Factors Influencing Postoperative Prognosis in Patients with Hypopharyngeal and Laryngeal Carcinoma

Abstract

Objectives

Despite the increasingly modern surgical techniques in the oncology field, the factors that influence postoperative prognosis in patients with hypopharyngeal and laryngeal carcinoma (HLC) remain unclear. The study aimed to evaluate the factors influencing the prognosis of HLC patients with pathological diagnosis of squamous cell carcinoma, and the findings are intended to direct follow-up management strategies.

Methods

A retrospective cohort study was performed. The study population included 407 postoperative patients with HLC from 2011 to 2015. Univariate and multivariate analyses were used to examine the prognostic factors identified.

Results

Based on univariate analysis results, smoking and alcohol history, tumor differentiation, preoperative radiotherapy, primary tumor sites, flap reconstruction, lymph node invasion (LNI), and preoperative albumin levels (PAL) significantly affects the prognosis of HLC patients (P < .05). Meanwhile, multivariate analysis revealed that smoking pack-year (OR = 1.002, 95% CI = 1.001 ∼ 1.003), primary tumor sites (OR = 6.241, 95% CI = 1.715 ∼ 18.433), LNI (OR = 2.869, 95% CI = 1.095 ∼ 8.743), and PAL (OR = .020, 95% CI = .004 ∼ 0.104) were associated with complications. Tumor differentiation (OR = 0.650, 95% CI = .383 ∼ 0.855), primary tumor sites (OR = 12.392, 95% CI = 3.290 ∼ 26.679), LNI (OR = 16.323, 95% CI = 2.726 ∼ 47.729), preoperative radiotherapy (OR = 9.300, 95% CI = 3.182 ∼ 27.181), and PAL (OR = .321, 95% CI = .141 ∼ .732) were associated with overall survival rates.

Conclusion

Smoking and alcohol history, tumor differentiation, LNI, primary tumor sites, flap reconstruction, PAL, and preoperative radiotherapy are crucial factors that influence the postoperative prognosis of patients with HLC. In addition, a monogram of five factors was established to predict the survival rates of HLC patients.

Introduction

Head and neck carcinomas (HNC) are malignant tumors that might occur in all head and neck organs, including the lips, floor of the mouth, tongue, thyroid, larynx, hypopharynx, nasal cavity, tonsils, paranasal sinuses, and skin of the head and neck. ¹ Hypopharyngeal and laryngeal carcinomas (HLC) are common head and neck carcinomas, of which pharyngeal malignancy is the second most common malignancy in the head and neck region. ² Hypopharyngeal malignancy represents 3%–4% of all HNC, with approximately 3400 new cases detected in the US annually. ³

Despite the availability of several treatment modalities, including surgery, radiotherapy, and chemotherapy, patients with HNC still suffer from frequent recurrence after surgery with high complication rates and low three-year survival rates. In oral oncology, common factors influencing prognosis include the clinical stage of the tumor, quality of surgical resection, perineural growth, the size and number of metastatic neck nodes, patient age, and gender. ⁴ A previous study reported a linear relationship between smoking and the progression of HNC; they found smokers to have a 10- to 15-fold higher risk of developing HNC than nonsmokers. ⁵ Similarly, another research demonstrated a linear relationship between alcohol consumption and HNC. ⁶ However, the effects of alcohol and cigarette consumption on the postoperative prognosis of HNC patients, especially for hypopharyngeal and laryngeal carcinoma (HLC), remains unclear. In addition, it is equally unclear whether preoperative clinical characteristics and interventions impact the prognosis of HLC.

Our trial was a retrospective study on the complications and three-year survival of HLC patients with pathological diagnosis of squamous cell carcinoma. It aimed to analyze whether preoperative clinical characteristics and interventions could affect the prognosis of HLC. A total of 407 patients were included in this study as of December 1st, 2019, and possible influencing factors were subjected to analysis. This analysis’s results may help improve predictions for the prognosis of HLC.

Materials and methods

Data source and patient selection

The study was approved by the Medical Research Ethics Committee of Ningbo Medical Center Lihuili Hospital (KY2020PJ100). All data were obtained from the medical records of HLC patients who had undergone surgery from January 1st, 2011 to May 1st, 2015 in Ningbo Medical Center Lihuili Hospital. A total of 407 patients were enrolled in our study based on the following criteria. Inclusion criteria included patients who (a) signed informed consent; (b) received treatment from the same team; (c) had complete clinical records of diagnosis, treatment, and follow-up data; and (d) had pathological diagnoses of squamous cell carcinoma. Exclusion criteria comprised patients who (a) were diagnosed with other cancers, (b) had incomplete pathological diagnoses, and (c) were treated by other surgical teams. Preoperative and postoperative clinical information was obtained from patients with HLC, in which smoking was defined as at least one cigarette per day for more than six months. Drinking was defined based on the definition of National Institute on Alcohol Abuse and Alcoholism (NIAAA), that is, drinking a liquid containing 14 grams of pure alcohol for over six months. In addition, the clinical laboratory definition of hypoalbuminemia was based on the threshold for albumin ≤3.5 g/dL. Postoperative complications concluded were related to surgery rather than to the overall health status of HLC patients.

Patient follow-up

Patients were followed up from the completion of their operation to December 1st, 2019. The end of observation was calculated until the end of follow-up or alternatively loss or death during this period. The 3-year survival rate of patients and 30-day postoperative complications were recorded. Postoperative complications included pharyngocutaneous fistula (PCFs), pneumonia, wound infections, and hematoma. Additionally, 3.7% of patients were lost to follow-up.

Statistical analysis

Statistical analyses were conducted using Statistical Package for Social Science (SPSS) version 21.0 and R version 3.5.0. The differences were considered statistically significant when P < .05. Descriptive statistics were used to summarize the clinical preoperative and postoperative characteristics of the patients, such as tumor types, complications, survival status, and treatment histories. Chi-square test and logistic regressions were used to explore factors affecting postoperative complications. Kaplan–Meier survival analysis and Cox’s proportional hazards regression model were utilized to identify the independent risk factors for overall survival (OS). To build risk score systems for postoperative patients with HLC in complications and OS, multivariable logistic regression and Cox’s proportional hazards regression model were performed with stepwise elimination of covariates whose P-values were not less than .1. Additionally, the nomogram was utilized for predicting one-, two-, and three-year survival probability; C-statistic was used to examine the model fit.

Results

Profile of patients with HLC

Consistent with previous studies, most 407 cases comprised male patients (383, 94.1%). Second, the age of HLC patients was around 60 years old (Figure 1A), and the average preoperative albumin level was 4.0 ± 0.45 g/dL (Figure 1B). Next, 76.6% of patients had moderately or highly differentiated tumors (Figure 1C), and the primary tumor site was mainly focused on the larynx (319, 78.4%, Figure 1D). Postoperative complications occurred in 122 cases (30.0%), while 47 (11.5%) died within 3 years after surgery, as presented in the Venn diagram (Figure 1E). Specifically, the postoperative complications included 41 cases of pharyngeal fistula (10.1%), 40 cases of pneumonia (9.8%), 27 cases of hematoma (6.6%), and 14 cases of wound infection (3.4%) (Figure 1F) consistent with previous reports that state pharyngeal fistula to be the most common complication following surgical treatment in HLC patients.

Figure 1.

Profiles of patients with hypopharyngeal and laryngeal carcinoma (HLC). (A, B) Frequency histograms of patients with age and preoperative albumin levels. (C, D) Clinicopathological features in HLC patients. (E–G) The specifics of postoperative complications in HLC patients presented as a Wayne diagram or histogram. (H) Analysis of receiver operating characteristic (ROC) curve based on the logistic regression model.

Postoperative complications of HLC

Factors affecting postoperative complications were then explored. It is worth noting that all the postoperative complications concluded were due to surgery and were not related to the overall health status of HLC patients. Baseline characteristics of patients are shown in Table 1. The analysis of univariate chi-square test (Table 1) revealed that differentiation (χ² = 33.831,P < .05), preoperative radiotherapy (χ² = 72.333, P < .05), primary tumor sites (χ² = 21.448, P < .05), lymph node invasion (χ² = 23.720, P < .05), flaps reconstruction (χ² = 7.245, P < .05), and preoperative albumin levels (χ² = 122.713, P < .05) were significantly associated with postoperative complications. As expected, the underlying risk factors for postoperative complications in HLC patients were poorly differentiated tumor, lymph node invasion, and low preoperative albumin levels. Additionally, patients with hypopharyngeal cancer were more prone to complications than laryngeal cancer, with proportions of 50.0% and 24.5%, respectively. In addition, patients with preoperative radiotherapy had a higher rate of postoperative complications, while those with flap reconstruction showed a lower rate of postoperative complications. To be specific, no patient developed pharyngeal fistula or hematoma after surgical treatment with anterolateral thigh flaps (ATF), but some cases reported other complications such as pneumonia and wound infection. Meanwhile, a few patients experienced pharyngocutaneous fistula (PCF) or hematoma after undergoing the pectoralis flap (PF) treatment (Figure 1G), suggesting that ATF was favorable for managing complications of PCF and hematoma.

Table 1.

Univariate chi-square test or Kaplan–Meier survival analysis (n = 407).

Characteristics	Category	Frequency	Percentage (%)	Complications (Univariate chi-square test)				Kaplan–Meier survival analysis (%)
				Count	Percentage (%)	χ2	P	1-year	2-year	3-year	P
Gender	Male	383	94.1	117	30.5	1.016	0.314	96.1	91.4	88.3	0.851
	Female	24	5.9	5	20.8			95.8	87.5	87.5
Drinking	No	224	55.0	39	17.4	37.470	<0.001*	98.2	96.4	94.5	<0.001*
	Yes	183	45.0	83	45.4			93.4	84.7	80.8
Smoking	No	171	42.0	12	7.0	74.050	<0.001*	95.3	93.0	90.5	0.261
	Yes	236	58.0	110	46.6			96.6	89.8	86.7
PAR	No	50	27.3	31	62.0	7.690	0.006*	89.5	78.9	71.1	0.001*
	Yes	133	72.7	52	39.1			98.0	91.9	89.8
PSC	No	38	16.1	22	57.9	2.318	0.128	92.0	82.0	73.0	0.145
	Yes	198	83.9	88	44.4			94.0	85.0	81.8
Differentiation	Poorly	95	23.4	46	48.4	33.831	<0.001*	91.6	78.9	74.2	<0.001*
	Moderately	156	38.3	53	34.0			100	97.4	92.8
	Well	156	38.3	23	14.7			94.9	94.9	92.2
Preoperative radiotherapy	No	376	92.4	92	24.5	72.333	<0.001*	96.8	94.7	92.7	<0.001*
	Yes	31	7.6	30	96.8			87.1	48.4	35.5
Primary tumor sites	Larynx	319	78.4	78	24.5	21.448	0.001*	99.4	98.7	98.7	<0.001*
	Hypopharynx	88	21.6	44	50.0			84.1	63.6	50.0
LNI	No	228	56.0	46	20.2	23.720	<0.001*	100	99.1	99.1	<0.001*
	Yes	179	44.0	76	42.5			91.1	79.9	74.6
Flap reconstruction	No	335	82.3	109	32.5	7.245	0.027*	95.8	90.4	87.9	0.709
	PF	32	7.9	8	25.0			96.9	93.8	86.7
	ATF	40	8.6	5	14.3			97.5	95.0	92.3
PAL (g/dL)	≤3.5	89	21.9	69	77.5	122.713	<0.001*	89.9	85.4	73.0	<0.001*
	>3.5	318	78.1	53	16.7			97.8	92.8	92.8

*Statistically significant (P < 0.05).

Abbreviations: PAR, postoperative alcohol restriction; PSC, postoperative smoking cessation; LNI, lymph node invasion; PF, pectoralis flaps; ATF, anterolateral thigh flaps; PAL, preoperative albumin levels.

The associations between lifestyle habits and postoperative complications in HLC patients were also investigated. Based on the results, alcohol consumption and tobacco smoking had striking risks of postoperative complications, while the limitation of alcohol had a statistically significant reduction in complications (χ² = 7.690, P < .05).

Risk score system of postoperative complications for HLC patients

To establish a risk score system for postoperative complications in HLC patients, the logistic regression model was used with the occurrence of postoperative complications as the dependent variable (No = 0, Yes = 1). The obtained variables with P-value < .05 were transferred into a multiple logistic regression model. Specific results are shown in Table 2. The analysis of the final model indicated that smoking pack-year (OR = 1.002, 95% CI = 1.001 ∼ 1.003) and lymph node invasion (OR = 2.869, 95% CI = 1.095 ∼ 8.743) demonstrated statistical significance in affecting postoperative complications, and low preoperative albumin levels (OR = .020, 95% CI = .004 ∼ .104) might potentially increase the risk of postoperative complications (Table 3). Besides, a high risk of complications was noted after surgical treatment of hypopharyngeal cancer (OR = 6.241, 95% CI = 1.715 ∼ 18.433). Through stepwise selection and elimination, the receiver operating characteristic (ROC) was generated based on the final logistic regression model. The area under the curve (AUC) was 0.971 (P < .05, 95% CI = .942 ∼ 1.000), thus verifying a good model fit and can be used to predict the occurrence of postoperative complications (Figure 1H).

Table 2.

Univariate regression analysis (n = 407).

Characteristics	Category (Zero as reference group)	Complications (Logistic)			Three-year survival analysis (Cox)
		P value	OR value	OR 95% CI	P-value	OR value	OR 95% CI
Gender	Male (0)	0.318	0.598	0.218∼1.641	0.852	1.118	0.347∼3.601
	Female (1)
Age (years)		0.055	1.016	0.972∼1.063	0.807	1.004	0.974∼1.035
Drinking	No (0)	<0.001*	3.937	2.506∼6.185	<0.001*	3.740	1.941∼7.205
	Yes (1)
Intake of alcohol (g)		0.013*	1.006	1.004∼1.008	<0.001*	1.005	1.003∼1.006
Smoking	No (0)	<0.001*	11.567	6.099∼21.939	0.290	1.008	0.993∼1.024
	Yes (1)
Years of tobacco use		<0.001*	1.065	1.034∼1.096	0.551	1.009	0.979∼1.041
Cigarettes smoked per day		<0.001*	1.074	1.043∼1.107	0.013*	1.020	1.004∼1.035
Smoking pack-year		<0.001*	1.002	1.001∼1.003	0.476	1.000	0.999∼1.001
PAR	No (0)	0.006*	0.393	0.202∼0.768	0.002*	0.317	0.152∼0.662
	Yes (1)
PSC	No (0)	0.131	0.582	0.288∼1.174	0.047*	0.804	0.694∼0.941
	Yes (1)
Differentiation	Poorly (0)	<0.001*	0.423	0.315∼0.568	<0.001*	0.483	0.329∼0.709
	Moderately (1)
	Well (2)
Preoperative radiotherapy	No (0)	—	—	—	<0.001*	12.553	7.014∼22.466
	Yes (1)
Primary tumor sites	Larynx (0)	<0.001*	3.090	1.893∼5.042	<0.001*	51.457	18.442∼143.577
	Hypopharynx (1)
LNI	No (0)	<0.001*	2.919	1.883∼4.527	<0.001*	32.629	7.914∼134.525
	Yes (1)
Flap reconstruction	No (0)	0.037*			0.716
	PF (1)	0.384	0.691	0.301∼1.588	0.933	1.045	0.374∼2.921
	ATF (2)	0.013*	0.296	0.113∼0.777	0.421	0.618	0.191∼1.998
PAL (g/dl)		<0.001*	0.019	0.009∼0.040	0.001*	0.131	0.067∼0.257

*Statistically significant (P < 0.05).

Abbreviations: PAR, postoperative alcohol restriction; PSC, postoperative smoking cessation; LNI, lymph node invasion; PF, pectoralis flaps; ATF, anterolateral thigh flaps; PAL, preoperative albumin levels; CI, confidence interval; OR, odds ratio.

Smoking pack-year = years of tobacco use × packs of cigarettes smoked daily (a pack comes with 20 cigarettes).

Table 3.

Multivariate regression analysis (n = 407).

Group	Characteristics	B	P-value	OR value	OR 95% CI
Complications (Logistic)	Smoking pack-year	0.002	<0.001*	1.002	1.001∼1.003
	Primary tumor sites	1.895	0.010*	6.241	1.715∼18.433
	LNI	1.158	0.034*	2.869	1.095∼8.734
	PAL (g/dL)	−3.918	<0.001*	0.020	0.004∼0.104
Three-year survival rate (Cox)	Differentiation	−1.295	0.002*	0.650	0.383∼0.855
	Primary tumor sites	2.517	<0.001*	12.392	3.290∼26.679
	LNI	2.793	0.002*	16.323	2.726∼47.729
	Preoperative radiotherapy	2.230	<0.001*	9.300	3.182∼27.181
	PAL (g/dL)	−1.137	0.007	0.321	0.141∼0.732

*Statistically significant (P < 0.05).

Abbreviations: LNI, lymph node invasion; PAL, preoperative albumin levels; CI, confidence interval; OR, odds ratio.

Smoking pack-year = years of tobacco use × packs of cigarettes smoked daily (a pack comes with 20 cigarettes).

Postoperative three-year survival rate of HLC

The three-year overall survival (OS) rate for HLC patients was 88.8% (Figure 2A). Based on Kaplan–Meier survival analysis results, seven clinicopathological characteristics were associated with OS. Specifically, alcohol consumption before (Figure 2B) or after (Figure 2C) operation, poor differentiation of tumor (Figure 2D), preoperative radiotherapy (Figure 2E), hypopharyngeal cancer (Figure 2F), lymph node invasion (Figure 2G), and low preoperative albumin levels (Figure 2H) presented as risk factors affecting overall survival rate.

Figure 2.

Postoperative three-year survival rates of HLC. (A) Postoperative three-year survival rates of HLC patients. (B-H) Analysis of Kaplan-Meier survival with seven clinicopathological features.

Cox’s proportional hazards regression model of overall survival for HLC

To further investigate the effects of factors on overall survival for HLC patients, a Cox proportional hazards regression model was constructed with survival outcome (Death = 1) and survival time as dependent variables. Univariate analysis was performed using Cox proportional hazards and logistic regression model (Table 2). Then, the selected potential factors were introduced into Cox regression model for multivariate analysis. Table 3 lists the independent factors influencing the prognosis of HLC patients, namely, tumor differentiation (OR = .650, 95% CI = .383 ∼ .855), primary tumor site (OR = 12.392, 95% CI = 3.290 ∼ 26.679), lymph node invasion (OR = 16.323, 95% CI = 2.726 ∼ 47.729), preoperative radiotherapy (OR = 9.300, 95% CI = 3.182 ∼ 27.181), and preoperative albumin levels (OR = .321, 95% CI = .141 ∼ .732). Specifically, poor tumor differentiation, lymph node invasion, preoperative radiotherapy, and lower preoperative albumin levels were independent predictors of increased risk of death. Additionally, hypopharyngeal cancer had a worse prognosis compared to laryngeal cancer. These five factors were then incorporated into the nomogram to calculate patients’ one-, two- and three-year survival probabilities (Figure 3A). The C-statistic for the nomogram was 0.900 (95% CI = .842 ∼ .0.959), thus confirming a good fit to predict prognosis for HLC patients. The associated calibration curves from the nomogram are depicted in Figure 3B.

Figure 3.

Cox’s proportional hazards regression model of overall survival for HLC patients. Analysis of monogram including five clinicopathological features (A) and the obtained associated calibration curves (B).

Discussion

Disease prognosis is an important indicator of the quality of life. Identifying the prognostic factors of HLC remains a huge challenge for clinical researchers. This study evaluated 30-day postoperative complications and three-year survival rates as prognostic indicators for patients with HLC after operation.

Based on the findings, tobacco and alcohol consumption increased the risk of poor prognosis through separate and synergistic effects. Recent research has revealed a strong link between smoking and recovery, with smoking having a negative effect on wound healing. ⁷ Wound healing and resistance to infection have been found to partly depend on oxygen content or oxygen partial pressure (Po2) in tissue; smoking can significantly decrease subcutaneous wound-tissue oxygen levels and keep them low for 30–50 min. ⁸ Our findings revealed that smokers had a higher risk of complications, suggesting that smoking cessation before and after surgery is necessary. Besides, alcohol consumption is known to reduce levels of pro-inflammatory cytokines (IL-1β) and cathelicidin-related antimicrobial peptides (CRAMP) as well as affect macrophage infiltration, which would reduce the efficacy of pathogen clearance. ⁹ Meanwhile, alcohol consumption negatively affected the later stages of wound healing, including reduced epithelialization, decreased collagen deposition, and slower wound vascularization. ¹⁰ Furthermore, the cumulative consequences of alcohol exposure on wound repair are not merely transient, as these changes can cause wound complications even several days after surgery. Koskinen et al. found that laryngeal carcinoma patients who consumed large amounts of alcohol had a poorer prognosis, probably due to alcohol-related conditions. ¹¹ Overall survival was significantly reduced when alcohol intake exceeded 100 g per day. Therefore, smoking cessation and alcohol restriction could improve postoperative prognosis and survival time in HLC patients.

According to keratinization ability and cellular atypia, tumor tissues can be classified into well-differentiated, moderately differentiated, and poorly differentiated carcinoma. Well-differentiated tumor cells resemble normal squamous epithelial cells with high keratinization and mild cellular heterogeneity; moderately differentiated tumor cells have obvious nuclear pleomorphism and mitotic activity with lower keratinization; poorly differentiated cancer cells contain abundant immature cells with significant mitotic activity and are nearly non-keratinizing. ¹² Our findings indicated that the prognosis of poorly differentiated HLC was worse than the other two types, which might be related to the high recurrence risk for poorly differentiated tumors.

Generally, preoperative radiotherapy patients have a higher risk of complications and a lower three-year survival rate. Relapsed carcinomas with prior exposure to radiotherapy often show aggressive behaviors and are associated with poor control rates due to radiation sequelae edema. ¹³ Compared with primary cancer, recurrent HLC after radiotherapy was associated with more aggressive growth patterns as shown by multiple invasive lesions, higher extra-laryngeal and sublingual dilation tendencies, lower degrees of differentiation, and more frequent intravascular and perineural invasion. ¹³ Therefore, surgery is the most effective treatment for cases with dysfunctional larynx after radiotherapy.

Lymph node invasion can also affect the prognosis of postoperative HLC patients. Previous studies reported that the number, location, and volume of metastatic lymph nodes adversely affected the prognosis. ¹⁴ Lymph node metastases often occur in patients who have experienced lymph node invasion. Therefore, these patients tend to have a higher risk of complications and a lower three-year survival rate.

The prognosis of postoperative HLC patients is highly associated with preoperative albumin levels. Patients with a 10% postoperative weight loss or low albumin levels have been reported to show a higher postoperative recurrence rate, demonstrating the negative effects of poor nutrition on prognosis. ¹⁵ Since preoperative nutritional status is one of the main factors influencing preoperative albumin levels, preoperative nutritional interventions for surgical patients could effectively reduce the incidence of complications and mortality.

The tumor site is another vital factor influencing the prognosis of HLC patients. In our follow-up study, the tumor was present in two locations, that is, hypopharynx and larynx. Hypopharyngeal carcinomas have the worst 5-year survival rate (approximately 30%–40%) of all cases (Netherlands, 1989–2011). ¹⁶ The vast majority of hypopharyngeal carcinomas are squamous cell carcinomas with high metastatic potential, and the mortality rates of patients with distant metastases from hypopharyngeal cancer exceed 90%. Compared with hypopharyngeal carcinomas, laryngeal cancer has a better prognosis with a five-year survival rate of 63%–66%. ¹⁷

Based on a previous report, ¹⁸ flap reconstruction can greatly reduce the incidence of wound complications, which is consistent with our findings. There are two reconstruction modalities, namely, ATF and pectoralis flaps. From the perspective of complications (Figure 1), ATF was superior to the pectoralis flap, especially in reducing PCFs and hematoma complications. The pectoralis flap has been a common choice for primary reconstruction since 1979. ¹⁹ However, it is accompanied by several significant drawbacks, including minimal pliability, a restrictive pedicle length, significant tissue bulk limiting the reconstruction of 3-dimensional defects, and long-term side effects of contracture with poor aesthetics. ²⁰ As one of the most preferred options for soft-tissue reconstruction, ATF has helped shift focus away from simple coverage of defects towards minimizing morbidity at the donor site, refining the flap, and reducing bulk. Clinically, ATF is recommended as the first choice for flap reconstruction.

Conclusions

In summary, smoking and alcohol history, tumor differentiation, lymph node invasion, flap reconstruction, preoperative albumin levels, and prior interventions (radiotherapy) are crucial factors affecting the postoperative prognosis of HLC patients with pathological diagnosis of squamous cell carcinoma. Based on the above findings, we propose the following recommendations to improve the quality of prognosis:

(1) Early detection, diagnosis, and treatment should be prioritized and implemented promptly in the absence of lymph node metastasis.

(2) Smoking cessation and alcohol restriction are necessary measures before and after surgery.

(3) Preoperative nutritional assessment and intervention from a clinical dietitian should be considered.

(4) Surgical treatment should be chosen for previously recurring tumors.