Introduction

Inflammationplays a critical role in both the progression of cancer and its response to therapies [1, 2]. There has been a recent focus on exploring inflammatory markers as a prognostic factor for cancer-related outcomes as they are inexpensive, non-invasive, and minimize complications for the patient [3]. These markers are of particular interest in human papillomavirus (HPV)-negative head and neck cancers, where no widely accepted prognostic biomarkers exist [3]. One such marker is lymphocyte-monocyte ratio (LMR). The use of LMR as a prognostic factor in head and neck cancer is equivocal in a recent meta-analysis displaying conflicting findings [4].

To date, there have been no studies evaluating the utility of LMR as a prognostic factor for head and neck cancer within North America. The majority of studies were performed in China, Japan, and the United Kingdom, with inconsistent use of smoking history as a pertinent risk factor in a recent meta-analysis [4]. Current studies may not be fully applicable to North America due to differential HPV distribution and differences in prevalence of other risk factors such as smoking and alcohol use [5,6,7]. Recent in-vitro studies have found p16-mediated inflammatory microenvironments in models of HPV positive cancer which may contribute to differential inflammatory profiles between HPV positive and HPV negative cohorts [8]. In addition, there has been no subset analysis of HPV positive and HPV negative head and neck cancers, which vary greatly in their outcomes [9]. To address this knowledge gap, we performed an observational cohort study to evaluate the association of LMR and survival outcomes in North American patients with head and neck cancer.

Materials and methods

Roswell Park Comprehensive Cancer Center institutional review board approved our study (EDR 103707). Our study complies with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

A single-institution, retrospective database was queried for patients with non-metastatic head and neck cancer who underwent curative-intent definitive chemoradiation from June 2007 to April 2021 at the Roswell Park Comprehensive Cancer Center. Intensity modulated radiation therapy (IMRT) with 70 Gy to gross disease and 56 Gy to elective neck lymph nodes in 35 fractions [10]. Patients were excluded if they underwent radiation alone, induction chemotherapy, postoperative radiation, or did not have LMR or survival data.

Variables of interest used in this study included pre-treatment LMR, age, race, gender, smoking status, Karnofsky Performance Status (KPS), number of comorbidities, primary disease site, tumor T and N staging based on the American Joint Committee on Cancer (AJCC) 7th edition, HPV status based on p16 status, and chemotherapy agent. Comorbidities included respiratory (e.g., chronic obstructive pulmonary disease), genitourinary (e.g., chronic kidney disease), endocrine (e.g., diabetes, hypothyroidism), cardiovascular (e.g., hypertension, stroke), and gastrointestinal systems (e.g., gastroesophageal reflux disease). For analysis, missing values were coded as unknown. Races are self-identified as African American, American Indian/Alaska Native, Asian, Hispanic, unknown or declined to answer, and White. Given the small subgroup sample sizes, non-White patients were grouped together as a single category.

Primary endpoints were overall survival (OS) and cancer-specific survival (CSS), defined as time intervals from diagnosis to death from any cause or cancer-related death respectively. Other endpoints included progression-free survival (PFS), locoregional failure (LRF), and distant failure (DF). PFS was defined as time interval from diagnosis to either death from any cause or tumor progression. LRF and DF were defined as time intervals from diagnosis to tumor recurrences in head and neck or outside the head and neck, respectively. All tumor recurrences were confirmed based on multidisciplinary discussion using radiographic findings and, if applicable, biopsy results of metastatic sites. For those with multiple failure events either synchronously or metachronously during their follow up period, all failure events were counted separately for analysis.

Statistical analysis

Peripheral complete blood count data was used to calculate pre-treatment LMR. The association of LMR with OS and CSS was examined using nonlinear Cox proportional hazard model using restricted cubic splines (RCS) with 3 knots at the 10th, 50th, and 90th percentiles based on the lowest Akaike information criterion [11, 12] as previously shown [13].

Cox multivariable analysis (MVA) and Kaplan–Meier method were used to analyze OS, CSS, and PFS using LMR as a continuous variable. Pre-radiation LMR was then stratified into high and low based on its median value. Logistic MVA was performed to identify variables associated with low LMR below its median value. Fine-Gray MVA was performed to analyze LRF and DF outcomes with death as a competing event. MVA models included all of the variables listed previously. Among those with available HPV data for oropharyngeal cancer, subgroup analysis was performed. In addition, given the prognostic role of neutrophil counts from peripheral blood on treatment outcomes [14, 15], another subgroup analysis including absolute neutrophil count (ANC) was performed. Propensity scored matching between high versus low LMR based on its median value was performed to construct matched pairs based on nearest neighbor method in a 1:1 ratio with no replacement using a caliper distance of 0.2 [16]. Standardized means differences for all matched variables were less than 0.1, suggesting negligible differences [17]. Matched variables included all variables previously included for MVA. Cox regression model was used to evaluate OS and CSS after matching.

P values less than or equal to 0.05 were considered statistically significant. All p values were two-sided. Analyses was performed using R (version 4.1.2, R Project for Statistical Computing, Vienna, Austria).

Results

A total of 476 patients (391 male [82.1%], median [interquartile range] age, 61 [55–67] years) met our criteria (Table 1). Median follow up was 45.3 months (interquartile range 22.8–74.0). Most patients were White (n = 414, 87.0%) with favorable performance status (KPS 90–100: n = 348, 73.1%) and had HPV-associated squamous cell carcinoma (n = 231, 48.5%) in oropharynx (n = 272, 57.1%) treated with cisplatin as concurrent chemotherapy regimen (n = 403, 84.7%).

Table 1 Baseline characteristics
Full size table

The nonlinear Cox regression model using RCS method showed that low LMR was associated with worse OS and CSS in a continuous fashion without plateau and crossed the hazard ratio of 1 at LMR 3.4 for both OS and CSS outcomes (Fig. 1). On Cox MVA, higher LMR was associated with improved OS (adjusted hazard ratio [aHR] 0.90, 95% confidence interval [CI] 0.82–0.99, p = 0.03) and CSS (aHR 0.83, 95% CI 0.72–0.95, p = 0.009; Table 2). However, it was not associated with PFS (aHR 0.93, 95% CI 0.86–1.01, p = 0.09), LRF (aHR 0.89, 95% CI 0.75–1.05, p = 0.18), or DF (aHR 0.94, 95% CI 0.81–1.08, p = 0.39; Table 3).

Fig. 1
figure 1

Nonlinear Cox regression model using restricted cubic spline for the association between lymphocyte-monocyte ratio and survival outcomes

Full size image
Table 2 Cox multivariable analysis for overall survival and cancer-specific survival
Full size table
Table 3 Cox multivariable analysis for progression-free survival and Fine-Gray multivariable analysis for locoregional and distant failures
Full size table

The median value of LMR was 3.8. On logistic MVA (Table 4), patients with other racial background (adjusted odds ratio [aOR] 0.85, 95% CI 0.74–0.97, p = 0.02) and positive HPV status (aOR 0.82, 95% CI 0.72–0.94, p = 0.005) were less likely to have low LMR. Higher T staging was associated with low LMR (aOR 1.15, 95% CI 1.04–1.27, p = 0.005).

Table 4 Logistic multivariable analysis for Lymphocyte-Monocyte Ratio
Full size table

After propensity score matching, a total of 186 pairs were matched, and their baseline characteristics were well balanced (Table 1). Lower LMR remained to be associated with worse OS (HR 1.59, 95% CI 1.12–2.26, p = 0.009; Fig. 2) and CSS (HR 1.68, 95% CI 1.08–2.63, p = 0.02; Fig. 2). However, it was not associated with PFS (aHR 1.35, 95% CI 0.97–1.86, p = 0.07), LRF (aHR 1.06, 95% CI 0.58–1.94, p = 0.85), or DF (aHR 1.30, 95% CI 0.78–2.17, p = 0.31; Fig. 3).

Fig. 2
figure 2

Kaplan–Meier curves for overall and cancer-specific survival outcomes for low versus high lymphocyte-monocyte ratio. LMR: lymphocyte-monocyte ratio

Full size image
Fig. 3
figure 3

Kaplan–Meier curves for progression-free survival and cumulative incidence of locoregional and distant failure outcomes for low versus high lymphocyte-monocyte ratio. LMR: lymphocyte-monocyte ratio

Full size image

For the entire cohort, median ANC was 4750 cells/microliter (interquartile range 3607–6282). When the absolute neutrophil count as a continuous variable was adjusted in the MVA, similar findings for the LMR were noted on MVA. Higher LMR was associated with improved OS (aHR 0.91, 95% CI 0.83–1.00, p = 0.047) and CSS (aHR 0.85, 95% CI 0.74–0.98, p = 0.02), while it was not associated with PFS (aHR 0.95, 95% CI 0.88–1.02, p = 0.17), LRF (aHR 0.91, 95% CI 0.77–1.07, p = 0.25), or DF (aHR 0.95, 95% CI 0.83–1.09, p = 0.48). In the subgroup of 319 patients (67.0%) with available HPV data for oropharyngeal cancer, 231 patients (48.5%) had HPV-associated head and neck cancer. LMR status was not associated with both OS and CSS regardless of HPV status (Table 5).

Table 5 Cox multivariable analysis for overall survival and cancer-specific survival stratified by p16 status
Full size table

Discussion

To our knowledge, this is the first study of a North American head and neck cancer patient cohort to evaluate the prognostic value of LMR. Low LMR, both as a continuous variable and dichotomized variable below the median value, was associated with worse OS and CSS. Low LMR was associated with higher T staging and negative HPV status.

The association of LMR with survival outcomes and higher T staging in our study is inconsistent with a recent meta-analysis evaluating the role of LMR as a prognostic factor among patients with head and neck cancer [4]. Many studies included in the meta-analysis were performed outside the North America, and a recent Korean study showed different average LMR across age and sex groups in healthy subjects, suggesting varied degrees of the prognostic role for LMR based on different patient demographics [18].

Our finding on low LMR as an adverse prognostic factor supports a growing body of literature that systemic inflammation, as indicated by inflammatory markers, has been demonstrated to result in worse prognosis [19]. Recent studies have emphasized that host inflammatory response greatly influences the development of cancer, as it has been suggested that inflammatory cells and cytokines are increasingly likely to impact cancer growth and metastasis, while contributing to immunosuppression associated with malignancy [20, 21]. Peripheral blood biomarkers have been used to capture the magnitude of such inflammation, and several studies have demonstrated their prognostic value across cancer types [22]. An insufficient count of lymphocytes can result in inadequate immunological response to a tumour present, promoting progression and spread; specifically, it has been reported that types of tumor infiltrating lymphocytes, including CD8 + T cells and memory T cells, are associated with positive prognosis of tumors [23]. Increased monocyte number, however, has been associated with unfavorable outcomes of a variety of tumors, differentiating into tumor-associated macrophages and promoting tumor angiogenesis, growth, invasion, and migration [23]. Our cutoff of 3.8 as a median value in this study is consistent with previous studies incorporating cutoff values ranging from 2.35 to 5.22 [24].

Low LMR was also associated with HPV-negative cancer. HPV positive cancers have a distinct molecular pathogenesis from HPV negative cancers facilitated by upregulation of p16 [8, 25]. One study found increased CD8 + T cell tumor infiltration in HPV positive cancer compared to HPV negative tumors [26]. The different tumor microenvironments between the head and neck cancer subgroups may in part explain our findings. Another study found that HPV can inhibit monocyte differentiation to Langerhans cells, thereby evading immune surveillance [27]. It is possible that through this mechanism, a higher proportion of monocytes would be insignificant in affecting outcome.

Limitations

The limitations of this study are those inherent to single-institution retrospective studies including potential for selection bias. In addition, our analysis did not include address change in pre-treatment compared to post-treatment LMR (delta LMR), which may better account for baseline LMR and be a stronger predictor of prognosis [28]. Since only those with definitive chemoradiation were included in this study, our findings may not be generalizable to other patient populations treated with surgery, postoperative radiation, surgery or radiation alone, and palliative radiation.

Conclusion

Low LMR, both as a continuous variable and dichotomized variable below 3.8 in our study, was associated with worse overall survival and cancer-specific survival. Low LMR was associated with higher T staging and HPV negative cancer. Further studies are warranted to elucidate the role of inflammatory markers in head and neck cancer management.