If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Thoracic Surgery, Shanghai Pulmonary Hospital, Tongji University School of Medicine, Shanghai, People’s Republic of ChinaDivision of Epidemiology, Department of Health Sciences Research, Mayo Clinic College of Medicine, Rochester, Minnesota
We sought to build prognostic nomograms and identify novel prognostic factors in small-cell lung cancer (SCLC) incorporating both clinical data and peripheral blood markers.
Methods:
We analyzed 938 patients with SCLC (555 extensive stage SCLC [ES-SCLC] and 383 limited stage SCLC [LS-SCLC]) diagnosed between 1997 and 2012 from a single institution. We investigated the prognostic value of pretreatment neutrophil to lymphocyte ratio, platelet to lymphocyte ratio, red cell distribution width, hemoglobin, and other clinicopathological factors. Cox proportional hazards models determined the effects of multiple factors on overall survival (OS). Two nomograms were developed to predict median survival and 6- and 12-month OS for ES-SCLC, and 1- and 2-year OS for LS-SCLC.
Results:
In ES-SCLC, the multivariate Cox model identified neutrophil to lymphocyte ratio and red cell distribution width as significant prognostic factors for OS independent of age, Eastern Cooperative Oncology Group performance score, chest radiation, chemotherapy, liver metastases, and numbers of metastatic sites. In LS-SCLC, significant prognostic variables included platelet to lymphocyte ratio, age, smoking cessation, chest radiation, chemotherapy, surgery, and prophylactic cranial irradiation. The two nomograms show good accuracy in predicting OS, with a concordance index of 0.73 in both ES- and LS-SCLC.
Conclusion:
The two nomograms incorporating hematological markers could more accurately predict individualized survival probability of SCLC than the existing models.
Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: Analysis of the surveillance, epidemiologic, and end results database.
SCLC treatment remains unsatisfying, as minimal breakthroughs have occurred in the past decade. Despite high initial responses to therapy, most patients die from recurrent disease, and the median survival after diagnosis is estimated to be 8–20 months. To better predict the SCLC patients’ outcomes, various prognostic factors and models have been investigated, such as age, gender, performance score (PS), serum neuron-specific enolase (NSE), serum lactate dehydrogenase, the Spain prognostic index,
Pretreatment prognostic factors for survival in small-cell lung cancer: A new prognostic index and validation of three known prognostic indices on 341 patients.
The development of novel prognostic factors and models will enable a better treatment stratification for patients with SCLC.
Statistical prediction models are widely used for predicting cancer outcomes. Among those, the nomogram is a graphical presentation of the results from a statistical model, which utilizes combined prognostic factors in predicting outcome for a given patient. Individualized estimation of survival among patients with cancer could be useful for counseling patients in making treatment decisions and optimizing therapeutic approaches. However, to the best of our knowledge, no nomogram has been reported for SCLC.
Inflammation is a known critical component of cancer progression.
Neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) in peripheral blood have been proposed as reliable indicators of the host’s inflammatory status; they have been identified as both prognostic and predictive biomarkers in many types of cancer including non–small-cell lung cancer.
Neutrophil:lymphocyte ratio and intraoperative use of ketorolac or diclofenac are prognostic factors in different cohorts of patients undergoing breast, lung, and kidney cancer surgery.
Prognostic significance of neutrophil lymphocyte ratio and platelet lymphocyte ratio in advanced gastric cancer patients treated with FOLFOX chemotherapy.
Recently, two studies have shown that elevated red cell distribution width (RDW) level is also a marker of poor prognosis in non–small-cell lung cancer.
However, there is little information on prognostic relevance of the pretreatment NLR, PLR, and RDW in SCLC.
Therefore, we conducted this study to investigate the prognostic relevance of NLR, PLR, and RDW with regard to the overall survival (OS), stratified for patients with extensive stage SCLC (ES-SCLC) and limited stage SCLC (LS-SCLC). We also sought to develop two prognostic nomograms that incorporate NLR, PLR, RDW, and other important clinicopathological variables.
MATERIALS AND METHODS
Study Cohort and Data Collection
Since 1997, all patients with a pathologic diagnosis of primary lung cancer evaluated and treated at Mayo Clinic, Rochester, MN, have been prospectively enrolled and followed for outcome research, using protocols approved by the Mayo Clinic Institutional Review Board (IRB Number: 225–99), and all participants have provided written informed consent. Detailed procedures of patient enrollment, diagnosis, data collection, and follow-up have been described in previous publications.
Between January 1, 1997 and December 31, 2012, a total of 1558 patients with a pathologically confirmed diagnosis of SCLC have been enrolled. Of these, 938 SCLC cases met our study inclusion criteria as they had a complete blood count with leukocyte differential performed before any treatment. Excluded from consideration were patients with leukemia or lymphoma. We excluded all atypical NETs and mixed types of SCLC. A full medical record abstraction was carried out to obtain demographics, history of tobacco exposure, lung cancer pathologic type, clinical staging, and treatment. All patients were actively followed up. Annual verification of patients’ vital status was accomplished through Mayo Clinic’s electronic medical records and registration database, next-of-kin reports, death certificates, and obituary documents filed in the patients’ medical records, as well as through the Mayo Clinic Tumor Registry and Social Security Death Index website.
Analyses were applied separately in ES- and LS-SCLC. The NLR was categorized into two groups based on a cut-off value of 5
When necessary, natural log transformations of continuous variables were calculated for several laboratory values including NLR, PLR, and RDW to reduce the distribution skewness. A simplified comorbidity score was calculated to evaluate comorbidity conditions.
oncoLR health network. A new simplified comorbidity score as a prognostic factor in non-small-cell lung cancer patients: Description and comparison with the Charlson's index.
Clinical data are reported as mean ± standard deviation or median (full range). Cumulative survival is estimated with a Kaplan–Meier model using the time of diagnosis as the starting point. Univariate and multivariate Cox proportional hazards models are used to assess prognostic factors including clinicopathological variables and pretreatment hematologic markers (i.e., continuous or dichotomized levels when appropriate).
Cox Proportional Hazards models determined the effects of multiple factors on a nomogram, and only the factors with a p value less than 0.05 were incorporated into the nomogram. Two separate nomograms were developed, one for ES-SCLC and another for LS-SCLC, to predict median survival and 6- and 12-month OS, or 1- and 2-year OS, respectively.
The performance of the nomogram is assessed using the concordance index (C-index) and calibration curve. The predictive accuracy of OS is estimated using the C-index. A larger C-index is associated with a more accurate prognostic prediction. Two hundred bootstrap resamples were used for internal validation of the accuracy of predictions and to avoid overfitting the model. Calibration refers to whether the predicted probabilities agree with observed probabilities, which is generated by plotting the predicted survival probabilities against the actuarial outcome. In a well-calibrated model, the calibration curve should be close to 45°.
All statistical analyses were carried out using SAS 9.3 (SAS Institute Inc., Cary, NC) and R version 3.0.2 (The R Foundation for Statistical Computing, Vanderbilt University, Nashville, TN) with the rms and survival libraries. All p values were two-tailed.
RESULTS
Characteristics of All Patients
With a median follow-up time of 10.8 months, 856 deaths (91.3% of the 938 total) have been observed. The median age at the time of diagnosis was 68 years (range 27–91 years). The median follow-up time for the surviving patients was 7.8 years. Nine hundred twenty-one (921) patients (98.8%) were former or current smokers.
Prognostic Effect of Blood Markers
The demographics and clinical information of 555 ES-SCLC and 383 LS-SCLC patients are summarized in Table 1. Elevated PLR, NLR, and RDW were associated with extensive stage disease. Supplementary Table 1 (SDC 1, http://links.lww.com/JTO/A846) shows the 6 month, 1-, 2-, and 3-year survival by hematological markers. Low hemoglobin (p = 0.008) and elevated PLR (p < 0.0001), NLR (p < 0.0001) and RDW (p < 0.0001) were significantly associated with a worse prognosis (Supplementary Table 1, SDC 1, http://links.lww.com/JTO/A846).
TABLE 1Characteristics of All SCLC by Stage
ES-SCLC (n = 555)
LS-SCLC (n = 383)
Characteristic
No. (%)
No. (%)
p
Age at diagnosis
0.9499
Mean (SD)
66.7 (10.3)
66.7 (10.0)
Gender
0.0032
Female
237 (42.7%)
201 (52.5%)
Male
318 (57.3%)
182 (47.5%)
Smoking status
0.3384
Never
12 (2.2%)
5 (1.3%)
Former
200 (36.0%)
151 (39.4%)
Current
343 (61.8%)
227 (59.3%)
Pack-year
0.0025
Missing
49 (0.0%)
21 (0.0%)
0–19
47 (9.3%)
16 (4.4%)
20–39
132 (26.1%)
74 (20.4%)
40–59
138 (27.3%)
127 (35.1%)
>60
189 (37.4%)
145 (40.1%)
Smoking cessation
0.0693
Quit or never smoker
375 (67.6%)
280 (73.1%)
Never quit
180 (32.4%)
103 (26.9%)
ECOG performance status
<0.0001
<2
399 (71.9%)
331 (86.4%)
>2
156 (28.1%)
52 (13.6%)
BMI
0.4278
Missing
14 (0.0%)
7 (0.0%)
<25
187 (34.6%)
142 (37.8%)
25–30
230 (42.5%)
144 (38.3%)
>30
124 (22.9%)
90 (23.9%)
Therapy
<0.0001
No treatment
114 (20.5%)
19 (5.0%)
Surgery with adjuvant therapy
6 (1.1%)
52 (13.6%)
Chemotherapy or chest radiation only
292 (52.6%)
75 (19.6%)
Chemotherapy plus chest radiation
143 (25.8%)
237 (61.9%)
Chemotherapy
<0.0001
No
125 (22.5%)
36 (9.4%)
Yes
430 (77.5%)
347 (90.6%)
Chest radiation
<0.0001
No
399 (71.9%)
126 (32.9%)
Yes
156 (28.1%)
257 (67.1%)
PCI
<0.0001
No
536 (96.6%)
293 (76.5%)
Yes
19 (3.4%)
90 (23.5%)
Platinum agent
0.0042
No chemotherapy
125 (0.0%)
36 (0.0%)
No
41 (9.5%)
13 (3.7%)
Yes
376 (87.4%)
318 (91.6%)
Unknown
13 (3.0%)
16 (4.6%)
Chemotherapy-agent combination
0.0022
No chemotherapy
125 (0.0%)
36 (0.0%)
VP16 + CDDP/CBP
351 (81.6%)
305 (87.9%)
Other combination
66 (15.3%)
26 (7.5%)
Unknown
13 (3.0%)
16 (4.6%)
Liver metastases at baseline
No
328 (59.1%)
Yes
227 (40.9%)
Numbers of metastatic sites at baseline
<2
369 (66.5%)
>2
186 (33.5%)
NLR
<0.0001
Median (range)
4.4 (0.2–60.3)
3.1 (0.2–56.7)
PLR
0.0001
Median (range)
190.0 (2.3–3944.4)
160.1 (23.4–1034.8)
RDW
0.0189
Median (range)
13.5 (10.1–24.5)
13.3 (11.3–23.3)
Hemoglobin (g/dl)
0.8319
Median (range)
13.4 (4.3–20.4)
13.3 (8.0–18.1)
Any other cancer
0.4722
No
458 (82.5%)
309 (80.7%)
Yes
97 (17.5%)
74 (19.3%)
COPD
0.0104
No
415 (74.8%)
257 (67.1%)
Yes
140 (25.2%)
126 (32.9%)
Diabetes
0.8326
No
488 (87.9%)
335 (87.5%)
Yes
67 (12.1%)
48 (12.5%)
Cardiovascular disease
0.1060
No
408 (73.5%)
263 (68.7%)
Yes
147 (26.5%)
120 (31.3%)
SCS
0.0415
<9
473 (85.2%)
307 (80.2%)
≥9
82 (14.8%)
76 (19.8%)
ES-SCLC, extensive stage small-cell lung cancer; LS-SCLC, limited stage small-cell lung cancer; NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; RDW, red cell distribution width; ECOG, Eastern Cooperative Oncology Group; BMI, body mass index; PCI, prophylactic cranial irradiation; CDDP, cisplatinum; CBP, carboplatinum; COPD, chronic obstructive pulmonary disease; SCS, simplified comorbidity score.
Kaplan–Meier survival estimates for NLR, PLR, RDW, and hemoglobin are shown in Supplementary Figures 1–4 (SDC 2, http://links.lww.com/JTO/A847). In ES-SCLC, elevated NLR (p < 0.001) and RDW (p < 0.001) were associated with poor prognosis. In LS-SCLC, low hemoglobin (p = 0.048), elevated PLR (p = 0.001), and NLR (p = 0.044) were associated with poor prognosis.
Nomograms Development
In univariate analysis, NLR, RDW, PLR, age, gender, Eastern Cooperative Oncology Group (ECOG) PS, chest radiation, chemotherapy, liver metastases, and numbers of metastatic sites at diagnosis were significantly associated with OS in ES-SCLC (Supplementary Table 2, SDC 1, http://links.lww.com/JTO/A846). For patients with LS-SCLC, possible predictors for OS included PLR, age, smoking cessation, chest radiation, chemotherapy, ECOG PS, surgery and prophylactic cranial irradiation (PCI) (Supplementary Table 3, SDC 1, http://links.lww.com/JTO/A846).
Tables 2 and 3, respectively, summarize the findings of the multivariate Cox proportional hazards analyses in ES- and LS-SCLC. All significant prognostic variables were used to build the nomograms.
TABLE 2Multicovariate Cox Regression Model for Overall Survival in Extensive Stage SCLC
Variable
HR (95% CI)
p
Loge (RDW)
2.81 (1.32–6.01)
0.0093
Loge (NLR)
1.41 (1.24–1.59)
<0.0001
PLR (vs. <210)
≥210
0.827 (0.672–1.017)
0.0718
Age at diagnosis
1.01 (1.001–1.02)
0.0304
Gender (vs. female)
Male
1.15 (0.97–1.37)
0.1128
ECOG performance status (vs. <2)
≥2
1.68 (1.25–2.24)
0.0008
Chest radiation (vs. no)
Yes
0.809 (0.66–0.99)
0.0376
Chemotherapy (vs. no)
Yes
0.24 (0.18–0.34)
<0.0001
Liver metastases (vs. no)
Yes
1.23 (1.03–1.48)
0.0263
Numbers of metastatic sites (vs. <2)
≥2
1.39 (1.15–1.67)
0.0007
NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; RDW, red cell distribution width; ECOG, Eastern Cooperative Oncology Group; HR, hazard ratio; SCLC, small-cell lung cancer; CI, confidence interval.
The nomogram of ES-SCLC included the following variables: NLR, RDW, age at diagnosis, ECOG PS, chemotherapy, chest radiation, liver metastases, and numbers of metastatic sites (Fig. 1). The nomogram assigned points based on NLR and RDW in a continuous but nonlinear fashion. Outcomes were reported as 6 and 12 months OS and median survival. The nomogram of LS-SCLC included the following variables: PLR, age at diagnosis, smoking cessation, chemotherapy, chest radiation, surgery, and PCI (Fig. 2). Outcomes were reported as 1- and 2-year OS and median survival.
FIGURE 1Nomogram for 6- and 12-month survival and median survival for extensive stage SCLC patients, including data derived from 555 patients and 547 mortality events. The nomogram is used by adding up the points identified on the points scale for each variable. The total projected on the bottom scale indicates the probability of 6- and 12-month survival and median survival. SCLC, small-cell lung cancer; NLR, neutrophil to lymphocyte ratio; RDW, red cell distribution width; PS, Eastern Cooperative Oncology Group performance status; Liver metastases, liver metastases at baseline; Metastatic sites, numbers of metastatic sites at baseline.
FIGURE 2Nomogram for 1- or 2-year survival and median survival for limited stage SCLC patients, including data derived from 383 patients and 314 mortality events. The nomogram is used by adding up the points identified on the points scale for each variable. The total projected on the bottom scale indicates the probability of 1- or 2-year survival and median survival. SCLC, small-cell lung cancer; PLR, platelet to lymphocyte ratio; PCI, prophylactic cranial irradiation; surgery, surgery with adjuvant therapy.
The nomograms that predicted OS were well calibrated with the Kaplan–Meier observed OS at 6 and 12 months in ES-SCLC (Supplementary Figure 5, SDC 2, http://links.lww.com/JTO/A847) and at 1 and 2 years in LS-SCLC (Supplementary Figure 6, SDC 2, http://links.lww.com/JTO/A847). The bootstrap C-index of the nomogram were both 0.73. A histogram of nomogram-predicted probabilities is shown in Figure 3 and illustrates the heterogeneity in predicted outcome within two stages.
FIGURE 3Histogram of nomogram-predicted overall survival. Note: the heterogeneity of predicted probabilities of time to recurrence within each stage.
We have developed and internally validated two nomograms that assign predictions for OS based on NLR, PLR, RDW, and other clinicopathological variables in a series of 938 patients from a single institution. We propose that the two nomograms provide more individualized OS predictions and could help patients and clinicians in the treatment decision-making process.
Neutrophil:lymphocyte ratio and intraoperative use of ketorolac or diclofenac are prognostic factors in different cohorts of patients undergoing breast, lung, and kidney cancer surgery.
Prognostic significance of neutrophil lymphocyte ratio and platelet lymphocyte ratio in advanced gastric cancer patients treated with FOLFOX chemotherapy.
Pretreatment neutrophil count as an independent prognostic factor in advanced non-small-cell lung cancer: An analysis of Japan Multinational Trial Organisation LC00-03.
in the peripheral blood of cancer patients may reflect the extent of systemic inflammation elicited by cancer cells, which have been identified as poor prognostic markers in many types of cancer. A definitive explanation underlying these findings has not been clearly elucidated yet. The systemic inflammatory reaction results in neutrophilia, thrombocytosis, and relative lymphocytopenia. Elevated NLR and PLR may reflect relatively depleted lymphocyte populations, possibly impairing the host immune response to malignancy. In SCLC, thrombocytosis was also found to indicate a poor prognosis.
In our study, elevated NLR and RDW represent significant independent prognostic indicators in ES-SCLC (loge NLR, hazard ratio [HR] = 1.41, 95% confidence interval [CI]: 1.24–1.59, p < 0.0001; loge RDW, HR = 2.81, 95% CI: 1.32–6.01, p = 0.0093). Elevated PLR was associated with significantly worse outcomes (HR = 1.60; 95% CI: 1.18–2.18; p = 0.0028) in LS-SCLC. Most of the former studies use categorical variables of NLR or RDW to assess the prognosis. We found that both continuous and categorical variables of NLR and RDW are significant independent prognostic indicators (results not shown). When building a nomogram in ES-SCLC, we modeled NLR and RDW as continuous variables because continuous variables could preserve more information than categorical variables.
In addition to NLR, PLR, and RDW, we identified that age, chemotherapy, and chest radiation were independent prognostic factors in both ES- and LS-SCLC. Other independent prognostic factors included smoking cessation, PCI, and surgery in LS-SCLC, and PS, liver metastases and number of metastatic sites in ES-SCLC. These are consistent with previous reports.
SCLC is relatively homogeneous as most patients are treated with chemotherapy and/or radiation, yet survival outcomes vary from one individual to another. Patient prognosis is currently estimated on the basis of the old Veterans’ Administration and AJCC TNM staging classification,
International Association for the Study of Lung Cancer International Staging Committee and Participating Institutions. The IASLC Lung Cancer Staging Project: Proposals regarding the relevance of TNM in the pathologic staging of small cell lung cancer in the forthcoming (seventh) edition of the TNM classification for lung cancer.
not on other factors like age, gender, smoking, or PS. By integrating additional significant prognostic factors, a nomogram could be applied to more accurately estimate an individual patient’s survival. Based on statistical models, our nomograms allow for individualized survival probability estimation for ES- and LS-SCLC, which discriminate better than the older Veterans’ Administration staging system (Fig. 3). Several scoring systems have been established for the prognosis of SCLC, such as the Spain prognostic index
Pretreatment prognostic factors for survival in small-cell lung cancer: A new prognostic index and validation of three known prognostic indices on 341 patients.
Moreover, our nomograms are better able to predict OS for individual patients than the scoring systems that stratify patients into a few risk groups. In addition, the nomograms have great potential of estimating risk in clinical trial design, which could be used for stratification in randomized studies based on their survival probability.
The performance of a nomogram needs to be assessed by calibration and discrimination. The C-index reflects the predictive accuracy of a nomogram. In this study, internal validation demonstrated good discrimination power (C-index, 0.73 in both ES- and LS-SCLC). The nomograms were well calibrated for predictions of OS (Supplementary Figures 5–6, SDC 2, http://links.lww.com/JTO/A847).
Peripheral blood markers would be valuable in SCLC, given that most patients with SCLC are not operative candidates and their primary tumors are rarely available for extensive analyses. Assessment of the peripheral blood markers may be easier and more cost–effective than conventional tumor markers, such as serum NSE and carcinoembryonic antigen in clinical practice. We used peripheral blood markers to build nomograms, which could be readily available for validation in any other clinical settings.
There are several limitations to this study. These models are based on a specific population treated at a tertiary medical center. Our nomograms were built and validated internally, and they should be externally validated in a larger number of patients at multiple institutions. Finally, we did not include some known prognostic factors, such as the level of lactate dehydrogenase, albumin, NSE, carcinoembryonic antigen, and other prognostic factors. The addition of these markers in future studies may improve the predictive ability of the two nomograms, which is also one of the benefits of this type of prognostic model.
CONCLUSION
In summary, we have identified that elevated NLR and RDW in ES-SCLC, and elevated PLR in LS-SCLC are poor prognostic factors. Our study constitutes the first two nomograms to accurately predict individualized survival probability in SCLC. These models could assist clinicians and patients in clinical decision-making and treatment tailoring. These results could be used to define proper stratification factors in future clinical trials.
ACKNOWLEDGMENTS
The authors appreciate Ms. Monique E. Smith, for her professional editing and technical assistance with the manuscript.a
Changing epidemiology of small-cell lung cancer in the United States over the last 30 years: Analysis of the surveillance, epidemiologic, and end results database.
Pretreatment prognostic factors for survival in small-cell lung cancer: A new prognostic index and validation of three known prognostic indices on 341 patients.
Neutrophil:lymphocyte ratio and intraoperative use of ketorolac or diclofenac are prognostic factors in different cohorts of patients undergoing breast, lung, and kidney cancer surgery.
Prognostic significance of neutrophil lymphocyte ratio and platelet lymphocyte ratio in advanced gastric cancer patients treated with FOLFOX chemotherapy.
oncoLR health network. A new simplified comorbidity score as a prognostic factor in non-small-cell lung cancer patients: Description and comparison with the Charlson's index.
Pretreatment neutrophil count as an independent prognostic factor in advanced non-small-cell lung cancer: An analysis of Japan Multinational Trial Organisation LC00-03.
International Association for the Study of Lung Cancer International Staging Committee and Participating Institutions. The IASLC Lung Cancer Staging Project: Proposals regarding the relevance of TNM in the pathologic staging of small cell lung cancer in the forthcoming (seventh) edition of the TNM classification for lung cancer.
The work was supported by the USA National Institutes of Health Grants (R03 CA77118, R01 CA80127, R01 CA84354, and R01CA115857) and Mayo Clinic Foundation, Foundation for Youths of Shanghai Municipal Health Bureau (2012093), and Shanghai Hospital Development Center Grant (Grant No. shdc12012111).
Disclosure: The authors declare no conflict of interest.
33550-4&id=gr1.jpg){kind=link}
33550-4&id=gr2.jpg){kind=link}
33550-4&id=gr3.jpg){kind=link}