Introduction

The standard of care in postoperative therapy for patients with locally advanced squamous cell carcinoma of the head and neck (LASCCHN) is concurrent chemoradiotherapy (CCRT) with high-dose cisplatin (CDDP) [1,2,3,4,5,6].

Approximately half of patients (34–57%) receiving CCRT experience grade ≥ 3 mucositis induced by CCRT [7], which is often accompanied by dysphagia, dry mouth, and dysgeusia. These adverse events often lead to malnutrition owing to inadequate oral intake, treatment interruption, and a negative impact on treatment outcomes [8,9,10]. Therefore, nutrition management is important during CCRT for LASCCHN. The American Society for Parenteral and Enteral Nutrition (A.S.P.E.N.) clinical guidelines recommend enteral nutrition (EN) rather than intravenous nutrition as nutritional support in patients unable to maintain oral intake [11,12,13]. Thus, prophylactic percutaneous endoscopic gastrostomy (PEG) has been widely performed before CCRT for LASCCHN [14,15,16,17]. However, PEG placement often causes certain complications, such as bleeding or infection [18, 19]. Furthermore, some patients receiving CCRT after total laryngectomy maintain oral intake during and after CCRT and do not necessarily need EN support in clinical practice. Therefore, it is controversial whether all patients require prophylactic PEG for adjuvant CCRT.

The aim of this study is to evaluate predictive factors for the feasibility of oral intake in adjuvant CCRT for patients with LASCCHN.

Materials and methods

Patients

This study retrospectively analyzed LASCCHN patients who had undergone surgery followed by adjuvant CCRT with CDDP at Shizuoka Cancer Center between April 2008 and December 2018. Inclusion criteria were as follows: histologically proven squamous cell carcinoma located in the oral cavity, oropharynx, hypopharynx, or larynx; age 20–80 years; Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0–2; dysphagia score (DS) ≤ 2 on the Ogilvie scale [20, 21] without nutritional support before CCRT initiation; total radiotherapy (RT) tumor dose of 50–70 Gray (Gy); pathological stage III, IVA, or IVB (UICC eighth edition); no distant metastasis identified by computed tomography (CT); no prior RT or chemotherapy; and adequate organ function. Exclusion criteria were as follows: double cancers except for carcinoma in situ or intramucosal tumor; patients who depended on EN or intravenous nutrition when starting CCRT; and patients also exhibiting active infection. All clinical data were retrospectively obtained from medical records. This study was carried out in accordance with the recommendations of the institutional review committee of Shizuoka Cancer Center (Shizuoka, Japan). The study was approved by the institutional review committee of Shizuoka Cancer Center (institutional ID: 30-J125-30-1-3) and met the standards set forth in the Declaration of Helsinki. Written informed consent was obtained from all patients. If it was difficult to get, authors provided the information of this study and patients’ right on the website or in a notice board at the hospital.

Adjuvant chemoradiotherapy

The decision to choose adjuvant therapy for each patient was made during a multidisciplinary tumor board discussion, depending on the presence of extracapsular extension, microscopically involved surgical margins, the extent of lymph node disease, age, PS, organ function, and patient preference, among others.

The RT was performed by either three-dimensional conformal radiotherapy (3D-CRT) or intensity-modulated radiotherapy (IMRT) [22] up to a dose of 50–70 Gy. RT was administered in 25–35 fractions with 2.0 Gy given once a day. The radiation dose was decided in accordance with the International Commission on Radiation Units and Measurements. Volume definition and dose calculation were CT-based. There were no planned treatment interruptions, except for weekends.

Concurrent chemotherapy with CDDP was administered tri-weekly at 80–100 mg/m2 on days 1, 22, and 43. Some patients received weekly cisplatin at 40 mg/m2 on days 1, 8, 15, 22, 29, 36, and 43, or tri-weekly split-dose cisplatin at 20 mg/m2 on 4 consecutive days.

Prophylactic PEG

The determination to perform prophylactic PEG before CCRT depended on each physician’s discretion, considering pre-treatment PS, DS, and patient preference. PEG placement was carried out by the direct method (Direct Ideal PEG Kit; Olympus Corp., Tokyo, Japan) under endoscopic support. All patients received local anesthesia, analgesia, and sedation with midazolam. Prophylactic antibiotics were given for 3 days.

Evaluation of adverse events related to CCRT and nutritional support

Adverse events related to CCRT were assessed using the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0. The cumulative incidence of aspiration pneumonia from the initiation of CCRT until 1 year after its completion was also evaluated. Aspiration pneumonia was defined as a clinical condition that met all of the criteria presented in a previous paper [23]. The rates of patients with difficulty obtaining adequate nutrition via oral intake from the initiation of CCRT until 1 year after its completion and nutrition-support-free survival were evaluated. Feasibility of oral intake was defined as ability to orally consume adequate amounts of water and calories without any nutritional support.

Statistical analysis

The candidates for predictive factors for feasibility of oral intake were as follows: age; sex; primary tumor location; ECOG PS; total laryngectomy or other surgery; pre-treatment DS; pre-treatment body mass index (BMI) (kg/m2); pre-treatment serum albumin (Alb) (g/ml); total dose of CDDP; method of CDDP administration (tri-weekly, weekly, or split dose); method of RT (3D-CRT or IMRT); irradiation field (bilateral-neck RT or ipsilateral-neck RT); total dose of RT; mean RT dose to the constrictor muscles; and mean RT dose to the oral cavity. Regarding mean RT dose to the pharyngeal constrictor muscles and that to the oral cavity, the cut-off values were decided on basis of a receiver operating characteristic (ROC) curve. All variables in univariate analysis were included for the multivariate analysis. Hazard ratios (HR) and 95% confidence intervals (95% CI) were calculated using the Cox proportional hazards model to identify predictive factors for the feasibility of oral intake. The nutrition-support-free survival rate and cumulative incidence of aspiration pneumonia were calculated using the Kaplan–Meier method, and significance was evaluated using the log-rank test. The nutrition-support-free survival rate was calculated from the first day of CCRT until the onset of difficulty obtaining adequate nutrition via oral intake and was censored at the last follow-up visit or death from any cause. We used the Mann–Whitney U test for comparisons of continuous variables and Fisher’s exact test for comparisons of categorical variables between the groups. P < 0.05 was considered statistically significant. All statistical analyses were conducted using EZR (version 1.37; Saitama Medical Center, Jichi Medical University, Saitama, Japan) [24].

Results

Patients’ characteristics

Of 134 patients who received adjuvant CCRT, 2 patients with dysphagia score ≥ 3, 9 patients who depended on enteral nutrition or intravenous nutrition when starting CCRT, 4 patients who received Cetuximab or Carboplatin, and 2 patients with RT dose < 50 Gy were excluded (Supplement Fig. 1). Finally, a total of 117 patients were analyzed in this study (Table 1). Twenty-five patients received total laryngectomy, while 92 received other surgery. Forty-nine patients (42%) received IMRT and 34 patients (29%) underwent ipsilateral-neck RT. The median overall treatment time of RT was 44 days (range 32–56). The median total doses of RT and CDDP were 66 Gy and 240 mg/m2, respectively. Although RT was interrupted in four patients (3.4%), all patients completed the planned RT. More than half of the patients received the planned dose of chemotherapy without dose reduction.

Fig. 1
figure 1

Kaplan–Meier plot showing nutrition-free survival in the total laryngectomy group (n = 25) and the other surgery group (n = 92)

Full size image
Table 1 Patients’ characteristics
Full size table

Predictive factors for feasibility of oral intake

Predictive factors for feasibility of oral intake are presented in Table 2. On basis of the ROC curve, mean RT dose of 30 Gy to the oral cavity was determined as a cut-off value with a sensitivity of 83.9%, the specificity of 49.2%, and area under the curve (AUC) of 0.793. Similarly, mean RT dose of 37 Gy to pharyngeal constrictor muscles was determined as a cut-off value with a sensitivity of 76.8%, specificity of 37.7%, and area under the curve (AUC) of 0.522. In univariate analysis, total laryngectomy, ipsilateral-neck RT, pre-treatment DS, male sex, mean RT dose to the oral cavity < 30 Gy, and BMI ≥ 18.5 were significantly associated with the feasibility of oral intake. In multivariate analysis, total laryngectomy (HR 0.09, 95% CI 0.020–0.37, P = 0.001) and oral cavity of primary tumor location (HR 0.21, 95% CI 0.051–0.87, P = 0.031) were significantly associated with the feasibility of oral intake. Of these factors, in accordance with HR, total laryngectomy was identified as one of the strongest independent predictive factors for the feasibility of oral intake.

Table 2 The assessment of predictive factors for the feasibility of oral intake
Full size table

Nutritional support, opioid use, and PEG-related adverse events

Because total laryngectomy was identified as one of the strongest independent predictive factors for the feasibility of oral intake in accordance with HR, we compared the nutritional support and adverse events between the total laryngectomy group and the other surgery group.

Nutritional support in both groups is summarized in Table 3. The rate of patients with difficulty obtaining adequate nutrition via oral intake was significantly lower in the total laryngectomy group than in the other surgery group (16% vs. 57%, P < 0.001). In fact, although prophylactic PEG was performed in 10 patients in the total laryngectomy, only 3 patients used PEG. The main reason for this difficulty in both groups was mucositis. The rate of patients using opioids was significantly lower (9% vs. 55%, P = 0.04) and the median maximum dose of opioids per day tended to be lower in the total laryngectomy group than in the other surgery group (0 mg/day vs. 15 mg/day, P = 0.07). Consistent with the findings on the feasibility of oral intake, the rate of usage of enteral nutrition was statistically significantly lower in the total laryngectomy group than in the other surgery group (8% vs. 48%, P = 0.001). Furthermore, the 1-year nutrition-support-free survival rate was significantly higher in the total laryngectomy group than in the other surgery group (84%, 95% CI 62.8–93.7% vs. 44%, 95% CI 33.2–53.3%, P < 0.001, Fig. 1). Adverse events related to PEG occurred in 5/76 patients (6.5%), including two cases of major wound infection (2.6%), 1 of major bleeding (1.3%), and 2 of minor bleeding (2.6%).

Table 3 Summary of nutritional support
Full size table

Adverse events related to CCRT

The incidences of grade ≥ 3 mucositis and dysphagia were significantly lower in the total laryngectomy group than in the other group (12% vs. 51%, P < 0.001, 12% vs. 51%, P < 0.001, respectively). There were no significant differences with regard to the incidences of neutropenia, anemia, creatinine increase, dry mouth, and radiation dermatitis (Table 4). One-year cumulative incidences of aspiration pneumonia were 0% in the total laryngectomy group and 13.5% in the other surgery group (P = 0.06) (Fig. 2). No deaths occurred within 30 days after treatment.

Table 4 Adverse events related to CCRT
Full size table
Fig. 2
figure 2

Kaplan–Meier plot showing cumulative incidences of aspiration pneumonia in the total laryngectomy group (n = 25) and the other surgery group (n = 92)

Full size image

Discussion

Postoperative CCRT is the standard treatment for patients with LASCCHN. However, it is associated with severe mucositis, dry mouth, and dysphagia, resulting in difficulty obtaining adequate nutrition via oral intake. Therefore, nutritional support is required during CRT. As methods of nutritional support, intravenous nutrition, nasogastric tube, and PEG are available. Among these, long-term intravenous nutrition has a higher risk of infectious morbidity than EN, which is not recommended in the A.S.P.E.N. clinical guidelines [11,12,13]. Continuous nasogastric tube insertion could increase the risk of dysphagia and aspiration pneumonia, which is not suitable for long-term use during CRT. Thus, the need for prophylactic PEG has been widely recognized in the treatment with CCRT [14,15,16,17]. However, PEG-related adverse events sometimes lead to not just minor but also major complications. In a previous study, there were four major complications (3.3%) in 121 patients receiving PEG, including two cases of bleeding and two of wound infection [18]. Our patients treated by postoperative CCRT also experienced major PEG-related complications. Therefore, it is critical to avoid complications caused by prophylactic PEG as much as possible. Our hypothesis is that primary surgical procedures may play a major role in the ability of oral intake in the adjuvant CCRT.

Our multivariate analysis identified total laryngectomy as one of the strongest independent predictive factors for the feasibility of oral intake. In accordance with this finding, the rate of patients with difficulty obtaining adequate nutrition via oral intake was significantly lower in the total laryngectomy group than in the other surgery group for LASCCHN in adjuvant CCRT. Furthermore, the incidences of grade ≥ 3 mucositis and dysphagia were significantly lower in the total laryngectomy group than in the other group. These results suggest that surgical procedures with laryngectomy specifically influence the ability to perform oral intake in postoperative CCRT and that prophylactic PEG may not be required for patients who have undergone laryngectomy.

Several factors may explain why oral intake was feasible in patients after laryngectomy. First, the sensory functions of the larynx and hypopharynx decrease due to removal of the superior laryngeal nerve after total laryngectomy. Indeed, the rate of patients who needed opioids and the dose of opioids were lower in the total laryngectomy group than in the other surgery group. Second, the risk of aspiration is extremely low in the total laryngectomy group since the airway is completely separated from the gastrointestinal tract through total laryngectomy [25,26,27,28]. Indeed, our results revealed that no aspiration pneumonia occurred in the total laryngectomy group. Because of the difference in sensory functions of the larynx and hypopharynx and that in the risk of aspiration, oral intake could be more feasible in patients receiving total laryngectomy followed by CCRT than in those receiving other surgery followed by CCRT. Therefore, prophylactic PEG for adjuvant CCRT may not necessarily be required in patients who have undergone total laryngectomy.

Our multivariate analysis also revealed that ipsilateral-neck RT and mean RT dose to the oral cavity < 30 Gy tended to be associated with the feasibility of oral intake. As for the irradiation field, previous studies reported that the incidences of oral mucositis and dysphagia were significantly lower in ipsilateral-neck RT than in bilateral-neck RT for LASCCHN [29, 30], which may reflect the difference in the feasibility of oral intake. A previous study also reported that risk of oral mucositis increased as RT dose to the oral cavity increased [31]. This supports our results. Although oral cavity of primary tumor location was significantly associated with feasibility of oral intake, this mechanism for feasibility of oral intake is unknown and cannot be explained as a hypothesis, such as that of total laryngectomy. In addition, no previous studies reported association between oral cavity of primary tumor location and feasibility of oral intake during and after adjuvant CCRT. Therefore, this result should be interpreted carefully at this moment.

Our study has several limitations. First, this is a retrospective study at a single medical center with a small number of patients. Further multicenter cohort study is thus needed to confirm our findings. Second, there were the unbalance in patients’ baseline characteristics between the total laryngectomy group and the other surgery group. However, to solve the issue on the heterogeneity of the populations as much as possible, multivariate analysis was firstly performed to exclude confounding with each predictive factor. Importantly, our multivariate analysis identified total laryngectomy as a strongest and independent predictive factor for the feasibility of oral intake. Third, we used DS on the Ogilvie scale for inclusion criteria and analysis of predictive factors for the feasibility of oral intake. Although Ogilvie scale is simple and widely used for assessment of dysphasia in the area of esophageal cancer [32, 33], Functional Outcome Swallowing Scale (FOSS) and Food Intake LEVEL Scale (FILS) may be more standardized and sophisticated in the area of head and neck. Despite these limitations, our study suggests the low need for prophylactic PEG before adjuvant CCRT for patients with LASCCHN after total laryngectomy.

In conclusion, our study suggests that the majority of patients who underwent total laryngectomy are able to maintain oral intake and may not need nutritional support. Therefore, prophylactic PEG for adjuvant CCRT may not necessarily be required for LASCCHN patients who have undergone total laryngectomy. These information from our study is of great significance in terms not just of clinical convenience, avoidance of PEG-related adverse events, but also of cost reduction.