were statistically similar between the two groups of patients. The neuro-specific enolase (NSE) levels were lower in the sensitive patients compared to refractory patients (44.44
We then compared mean NLR and platelet-tolymphocyte ratio (PLR) to disease status (Table 4). In platinum-sensitive group, NLR decreased from 3.83 to 2.45 (p < 0.001) after four cycles of chemotherapy, and then increased to 4.12 (p < 0.001) at progression. In the refractory group, NLR decreased from 4.66 to 4.10 (p = 0.03), but sharply increased to 5.11 at progression (p = 0.006). PLR also demonstrated similar trend in the two patient groups, although the differences were not significant when compared in terms of disease status (Table 4).

Prognostic factors for LS-SCLC survival
The median overall survival (OS) of LS-SCLC patients was 19.7 months and median progression-free survival (PFS) was 10.4 months. Platinum-sensitivity status affected survival and PFS rates differentially. The platinum sensitive group demonstrated better survival than the refractory group (21.1 months vs.16.8 months, log rank p < 0.001, Figure 1). The PFS rates were also better for the sensitive group in comparison to the refractory patients (13.2 months vs. 8.4 months, log rank p < 0.001, Figure 2).

DISCUSSION
In this retrospective study of LS-SCLC patients, we evaluated the correlation between clinical test results and factors with platinum-sensitivity status and survival. Our analysis revealed that NSE, NLR and objective response were independent predictors of platinum-sensitivity status. Prognostic factors for survival included treatment modality, LDH, NLR and platinum-sensitivity status.
While the first-line treatment guidelines are clear, there is ambiguity regarding treatment for relapsed LS-SCLC patients. Ardizzoni et al. demonstrated that response to the second-line drug, topotecan, was 38% and 10% in sensitive and refractory patients, respectively [16]. In the latest meta-analysis, Horita et al. demonstrated that topotecan response rates for sensitive and refractive patients were 17% and 5% and the corresponding one-year OS rates were 27% and 9%, respectively [17]. Therefore, distinguishing sensitive and refractory LS-SCLC patients early could result in therapeutic benefits.
NSE levels are associated with response to therapy, disease stage and overall survival and it can be used to monitor disease progression and recurrence. Yan et al demonstrated that survival rates were significantly different between low and high NSE level groups in small cell carcinoma of esophagus [18]. The histological characteristic of small cell carcinoma of esophagus was similar to SCLC. Therefore, our data showed that the sensitivity of platinum-based chemoradiotherapy was associated with serum NSE levels before treatment.
Recently, a series of inflammatory factors, such as NLR, PLR and C-reactive protein (CRP) were shown to be correlated with a poor prognosis in various types of cancer [19][20][21]. NLR was involved with tumor growth, invasion and metastasis through neutrophil elastase activity and suppression of the adaptive immune system [22,23]. However, NLR was not a stable tumor-specific biomarker since it was easily affected by treatment and radiation-induced inflammation [24,25]. Therefore, some reports suggested that other inflammatory factors should be evaluated with NLR [19,26]. Furthermore, NLR levels always changed with treatment and disease courses [27]. Our results suggested that NLR was a critical parameter in evaluating tumor response and monitoring disease progression in LS-SCLC.
We also evaluated if therapeutic efficacy could distinguish sensitive and refractory patients. Nagy-Mignotte et al. demonstrated complete response in 60.1% sensitive patient group compared to 9.7% in the refractory patient category [12]. Johnson et al. demonstrated effective response in 12.5% sensitive SCLC patients [28]. Therefore, NSE, NLR and objective response provide reliable and important predictive information about platinum-sensitivity status that would be helpful to weigh benefits of different treatments in the two groups.
Many studies have reported that good PS score, younger age and early stage are associated with survival benefit in LS-SCLC [29][30][31]. However, there is ambiguity due to variable biomarkers usage and analysis limitations. Some studies have suggested that radiotherapy in LS-SCLC patients yields a 5-10% improvement in two-year survival rate, and a corresponding 20%-35% increase in local control compared to chemotherapy alone [32,33]. Our results indicated that patients with combined therapy were associated with decreased hazard ratios (HRs) compared to chemotherapy alone. Furthermore, many studies also suggested that early concurrent thoracic radiation was superior to late concurrent and sequential radiotherapy [34][35][36]. There was no association between sequential thoracic therapy and survival in our study,   [41]. Moreover, Stokkel reported that LDH was an independent prognostic factor for tumor progression and patient survival [42]. However, Li et al. showed that high LDH levels were more frequent in advanced-stage SCLC patients and did not correlate with long-term survival [43]. Similarly, Brueckl et al. demonstrated that LDH was not associated with OS due to its high correlation with WBC [44]. The differences among the results in these studies might be due to different approaches to measuring survival, disparate study designs or variables and sample sizes. Establishing an association between serum LDH level and SCLC progression would be beneficial to effectively monitoring therapy response, as previously shown [45,46].
The prognostic role of NLR in LS-SCLC has not been investigated in detail. Kang et al. demonstrated that patients with low NLR (NLR < 4) at diagnosis had relatively longer OS and PFS compared to those with high NLR (≥ 4) [27]. However, Wang et al. did not find significant association between NLR and OS in LS-SCLC [47]. Apart from assessing the prognostic role of NLR, the cut-off value of NLR is worth to be discussed. Some studies demonstrated significant correlation between NLR and overall survival with different cutoff values of NLR [20,21]. Yamanaka et al. showed that advanced gastric cancer patients with low NLR (< 2.5) had better OS [48].  [50]. However, there are some contradictory reports. Lara et al. investigated 3 clinical trials based on advanced stage SCLC and did not find association between platinum-sensitivity status and progression-free survival (PFS) (p = 0.49) or overall survival (OS) (p = 0.14) [51]. This study used different methodologies for clinical staging and data collection compared to our study. Based on our data, we postulate that response to platinum-based chemotherapy is prognostic.
Adequate chemotherapy doses can significantly improve the short-term benefits and PFS of LS-SCLC patients. In the present study, we observed that patients treated with no less than four cycles of chemotherapy showed higher median survival although the data was not statistically significant in multivariate analysis. This could be because most of patients cannot tolerate six cycles of chemotherapy, especially concurrent chemoradiotherapy    due to serious side effects like gastrointestinal irritation and bone marrow suppression. Furthermore, long term chemotherapy before radiotherapy led to drug-resistant cancer cells that formed small distance metastases and reduced the efficiency of radiotherapy [52]. Subgroup analysis demonstrated that more than four cycles of chemotherapy and objective response to initial treatment were favorable prognostic factors only in refractory patients. This was probably due to different patient characteristics and response to successive lines of chemotherapy in two groups. The refractory group was mostly made up of elderly and poor PS patients, with greater reliance on chemotherapy cycle numbers and the drug dose. Nagy-Mignotte et al. showed that sensitive patients benefit more than refractory patients from subsequent therapy [12]. As a consequence, the response to initial therapy was critical in improving disease prognosis.
Serum LDH levels were related with PFS in the sensitive group, but not in the refractory group. Previous studies demonstrated that elevated LDH levels were associated with poor survival and chemo-/radio resistance in malignancies as they signified hypoxia inducible factor induced tumor aggressiveness [53].
Apart from basic clinical and laboratory factors, many molecular and genetic studies have explored novel prognostic factors that require invasive examinations and increase the economic burden on patients [54,55]. Our study suggested that routine laboratory test results are prognostic and can provide information for precision treatment. By identifying patients with sensitive disease, patients can be treated with lower total radiotherapy and chemotherapy doses to achieve good results. At the same time, patients that are more likely to progress can be switched to alternative treatments earlier to reduce high drug toxicity.
This study is limited because it is retrospective and due to heterogeneity of the characteristics and therapies administered in individual trials. Furthermore, confounding factors like race and socioeconomic status [56,57], which were not taken into consideration in  this study. Also, the small number of patients and the single center study can result in intrinsic bias. Finally, since the radiotherapy technique is undergoing constant improvements, the platinum-sensitivity status between patients at different times is hard to compare in a retrospective study. Therefore, multi-center prospective study with larger patient pools is needed to confirm our results. In conclusion, our study demonstrates that platinum-sensitivity status is a critical prognostic factor that determines the survival rates in LS-SCLC patients.

Patient enrollment criteria
We retrospectively enrolled LS-SCLC patients that were diagnosed by cytology or histology between January 2005 and December 2010 at Shandong Cancer Hospital and received at least one cycle of chemotherapy. The cytological or histological diagnosis of disease was performed by mediastinoscopy, bronchofiberoscopy and biopsy of lymph nodes. For all patients, standard evaluation before treatment included brain magnetic resonance imaging (MR), bone scintigraphy and CT imaging of chest and abdomen. Positron emission tomography-computed tomography (PET/CT) was not routinely performed. Tumor stage was determined according to the Veterans Administration Lung Study Group system [58]. This study was approved by the Research Ethics Committee of Shandong Cancer Hospital, China. Informed consent was obtained from all participants.

Therapeutic treatment and tumor response
Patients that did not receive standard therapeutic strategy were excluded from this study. The chemotherapy regimens included etoposide plus platinum and irinotecan plus platinum. Chemotherapy doses were modified based on individual toxicity levels and blood counts. Patients underwent radiotherapy either by 3-dimensional conformal radiotherapy (3D-CRT) or intensity modulated radiotherapy (IMRT). The gross tumor volume (GTV) included primary tumor and positive lymph nodes. The clinical tumor volume (CTV) was drawn from GTV with 8 mm margin. The planning target volume (PTV) was determined from GTV with less than 1.5 cm in 3D. Patients with disease progression in less than 90 days after treatment were denoted as platinum refractory, whereas patients with disease progression at 90 or more days were identified as platinum sensitive. During chemotherapy, CT scan and the specific biomarkers of tumor were accessed and analyzed every two or three cycles to determine tumor response. Based on Response Evaluation Criteria in Solid Tumor (RECST) guidelines [59], tumor response to first-line treatment was subdivided into complete response (CR), partial response (PR), stable disease (SD) and progression disease (PD).

Follow-up and data collection
The clinical data was obtained from medical records. Re-examinations included physical examination, tumor biomarkers, routine laboratory test, medication history and CT scan. Chest and abdomen CT imaging was performed one month after initial treatment, with follow-up every 2-3 months in the first year, then every sixth months, second year onwards. Brain MR and bone scintigraphy were not compulsory and were administered according to clinical requirements.
In this study, clinical and demographic indicators including age, gender, smoking status, ECOG PS and treatment modality were collected using medical record system at baseline. Also, routine laboratory tests that were obtained from patient records included WBC, hemoglobin (HB), PLT, LDH, sodium (Na), γ-glutamyl transferase (GLUT) at diagnose. Cyfra21-1, NSE, carcinoembryonic antigen (CEA), NLR and PLR were recorded at diagnosis, after four cycles of chemotherapy, and at the time of progression. The NLR was calculated from the differential counts by dividing the neutrophil number by the lymphocyte number. The PLR was calculated by dividing the platelet count by the lymphocyte count. The cutoff values of variables were determined using the clinical normal range and the related research results [27]. Parameters with more than 20% missing observations were excluded from the study.

Statistical analysis
Survival time was measured from the time of diagnosis to death or the last follow-up date. Progressionfree survival was determined from the time of therapy initiation to the time of disease progression or death. Date of recurrence was determined by the date of positive CT imaging results. The platinum-sensitivity status was determined based on the time between last platinum administrations to date of recurrence.
All statistical analyses were performed using SPSS software version 19.0 for windows (SPSS Inc., Chicago, IL, USA). Categorical variables were expressed as percentage and compared between groups using chi-square test. Continuous variables were presented as mean ± standard deviation (s.d.) or median and range with a Mann-Whitey U test performed for comparison. Variables with p < 0.1 were considered statistically significant (two-sided). Logistic regression with backward stepwise method was used for multivariate analysis. Patient survival curves were measured by the Kaplan-Meier method and compared by log-rank test. Univariate Cox proportional hazard models were used to identify factors associated with survival and PFS. Multivariate Cox regression analyses were performed to determine the factors that were significant based on univariate analysis. Variables with p < 0.05 were considered statistically significant (two-sided).