Incidences are based on the National Cancer Institute’s Surveillance, Epidemiology, and End Results 9 registry, 1973-2009.
Incidence of synchronous esophageal malignant neoplasms in patients with index head and neck squamous cell carcinoma detected on staging flexible or rigid esophagoscopy based on literature review, 1979-2014. Results are stratified by studies that examined patients from North America and the World (excluding North America). Exponential line of best fit for North America values is shown.
McGarey PO, O’Rourke AK, Owen SR, Shonka DC, Reibel JF, Levine PA, Jameson MJ. Rigid Esophagoscopy for Head and Neck Cancer Staging and the Incidence of Synchronous Esophageal Malignant Neoplasms. JAMA Otolaryngol Head Neck Surg. 2016;142(1):40-45. doi:10.1001/jamaoto.2015.2815
Rigid esophagoscopy (RE) was once an essential part of the evaluation of patients with head and neck squamous cell carcinoma (HNSCC) due to the high likelihood of identifying a synchronous malignant neoplasm in the esophagus. Given recent advances in imaging and endoscopic techniques and changes in the incidence of esophageal cancer, the current role for RE in HNSCC staging is unclear.
To analyze the current role of RE in evaluating patients with HNSCC, and to determine the incidence of synchronous esophageal malignant neoplasms in patients with HNSCC.
Design, Setting, and Participants
In this retrospective study performed at an academic tertiary care center, 582 patients were studied who had undergone RE for HNSCC staging from July 1, 2004, through October 31, 2012. To assess the incidence of synchronous esophageal malignant neoplasms, a literature review was performed, and the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) data set was queried.
Main Outcomes and Measures
The primary outcome measure was the incidence of synchronous esophageal malignant neoplasms, as measured by retrospective review at our institution, SEER data set analysis, and literature review. Secondary outcome measures were RE complications and nonmalignant findings during RE.
A total of 601 staging REs were performed in 582 patients. The mean age was 60.2 years and 454 (78.0%) were men. There were 9 complications (1.5%), including 1 esophageal perforation (0.2%). Rigid esophagoscopy was aborted in 50 cases. Of the 551 completed REs, no abnormal findings were noted in 523 patients (94.9%), and nonmalignant pathologic findings were identified in 28 patients (5.1%). No synchronous primary esophageal carcinomas were detected. The incidence of synchronous esophageal malignant neoplasms found on screening endoscopy based on literature review and on SEER data set analysis was very low and has decreased from 1980 to 2010 in North America. The incidence reported in South America and Asia was relatively high.
Conclusions and Relevance
Rigid esophagoscopy is safe, but the utility is low for cancer staging and for detection of nonmalignant esophageal disease. Review of the literature and analysis of a large national cancer data set indicate that the incidence of synchronous esophageal malignant neoplasms in patients with HNSCC is low and has been decreasing during the past 3 decades. Thus, screening esophagoscopy should be limited to patients with HNSCC who are at high risk for synchronous esophageal malignant neoplasms.
Panendoscopy, including laryngoscopy, bronchoscopy, and esophagoscopy, has historically served as the standard in the evaluation of patients with head and neck squamous cell carcinoma (HNSCC) to stage the primary tumor, assess for synchronous malignant neoplasms, and/or locate unknown primary tumors.1 Bronchoscopy has been eliminated as part of the routine evaluation because of clear demonstration that, in the absence of pulmonary symptoms and with normal chest radiography results, finding a synchronous lung cancer is rare.2
The approach to esophageal evaluation has also evolved. Although esophageal evaluation was historically performed using rigid esophagoscopy (RE), barium esophagography and flexible esophagoscopy, including transnasal esophagoscopy (TNE), are current alternative methods to detect synchronous esophageal malignant neoplasms. In addition, high-resolution computed tomography (CT) and positron emission tomography (PET)/CT are sensitive for detecting esophageal disease. Despite these advances, many institutions still use RE during initial workup of HNSCC.3
As alternatives to staging RE become more popular, the number of cases of RE will decrease. It is important to understand this change because RE has utility in other settings, such as esophageal dilation and foreign-body removal, and fewer REs may alter resident training. In the present study, we sought to ascertain the real value of RE by assessing a large number of RE procedures for trends in detection of synchronous esophageal malignant neoplasms in patients with HNSCC. We hypothesized that, as an element of staging endoscopy for HNSCC, rigid esophagoscopy is safe but has little impact on patient management and that this may, in part, be owing to a reduction in the number of synchronous esophageal malignant neoplasms over time.
This study was approved by the University of Virginia Institutional Review Board. Patients who underwent RE for HNSCC staging from July 1, 2004, through October 31, 2012, were identified using the University of Virginia Head and Neck Cancer Research Database. During this time, standard staging examination included RE; no flexible esophagoscopy was performed for HNSCC staging. Operative notes, postoperative progress notes, discharge summaries, and clinic notes were reviewed. It was noted when patients had insufficient documentation or if their procedure was aborted. In addition, the medical record was searched for emergency department visits or readmissions within 1 week after surgery.
A literature search was conducted for studies documenting synchronous esophageal malignant neoplasms in patients with head and neck cancer. The MEDLINE database was queried using medical subject headings (MeSH) and keyword searches: head & neck cancer (MeSH) + esophagoscopy (MeSH) [second primary neoplasm or multiple primary neoplasm] and head & neck cancer (MeSH) and esophageal neoplasm (MeSH). Studies were included if the study population included patients initially diagnosed as having HNSCC and if the study documented the incidence of synchronous esophageal malignant neoplasms detected by rigid or flexible screening endoscopy. A synchronous second primary neoplasm was defined as a neoplasm diagnosed simultaneously or within 6 months of the index cancer diagnosis. Appropriate studies referenced by review articles identified by the MEDLINE search were included. Studies were excluded if they used registries or included limited head and neck subsites.
The calendar year incidence of synchronous esophageal malignant neoplasms in patients with HNSCC was calculated using the Multiple Primary – Standard Incidence Ratio session of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) 9 registry between January 1973 and December 2009.4 Patients included were diagnosed as having oral cavity or pharynx cancer and subsequently diagnosed as having malignant esophageal neoplasm within 6 months. To maximize inclusion, patients diagnosed as having an index malignant esophageal neoplasm and subsequently diagnosed as having oral cavity or pharynx cancer within 6 months were also included.
A total of 601 REs were performed for staging purposes in 582 patients (Table 1); all were performed in conjunction with direct laryngoscopy. There were 454 men (78.0%) and 128 women (22.0%). The mean age was 60.2 years. Of the 50 aborted staging REs (Table 1), 25 patients underwent alternative studies that revealed no esophageal abnormalities before treatment: 5 underwent barium esophagography, 16 underwent chest CT, and 4 underwent whole-body PET/CT.
Of the 551 completed examinations, 523 (94.9%) detected no disease, whereas nonmalignant disease was found in 28 (5.1%) (Table 2). There were 9 complications (1.5%); 1 major complication occurred (esophageal perforation) for a rate of 0.2%. No pneumothorax, postoperative bleeding, or postoperative death occurred. Two minor anesthesia-related complications (prolonged sedation) occurred, and 1 patient had temporomandibular joint dislocation after direct laryngoscopy and esophagoscopy (Table 2).
Figure 1 shows the annual and decade incidence of synchronous esophageal malignant neoplasms in patients with HNSCC based on assessment of the SEER database; the incidence has consistently decreased during the past 3 decades. The mean rate of synchronous esophageal malignant neoplasms in the 1980 to 1989 decade was 0.36%; this trended down to 0.19% for the 2000 to 2010 decade, representing a nearly 50% reduction in incidence.
Table 3 summarizes the results of a literature search performed to evaluate the rate of synchronous esophageal malignant neoplasms detected by staging esophagoscopy and reveals distinct trends in North America2,5- 20 compared with the remainder of the world.21- 36 These results are shown graphically in Figure 2. In North American studies, the rate of synchronous esophageal malignant neoplasm detection has decreased from 1% to 8% in the late 1970s and early 1980s to 0% in the last decade. In studies from the remainder of the world, the rate has remained varied without a clear trend, ranging from 0% to 23%.
In our experience, RE as part of a staging examination for patients with HNSCC did not change patient management—no synchronous esophageal malignant neoplasms were identified. Literature review reveals a trend of decreasing prevalence of synchronous esophageal malignant neoplasms in patients undergoing staging endoscopy for HNSCC in the United States and Canada (Table 3 and Figure 2). Literature from the 1980s reveals a rate of synchronous esophageal malignant neoplasms found during staging endoscopy for HNSCC ranging from 0% to 8%, with most studies reporting ranges of 1% to 3%. Since 2000, no North American study in our review reported a single synchronous esophageal malignant neoplasm found on staging esophagoscopy. Our SEER analysis (Figure 1) corroborates this trend. The decreasing rate of synchronous esophageal malignant neoplasms detected on endoscopic staging of HNSCC may be owing to several factors. First, as use of thoracic CT, whole-body PET/CT, and flexible TNE increases, a higher percentage of synchronous esophageal malignant neoplasms may already be identified before formal staging endoscopy. However, thoracic CT and PET/CT miss esophageal cancers that are identified by esophagoscopy, and it is not clear that these techniques can substitute for visualization of the esophageal mucosa.30,37
Second, the incidence of esophageal squamous cell carcinoma (ESCC) in the United States has decreased during the past several decades, probably because of a decrease in tobacco and alcohol use combined with an increased intake of fresh fruits and vegetables. This trend may not apply to patients with HNSCC, given that tobacco and alcohol abuse are highly prevalent in this population.38 In stark contrast to ESCC, the incidence of esophageal adenocarcinoma in white Americans has increased 300% to 400% from 1975 to 2004, likely related to the increased incidence of gastroesophageal reflux disease.39 There is substantial geographic variation in the incidence of esophageal cancer, with Chinese and Japanese populations being at much higher risk.40 Wang et al41 published their results in 315 patients with newly diagnosed HNSCC in China; they found synchronous esophageal neoplasia (ranging from squamous dysplasia to invasive ESCC) in 69 of 315 patients (21.9%). Thus, although the prevalence of synchronous esophageal malignant neoplasms may now be extremely low in populations such as those outlined in Table 3 (eg, American, Canadian, and Swiss), high-risk populations may continue to warrant careful esophageal examination.40,41
Third, the incidence of human papillomavirus (HPV)–related oropharyngeal squamous cell carcinoma (OPSCC) has increased significantly from 1973 to 2004, whereas the incidence of HPV-unrelated OPSCC has significantly decreased during the same period.42 At our institution, 62% to 68% of OPSCC cases are HPV related.43- 45 Patients with HPV-related HNSCC have a significantly lower rate of tobacco and alcohol use,43 and there is a very low prevalence of HPV DNA in ESCC samples.41,46 Studies47,48 have found a decreasing rate of synchronous and metachronous esophageal malignant disease in OPSCC over time, whereas other head and neck subsites remained without significant change. It therefore follows that patients with HPV-related HNSCCs have a low propensity for synchronous esophageal malignant neoplasms and that, as this population has increased, the overall rate of synchronous esophageal malignant neoplasms has decreased.
The present study reveals a low incidence of nonmalignant findings compared with other published reports: reflux esophagitis (2.7%), Barrett esophagus (0.5%), and esophageal candidiasis (0.2%). Farwell et al49 published findings from TNE in 100 patients with HNSCC recruited during routine posttreatment follow-up. The rate of peptic esophagitis was 63% (37% grade 1-3), the rate of Barrett esophagus was 8%, and the rate of esophageal candidiasis was 9%. Postma et al50 reported a series of more than 700 TNEs for all indications in the otolaryngology clinic setting. The rates of esophagitis, Barrett esophagus, and esophageal candidiasis were 16.5%, 4.6%, and 3.9%, respectively. There is little doubt that the narrow field of view afforded by the RE hinders the detection of subtle mucosal change compared with the wider and better illuminated field of view of modern flexible esophagoscopes. This finding likely explains the large difference in rates of nonmalignant disease detected by TNE compared with RE.
We had a low incidence of complications in our series. Although there were 4 dental injuries and 1 temporomandibular joint dislocation, all esophagoscopies were completed in conjunction with direct laryngoscopy, so it is uncertain that these complications are directly related to the esophagoscope. One patient (0.2%) had the major complication of esophageal perforation. Perforation rates in the literature are typically low for staging endoscopy but higher in patients undergoing dilation or biopsy of esophageal lesions. In a retrospective review by Kubba et al51 of 434 esophagoscopies completed for diagnostic purposes, 5 perforations (1.2%) occurred; 2 of these patients with perforations exhibited no signs or symptoms within 8 hours of the surgery. The mortality after perforation was 40%. This finding highlights the fact that although the risk of esophageal perforation during staging procedures is low, the consequences are grave. It also emphasizes the need for trainees to be adequately experienced with this technique. Tao and Damrose19 performed 99 REs as part of staging endoscopy, and 3 of these patients (3%) experienced esophageal perforation. In the same study, none of the 276 patients who underwent flexible esophagoscopy experienced a perforation. The rate of perforation is also extremely low with flexible TNE; to our knowledge, there is only 1 published case of perforation with awake in-office esophagoscopy.52 No perforations were noted in more than 700 TNEs published by Postma et al.50
It would seem that there is now little need for rigid endoscopic evaluation of the esophagus in the staging of HNSCC in North American populations. However, a subset of patients with HNSCC are likely at higher risk, and it may be prudent to continue to survey the esophagus when staging the disease of these patients. Studies22,23,28,34- 36 reviewed from Japan, China, and Taiwan found rates of synchronous esophageal malignant neoplasms of 4.3% to 23% (Table 3), which are significantly higher than in North American HNSCC populations. In addition, several studies34,41 have reported risk factors for synchronous esophageal malignant neoplasms, including degree of alcohol abuse, cigarette smoking status, and primary tumor subsite (HPV-unrelated oropharynx and hypopharynx). Such risk factors, along with East Asian ethnicity or recent immigration from countries with documented high rates of synchronous esophageal malignant neoplasms in HNSCC, should be considered to guide patient selection for screening endoscopy.
Limiting the use of RE for cancer staging will clearly result in fewer training opportunities for otolaryngology residents, as we have seen in recent years in our training program. Rigid esophagoscopy remains a useful skill when performing esophageal dilation and foreign-body retrieval. Although it has been found that esophageal foreign bodies can be managed safely using flexible TNE in select cases, TNE has not been recommended for this indication.50 Thus, RE remains a necessary component of resident training. Simulation may supplement training in the face of decreasing case numbers; work is currently under way at our institution to develop, refine, and validate an RE simulator.
Staging RE for cancer staging can be performed safely in patients with HNSCC, but the utility is low. Routine use of RE imparts unnecessary costs, and serious complications can occur. In North America, the incidence of synchronous esophageal malignant neoplasms has been decreasing during the past 3 decades, and routine esophagoscopy is no longer warranted, except in high-risk patient groups. On the basis of current literature, factors to consider may include esophageal symptoms, degree of tobacco and alcohol use, HNSCC subsite, and patient ethnicity, but further study is required to clearly delineate these parameters. More selective use of RE will result in a reduction in training opportunities that may be addressable with simulation to preserve this necessary skill set.
Submitted for Publication: June 25, 2015; final revision received September 4, 2015; accepted October 12, 2015.
Corresponding Author: Mark J. Jameson, MD, PhD, Department of Otolaryngology–Head and Neck Surgery, University of Virginia Health System, PO Box 800713, Charlottesville, VA 22908-0713 (email@example.com).
Published Online: December 3, 2015. doi:10.1001/jamaoto.2015.2815.
Author Contributions: Dr Jameson had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: McGarey, Shonka, Jameson.
Acquisition, analysis, or interpretation of data: McGarey, O’Rourke, Owen, Reibel, Levine, Jameson.
Drafting of the manuscript: O’Rourke, Owen.
Critical revision of the manuscript for important intellectual content: McGarey, Shonka, Reibel, Levine, Jameson.
Statistical analysis: McGarey, O’Rourke, Jameson.
Administrative, technical, or material support: O’Rourke, Owen, Jameson.
Study supervision: Shonka, Levine, Jameson.
Conflict of Interest Disclosures: None reported.
Disclaimer: Dr Levine is the editor of JAMA Otolaryngology—Head & Neck Surgery but was not involved in the editorial review or the decision to accept the manuscript for publication.
Previous Presentation: This study was presented at the Virginia Society of Otolaryngology Annual Meeting; May 30, 2015; Wintergreen, Virginia.