There was a statistically significant association of laryngeal preservation with higher levels of speech intelligibility (A) and with the retained ability to phonate (C). There was no statistically significant association of the performance of laryngectomy with feeding tube dependence (B) or with levels of oral intake (D).
eTable. Literature Review of Functional Outcomes After Total Glossectomy With and Without Total Laryngectomy
Lin DT, Yarlagadda BB, Sethi RKV, Feng AL, Shnayder Y, Ledgerwood LG, Diaz JA, Sinha P, Hanasono MM, Yu P, Skoracki RJ, Lian TS, Patel UA, Leibowitz J, Purdy N, Starmer H, Richmon JD. Long-term Functional Outcomes of Total Glossectomy With or Without Total Laryngectomy. JAMA Otolaryngol Head Neck Surg. 2015;141(9):797-803. doi:10.1001/jamaoto.2015.1463
The optimal reconstruction of total glossectomy defects with or without total laryngectomy is controversial. Various pedicled and free tissue flaps have been advocated, but long-term data on functional outcomes are not available to date.
To compare various total glossectomy defect reconstructive techniques used by multiple institutions and to identify factors that may lead to improved long-term speech and swallowing function.
Design, Setting, and Participants
A multi-institutional, retrospective review of electronic medical records of patients undergoing total glossectomy at 8 participating institutions between June 1, 2001, and June 30, 2011, who had a minimal survival of 2 years.
Total glossectomy with or without total laryngectomy.
Main Outcomes and Measures
Demographic and surgical factors were compiled and correlated with speech and swallowing outcomes.
At the time of the last follow-up, 45% (25 of 55) of patients did not have a gastrostomy tube, and 76% (42 of 55) retained the ability to verbally communicate. Overall, 75% (41 of 55) of patients were tolerating at least minimal nutritional oral intake. Feeding tube dependence was not associated with laryngeal preservation or the reconstructive techniques used, including flap suspension, flap innervation, or type of flap used. Laryngeal preservation was associated with favorable speech outcomes, such as the retained ability to verbally communicate in 97% of those not undergoing total laryngectomy (35 of 36 patients) vs 44% (7 of 16) in those undergoing total laryngectomy (P < .001), as well as those not undergoing total laryngectomy achieving some or all intelligible speech in 85% (29 of 34 patients) compared with 31% (4 of 13) undergoing total laryngectomy achieving the same intelligibility (P < .001).
Conclusions and Relevance
In patients with total glossectomy, feeding tube dependence was not associated with laryngeal preservation or the reconstructive technique, including flap innervation and type of flap used. Laryngeal preservation was associated with favorable speech outcomes such as the retained ability to verbally communicate and higher levels of speech intelligibility.
Total glossectomy (TG) remains a surgical option for patients with advanced or recurrent tongue cancer. While most commonly performed for oncological reasons, it also has a palliative role in patients with intractable pain, bleeding, or airway compromise. Despite radical surgery, more than half of the patients will die within 1 to 2 years of progressive or distant disease.1,2
Reconstruction of TG defects remains challenging in a patient population that is often older, has undergone prior radiation therapy, and carries significant comorbidities. Short-term goals of reconstruction include safe separation of the neck from the oral cavity and oropharynx, prevention of aspiration, and creation of a healthy wound for timely administration of adjuvant therapy. Long-term goals include optimizing oral intake, obtaining intelligible speech, and maintaining the larynx if total laryngectomy (TL) has not been performed. While attaining the short-term goals is expected, successful achievement of the long-term goals is often precluded by progressive flap atrophy, fibrosis, neuromuscular alterations, and loss of dentition that continues to evolve over a period of years.
The long-term functional outcomes for patients undergoing TG have traditionally been thought to be poor. Surgeons have debated the optimal reconstructive techniques for managing this challenging defect. Functional outcomes remain difficult to measure because of the poor overall prognosis of these patients. Reconstructive options include regional myocutaneous flaps and free flaps, with no clear method leading to improved functional results, especially because large single-institution experience is lacking.3- 9 Furthermore, various technical modifications have been described, including innervated flaps, suspension of the larynx, epiglottoplasty, and the use of oral prostheses to enhance function.3,6,7,10,11
In this study, we report the results of our multi-institutional, retrospective analysis evaluating the long-term functional outcomes of patients undergoing TG for advanced disease. We sought to determine which interventions predict improved speech and swallowing function.
A multi-institutional, retrospective review supported by the American Head and Neck Society Reconstructive Committee was performed of patients undergoing TG with or without TL between June 1, 2001, and June 30, 2011. The 8 participating institutions included the Massachusetts Eye and Ear Infirmary (Boston), The Johns Hopkins Hospital (Baltimore, Maryland), The University of Texas MD Anderson Cancer Center (Houston), Louisiana State University Health Sciences Center (Shreveport), Kansas University Medical Center (Kansas City), Cook County Hospital (Chicago, Illinois), University of Miami Miller School of Medicine (Miami, Florida), and Washington University School of Medicine (St Louis, Missouri). Institutional review board approval was obtained at each institution. Given the retrospective nature of the study and the presence of no more than minimal risk to patients, a waiver of informed consent was granted by each institution. For this study, TG is defined as removal of the entire tongue, with sacrifice of both hypoglossal nerves. Other study inclusion criteria were the performance of regional or free flap reconstruction and a minimal survival of 2 years, with documented follow-up. Patients who had a TL performed for oncological reasons at the time of TG were excluded from the study.
The following independent variables were collected from each institution’s electronic medical record and compiled in a master database: patient sex, marital status, primary vs salvage surgery, mandible split, type of flap reconstruction, flap innervation, flap suspension, laryngeal suspension when applicable, the performance of TL for prevention of aspiration, and the performance of later surgery to improve function. Patients were noted to have subsequent functional surgery if procedures were performed strictly for improvement in speech and swallowing. Subsequent oncological procedures were not considered for this parameter. Other collected data that were not included in univariate analysis were histopathological diagnosis, the performance of epiglottoplasty, and the use of an oral appliance after surgery. Regarding functional outcomes, the following dependent variables were recorded: gastrostomy tube dependence, the ability to verbally communicate, approximate percentage of intelligible speech as determined by postoperative speech language pathologist’s evaluations, oral diet status and amount taken by mouth as determined by surgeon report or speech language pathologist’s report, and primary method of communication.
Demographic descriptors were summarized by the mean (SD) for continuous variables and by the percentage of total patients for categorical variables. Functional outcomes were separated into groups with respect to the main dependent variables of gastrostomy tube dependence and the ability to verbally communicate. Association of a functional variable with gastrostomy tube dependence and phonation was assessed by χ2 test or Fisher exact test (as appropriate) for categorical variables and by t test for continuous variables.
In total, 55 patients (36 men and 19 women) were included in the study, with a mean age of 56.7 (10.9) years (age range, 23-78 years). Other patient demographics and independent variables related to the surgical procedure are summarized in Table 1. Most patients (84% [46 of 55]) were treated for squamous cell carcinoma. There was an almost equal split between primary TG and salvage TG (55% [30 of 55] and 45% [25 of 55] of patients, respectively). At the time of TG, 31% (17 of 55) of patients underwent TL. A greater percentage of patients underwent concurrent TL in the salvage surgical setting (43% [13 of 30]) compared with the primary surgical setting (16% [4 of 25]) (P < .05, 2-tailed Fisher exact test). The mean length of follow-up was 52.2 (28.1) months (range, 24-141 months).
Characteristics of patients undergoing TG are summarized in Table 1. The most frequently used reconstruction was the anterolateral thigh flap, used in 45% (25 of 55). Flap innervation was performed in 29% (16 of 55) of patients and consisted of innervation with motor, sensory, or both nerve types. No patients underwent epiglottoplasty. Subsequent surgery for improvement in speech and swallowing function was performed in 13% (7 of 55) of patients, including one TL because of recurrent aspiration pneumonia. Two patients underwent partial supraglottic laryngectomy at the time of TG. These patients were categorized under laryngeal preservation because they lack aerodigestive separation. At the time of the most recent follow-up, 24% (13 of 55) of patients reported the use of palatal drop prostheses. Baseline characteristics comparing patients who did or did not undergo concomitant TL are summarized in Table 2.
Long-term primary outcomes demonstrated that 45% (25 of 55) of patients did not have a gastrostomy tube at the time of the last follow-up (Table 3). While 22% (12 of 55) of patients were unable to tolerate any oral intake, 75% (41 of 55) of patients were tolerating at least minimal nutritional oral intake. Data were not available for the remaining 4% (2 of 55) of patients. Gastrostomy tube dependence was not associated with the performance of TL vs laryngeal preservation or with the reconstructive techniques, including flap innervation or type of flap used (Table 4).
The majority of patients (76.4%) retained the ability to verbally communicate which, as expected, was associated with laryngeal preservation (P < .001), as was higher levels of speech intelligibility (P < .001) (Table 4 and Figure). Of those not undergoing TL, 97% of patients (35 of 36) were found to retain the ability to phonate compared with 44% (7 of 16) of those who underwent TL (P < .001). In regard to the quality of speech, of those undergoing TL, 31% (4 of 13 patients with available data) were found to have speech intelligibility of some or all of expressed speech per physician or speech-language pathologist evaluation compared with 85% (29 of 34) of those who underwent TG without TL (P < .001).
Total glossectomy with or without laryngectomy is associated with significant functional morbidity and is often reserved for very advanced primary cancers or in the salvage surgical setting after chemoradiation failure. With 5-year survival reported between 25% and 35%, long-term functional outcomes after TG have been variable and not previously well described.1,2 Single-institution experience is generally limited regarding such outcomes. The objective of our study was to compile multi-institutional data to evaluate the long-term functional outcomes of the subgroup of patients undergoing TG who survived a minimum of 2 years.
In our multi-institutional study, overall long-term functional outcomes were promising, with 45% (25 of 55) of patients not requiring a gastrostomy tube and 76% (42 of 55) of patients retaining the ability to verbally communicate. These outcomes compare favorably with other studies having variable long-term end points (eTable in the Supplement). For example, in their studies of TG with laryngeal preservation, Dziegielewski et al12 reported 50% gastrostomy tube dependence, and Rihani et al13 reported 71% gastrostomy tube dependence.
One goal of TG reconstruction is to create a large, bulky neotongue to assist with speech articulation and oral bolus propulsion into the pharynx. The relationship between tongue bulk and function has been substantiated.14,15 Kimata and colleagues15 noted that patients with “protuberant” reconstructed tongues experienced better speech and swallowing function than those with flat or depressed neotongues. Although the literature does not clearly define which flap choice is optimal for TG reconstruction, multiple research groups agree that the donor site should be based on a patient’s body habitus and the ability to provide bulky tissue for favorable postoperative anatomy.12- 15
To this end, numerous reconstructive techniques have been described to maintain bulk or optimize function, including the use of innervated flaps, secondary onlay flaps, laryngeal suspension, and epiglottoplasty.6,7,10,11,16 However, most studies had a small study cohort or a short follow-up period, precluding definitive conclusions regarding long-term outcomes. In our study, we found that gastrostomy tube dependence was not associated with the performance of TL or reconstructive techniques, including laryngeal suspension, flap innervation, or choice of flap. One potential explanation for these findings is that, over longer follow-up periods in the minority of patients who survive, the technical modifications listed at the beginning of this paragraph no longer provide a functional advantage in the face of persistent gravity and fibrosis. This rationale may be particularly noteworthy given that our subgroup of patients undergoing TG without TL represents the largest cohort reported to date with a minimal follow-up period of 2 years (eTable in the Supplement). It is also possible that reconstructive techniques may provide differing results on more specific swallowing-related outcomes such as diet intake and oral and pharyngeal clearance. However, this data set did not allow for analysis of more specific swallowing parameters and outcomes.
When comparing the most frequent reconstructive method (the anterolateral thigh flap) with all other flap types, no significant difference in functional outcomes was noted. However, when comparing the use of the pectoralis flap vs all free tissue transfer, the pectoralis flap was associated with poorer outcomes in speech intelligibility (with 55% [6 of 11] achieving minimal or no intelligibility vs 22% [8 of 37] in those undergoing free flap transfer) and the ability to phonate (with 50% [6 of 12] achieving no phonation vs 10% [4 of 40] in those undergoing free flap transfer) (P < .05 for the ability to phonate only, Fisher exact test). These results can be explained by the finding that pectoralis flaps were used more often when concurrent TL was performed (47% [8 of 17] vs 13% [5 of 38] of laryngeal preservation cases; P < .05, Fisher exact test). No difference was noted in swallowing outcomes (feeding tube dependence or oral intake) when comparing pectoralis flaps with free tissue reconstructions.
Concurrent TL to prevent chronic aspiration and pneumonia has long been a contentious topic in the treatment of patients undergoing TG. Our study supports previous reports that TG without TL can be a safe practice, with a high rate of oral diet maintenance. These results compare favorably with those reported by Dziegielewski et al12 and Sinclair et al.17 However, our study may have had a potential selection bias by inclusion of only those patients who were still alive at 2 years because some patients receiving TG with laryngeal preservation may have died secondary to complications of a retained larynx before this follow-up period. Furthermore, the degree of swallowing function that was present before surgery is unknown. Additional selection bias may exist in that only those with optimal baseline function were chosen for TG without laryngectomy.
The results of this investigation are especially pertinent given the increasing incidence of human papillomavirus (HPV)–associated oropharyngeal cancer. Despite an overall favorable prognosis, up to 18% of patients with HPV-positive head and neck carcinoma have locoregional recurrence within 3 years of treatment.18,19 Patients with HPV-associated recurrence have improved prognosis compared with their HPV-negative counterparts and have demonstrated higher survival rates after salvage surgery.20,21 Therefore, functional outcomes after TG in this patient population that is healthier and has a longer life expectancy than HPV-negative patients is an important issue.22 Unfortunately, we were unable to stratify our results by HPV status in the present study.
There are several limitations of our study. First is its retrospective nature, as well as the fact that the patients were from multiple institutions and were operated on by a cohort of surgeons with their own inherent biases. This variability may have led to the stratification of more functional patients to TG vs TG and TL options because the ideal candidates for TG with laryngeal preservation are patients without prior radiation therapy with intact laryngeal function who can successfully undergo reconstruction with free tissue transfer. We noted that significantly more patients underwent TG and TL in the salvage surgery setting than in the primary surgical setting. There may have been preexisting laryngeal dysfunction from prior chemoradiation in the salvage surgery group that led to the patient and surgeon decision to perform TG and TL in an effort to prevent postoperative aspiration. Our data do not indicate that the performance of TL in the primary or salvage surgical setting correlated with superior functional outcomes, but the results are limited by the small sample size. We found no other baseline characteristic differences between those undergoing TL vs those who had laryngeal preservation. Because the baseline status of swallowing function before surgery was not available, we acknowledge that the favorable outcomes of TG with laryngeal preservation observed herein may not be generalized to all patients considering TG.
Second, as previously mentioned, the 2-year minimal survival criterion in our study may represent a possible limitation as well. Because we did not examine cause of death before 2 years, we may have excluded patients with death from aspiration secondary to laryngeal preservation, thus weakening our conclusion that laryngeal preservation for function is a safe endeavor.
Third, given the retrospective nature of this investigation, quantitative speech and swallowing metrics before and after surgery were not available. Ideally, future studies should include quantitative metrics achieved through videofluoroscopic swallowing assessment, application of consistent diet scale ratings, speech and swallowing–related quality-of-life scores, and speech assessment by a qualified speech language pathologist. Inclusion of such a global patient assessment would determine if reconstructive techniques offer subtle benefits regarding functional outcomes.
Total glossectomy defects remain a challenge for reconstructive surgeons. Balancing quality of life with cure is difficult in such situations. Our multi-institutional study shows that long-term speech and swallowing function is a reasonable outcome in these patients. Success was not dependent on the performance of TL or reconstructive techniques, including laryngeal suspension, flap innervation, or choice of flap. Furthermore, laryngeal preservation can be successfully accomplished when TG is performed, with expected outcomes of the ability to verbally communicate and approximately a 50% likelihood of maintaining an oral diet.
Submitted for Publication: November 19, 2014; final revision received May 31, 2015; accepted June 8, 2015.
Corresponding Author: Jeremy D. Richmon, MD, Department of Otolaryngology–Head and Neck Surgery, The Johns Hopkins University, 601 N Caroline St, Sixth Floor, Baltimore, MD 21287-0910 (firstname.lastname@example.org).
Published Online: August 20, 2015. doi:10.1001/jamaoto.2015.1463.
Author Contributions: Drs Yarlagadda and Richmon had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Richmon.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Lin, Yarlagadda, Richmon.
Critical revision of the manuscript for important intellectual content: Lin, Yarlagadda, Richmon.
Statistical analysis: Yarlagadda.
Administrative, technical, or material support: Richmon.
Study supervision: Richmon.
Conflict of Interest Disclosures: None reported.
Previous Presentation: These data were presented at Fifth World Congress of the International Federation of Head and Neck Oncologic Societies and 2014 Annual Meeting of the American Head and Neck Society; July 18, 2014; New York, New York.