Surgical Management of Carcinoma of the Hypopharynx and Cervical Esophagus: Analysis of 209 Cases

Background  Free jejunal transfer has become the standard technique for reconstruction of the pharynx and hypopharynx, especially with proximal neoplastic lesions, whereas gastric tube interposition is the technique of choice for reconstruction of the hypopharynx and cervical esophagus when resection extends below the thoracic inlet.

Hypothesis  Surgical ablation is a viable option for advanced hypopharyngeal and cervical esophageal neoplasms, with stomach interposition a safe and preferred method of reconstruction.

Design  Retrospective analysis.

Setting  University hospital that is a regional referral institution for esophageal cancer treatment and complex digestive reconstructions after esophagectomy.

Patients  We reviewed the records of 209 patients who underwent total pharyngolaryngectomy between May 1982 and July 1999. The majority of patients had advanced cancer: hypopharyngeal in 131 cases and cervical esophageal in 78 cases.

Interventions  Pharyngolaryngectomy and total esophagectomy with pharyngogastric anastomoses (n = 127); pharyngolaryngectomy, cervical esophagectomy, and reconstruction with free jejunal transplant (n = 77); and pharyngolaryngectomy and total esophagectomy with pharyngocolic anastomoses (n = 5).

Main Outcome Measures  Postoperative mortality and morbidity, long-term survival, and prognostic factors influencing survival.

Results  The postoperative in-hospital mortality rate was 4.8% (10 patients), with a postoperative morbidity rate of 38.3%. Alimentary continuity was achieved using the stomach (127 patients), colon (5 patients), or free jejunal autograft (77 patients). The 1-year and 5-year survival rates were 62% and 24%, respectively. There was no significant difference with regard to the survival between gastric transposition and free jejunal autograft, but there were fewer complications in the gastric pull-up group (33% vs 47%, P<.05). The significant adverse factors affecting survival were tumor cervical localization, postoperative complications, disease stages pT3 and pT4 for the cervical esophageal tumors, microscopic pharyngeal penetration, or incomplete resection. The significant beneficial factors were tumor hypopharyngeal localization and postoperative radiotherapy.

Conclusions  Surgical ablation is a viable option for advanced hypopharyngeal and cervical esophageal neoplasms, with stomach interposition the preferred method of reconstruction. Although the prognosis is poor, satisfactory short-term palliation can be achieved. The significant adverse factors affecting survival should be taken into account to select the candidates for surgery.

HYPOPHARYNGEAL and cervical esophageal squamous cell carcinomas are often known to have late onset and a bad prognosis. In the past, radiotherapy was the gold standard of treatment of upper esophageal cancer, but long-term follow-up was disappointing, especially with regard to quality of life and cancer control. Many patients needed further operations for persistent dysphagia or for complications related to the radiation, chemotherapy, or both.^1,2

Surgical ablation of advanced tumors of the hypopharyngeal and upper esophageal regions, together with optimal reconstruction of the resultant defect, has remained a surgical challenge. The goal of this surgery is a single-stage reconstruction with low morbidity and mortality, short hospital stay, and early restoration of swallowing.

Historically, various methods of reconstruction have been used after resection of the pharynx, hypopharynx, and cervical esophagus. These have included local skin flaps; deltopectoral flaps; reversed gastric tube esophagoplasty; pectoralis major myocutaneous flaps; visceral interposition with stomach or colon; and free tissue autografts utilizing colon, jejunum, and tubed radial forearm flaps. Free jejunal transfer has become the standard technique for reconstruction of the pharynx and hypopharynx, especially with proximal lesions,^3-5 whereas gastric tube interposition is the technique of choice for reconstruction of the hypopharynx and cervical esophagus when the resection extends below the thoracic inlet.^6-11

In this study, we have reviewed the results of surgical resection of advanced hypopharyngeal or cervical esophageal tumors with visceral interposition in 209 patients. We have also attempted to identify the principal factors affecting survival rate and the best technique of reconstruction.

Between May 1982 and July 1999, 209 patients underwent total pharyngolaryngectomy with visceral interposition in our institution. Most had locally advanced tumors of the hypopharynx (n = 131), with the remainder having cervical esophageal carcinomas (n = 78). Most tumors of the hypopharynx involved the upper esophageal sphincter and the cervical esophagus, and it was therefore often difficult to define the exact site of origin.

In addition to demographic data, patient records were reviewed to determine the indications for radical resection, location and stage of the primary tumor, gross and pathological surgical resection margins, and operative complications and mortality. Functional results, specifically the ability to resume oral intake, were determined using outpatient medical history and telephone surveys. Follow-up was complete for all 209 patients.

In analyzing survival time, we used LIFETEST in the Statistical Analysis System (SAS Institute Inc, Cary, NC). The survival function has been estimated by the actuarial method without excluding postoperative deaths. The log-rank test was used for comparison of the survival curves, with a limit of significance of P<.05. The factors used in the univariate analysis are presented in Table 1. Prognostic factors were analyzed with a proportion hazard regression analysis according to Cox,¹² using a stepwise procedure. The 0.2 level was defined for entry into the model. Multivariate χ² and P values were used to characterize the independence of these factors; the relative risk was used to quantify the relationship between survival time and each independent factor.

Mean age was 55 years (range, 33-81 years). The male-female ratio was 193:16 (92.3% and 7.7%, respectively). The most common complaint at initial evaluation was dysphagia (72%). A total of 42 patients (20%) had received some form of treatment before undergoing surgery: 22 patients underwent primary chemoradiotherapy, 15 had chemotherapy, and 5 had radiotherapy. The most common chemotherapeutic regimens included 5-fluorouracil and cisplatin.

Forty-eight patients (23%) underwent primary treatment for carcinoma of the hypopharynx or cervical esophagus at or below the thoracic inlet. Resection for recurrent tumors was performed in 11 patients (5.3%). No patient underwent this type of surgery for benign strictures.

Six (2.8%) of these patients had one or more synchronous tumors in the upper gastrointestinal tract, tracheobronchial tree, or oral cavity, and this significantly influenced the therapeutic strategy. This fact underlines the need for a careful preoperative workup that has to include endoscopy of these regions.

One hundred forty-five patients received postoperative radiotherapy (73%). Mean duration of follow-up was 28.5 months (range, 0-157 months).

For the hypopharyngeal tumors, clinical preoperative staging was according to the 1983 American Joint Committee on Cancer Classification.¹³ The tumors were staged T1 (n = 15), T2 (n = 62), T3 (n = 58), T4 (n = 7), N0 (n = 94), N1 (n = 28), N2a (n = 11), N2b (n = 1), and N2c (n = 8). All patients with a tumor staged T4 had thyroid invasion. The T1 stage was represented by patients with recurrent hypopharyngeal tumor or associated tumor with a cervicothoracic one. Lymph node stages N2b and N2c included small bilateral lymph nodes sometimes found during the operation, and 1 patient had a large tumor that was unlikely to be reduced in size by radiotherapy.

For the cervicothoracic esophageal tumors, staging was according to the modified Wurtz classification.¹⁴ Tumors were staged T1 (n = 20), T2 (n = 31), T3 (n = 32), N0 (n = 47), N1 (n = 25), and NX (n = 11).

Six patients had 1 or more synchronous tumors, which explains why there are 142 hypopharyngeal tumors in 131 patients and 83 cervicothoracic tumors in 78 patients.

Three different operative procedures were performed in this study: 127 patients (61%) had pharyngolaryngectomy and total esophagectomy with pharyngogastric anastomoses; 77 patients (37%) had pharyngolaryngectomy and cervical esophagectomy, and reconstruction with free jejunal transplant; and 5 patients (2%) had pharyngolaryngectomy and total esophagectomy with pharyngocolic anastomoses.

A free jejunal transplant was used for reconstruction in the hypopharyngeal carcinomas in 74 cases and in 3 of the cervical esophageal carcinomas in which we obtained adequate distal resection margins of more than 2 cm. Total esophagectomy with gastric replacement was done in 68 patients with cervical esophageal carcinomas, in 53 patients with hypopharyngeal carcinomas, and in 6 patients with thoracic carcinomas.

After the 10th postoperative day, a gastrograffin swallow was performed prior to beginning oral feeding. If this showed a leak, oral feeding was delayed for 1 or 2 weeks.

Ten patients (4.8%) died during the postoperative period. The causes of death were pulmonary infection (4 patients), cardiovascular disease (3 patients), cirrhosis (1 patient), sepsis (1 patient), and carotid blowout (1 patient).

Significant complications occurred in 80 patients (38.3%) (Table 1). Several patients had more than one complication.

Forty-seven anastomotic leaks (22.5%) were observed, including 11 that were clinically evident and 36 that were asymptomatic but detected radiologically, and that closed without surgical correction. Where free jejunal grafts had been used, the fistulas always occurred proximally.

Graft necrosis occurred in 5 patients. This was treated by further jejunal reconstruction in 1, by the use of a deltopectoralis myocutaneous flap in 1, and by only local treatment and jejunostomy in 3 instances (1 of these had had a gastric pull-up before pharyngolaryngectomy for esophageal cancer). Of the 3 patients in whom the stomach was anastomosed to the pharynx, necrosis was limited and treated by resection of the necrotic stomach in 1, and temporary cervical diversion and delayed reestablishment of the continuity of the alimentary tract was achieved by tubed radial forearm flaps and deltopectoralis myocutaneous flap in another. The third patient died in the postoperative period of an associated posterior wall tracheal necrosis.

Necrosis and fistula rates for free jejunal transfer (7% and 31%, respectively) were significantly higher than rates for gastric interposition (P = .01).

The average time of resumption of feeding was 19.7 days following the operation (range, 0-700 days). Excluding the postoperative deaths, swallowing was achieved in 98.4% of patients. Only 3 patients were not able to swallow after recovering from the procedure: 2 because of necrosis of the jejunal graft and 1 because of a major swallowing disorder.

The average hospital stay for the whole population was 23.6 days (range, 1-240 days): 24.7 days (range, 1-240 days) for the pharyngogastric and pharyngocolic anastomoses, and 21.7 days (range, 11-60 days) for the free jejunal grafts (P = .02).

All patients were found to have squamous cell carcinomas; 130 (62.2%) had positive lymph nodes, and 56 had capsular penetration.

Stricture was a late complication occurring in 17 patients (8.5%): 8 patients (6.7%) after pharyngogastric anastomoses and 9 patients (12.2%) after free jejunal graft (P = .23). Stricture was due to local recurrence in 12 patients. This complication was treated by esophageal dilatation except where mechanical dilatation was impossible, in which case a jejunostomy was performed. In 1 patient with a benign stricture, a gastric pull-up was performed to reconstruct the pharynx.

The mean hospital stay in this series was 23.6 days (range, 1-240 days) for the whole population, 24.7 days (range, 1-240 days) for pharyngogastric and pharyngocolic anastomoses, and 21.7 days (range, 11-60 days) for free jejunal grafts. These differences did not reach statistical significance.

The average time until resumption of feeding was of 19.7 days (range, 0-700 days) after the procedure. Swallowing was achieved in 98.4% of patients. In addition, the survival rate without dysphagia was higher for the gastric interposition (89% of patients) than for the jejunal grafts (76% of patients).

Tumor recurrences occurred in 118 patients (59.3%), with a mean delay of 19 months (range, 1-104 months). These patients were treated by chemotherapy, radiotherapy, or neck dissection. Local recurrence alone occurred in 46 patients and was combined with neck recurrence in 48 patients. Distant metastases were found in 22 patients, and 2 patients had tumor recurrences at unknown sites.

A second primary cancer occurred in 28 patients, with 16 of their tumors being in the oropharyngeal region.

Overall, the 3-year and 5-year actuarial survival rates were 32% and 24%, respectively, with a median survival time of 17 months (Figure 1). The 5-year survival rate was 29% for hypopharyngeal tumors and 14% for cervical esophageal tumors (P = .02) (Figure 2).

To identify prognostic factors, we evaluated the survival rates for each of the following parameters, as presented in Table 2. Only 6 of the possible prognostic factors were found to actually influence survival: neoplastic localization, postoperative complications, macroscopic or microscopic tumoral residues, pT3 or pT4 histopathological tumor stage for cervical esophageal tumors, and postoperative radiotherapy (Table 3). Lymph node involvement seemed to be a prognostic parameter after the univariate analysis (P = .04) but not after multivariate analysis. No statistical difference was observed between the 2 reconstructive procedures, pharyngolaryngectomy with stomach interposition, or free jejunal autograft (P = .12).

The overall 5-year survival rates for advanced squamous cell carcinomas of the larynx, pharynx, hypopharynx, and cervical esophagus have not changed significantly, despite advances in surgical techniques, chemotherapy, and radiation therapy. Actuarial 5-year survival rates reported in the literature range from 18% to 35% regardless of therapy,^6,15-21 and in our series, the overall 5-year survival rate (24%) was similar.

Optimal reconstructive procedures should provide the lowest mortality and morbidity, the shortest hospital stay, and the most rapid return to successful feeding.

Methods of reconstruction of the pharyngoesophagus are varied. Local skin flaps and deltopectoral flaps are rarely used because of the need for multistaged procedures; the limited length of potential reconstruction; and high morbidity rates resulting from fistula formation, strictures, and flap necrosis.^22,23 The pectoralis major musculocutaneous flap is advantageous because it can be a 1-stage procedure with shorter hospitalization, though morbidity and mortality rates remain high. It is valuable in reconstructing partial defects and secondary pharyngocutaneous fistulas when tumor bulk is not an issue.^23-25 Visceral pedicle grafts gained popularity because they had the advantage of being 1-stage procedures. With improvement in microsurgical techniques, free jejunal transfer has become the method of reconstruction for proximal lesions when resection includes the pharynx, larynx, hypopharynx, and cervical esophagus above the thoracic inlet. Several series have reported low morbidity and mortality rates, good long-term functional results, and successful reestablishment of gastrointestinal continuity in approximately 95% of patients.^3,26-28

Free jejunal transfer avoids a mediastinal dissection, can be used in irradiated fields or when the stomach is unavailable, and allows for reconstruction of more proximal defects. Disadvantages of this technique include the need for 3 intestinal anastomoses, 2 microvascular anastomoses, and difficulty in reconstruction of large defects extending below the thoracic inlet (or when total esophagectomy is required).^26-28

Gastric tube interposition has been shown to be a reliable method of reconstruction for circumferential defects of the pharyngoesophagus. Turner²⁹ first described a pull-through esophagectomy with use of the stomach as a method of reconstruction. Ong and Lee³⁰ revived this procedure in 1960 in the management of hypopharyngeal and cervical esophageal cancer using a thoracotomy. LeQuesne and Ranger³¹ performed the first transhiatal nonthoracic esophagectomy with stomach transposition in the neck. The gastric pull-up was later modified and optimized by Silver,³² Orringer,³³ and Spiro et al.¹⁸ Since that time, several series have reported successful reconstructions of the pharyngoesophagus using this technique, with acceptable morbidity and mortality rates ranging from 5% to 31% and from 5% to 55%,^3,6,11-18,21 respectively.

The incidence of clinically evident anastomotic leakage has decreased in the last 5 years to 10.5% for Bardini et al³⁴ and to 4.5% as quoted by Lam et al.⁶ Similar results were achieved in our study and were less commonly seen after gastric interposition. Pulmonary complications, which are rarely seen in free jejunal transplant, have been reported to occur commonly (40%) after gastric or colonic replacement,³⁵ and were seen in 21.6% of our patients, again being more likely to occur after gastric interposition. Following total esophagectomy, a pharyngogastrostomy is the preferred method for most surgeons because only one anastomosis is required, the stomach has an excellent blood supply (so that failure due to necrosis is almost unheard of), it is relatively easy to prepare, and swallowing after the operation is better because of low rates of stricture and fistula formation.³⁶ Whether the stomach can always easily reach the pharynx is a concern, but in our experience there was no difficulty in reaching the pharynx without significant manipulation of the stomach. If the pharyngeal resection extends to the oropharynx and nasopharynx, the gastric or colon pull-up procedures are compromised by the tension on the suture line and by ischemia of the replaced tissues as more mobilization is required.

Gastric tube interposition and free jejunal transfer are similar in that both can be used in irradiation fields, provide well-vascularized conduits, are single-stage procedures, and allow for return of oral intake. Schusterman et al⁵ and de Vries et al⁴ compared the 2 techniques and concluded that both treatments were equally successful and effective when used appropriately and that morbidity rates were similar. The nature of the defect clearly dictates the method of reconstruction. For short reconstructions above the thoracic inlet and those requiring total esophagectomy, gastric tube interposition is indicated.

In our series, the total mortality and morbidity rates of the operative procedure were 4.8% and 38.3%, respectively, which compares favorably with the results of others.^3,6-11,18,21 The overall decrease in mortality is attributed to improvements in surgical technique, and advances in critical care and nutritional support. The postoperative rate of complications of the jejunal transfer is higher than that of gastric tubes, but the difference is not significant. The gastric interposition is the most frequent cause of respiratory complications (21.6%), without significant difference with the graft (P = .08).

Consistent with the 7% to 37% leakage rates reported previously,^3,6-11,18,21 2.5% of patients developed clinically or radiologically significant anastomotic leakage. Necrosis and fistula rates for free jejunal transfer (7% and 31%, respectively) are significantly higher than for gastric interposition (P = .01).

Short hospital stay is important because of the poor 1-year survival rate, but we did not find any significant differences between jejunal transfers and gastric or colonic tubes. In addition, the survival rate without dysphagia is higher for the gastric interposition (89%) than for the jejunal grafts (76%).

On the whole, in our study, even if there was no significant difference in terms of survival between the 2 techniques of reconstruction, better results were obtained with the gastric interposition procedure because of fewer postoperative complications, faster refeeding, and better quality.

Overall 5-year actuarial survival in our series was 24%. Lam et al⁶ found a trend toward reduced survival in patients with regional lymph node metastasis and positive resection margin. Jacobs et al³⁷ showed that patients operated on for advanced squamous cell carcinoma of the head and neck with positive surgical margins had a survival rate approximately half that of patients with negative margins. In our series, patients with cervical tumor, postoperative complications, disease stages pT3 or pT4 for cervical esophageal tumors, microscopic pharyngeal penetration or incomplete resection, and no administration of postoperative radiotherapy, had a significantly reduced overall actuarial survival.

Preliminary results³⁸ suggest a new role for radiotherapy and chemotherapy in patients with advanced hypopharyngeal cancer and indicate that a treatment strategy involving induction chemotherapy and definitive radiation therapy can be effective in preserving the larynx in a high percentage of patients, without compromising overall survival.

Patients with advanced squamous cell carcinoma of the pharynx, larynx, hypopharynx, and cervical esophagus have a limited survival regardless of therapy. The significant adverse factors affecting survival were cervical tumor localization, postoperative complications, disease stages pT3 and pT4 for the cervical esophageal tumors, microscopic pharyngeal penetration, or incomplete resection. The significant beneficial factors were tumor hypopharyngeal localization and postoperative radiotherapy.

Gastric transposition is a safe and effective method for the immediate restoration of alimentary continuity after pharyngolaryngectomy. The procedure can be performed with low mortality, acceptable morbidity, and a short hospital stay. It has now become the preferred method of reconstruction after total pharyngolaryngectomy at our institution. Further studies will be necessary to compare the use of chemoradiotherapy, larynx preservation, and surgery in terms of overall survival and quality of life, especially the ability to continue oral feeding.

Corresponding author: Jean-Pierre Triboulet, MD, Service de Chirurgie Digestive et Generale, Clinique Chirurgicale Adultes est, Chru Lille Hopital Huriez, Place de Verdun, 59037 Lille CEDEX, France (e-mail: jp-triboulet@chru-lille.fr).