Negative Prognostic Impact of Smoking on Long-term Survival in Patients With Nasopharyngeal Carcinoma Treated With Curative (Chemo)radiotherapy

Abstract

Background/Aim: To evaluate the effectiveness of curative (chemo)radiotherapy in patients with nasopharyngeal carcinoma and to identify prognostic factors influencing treatment outcomes.

Patients and Methods: We conducted a retrospective study of 73 consecutive patients, treated with definitive (chemo)radiotherapy from 2002 to 2019 (median stage III/IV 78%). The median total dose of radiotherapy achieved was 70 Gy. Concomitant chemotherapy was given to 82% of patients.

Results: The five- and ten-year locoregional controls were 73% and 72%, respectively; the five- and ten-year distant controls were 93% and 93%, respectively. The five- and ten-year overall survival rates were 46% and 34%, respectively. A multivariate analysis identified age, smoking, and the initial response to treatment as the strongest prognostic factors in predicting survival.

Conclusion: Smoking ≤5 years before starting curative (chemo)radiotherapy for nasopharyngeal carcinoma was shown to be an independent negative prognostic factor for overall survival with a four-fold higher risk of death compared to non-smokers.

Nasopharyngeal carcinoma is a malignancy arising from the epithelium of the nasopharynx. Nasopharyngeal carcinoma is extremely rare in most parts of the world but is endemic in East and Southeast Asia. The incidence varies from less than one per 100,000 cases in non-endemic areas to higher than twenty per 100,000 cases in endemic regions (1,2). Due to the radiosensitivity of nasopharyngeal carcinoma and its deep-seated location, radiotherapy has been a critical part of its curative treatment for decades. Several meta-analyses have shown that concurrent chemoradiotherapy, either with or without adjuvant chemotherapy, is more effective than radiotherapy alone for treating nasopharyngeal carcinoma (3-8).

Several factors related to patients, tumors, and therapy affect the treatment results. Prognostic factors reported in clinical studies are: age (9-11); T-staging (12-14); N-staging (12-15); clinical stage (16-18); tumor volume (19-21); histological type (12,20,22); radiotherapy technique (22); total dose of radiotherapy (19,20,23); number of chemotherapy cycles (24); albumin level (25); hemoglobin level (13,26); thrombocyte-lymphocyte ratio (27,28); lymphocyte-monocyte ratio (21,29); neutrophil-lymphocyte ratio (29-31); and EBV status (32,33). The negative impact of cigarette smoking on survival has been reported for various types of tumors (34-43). The evidence on the role of smoking in the prognosis of patients with nasopharyngeal carcinoma is sparse, reported mainly from endemic areas of occurrence.

In this study, we aimed to analyze long-term treatment outcomes in a consecutive group of patients treated with definitive (chemo)radiotherapy at the Institute of Radiation Oncology and identify prognostic factors that affect treatment results.

Patients and Methods

Over the period of January, 2002 to November, 2019, 83 patients were treated for nasopharyngeal tumors at the Institute of Radiation Oncology. Seventy-three consecutive patients with nasopharyngeal carcinoma who started definitive radiotherapy with a curative intent were included in the study. The median follow-up of surviving patients was 106 months. The median age at the time of treatment initiation was 55 years (range=19-84 years). The female to male ratio was 1: 2.7. The vast majority of patients were smokers or former smokers (73%) with approximately one-fifth of patients admitting to daily alcohol consumption. All tumors were retrospectively reclassified according to the seventh version of the TNM classification. The majority of patients were treated for a locally advanced disease (78% clinical stage ≥III). Regional cervical metastases were initially diagnosed in 88% of patients. Squamous cell carcinoma was the most frequent histology (Table I).

Table I. Demographics and tumor characteristics.

ACE: Adult comorbidity evaluation score.

Treatment. Radiotherapy. Before 2007, patients were treated with 2D and 3D conformal radiotherapy (24 cases). Patients were treated with intensity modulated radiation therapy (IMRT) from 2007 onwards (49 cases). The median total dose was 70 Gy (range=44-76 Gy) (Table II).

Table II. Treatment characteristics.

Chemotherapy. A total of 58 patients (82%) received concomitant systemic therapy (cisplatin 40 mg/m² qw 56 patients, cetuximab 450 → 300 mg/m² 2 patients), 36 of whom received adjuvant chemotherapy as well. The median cumulative cisplatin dose in concomitant chemotherapy was 200 mg/m². Two patients received neoadjuvant chemotherapy based on platinum derivatives (one of whom received concomitant cetuximab as well) (Table II).

Analysis. For the statistical analysis, all data were recorded and analyzed on XLSTAT software (Addinsoft) version 18.07. Kaplan–Meier methods were used to estimate locoregional control (LRC), distant metastasis-free interval (DMFI), disease-free interval (DFI), and overall survival (OS). The survival or disease-free periods began at the onset of radiation to the time of relapse (LRC, DMFI, DFI) or death (OS). The log-rank test was used to compare survival and recurrence rates between various parameters. We used the Cox regression hazard model to analyze multivariate data. All analyses were performed with a two-sided significance level of ≤0.05.

Results

Locoregional and distant control. A total of twenty locoregional failures (27%) were detected. In eight patients, there was persistence after the end of treatment; twelve patients failed during the follow-up within 5-79 months after the end of radiotherapy. The vast majority of locoregional failures (90%) were detected in the first 36 months. The five- and ten-year locoregional controls were 73% and 72%, respectively (Figure 1A). Distant failure was reported in five patients (7%) with four (80%) of the incidents occurring within 36 months of the completion of radiotherapy (range=3-45 months). The five- and ten-year distant controls were 93% and 93%, respectively.

Figure 1. Kaplan–Meier curves with 95% confidence interval for locoregional control (LRC) (A) and disease-free interval (DFI) (B). The dashed lines represent the 95% confidence interval.

A total of 23 locoregional and distant failures (32%) were detected (locoregional 18, distant three, locoregional & distant two). The five- and ten-year disease-free intervals were 69% and 68%, respectively (Figure 1B).

Survival. A total of 48 patients deceased. Tumor progression was the primary cause of death in 21 patients. In 24 patients, the cause of death was unrelated to cancer. Nine patients died within the first 180 days from initiation of radiotherapy, six of them with no known presence of cancer (two cases each of pulmonary embolism, sepsis, and epistaxis with massive blood aspiration, respectively). During the follow-up, nine metachronous tumors were diagnosed in six patients. Duplicate tumor progression was the cause of death in three of these cases. The five- and ten-year OS rates were 46% and 34%, respectively (Figure 2).

Figure 2. Kaplan–Meier curve with 95% confidence interval for overall survival (OS). The dashed lines represent the 95% confidence interval.

Univariate and multivariate analysis. Parameters that reached statistical significance in the univariate analysis were: age, smoking, T-staging, clinical stage, grading, the interval between histological confirmation and initiation of radiotherapy, total dose of radiotherapy, epoch, and response to treatment (Table III).

Table III. Univariate Cox proportional hazard regression analyses for overall survival (OS) and disease-free interval (DFI).

Statistically significant p-values are shown in bold.

The multivariate analysis of variables showed the following independent prognostic parameters: age for OS (HR=2.470; 95%CI=1.009-6.051; p=0.048), smoking for OS (HR=4.043; 95%CI=1.543-10.590, p=0.004), and response to treatment for OS (HR=7.181; 95%CI=2.056-25.086, p=0.002), and DFI (HR=7.115; 95%CI=2.726-18.571; p<0.0001) (Table IV).

Table IV. Multivariate analyses for disease-free interval (DFI) and overall survival (OS); only factors significant in univariate analysis were calculated.

Statistically significant p-values are shown in bold.

Discussion

The negative impact of cigarette smoking on survival has been reported for various types of tumors (34-36), including head and neck cancers (37-43). As cigarette smoking is an adverse lifestyle factor contributing to global cancer deaths (44), its prognostic value for nasopharyngeal carcinoma has also recently attracted research attention.

Shen et al. evaluated the relationship between lifestyle factors and survival in 1,533 patients with nasopharyngeal carcinoma in a prospective cohort study. As a result of a multivariate analysis, active smoking was shown to be an independent negative prognostic factor for survival with an almost two-fold increase in the risk of death (HR=1.88) compared to non-smokers (45). Chen et al. reported a significant reduction in locoregional control and OS in smokers in a cohort of men with nasopharyngeal carcinomas (n=2,450) treated with curative radiotherapy (HR=2,316). The risk of death increased in heavy smokers >15 pack-years (46). Ouyang et al. in a cohort of 1,849 patients with nasopharyngeal cancer recorded a higher risk of death, progression, locoregional failure, and distant failure for former or current smokers compared to non-smokers (HR for survival of heavy smokers=3.31). A similar difference in treatment outcomes between smokers and non-smokers was also noted in a subgroup of patients treated with IMRT (47).

The negative impact of smoking on OS has not been uniformly reported. In the analysis of a randomized study with induction chemotherapy followed by (chemo)radiotherapy, Guo et al. declared a statistically significant worsening of locoregional control in active/ex-smokers without a worsening of OS. The authors explained the discrepancy as a positive consequence of successful salvage treatment in recurrent tumors (48).

The negative effect of smoking can be explained by several mechanisms: i) Cigarette smoke is a complex mixture of many gaseous and solid substances created by the incomplete combustion of cigarettes. A rich component of cigarette smoke is also carbon monoxide, the inhalation of which is associated with an increased concentration of carboxyhemoglobin. Binding of carbon monoxide to hemoglobin iron reduces the binding capacity of erythrocytes for oxygen transfer (49-52). Peripheral vasoconstriction due to the adrenergic effect of nicotine may also contribute to the reduction in oxygen delivery (53). The resulting tumor hypoxia is co-responsible for tumor failure during curative radiotherapy. Evidence of the limited efficacy of radiotherapy in hypoxic tumor tissue has been repeatedly published (54-57).

ii) The association between EBV infection and the increased risk of nasopharyngeal cancer is known and has been repeatedly published (58-62). A high level of serum EBV antibodies or EBV viral DNA has repeatedly been shown to be an independent prognostic factor in patients from endemic areas undergoing treatment for this disease (32,33,63-65). The prognostic significance of serum levels of EBV DNA and VCA-Ig was also reported in a population from non-endemic areas (66). An increased level of EBV VCA-IgA antibodies was demonstrated in smokers with nasopharyngeal carcinoma. Cigarette smoking can thus lead to EBV reactivation (67). The combination of smoking and serum EBV DNA has repeatedly been shown to be a significant negative prognostic factor (65,68).

iii) Cigarette smoking affects a wide range of functions of both innate and adaptive immunity (69-71). Duffy et al. demonstrated that smoking was associated with increased pretreatment serum levels of interleukin 6 (IL-6) in patients treated for head and neck cancer (72). In their previous work, the authors demonstrated that a high pretreatment serum level of IL-6 reduced survival rate and tumor control in these patients. Increased IL-6 level was associated with age and tumor extent (73). In nasopharyngeal cancer, IL-6 is involved in the process of carcinogenesis through the activation of STAT3, and its high level correlates with the extent of the cancer. Thus, IL-6/STAT signaling can facilitate the malignant transformation of EBV-infected premalignant cells into invasive tumor cells (74).

iv) Cigarette smoking suppresses the activity of NK-cytotoxic lymphocytes. Reduced NK-cell activity results in the weakening of immune surveillance with a subsequent reduction of tumoricidal ability. The result can be an acceleration of cancer progression and metastasis (75,76).

v) Nicotine reduces the effect of treatment, due to DNA damage of the JNK1 pathway with its subsequent influence on the cell death pathway. The authors Omoda et al. reported a nicotine-induced reduction in the cytotoxic effect of cisplatin and ionizing radiation (77).

vi) Curative radiotherapy of advanced tumors is associated with significant acute toxicity, which can be life-threatening in some cases. The combination of radiotherapy and chemotherapy increases this acute toxicity and contributes to it with its systemic adverse effects. The toxicity of the treatment can lead to exacerbations of chronic diseases, which are common in long-term smokers. The association between increased toxicity of radiotherapy and smoking has been repeatedly reported in patients treated with radiotherapy (78,79). In our cohort, early grade 3/4 toxicity was expressed in 84% of patients. Nine patients died in the first six months following the initiation of treatment, six of them due to comorbidities without the known presence of tumors. All six patients had a history of smoking and five of them were active smokers.

vii) The occurrence of metachronous tumors, for which an association with smoking is known, may contribute to increased mortality in smokers. In patients with head and neck cancers, smoking has been repeatedly shown to be a negative factor increasing the risk of metachronous tumors (80,81). In our cohort, nine secondary metachronous tumors were diagnosed in six patients, all of whom had a history of smoking. The progression of the secondary tumor was the cause of death in three (6%) of the patients in the group.

Therefore, cigarette smoking has a negative role not only in promoting carcinogenesis in normal epithelial cells of the nasopharynx (67,82-85), but also in affecting survival rates in patients with nasopharyngeal cancer that has already arisen and is being treated. In our study, smoking proved to be a strong independent prognostic factor, and non-smokers/ex-smokers > five years achieved a statistically significant increase in OS compared to smokers (HR=4.043; 95%CI=1.543-10.590, p=0.004) (Figure 3). According to our knowledge, the prognostic impact of smoking in a population outside of endemic areas has yet to be published.

Figure 3. Overall survival related to smoking status.

It is not surprising that the initial response to treatment was the main negative prognostic factor. For tumors, in which we do not achieve complete remission after primary (chemo)radiotherapy, treatment options are limited and patients are mostly candidates for palliative treatment with limited capability of prolonging survival. In our cohort, patients with locoregionally advanced forms of the disease predominated. In eight patients, there was a locoregional persistence at the end of treatment; twelve patients failed during follow-up. Out of the total number of twenty locoregional failures, local salvage treatment was performed only in five patients: neck dissection in two patients with regional failures, salvage reirradiation in three patients with local failures. The remaining fifteen patients received only palliative treatment. The potential benefit of salvage procedures after the failure of the primary treatment on OS could thus be manifested to a very limited extent. Achieving complete remission was shown to be a strong negative prognostic factor for tumor control and OS (for DFI HR=7.115; 95%CI=2.726-18.571, p<0.0001; for OS HR=7.181; 95%CI=2.056-25.086, p=0.002). The result is in accordance with the work of Righini et al., where, in a group of 44 patients, the initial response to chemoradiotherapy was shown to be a strong negative prognostic factor in a multivariate analysis (86).

Accumulating studies have suggested that systemic inflammation promotes tumor progression and metastasis via the inhibition of apoptosis, promotion of angiogenesis, and damaging of DNA (87). Pretreatment hematological markers of inflammatory response, which can be measured by routine blood examinations, such as neutrophils, monocytes, lymphocytes, platelets, fibrinogen, and albumin, and several of these markers combined, have emerged as useful prognostic factors for patients with head and neck carcinomas (21,25,27-31,88-89).

Limitations of our study include its retrospective nature and heterogeneity. In the study, all consecutive patients with nasopharyngeal carcinoma who started radiotherapy with a curative intent were included. Patients were treated over the last two decades, during which radiation techniques have evolved. These changes might have contributed to the results observed. Consequently, the results should be interpreted within the context of the limitations of a retrospective study.

Conclusion

The results of this retrospective study demonstrate the long-term effectiveness of curative (chemo)radiotherapy in patients treated for nasopharyngeal carcinoma. The multivariate analysis identified age, smoking, and initial response to treatment as the strongest prognostic factors in predicting survival. Smoking ≤5 years before starting curative (chemo)radiotherapy for nasopharyngeal carcinoma was shown to be an independent negative prognostic factor for OS with a four-fold higher risk of death compared to non-smokers.

Conflicts of Interest

The Authors declare no conflicts of interest in relation to this study.

Authors’ Contributions

P.M.: Conceptualization, data collection, analysis, manuscript writing; N.P.: methodology, statistical analysis; D.T.: data collection, manuscript editing; V.A.: data collection; P.Z.: analysis; V.K.: data collection; S.A.: data collection; K.D.: data collection; H.P.: data collection; T.P.: manuscript editing, supervision; All Authors have read and agreed to the published version of the manuscript.