Survival Rates of Patients with Non-Small Cell Lung Cancer Depending on Lymph Node Metastasis: A Focus on Saliva

The aim of this study was to compare overall survival (OS) rates at different pN stages of NSCLC depending on tumor characteristics and to assess the applicability of saliva biochemical markers as prognostic signs. The study included 239 patients with NSCLC (pN0-120, pN1-51, pN2-68). Saliva was analyzed for 34 biochemical indicators before the start of treatment. For pN0, the tumor size does not have a prognostic effect, but the histological type should be taken into account. For pN1 and pN2, long-term results are significantly worse in squamous cell cancer with a large tumor size. A larger volume of surgical treatment reduces the differences between OS. The statistically significant factors of an unfavorable prognosis at pN0 are the lactate dehydrogenase activity <1294 U/L and the level of diene conjugates >3.97 c.u. (HR = 3.48, 95% CI 1.21–9.85, p = 0.01541); at pN1, the content of imidazole compounds >0.296 mmol/L (HR = 6.75, 95% CI 1.28–34.57, p = 0.00822); at pN2 levels of protein <0.583 g/L and Schiff bases >0.602 c.u., as well as protein >0.583 g/L and Schiff bases <0.602 c.u. (HR = 2.07, 95% CI 1.47–8.93, p = 0.04351). Using salivary biochemical indicators, it is possible to carry out stratification into prognostic groups depending on the lymph node metastasis.


Introduction
Approximately 85% of lung cancers are non-small cell lung cancer (NSCLC). The most important parameters that determine treatment and survival in this group are the stage of the disease and metastases in the lymph nodes [1,2]. The degree of lymphogenous metastasis of NSCLC not only affects the prognosis of patients, but also largely determines the optimal treatment tactics [1]. Therefore, at stage pN 0-1 the first and main stage of treatment is surgery, at pN 3 -chemotherapy and radiotherapy. The tactics of treating patients with pN 2 have not yet been fully determined and are the subject of active discussion in the literature [3,4]. Recent practice guidelines consider chemotherapy and radiation therapy to treat patients with pN 2 , and do not recommend isolated or primary surgery [5]. Some supporters of reducing the volume of surgery consider it possible to apply individual schemes of lymph node dissection, focusing on the frequency of metastasis, the size and location of the tumor, the form of growth, and topography of the lymph nodes [6,7]. However, in patients with an early clinical stage of NSCLC, metastases in the lymph nodes are sometimes found during histopathological examination [8]. At the same time, even within one stage pN, the survival rate varies greatly depending on the size, histological type, degree of differentiation of the tumor and several other factors [9].
We have previously shown that several biochemical indicators of saliva can act as prognostic signs in NSCLC [10]. The aim of this study was to compare overall survival rates at different pN stages of NSCLC, depending on the characteristics of the tumor, and to assess the applicability of saliva biochemical indicators as prognostic signs.

Saliva Analysis
Saliva samples were collected at baseline, right before the start of treatment. Collection of saliva samples was carried out on an empty stomach after rinsing the mouth with water in the interval of 8-10 am by spitting into sterile polypropylene tubes, the salivation rate (mL/ min) was calculated. Saliva samples were centrifuged (10,000× g for 10 min) (CLb-16, Moscow, Russia), after which biochemical analysis was immediately performed without storage and freezing using the StatFax 3300 semi-automatic biochemical analyzer [11]. In all saliva samples, 34 biochemical parameters were determined, including pH, electrolyte levels, parameters of protein and lipid metabolism, and activity of metabolic and antioxidant enzymes as described previously [10].

Statistical Analysis
The total follow-up time was 6 years; the median follow-up time was 42 months. The patient's overall survival (OS) was assessed from the date of hospitalization to the date of the last observation (censored) or the date of death of the patient (complete). OS was assessed using the Kaplan-Meier method with the presentation of survival curves and the calculation of the significance of differences by Log-rank (Statistica 10.0, StatSoft, Tulsa, OK, USA). Correction for uneven distribution according to the main initial criteria (gender, age, histological type, localization, tumor stage, treatment method) was performed using Cox regression. The description of the sample was made by calculating the median (Me) and interquartile range in the form of the 25th and 75th percentiles [LQ; UQ]. Differences were considered statistically significant at p < 0.05.
A univariate Cox proportional hazards regression analysis was initially variables carried out to investigate the relationships between salivary parameters and survival data. Finally, variables with p < 0.10 were chosen to formulate multivariate Cox proportional hazards regression models and determine the independent prognostic factors for OS. Hazard ratio (HR) was obtained with 95% confidence interval (CI). When evaluating the parameters of the regression model, those parameters for which the error was at least twice its standard error (t > 2.0) were considered statistically significant at the level of p < 0.05. A univariate Cox proportional hazards regression analysis was initially variables carried out to investigate the relationships between salivary parameters and survival data. Finally, variables with p < 0.10 were chosen to formulate multivariate Cox proportional hazards regression models and determine the independent prognostic factors for OS. Hazard ratio (HR) was obtained with 95% confidence interval (CI). When evaluating the parameters of the regression model, those parameters for which the error was at least twice its standard error (t > 2.0) were considered statistically significant at the level of p < 0.05.

Overall Survival Rates Depending on the Stage of the Disease, Histological Type, and Morphological Growth Forms of NSCLC
Median OS in the NSCLC group was 24.9 months. For patients without lymphogenous metastasis, the median OS was 36.1 months, with pN1 metastases, this value decreased to 18.2 months, and with pN2 to 14.3 months (Figure 1). The relative risk increases for stages pN0 vs. pN1 (HR = 5.85, 95% CI 2.71-12.31) and pN0 vs pN2 (HR = 10.55, 95% CI 4.73-22.80, p < 0.00001). At the next stage of the study, subgroups were identified taking into account the characteristics of the tumor. Thus, with an increase in tumor size, OS naturally decrease, but the differences are statistically insignificant (Table 1). For pN2, the differences between the pT3 and pT4 stages are the smallest. When compared for one tumor size, for small tumors (T2) the relative risk increases statistically significantly between pN0 and pN1, At the next stage of the study, subgroups were identified taking into account the characteristics of the tumor. Thus, with an increase in tumor size, OS naturally decrease, but the differences are statistically insignificant (Table 1). For pN 2 , the differences between the pT3 and pT4 stages are the smallest. When compared for one tumor size, for small tumors (T2) the relative risk increases statistically significantly between pN 0 and pN 1  Note. *-differences are statistically significant, p < 0.05; ADC-adenocarcinoma, SCC-squamous cell carcinoma; G1-highly, G2-moderately, and G3-poorly differentiated, G4-undifferentiated lung cancer. LNS-lymph node status, OS-overall survival.
Taking into account the histological type of NSCLC, it was shown that regardless of the presence / absence and degree of lymph node involvement, the survival rates for squamous cell carcinoma are worse than for adenocarcinoma (Table 1). It should be noted that adenocarcinoma is more often detected at the pN 0 stage (57.3 vs. 41.7%), while squamous cell carcinoma predominates at the pN 1 and pN 2 stages (16.8 vs. 26.9% and 25.9 vs. 31.5%, respectively). In general, patients with adenocarcinoma of the lung, even with lesions of the lymph nodes, have a more favorable prognosis than patients with squamous cell carcinoma (Table 1). For squamous cell lung cancer, the survival rate sharply decreases already at pN 1 , and then practically does not change. Apparently, it is for squamous cell lung cancer that metastatic lesions of the ipsilateral pulmonary, bronchopulmonary and/or lymph nodes of the lung root are a factor in the unfavorable prognosis of the disease.
In addition, we noted that regardless of the degree of damage to the lymph nodes, OS in central cancer is lower than in peripheral cancer (Table 1). This difference is most pronounced for pN 1 . Multiple lymph node lesions are characteristic of the mediastinal form of lung cancer; therefore, this subgroup is isolated only for pN 2 and is characterized by a minimum OS (Table 1).

The Predictive Value of the Type of Treatment
For pN 0 , radical surgical and combined treatment is used; in the second case, OS is statistically significantly worse ( Table 2). It is interesting to note that with lobectomy, including extended lobectomy, OS decreases with the transition from pN 0 to pN 1 , but remains at the same level for pN 2 ( Table 2). For pneumonectomy, OS changes are not significant regardless of the degree of lymph node involvement ( Table 2). Note. LNS-lymph node status, OS-overall survival, *-differences are statistically significant, p < 0.05.

Predictive Value of Saliva Biochemical Indicators
By constructing a Cox regression model, we selected indicators that have a potential prognostic value in NSCLC at various pN stages ( Figure 2). For pN 0 , such indicators include the activity of lactate dehydrogenase (LDH) and the level of diene conjugates (DC), for pN 1 -the level of imidazole compounds (ICs) and medium molecular weight toxins (MM), for pN 2 -the content of total protein and Schiff bases (SB) ( Table 3). Diagnostics 2021, 11, x FOR PEER REVIEW 8 of 13 It should be noted that with an LDH activity of more than 1900 U/L, the relative risk was 3.8 times lower than with an activity of less than 636 U/L. The resulting value is statistically significant and can be used as an independent option (Table 3).
It should be noted that with an LDH activity of more than 1900 U/L, the relative risk was 3.8 times lower than with an activity of less than 636 U/L. The resulting value is statistically significant and can be used as an independent option (Table 3).
We have presented the characteristics of cohorts depending on the differences in the biochemical composition of saliva with different status of lymph node involvement (Tables S1-S3). It was shown that in addition to the difference in the biochemical composition of saliva, there are no other statistically significant differences between subgroups, including age, gender, pT, histological type, growth form, type of treatment, smoking, and relapse status). The only identified difference is the lower recurrence rate in the group of patients with a favorable prognosis for the biochemical composition of saliva for pN 0 (Table S1).

Multivariate Survival Analysis Using the Cox Regression Model
Multivariate analysis, including the stage of the disease (pT), histological structure, growth form, type of treatment, as well as the studied biochemical indicators for each pN group, showed that in all cases the biochemical parameters of saliva are independent factors in predicting the overall survival of patients with lung cancer (Table 4).

Discussion
Traditional tumor characteristics such as differentiation, tumor invasion, lymph node metastasis, and TNM (Tumor, Nodes, and Metastasis) stage classification are not the only aspects that determine the prognosis of the disease [12][13][14]. For prognostic purposes, the use of several groups of biomarkers is described. Therefore, the most significant is the study of genetic, epigenetic, proteomic, metabolic markers, as well as the profile of synthesis and the level of microRNA [15][16][17][18]. These markers are detected in tumor tissue, serum and blood plasma, and exhaled air [19,20].
In the literature, there are sporadic data on the study of the composition of saliva in lung cancer, including for prognostic purposes [21][22][23][24]. We have shown for the first time the fundamental possibility of using saliva biochemical indicators for predicting the course of lung cancer [10,25]. Of the indicators that were selected in regression analysis, only LDH was previously mentioned in the literature as a prognostic sign for blood plasma in lung cancer [26][27][28][29][30]. In this regard, comparison with literature data is not possible. It should also be noted that in previous studies we have shown that for most biochemical markers of saliva correlations with the composition of blood plasma are weak or absent altogether, therefore the values of biochemical markers of saliva should be considered to be independent and set their own criteria for norm and pathology [31].
The use of biochemical indicators of saliva allows obtaining prognostic data comparable to those for the characteristics of the tumor. In particular, in the absence of lymph node metastasis, tumor size is not a significant prognostic sign. However, the histological type of NSCLC is prognostically important; poor prognosis is associated with squamous cell lung cancer and undifferentiated cancer [32]. The statistically significant factors of unfavorable prognosis are LDH activity less than 1294 U/L and a DC level of more than 3.97 c.u. (Table 4).
With pN 1 , unfavorable prognosis factors include large tumor size (pT 4 ), squamous histological type, central growth, as well as low differentiation or undifferentiated cancer (Table 1). An additional biochemical indicator in this case is the level of ICs. The concentration of ICs greater than 0.296 mmol/L is a statistically significant independent factor of poor prognosis (HR = 6.75, 95% CI 1.28-34.57, p = 0.00822). Multivariate analysis showed that the ICs level is the only independent prognostic factor for the group of patients with pN 1 (Table 4).
At pN 2 , a poor prognosis is associated with squamous cell carcinoma, mediastinal tumor growth, and any differentiation other than highly differentiated tumors. According to our data, the size of the primary tumor does not statistically significantly affect the prognosis, which may be the result of an insufficient sample size. In the literature, age and pT stage are considered prognostically important [33]. Of the biochemical indicators, only the combination of indicators "Protein + SB" can be attributed to independent prognostic signs (Tables 3 and 4). An unfavorable prognosis is typical for groups of patients with a protein content of less than 0.583 g/L and an SB of more than 0.602 c.u., as well as a protein level of more than 0.583 g/L and an SB of less than 0.602 c.u. (HR = 2.07, 95% CI 1.47-8.93, p = 0.04351).
In all cases, the type of treatment is a significant prognostic factor, which is quite natural. It is interesting to note that with an increase in the volume of surgery, the differences in OS medians decrease (Table 2). Thus, in the case of pneumonectomy, the median OS is 18.5, 17.8, and 17.1 months for stages pN 0 , pN 1 , and pN 2 , respectively.
An interesting result of our study is that with different degrees of damage to the lymph nodes, different biochemical parameters of saliva are used as prognostically important parameters. It can be assumed that this is due to the depth of the metabolic changes occurring in lung cancer. Therefore, in the absence of damage to the lymph nodes, the prognosis is determined by the activity of LDH as the main metabolic enzyme, the activity of which changes in many types of cancer, including lung cancer. The level of diene conjugates, which characterize the content of primary lipid peroxidation products, is also prognostically important. At pN 1 , the predictive factors include the total content of protein toxins and imidazole compounds, while at pN 2 toxic Schiff bases, which are the end products of lipid peroxidation, are prognostically important. However, this hypothesis requires additional verification in the course of further research.
The limitations of the study are related to the fact that it was not assessed whether the lesions of the pN 1 and pN 2 lymph nodes are single or multiple [34]. The limitations should also include the small sample size, which reduces the statistical significance of the data obtained and limits the possibility of dividing into subgroups. Further studies are warranted to confirm our observation.

Conclusions
In the absence of metastases in regional lymph nodes, the size of the primary tumor has no significant prognostic effect; however, the histological type of tumor should be taken into account. For stages pN 1 and pN 2 , long-term results are significantly worse with a large tumor size, and the presence of histology of squamous cell lung cancer critically decreases the median OS in these groups. The degree of tumor differentiation at pN 0 has practically no effect on OS, whereas for all types of tumors, except for highly differentiated ones, the median OS sharply decreases at pN 1 and pN 2 . A larger volume of surgical treatment reduces the differences between OS in the study groups. It has been shown for the first time that using biochemical indicators of saliva, additional stratification into prognostic groups can be carried out, depending on the presence / absence and the prevalence of regional metastasis.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/diagnostics11050912/s1, Table S1: Characteristics of cohorts depending on the differences in the chemical composition of saliva in pN 0 , Table S2: Characteristics of cohorts depending on the differences in the chemical composition of saliva in pN 1 , Table S3: Characteristics of cohorts depending on the differences in the chemical composition of saliva in pN 2 .