Stricture of the proximal esophagus in head and neck carcinoma patients after radiotherapy




It is well recognized that many patients with head and neck carcinoma have problems with food intake and malnutrition. The objective of the current study was to determine the clinical pattern of patients with nonneoplastic stricture of the upper esophagus after radiotherapy for head and neck carcinoma.


A retrospective chart study of 22 patients with stricture of the proximal esophagus diagnosed between 1993 and 1999 at Karolinska Hospital was performed. The dose volume histograms of the first 2 cm and 5 cm, respectively, of the proximal esophagus were calculated.


Five of the patients (23%) had total obliteration. The first 2 cm of the esophagus received at least 60 grays (Gy) in > 80% of the volume. Radiation injury was not reported to occur at doses < 60 Gy. There was a correlation found between dysphagia during radiotherapy and the development of proximal esophageal stricture. Stricture was diagnosed 1–60 months (median, 6 months) after radiotherapy. In 18 patients, the stricture was treated with single or repeated endoscopic dilation. These treatments allowed a nearly normal diet in 78% of the patients.


Stricture of the upper esophagus is one deglutition disorder that is reported to occur after radiotherapy for head and neck carcinoma. In the current study, the authors emphasize the importance of knowing the tolerance of the normal esophagus to irradiation as well as early diagnosis of stricture of the proximal esophagus because this condition may lead to physical and emotional distress. Cancer 2003;97:1693–700. © 2003 American Cancer Society.

DOI 10.1002/cncr.11236

Dysphagia is commonly observed in patients with head and neck carcinoma. A wide range of swallowing problems induced by the tumor itself or by any therapeutic intervention generally are considered to be problems that must be accepted. Particular emphasis, however, must be placed on the pathogenesis of such problems to improve the situation for the patient. Even though the nutritional status of the patient might be restored by tube feeding, swallowing problems generally have a considerable impact on quality of life1 and also might lead to social isolation. There are certain conditions that should be recognized by the physician to avoid malnutrition and unnecessary suffering. One such condition is injury to the proximal esophagus induced by radiotherapy.

Injury to the gastrointestinal tract is a well recognized complication after radiotherapy for pelvic and abdominal malignant tumors. The ileum and rectum are parts of the gastrointestinal viscera that quite frequently are damaged because of the wide use of radiotherapy for gynecologic and urologic malignancies.2 The incidence of gastrointestinal irradiation injury is dependent on different factors but knowledge of the tolerance of normal tissue is limited for the majority of sites.3 The esophagus at the upper end of the gastrointestinal tract is reported to be fragile and radiosensitive.4 Acute radiation esophagitis commonly is observed in connection with radiotherapy for operable breast carcinoma, inoperable carcinoma of the lung, and esophageal carcinoma.5–7 It has been suggested that there is no clear relation between acute esophagitis observed during radiotherapy and late side effects such as stricture of the esophagus.8 The pathophysiology of the acute and late side effects on the esophagus therefore might differ. To our knowledge there have been a few previous reports in the medical literature of esophageal irradiation injury in patients with head and neck carcinoma.

To our knowledge, uncertainty exists regarding the predisposing factors and the incidence of stricture of the esophagus. Early signs of radiation reactions frequently are observed in some patients whereas others appear to be resistant. The general aim of the current study was to describe 22 patients with nonneoplastic stricture of the upper esophagus after radiotherapy who were diagnosed at the Department of Otolaryngology at Karolinska Hospital between 1993 and 1999 and to correlate further the absorbed dose in the esophagus with clinical findings. Specific questions to be elucidated include: 1) is it possible to predict the risk of a stricture of the upper esophagus from the planned absorbed dose in this segment of the esophagus? and 2) what is the treatment outcome in patients with proximal esophageal stricture after radiotherapy?



Between 1993–1999, 22 patients with stricture of the proximal esophagus after treatment for head and neck carcinoma were diagnosed at the Department of Otolaryngology at Karolinska Hospital. All patients were given external radiotherapy at the Department of Oncology, Radiumhemmet, amounting to a total dose of 64 grays (Gy) in 32 fractions. The patients had been treated between 1992 and 1998 and the total number of patients who were given radiotherapy during this period was 642. The treatment modality was decided by a joint group of oncologists and head and neck surgeons prior to therapy. Ten patients who achieved a complete response received radiotherapy as the only therapeutic modality. Combined treatment with preoperative radiotherapy followed by surgery was given to seven patients and five patients underwent surgical treatment and postoperative radiotherapy. The most commonly performed surgery was standard neck dissection. No patient underwent any surgical procedure involving the hypopharynx, larynx, or proximal esophagus. All patients except one had squamous cell carcinoma. One patient was given postoperative radiotherapy after resection of an adenocarcinoma in the ethmoid sinus. Presentation of the patients according to gender, age, tumor site, TNM classification, and treatment is shown in Table 1.

Table 1. Presentation of 22 Patients with a Proximal Esophageal Stricture According to Gender, Age, Tumor Site, TNM Classification, and Treatment
Patient no.GenderAge (yrs)Tumor siteTNMTreatment
  1. Preop: preoperative; RT: radiotherapy; postop: postoperative.

  2. Type of surgical procedure performed: 1: neck dissection; 2: hemiglossectomy; 3: excision of the floor of the mouth; 4: resection of the base of the tongue; 5: resection of tumor of the oropharynx; 6: total ethmoidectomy; 7: radial forearm free flap.

1Male66OralT3N0M0Preop RT + surgery (1,2,3,7)
3Male59OropharynxT3N0M0Preop RT + surgery (1,5)
4Female70OropharynxT4N2bM0Preop RT + surgery (1,5)
6Male62SinusT1N0M0Surgery (6) + postop RT
8Female58OralT1N0M0Surgery (1,2) + postop RT
14Male63PharynxT1N1M0Preop RT + surgery (1)
15Female61OralT2N0M0Surgery (3) + postop RT
16Male65OropharynxT2N2bM0Preop RT + surgery (1,4,7)
18Male40OralT3N0M0Surgery (2) + postop RT
19Female59OropharynxT2N2M0Surgery (1,4) + postop RT
21Male73EpipharynxT1N2AM0Preop RT + surgery (1)
22Male44EpipharynxT1N2cM0Preop RT + surgery (1)


Treatment technique and dose planning

All patients were set up in the supine position and immobilized in a mask (Posifix; Sinmed BV, Reeuwijk, The Netherlands). Treatment planning was performed on a three-dimensional conformal treatment planning system (TMS; Helax MDS-Nordion, Ottawa, Ontario, Canada). The planning was based on approximately 30 computed tomography (CT) slices with a thickness of 0.5–1 cm and contiguous acquisition. In the calculation process of the treatment plan correction for tissue inhomogeneity was performed. All patients in the study were treated on a 6-megavolt accelerator. The treatment technique for bilateral clinical target volume (CTV) including the primary tumor, bilateral cervical lymph nodes, and supraclavicular lymph nodes comprised two opposed coplanar or noncoplanar lateral conformal fields. Treatment in patients with a lateral target was planned with one frontal and one oblique posterior field.

Dose and fraction: patient dosimetry

All treatments were performed with 5 daily fractions per week and 2 Gy per fraction. The volume anterior to the spinal cord then was boosted to 64 Gy in 9 fractions. The absorbed dose to the spinal cord did not exceed 50 Gy. For patients with lateral or small frontal CTV, the complete CTV was covered with 64 Gy and the dose to the spinal cord still did not exceed 50 Gy. In vivo dosimetry was performed according to the ordinary routine of the department with a commercial dosimetric system (DPD 510; Scanditronix, Uppsala, Sweden). For practical reasons, the absorbed dose was checked in one field only per fraction.

Dose to the esophagus

The dose planning of 20 of the 22 patients with stricture of the upper esophagus could be studied retrospectively. On the dose planning CT scan, the absorbed dose in the upper esophageal region was calculated using a 1-cm CT scan cut, starting from a level approximately 1 cm below the vocal cords and going downward. In each cut the absorbed dose in five spots corresponding to the esophagus was calculated. The mean of the five spots was calculated as the absorbed dose in each cut within the upper esophagus. Two calculations were made. The dose delivered to the first 3 slices, which corresponded with the first 2 cm of the proximal esophagus, was calculated. Furthermore, the dose delivered to the first 6 slices corresponding to the first 5 cm of the esophagus was derived and correlated with clinical findings. This approach was followed for two reasons. First, the strictures appear in the upper part of the esophagus. Second, the irradiation techniques applied are such that the dose to the lower parts of the esophagus drops rapidly. In the analysis the full-dose volume histograms of the first 2 cm and 5 cm, respectively, of the proximal esophagus of each patient together with their corresponding mean dose and standard deviation have been employed.


After the final treatment the patients were seen every 1–3 months by the surgeon and oncologist, respectively, for 5 years. All patients had a nursing visit for weight control every second week during and after radiotherapy at the ward for head and neck carcinoma patients in the Department of Otolaryngology for a period of 3 months. Weight was taken at each visit. In case of a 5% weight loss compared with the weight before radiotherapy, the patient was referred to a physician for nutritional support.

Diagnostic Criteria and Treatment

Clinical symptoms that lead to suspicion of a stricture were impaired swallowing function and weight loss that was either persistent after radiotherapy or of late onset. Conventional swallow X-ray using barium contrast material was performed as the initial examination. Endoscopic examination was performed under general anesthesia to identify the cause of deglutition disorder and to diagnose the degree of stricture. Repeated dilation using Savary dilators generally was used to dilate the esophageal stricture. In appropriate patients this maneuver was followed by repeated dilations as an outpatient procedure using mercury bougies. In addition, rigid endoscopes were used to dilate mild strictures by placing endoscopes of different sizes, starting with a 7 mm × 10 mm endoscope and ending with a 14 mm × 16 mm endoscope. Alternatively, resection of the esophageal stricture was performed with reconstruction using a free microvascular forearm flap.

Reference Group: Patients without Dysphagia

The radiotherapeutic data for the patients with esophageal stricture were compared with 22 head and neck carcinoma patients without dysphagia after radiotherapy. A survey using a questionnaire that was mailed to all surviving patients who received radiotherapy during 1994 at the Department of Oncology, Radiumhemmet, for a head and neck tumor selected the patients without any swallowing problems. The data files showed that irradiation had been administered to 105 patients with head and neck carcinoma during 1994. In January 1996, 65 of these patients still were alive. Of the 65 surviving patients, 51 answered the questionnaire regarding swallowing problems. Of these, 24 had no deglutition disorders. It was only possible to investigate the radiotherapeutic dose planning of 22 survivors without swallowing problems retrospectively. The number of patients in each tumor stage grouping (TNM staging system) are as follows: Stage 0, two patients; Stage I, three patients; Stage II, seven patients; Stage III, two patients; and Stage IV, four patients. Four patients could not be classified retrospectively. The corresponding dosimetric information to the one that was derived for the patients with esophageal stricture was used for the patient group without dysphagia to make dosimetric comparisons between the two groups of patients.


Characteristics of Patients Who Develop Esophageal Stricture

Seventeen patients were severely affected by oral and pharyngeal mucositis in connection with radiotherapy and could not maintain sufficient oral intake. In 12 of these patients, hospitalization was required during radiotherapy for supportive care because of dysphagia and odynophagia. Introduction of a nasogastric tube or percutaneous endoscopic gastrostomy (PEG) was performed in 13 patients during radiotherapy (in 2 patients before 30 Gy of radiation, in 7 patients between 30–50 Gy of radiation, and in an additional 4 patients after 50 Gy of radiation). Another 4 patients experienced severe swallowing problems at the end of radiotherapy and were hospitalized within 3 weeks after the termination of radiotherapy for tube feeding. One additional patient who experienced dysphagia and considerable weight loss was given a PEG 5 months after radiotherapy. Altogether, 13 patients received a PEG and 5 patients were fed via a nasogastric tube. In all patients in whom the onset of dysphagia occurred during radiotherapy, swallowing problems were persistent or progressive after their first appearance. Only four patients were able to maintain sufficient oral intake. The interval between the end of radiotherapy and the diagnosis of stricture and the duration of tube feeding before the diagnosis of stricture was 1–60 months (median, 6 months) and 2–25 months (median, 5 months), respectively.

Comparison between X-Ray Findings and Endoscopy in Patients with Stricture

Diagnosis of deglutition disorder initially was based on X-ray using barium swallow in 16 patients. The examinations were comprised of radiographic and fluoroscopic studies. Radiographic evaluation was based on anatomic and functional changes. The anatomic abnormalities found were classified in four cases as total obstruction, as severe stricture in five cases, and as moderate to mild narrowing of the esophageal inlet in seven patients. Figure 1 shows the three different categories of stricture in the proximal esophagus. Endoscopy was performed in all patients except one. The patient who did not undergo an endoscopic examination was hospitalized for such a procedure when cerebral metastasis were diagnosed and he died within 12 days. Using endoscopy, the remaining 21 patients with stricture could be categorized into 3 groups by the appearance of their esophageal stricture. Grade 1 stricture indicated moderate stricture with a fibrous membranous ring or moderate fibrosis. The stricture initially could be passed by a rigid esophagoscope (7 mm × 10 mm) and dilated (12 patients). Grade 2 indicated severe stricture with severe fibrosis of the esophageal inlet. It was not possible to pass the stricture using the smallest endoscope (7 mm × 10 mm) without dilation (4 patients). Grade 3 indicated total obliteration. No communication was visible between the hypopharynx and the esophagus (5 patients).

Figure 1.

Radiographs demonstrating three representative cases with radiation-induced narrowing of the proximal esophagus. Left: moderate stricture (fibrous membranous ring or moderate fibrosis). Center: severe stricture (severe fibrosis of the esophageal inlet). Right: total obliteration (no communication visible between the hypopharynx and esophagus). The arrow in the graph indicates where the stricture is located in each case.

Treatment Results of Esophageal Stricture

A total of 21 patients underwent endoscopic dilations under general anesthesia. Dilation using Savary dilators was performed in 14 patients with moderate or severe stricture (Grade 1-2) between 1–8 times (median, 3 times). Successful dilation was reported to occur in six of these patients followed later by outpatient procedures using peroral mercury-filled rubber dilators to restore swallowing function. In addition, dilation with rigid endoscopes was performed in four patients with moderate stricture (Grade 1) once or twice. Initially, it was impossible to perform endoscopic dilation in the three patients diagnosed with a Grade 3 stricture (no visible communication between the hypopharynx and esophagus). Endoscopic dilation as a single treatment modality was helpful in 14 patients (i.e., the nasogastric tube or PEG could be removed and the patient was able to receive sufficient oral feeding). Endoscopic dilation failed to offer any improvement in four patients. In two of these patients the initial examination revealed a stricture suitable for dilation according to Savary but at later examination a total obliteration of the esophageal inlet was found. In one of these patients, a hospital stay was complicated by a perforation of the esophagus and the patient died 1 month later, indicating a mortality rate of 5%. Resectional surgery with larynx preservation, including reconstruction with a free microvascular forearm flap, was performed in the two patients with progressive stricture. In a 76-year-old man with severe stricture, the maneuver failed because of necrosis of the free flap. The second patient, a 40-year-old man, had a total obliteration of the proximal esophagus. He demonstrated considerable improvement in swallowing after surgery and deglutition thereafter could be maintained by repeated endoscopic dilations.

Dose to the Esophagus

The proximal esophagus was in the high-dose region for all the applied irradiation techniques. This can be seen clearly in Figure 2, showing the dose volume histograms of the stricture group and reference group, respectively. The dose volume histograms represent the distribution of dose to the volume of the organ. In other words, they show the proportion of the organ that receives a certain dose. The volume (V) can be in units of absolute value (m3) or of relative values (%) as a proportion of the whole organ volume (V0). The first 2 cm of the esophagus in particular received 60 Gy in > 80% of the volume. It was observed that the dose to patients with esophageal stricture was higher than that to patients without dysphagia. The difference between the doses was small, especially for the high doses (i.e. doses > 60 Gy). The dose for the first 5 cm of the esophagus was reduced compared with the first 2 cm, although the main characteristics of the cumulative dose volume histograms remain the same. In Figure 3, in which the biologically effective uniform dose9 was used as the dose unit, there was a clear distinction between the patients who developed stricture of the proximal esophagus and those without dysphagia. The biologically effective dose is the dose that causes exactly the same normal tissue complication rate as the delivered dose distribution. The dose response curve is characterized by the values D50 = 68.9 Gy, signifying the uniform dose that causes complications to 50% of the patients and γ = 4.9, which is the maximum normalized gradient of the curve demonstrating the change in the response probability for a certain change in dose. It appears that the mean dose of 65 Gy could be a good choice as a tolerance dose for the endpoint of esophageal stricture and reference length of the first 2 cm of the proximal esophagus, although a whole shape of the delivered dose distribution is of importance in making such an assessment. If the dose to the first 5 cm of the proximal esophagus is to be used as dose reference, the mean dose value of 60 Gy should be used.

Figure 2.

The cumulative dose volume histogram for the upper 2 cm and upper 5 cm of the esophagus in the entire study population, divided among the group of patients with stricture of the proximal esophagus (complications) and the group of patients without dysphagia (complication-free). The corresponding mean dose and mean maximum dose to the esophagus (Dmean ± standard deviation [SD] and Dmax ± SD, respectively) for the two groups of patients also are included. V: volume; VO: whole organ volume.

Figure 3.

The dose response curve of the esophagus using the patient data for the 2-cm reference segment. The dose unit is the biologically effective uniform dose (equation image;) and it is used to determine the position of each patient (solid squares for the patients with esophageal stricture and open circles for the patients without dysphagia) on the theoretic dose response curve (solid line), which is based on the dose distribution. The dose of each patient was converted to the biologically effective uniform dose. A significant positive association between radiation stricture induction and dose (cutoff value of 65 grays [Gy]) was found (odds ratio = 3.93; 95% confidence interval, 1.03–15.00). CR: complete response.


The application of radiotherapy to the management of carcinoma of the head and neck involves a risk of acute and late side effects to surrounding tissues. The aim of the current study was to describe 22 patients treated with radiotherapy who had stricture of the proximal esophagus and were observed at the Department of Otolaryngology at Karolinska Hospital and to relate further the absorbed dose in the esophagus with clinical findings. Endoscopic confirmation of esophageal stricture was obtained in 96% of the cases. The most important symptoms in patients with proximal esophageal stricture after irradiation were dysphagia and weight loss. The severity of stricture varied from total obliteration of the esophageal inlet to a fibrous membranous ring.

Long-term radiation injury to the gastrointestinal tract is a common complication of irradiation for pelvic and abdominal tumors.4, 10 Less commonly reported symptoms include severe lesions to the upper digestive tract in patients receiving radiotherapy for head and neck malignancies. Obviously, there is a large interindividual variability in the long-term outcome of radiotherapy to normal tissue in the upper esophagus. With regard to the majority of other unwanted reactions, one has to ask whether the variability is strictly dose-dependent or mainly the result of the radiosensitivity of the patient. Using a questionnaire mailed to all surviving patients treated with radiotherapy during 1994 for head and neck carcinoma, a reference group of 22 patients who did not report swallowing disorders was identified and used for comparison. The current analysis suggests that a dose of < 60 Gy can be considered as a dose level corresponding to a low number of complications. The risk of esophageal injury is dependent on the shape of the applied dose distribution and possibly on other simultaneous treatment interventions in the region of the esophageal inlet. For those patients with head and neck carcinoma who are receiving radiotherapy, the proximal esophagus should be regarded as a dose-limiting organ. In conventional radiotherapy, shielding to the upper esophagus should be used whenever possible. For the future, intensity-modulated radiation therapy is one way to obtain lower doses to sensitive organs in the vicinity of target volumes.

Postirradiation complete esophageal obliteration was noted in 5 of 22 patients with stricture (23%). This is a rare condition but has been reported previously.11 The chronic and late phases of irradiation effect have been reported to correlate with progressive, obliterative endarteritis and ischemia.2, 12 The few earlier reports to our knowledge offered no clear consensus concerning the risk factors for radiotherapeutic injury to the esophageal inlet. Although other factors in addition to irradiation could be involved in the pathogenesis of an esophageal stricture, we could not identify any predisposing factor for esophageal injury besides the absorbed dose. In 91% of the patients with the most severe forms of esophageal stricture (severe stricture or total obliteration), an endoscopic placement of a permanent feeding tube was established during the period of irradiation or within 3 weeks after the end of treatment. It cannot be ignored that the introduction of the flexible endoscope during the PEG procedure might have injured the esophageal wall, thereby adding trauma to an inflammatory process. Several investigators have pointed out that concomitant use of chemotherapy increases the risk of chronic radiation injury to the esophagus.13, 14 No patient in the current study received chemotherapy. Another finding that has been reported to predispose patients to esophageal injury is chronic xerostomia,15 a risk factor that could not be determined retrospectively. There is scope for speculation that several other risk factors such as a poor nutritional status, intensive use of tobacco, and local infections could increase the incidence of radiotherapeutic complications in the upper digestive tract in patients with head and neck carcinoma.

In 78% of the cases with moderate to severe stricture, repeated endoscopic dilations were found to be significantly helpful. This finding is in agreement with earlier clinical experience demonstrating that in the majority of cases appropriate treatment involves endoscopic dilations.4, 8, 16–19 It should be emphasized that there is a considerable risk that the diagnosis might be overlooked and treatment thereby delayed. However, in recognizing the symptoms of stricture of the proximal esophagus in patients treated with irradiation, early treatment might prevent total obliteration and malnutrition. In patients with severe symptoms that do not respond to bougienage, surgical resection can be an alternative. Moghissi and Pender17 used conservative surgery including incision and transverse closure. Vanagunas et al.19 reported a few patients who underwent esophagectomy and gastric pull-up. In the current study, two patients were treated with surgical resection followed by reconstruction with a free forearm flap. Surgery was found to improve the situation for one patient considerably.

In some earlier reports, esophageal stricture has been described as a special condition that has a late and delayed onset after the termination of radiotherapy.8, 20 In the current study there was a remarkably high proportion of patients with persistent or progressive swallowing problems continuing since the final part of radiotherapy. Severe symptoms of dysphagia developed in 71% of the patients during radiotherapy or in connection with the termination of treatment. Altogether, 17 of 22 patients with stricture required tube feeding during radiotherapy or within 3 weeks after the end of therapy. Consequently, the findings in the current study cannot confirm that acute radiation esophagitis and esophageal stricture are two distinct entities. However, proximal esophageal stricture can have a late onset in some cases.


In the current study, a stricture of the proximal esophagus was found in approximately 3.4% of all patients with head and neck carcinoma who received radiotherapy during the observation period. The results of treatment of stricture of the proximal esophagus were dependent on the degree of stricture and the general condition of the patient. The majority of patients with moderate to severe stricture responded to repeated endoscopic dilations. The mortality rate from treatment was approximately 5%. The long-term complication that may result from an esophageal stricture demands special attention to the upper digestive tract when planning the dose to be given to patients with carcinoma of the head and neck.