Prognostic value of pretreatment circulating neutrophils , monocytes , and lymphocytes on outcomes in lung stereotactic body radiotherapy

Purpose In the present study, we determined the association of pretreatment circulating neutrophils, monocytes, and lymphocytes with clinical outcomes after lung stereotactic body radiotherapy (sbrt). Methods All patients with primary lung cancer and with a complete blood count within 3 months of lung sbrt from 2005 to 2012 were included. Overall survival (os) was calculated using the Kaplan–Meier method. Factors associated with os were investigated using univariable and multivariable Cox proportional hazards regression. Fine–Gray competing risk regression was performed to test the association of the neutrophil:lymphocyte (nlr) and monocyte:lymphocyte (mlr) ratios with two types of failure: disease-related failure and death, and death unrelated to disease. Results Of the 299 sbrt patients identified, 122 were eligible for analysis. The median and range of the nlr and mlr were 3.0 (0.3–22.0) and 0.4 (0.1–1.9) respectively. On multivariable analysis, sex (p = 0.02), T stage (p = 0.04), and nlr (p < 0.01) were associated with os. On multivariable analysis, T stage (p < 0.01) and mlr (p < 0.01) were associated with disease-related failure; mlr (p = 0.03), nlr (p < 0.01), and sbrt dose of 48 Gy in 4 fractions (p = 0.03) and 54 Gy or 60 Gy in 3 fractions (p = 0.02) were associated with disease-unrelated death. Median survival was 4.3 years in the nlr≤3 group (95% confidence interval: 3.5 to not reached) and 2.5 years in the nlr>3 group (95% confidence interval: 1.7 to 4.8; p < 0.01). Conclusions In lung sbrt patients, nlr and mlr are independently associated with os and disease-unrelated death. If validated, nlr and mlr could help to identify patients who would benefit most from sbrt.


INTRODUCTION
Stereotactic body radiotherapy (sbrt)-also called stereotactic ablative radiotherapy (sabr)-has become the standard of care for patients diagnosed with early-stage non-small-cell lung carcinoma (nsclc) who are medically inoperable or who refuse surgery 1 .Stereotactic body radiotherapy also has been explored in patients who are operable 2 .The technique can be used in patients of advanced age and with significant medical comorbidities, demonstrating good clinical outcomes and minimal toxicity 3 .As sbrt becomes readily available, selecting patients who will benefit from treatment in the setting of advanced medical comorbidities is an increasing clinical challenge 4 .
Given the advanced age of and prevalence of comorbidities in patients who commonly receive sbrt for nsclc, clinical selection metrics that are minimally invasive and readily available are most appropriate.Increasing evidence suggests that circulating levels of neutrophils, monocytes, and lymphocytes are associated with variable tumour control and survival in tumour sites, including head-and-neck cancer 5 , ovarian cancer 6 , colorectal cancer 7 , and cervical cancer a .A recent meta-analysis of 13,964 patients participating in twenty-six studies demonstrated that elevated platelet:lymphocyte ratio (plr) is a negative predictor for overall survival (os) with a hazard ratio of 1.60 [95% confidence interval (ci): 1.35 to 1.90; p < 0.01) 8 .High neutrophil:lymphocyte ratio (nlr) has also been shown to have poor prognostic implications in patients with chronic medical conditions such as coronary artery disease 9 and end-stage renal disease 10 .Utilizing prognostic indicators obtained from a complete blood count (cbc) to inform decision-making in patients with early-stage lung cancer considering sbrt is attractive, given that it is readily obtained, minimally invasive, and low cost.
The purpose of the present study was to determine the association of pre-treatment circulating neutrophils, monocytes, and lymphocytes with clinical outcomes after lung sbrt in patients with early-stage nsclc.

Patient Selection
Patients with clinical T1-2N0M0 nsclc who had received lung sbrt at our institution from January 2005 to January 2012 and who had cbc results within the 3 months before the start of sbrt were included 5,11 .Clinical and treatment data were abstracted from a prospective database that contains information on patient demographics, treatment, toxicities, patterns of recurrence, and cause of death.Patients with multiple courses of radiotherapy, metastatic lung tumours, or multiple synchronous primaries were excluded.The study was conducted with research ethics board approval.
In most cases, computed tomography (ct) imaging of the thorax, combined positron-emission tomography and ct imaging, and brain magnetic resonance imaging or ct imaging were used for pre-sbrt staging.The pathologic diagnosis was obtained where possible.Patients were treated according to the institutional sbrt protocol, mostly with 1 of 3 dose regimens: 48 Gy in 4 fractions (12 Gy per fraction delivered every other day) for peripheral tumours 3 cm or less in size, 54 Gy or 60 Gy in 3 fractions (without inhomogeneity correction; 18 Gy per fraction delivered every other day) for peripheral tumours larger than 3 cm in size, or 60 Gy in 8 fractions (7.5 Gy per fraction delivered once daily) for central tumours 3 .Patient follow-up included chest radiography 6 weeks after sbrt, followed by ct imaging of the thorax at 3, 6, 12, 18, and 24 months, and then annually thereafter.

Statistical Analyses
The os was calculated from date of radiation to date of death (from any cause) or was censored at the last follow-up date.The association of the nlr and the monocyte:lymphocyte ratio (mlr) with os was investigated using univariable and multivariable Cox proportional hazards regression.To further investigate any effect of those ratios, a competing risks analysis was carried out.The study considered two types of failure: failure type 1 (disease-related failure, including local, regional, or distant failure and death from disease) and failure type 2 (disease-unrelated death with no local, regional, or distant failure).Fine-Gray competing risk regression was performed to test the associations of nlr and mlr with the two types of failure.For the purposes of the foregoing analyses, nlr, mlr, and plr were treated as continuous variables.
Before all multivariable analyses, a stepwise variable selection procedure was performed, and variables at p ≤ 0.10 were included in the final multivariable analysis.The nlr and mlr, as the variables of interest, were included in all multivariable results regardless of statistical significance.The log-rank and Gray tests were used to compare the Kaplan-Meier estimates of os and the cumulative incidence functions of the two failure types, using the median nlr (3.0) as the cut-off value.

RESULTS
Of the 299 sbrt patients identified, 122 patients who had a cbc within the 3 months before sbrt were eligible for the analysis.Of those 122 patients, 47 had their cbc within 1 month before and 75 had their cbc 1-3 months before sbrt.Median age in the group was 75 ± 9 years, and 62 of the patients (51%) were women.In 52 patients (43%), histology was adenocarcinoma; in 23 (19%), it was squamous cell carcinoma; in 19 (16%), it was nsclc; and in 28 (23%), no pathologic diagnosis was made.The tumour in 91 patients (75%) was clinical stage T1, and in 31 (25%), it was T2.The sbrt dose in 28 patients (23%) was 54 Gy or 60 Gy in 3 fractions; in 72 (59%), it was 48 Gy in 4 fractions; and in 22 (18%) it was 60 Gy in 8 fractions.Based on the Radiation Therapy Oncology Group 0236 definition, 9 tumours (7%) were central.

DISCUSSION
Our data demonstrate that nlr and mlr are associated with both os and disease-unrelated death in patients treated with lung sbrt.Patients receiving sbrt are likely to achieve nsclc control and to die of other causes.Consideration of comorbid diseases such as chronic obstructive pulmonary disease (copd) is as relevant as tumour factors are in clinical decision-making for this patient population.The evidence suggests that withholding sbrt from elderly patients with advanced copd is not justified 12 and that there is no known lower limit for pulmonary function 13 .In the context of patients with previous copd diagnoses undergoing sbrt in increasing numbers, the implications of nlr and mlr in chronic disease are as important as are their roles in cancer prognosis.The nlr is associated with a number of chronic and comorbid conditions in this patient population, including copd, smoking, and vascular disease, and the nlr has been shown to be significantly higher in patients with copd than in healthy control subjects 14 .During acute exacerbation of copd, a further increase in nlr was demonstrated compared with the stable period in the same patients 14 .There is evidence that smoking has an effect on neutrophil count and inflammatory response markers, independent of genetic background and environmental variables 15 .Finally, an elevated nlr has also been shown to be associated with mortality in patients which coronary artery disease 16 .
The question therefore remains whether nlr is predictive of survival because of the tumour itself, because of an underlying poor medical status of the patient, or because of some interplay between the two.Until a better understanding of the association between nlr and nsclc outcomes emerges, the nlr could be used to inform decision-making in patients with early-stage lung cancer considering sbrt.Such patients, whose survival will be determined by their other comorbidities, could potentially be spared unnecessary treatment for their lung cancer.
Cells of myeloid a nd ly mphoid lineages might have different biologic effects on tumour progression, metastasis, and response to radiation therapy 17,18 .In a 2015 study of pre-treatment neutrophils, monocytes, and lymphocytes in oropharyngeal cancer, Huang et al. 5 demonstrate that higher circulating myeloid cells (for example, neutrophils and monocytes) independently predict inferior os and recurrence-free survival after chemoradiotherapy.Preclinical data suggest the presence of an interaction between lymphocytes and myeloid cells in regulating pro-tumour activity compared with antitumour activity in solid tumours, and B and T lymphocytes might play a role in regulating multiple protumour properties of myeloid cells 19,20 .Evidence suggests that neutrophils and monocytes are recruited to tumours during radiation therapy and promote vasculogenesis, offsetting the effectiveness of treatment and ultimately negatively inf luencing treatment outcomes [21][22][23] .The degree to which that hypothesis is relevant in situations such as sbrt, with its high dose per fraction, remains to be explored.
The nlr has been shown to be predictive of os and progression-free survival in patients with small-cell lung cancer 24 .Reports in patients with other chronic diseases have shown that nlr is predictive of survival 10,16 .To our knowledge, the present study represents the largest series of sbrt patients reported with respect to this topic to date; it is also the first to include mlr in the analyses.A previous study of 59 patients showed an association of nlr and plr with survival, and plr was associated with an increased risk of nonlocal failure 11 .The univariate regression analyses in that study demonstrated that os was associated with nlr and plr only when the latter variables were coded categorically; the association became statistically nonsignificant when nlr and plr were tested as continuous variables 11 .Using multivariate analyses, we demonstrated that low nlr and mlr are independently predictive of os and that high ratios are associated with disease-unrelated death.On univariate analysis, the plr was found to be significantly associated with os and disease-unrelated death, but not with disease-related failure.Together, both datasets confirm the importance of nlr in nsclc outcomes, and our data convey the importance of mlr in addition to nlr.
The nlr and mlr are simple, inexpensive, and reliable tests, and are thus clinically useful.Their application is still unclear, because nlr and mlr might not be consistent with tumour-specific markers.Comorbid inflammatory conditions and steroid treatments can confound results.Some authors recommend that these ratios be assessed with other inf lammatory markers such as C-reactive protein, red cell distribution width, and erythrocyte sedimentation rate 25,26 .Validation remains an ongoing consideration because numerous articles have reported on nlr and mlr using different cut-off levels [27][28][29][30] .
Our study has several limitations.First, the limited number of disease failures could represent a lack of power to determine any associations of nlr and mlr with disease-related outcomes such as local failure or disease-specific survival.In addition, as in many series about sbrt, confirming cause of death is challenging.Many patients have significant medical comorbidities, and death related to previously undiagnosed recurrence or related-to-treatment toxicity might not be identified in the face of sudden death in the presence of significant cardiac, pulmonary, or other chronic comorbidities.Future prospective studies to evaluate the effects of chronic versus acute sbrt on nlr and mlr are needed.Because high-dose radiation has been shown to alter cells associated with pro-and antitumour immunity 31 , investigating the interplay between dose and these cell ratios is also justified.Finally, cbc within 3 months of sbrt, as chosen for the present study, is consistent with other radiotherapy analyses of nlr and mlr 5,11 ; however, the degree to which neutrophils, monocytes, and lymphocytes can change within that time period is not known.

a
Glickman R, Chaudary N, Pintilie M, et al.Neutrophils modulate response to radiation and chemotherapy in locally advanced cervical cancer.Presented at: The Canadian Association of Radiation Oncology 28th Annual Scientific Meeting; St. John's, NL; 25-28 August 2014.

aA
stepwise variable selection was performed, and variables with p ≤ 0.10 were included in the multivariable result.b Per 10-unit increase in platelets or platelet:lymphocyte ratio.HR = hazard ratio; CL = confidence limits; ECOG = Eastern Cooperative Oncology Group; FEV1 = forced expiratory volume in 1 second; DLCO = diffusing capacity of the lungs for carbon monoxide.
failures, including local, regional, or distant failure and died of disease.b Disease-unrelated death.c Per 10-unit increase in platelets or platelet:lymphocyte ratio.HR = hazard ratio; CL = confidence limits; ECOG = Eastern Cooperative Oncology Group; FEV1 = forced expiratory volume in 1 second; DLCO = diffusing capacity of the lungs for carbon monoxide.

FIGURE 3
FIGURE 3 Cumulative incidence of disease-unrelated deaths and disease-related failure.NL = neutrophils:lymphocytes.

TABLE I
Patient demographics and characteristics FEV = forced expiratory volume; DLCO = diffusing capacity of the lungs for carbon monoxide; ECOG = Eastern Cooperative Oncology Group; PS = performance status.

TABLE II
Cox proportional hazards regression for overall survival

TABLE III
Fine-Gray competing risk regression for disease-related failure (type 1) and disease-unrelated death (type 2)