Dysphagia After Primary Transoral Robotic Surgery With Neck Dissection vs Nonsurgical Therapy in Patients With Low- to Intermediate-Risk Oropharyngeal Cancer

Key Points

Questions  What is the prevalence of acute dysphagia and the rate of dysphagia recovery after transoral robotic surgery for oropharyngeal cancer, and how do outcomes after transoral robotic surgery compare with those after primary radiotherapy?

Findings  In this cases series study of 257 patients with human papillomavirus–associated oropharyngeal cancer, the prevalence of moderate-severe acute dysphagia after transoral robotic surgery (per Dynamic Imaging Grade of Swallowing Toxicity grade ≥2) was 22.7%, with partial recovery that did not return to baseline by 3 to 6 months. Dysphagia profiles between patients with primary transoral robotic surgery and the groups that received primary radiotherapy differed in the short term but not by 3 to 6 months of subacute recovery.

Meaning  Subacute swallowing outcomes were similar regardless of primary treatment modality among patients with low- to intermediate-risk oropharyngeal squamous cell carcinoma.

Importance  A major goal of primary transoral robotic surgery (TORS) for oropharyngeal cancer is to optimize swallowing outcomes by personalized treatment based on pathologic staging. However, swallowing outcomes after TORS are uncertain, as are the outcomes compared with nonsurgical options.

Objectives  To estimate rates of acute dysphagia and recovery after TORS and to compare swallowing outcomes by primary treatment modality (TORS or radiotherapy).

Design, Setting, and Participants  This case series study was a secondary analysis of prospective registry data from 257 patients enrolled from March 1, 2015, to February 28, 2018, at a single academic institution who, according to the AJCC Staging Manual, 7th edition TNM classification, had low- to intermediate-risk human papillomavirus–related oropharyngeal squamous cell carcinoma possibly resectable by TORS.

Exposure  Patients were stratified by primary treatment (75 underwent TORS and 182 received radiotherapy).

Main Outcomes and Measures  Modified barium swallow (MBS) studies graded per Dynamic Imaging Grade of Swallowing Toxicity (DIGEST) and the MD Anderson Symptom Inventory–Head and Neck Module (MDASI-HN) questionnaires were administered at standard intervals. Prevalence and severity of dysphagia were estimated per DIGEST before and after TORS and 3 to 6 months after treatment. Moderate-severe dysphagia (DIGEST grade ≥2) was assessed using logistic regression and compared by primary treatment group. The MDASI swallowing symptom severity item scores during and after radiotherapy were compared using generalized estimating equations by treatment status at the start of radiotherapy, after induction, and after TORS.

Results  A total of 257 patients (mean [SD] age, 59.54 [9.07] years; 222 [86.4%] male) were included in the study. Dysphagia severity (per DIGEST) was significantly worse after TORS (r = −0.63; 95% CI, −0.78 to −0.44): 17 patients (22.7%; 95% CI, 13.8%-33.8%) had moderate-severe (DIGEST grade ≥2) acute post-TORS dysphagia significantly associated with primary tumor volume (odds ratio, 1.43; 95% CI, 1.11-1.84). DIGEST improved by 3 to 6 months but remained worse than that at baseline; at 3 to 6 months, the number of patients with DIGEST grade 2 or higher dysphagia was 5 (6.7%; 95% CI, 2.2%-14.9%) after primary TORS and 29 (15.9%; 95% CI, 10.9%-22.1%) after radiotherapy. At the start of radiotherapy, MDASI swallowing symptom severity item scores were significantly worse in the post-TORS group compared with postinduction (mean [SD] change, 2.6 [1.1]) and treatment-naive (mean [SD] change, 1.7 [0.3]) patients. This result inverted at radiotherapy end, and all groups converged at 3 to 6 months.

Conclusions and Relevance  Subacute swallowing outcomes were similar regardless of primary treatment modality among patients with low- to intermediate-risk oropharyngeal squamous cell carcinoma.

Transoral robotic surgery (TORS) is increasingly used as the frontline treatment for patients with low- to intermediate-risk oropharyngeal squamous cell carcinoma (OPSCC) who typically present with small-volume primary tumors and near-normal baseline function.^1,2 A major goal of a primary TORS strategy is to maintain high benchmarks for disease control while optimizing swallowing outcomes by personalized treatment based on pathologic rather than clinical staging. Dysphagia is highly prevalent in patients who survive OPSCC³ and is associated with adverse outcomes, including lower quality of life,⁴ decisional regret,⁵ aspiration pneumonia,⁶ and non–cancer-related death.⁷ A systematic review⁸ found that many studies report functional outcomes of TORS, typically in retrospective series. The most common functional outcome reported is feeding tube use, with highly variable rates of perioperative use (3%-100%), 18% to 39% gastrostomy tube insertion rates (typically during adjuvant therapy), and 0% to 7% long-term dependence.

Other measures of dysphagia commonly reported after TORS include patient-reported outcomes, such as the MD Anderson Dysphagia Inventory (MDADI)⁹ or Eating Assessment Tool 10 (EAT-10),¹⁰ and measures of oral intake or diet. Recent prospective patient-reported outcome data suggest that nearly all patients self-report dysphagia (per EAT-10 score >2) throughout the first month after TORS, but 92% started oral intake by discharge and 98% by 1 month.¹¹ Nonetheless, 89% had dysphagia symptoms above normal thresholds at 30 days, and many required compensatory strategies to resume oral intake.¹¹ The source of perceived dysphagia persisting 1 month postoperatively was not fully elucidated because uniform imaging tests (eg, fluoroscopy and endoscopy) were not performed alongside patient-reported outcomes, and the elevated EAT-10 responses appeared to correlate with pain.¹¹ Furthermore, painful swallowing had the highest EAT-10 item score. Thus, it remains unclear whether pharyngeal dysfunction or pain is the source of dysphagia symptoms in the early post-TORS period.

More objective data from instrumental swallowing evaluations are rarely reported after primary TORS.⁸ A single prospective report of post-TORS videofluoroscopy (or modified barium swallow [MBS] studies) found dysphagia in only 2 of 10 patients 1 month after TORS, in both of whom dysphagia was mild according to the Dynamic Imaging Grade of Swallowing Toxicity (DIGEST). Only inefficient bolus clearance was reported in both cases: one attributed to impaired tongue base retraction and the other to unilateral pharyngeal weakness.² No aspiration was detected. Although promising, many questions remain, such as whether these findings generalize; how function evolves through adjuvant therapy, which is often required; and how longer-term MBS outcomes compare with non–surgical therapy paradigms.

Adjuvant therapy is common after TORS and required in more than half of cases in most series.⁸ Radiotherapy and chemoradiotherapy cause short-term and long-term changes to salivary and neuromuscular function, with well-documented associations with oropharyngeal dysphagia.^12-15 Because most reports aggregate swallowing outcomes of all cases after TORS (regardless of adjuvant therapy), little is known about the cumulative effects of surgery followed by radiotherapy to the oropharynx. Likewise, it is unclear how surgery changes swallowing symptoms during adjuvant radiotherapy. That is, if most patients with low- to intermediate-risk OPSCC start radiotherapy when relatively dysphagia symptom free, how do patients who have had prior TORS compare? Thus, another gap in knowledge is how postoperative dysphagia symptoms that linger at the start of adjuvant radiotherapy evolve and compare with symptoms in treatment-naive counterparts.

The primary focus of the present study was to characterize the evolution of pharyngeal dysphagia per MBS studies before and after TORS (before adjuvant therapy) and the subacute recovery compared with primary radiotherapy using prospective registry data. Thus, longitudinal MBS DIGEST grades were evaluated for the following aims: (1) to estimate rates of acute postoperative dysphagia after TORS and recovery at 3 to 6 months after treatment; (2) to compare dysphagia grades at 3 to 6 months in patients treated with primary TORS vs primary radiotherapy; and (3) to compare swallow symptom trajectories associated with longitudinal patient-reported outcomes during and after radiotherapy.

This case series study was a secondary analysis of prospective registry data from the MD Anderson Oropharynx Cancer Registry Patient-Reported Outcomes and Functional Core. The registry has enrolled all consecutive consenting patients with oropharyngeal or unknown primary malignant tumor of the head and neck at the University of Texas MD Anderson Cancer Center (MDACC) since March 2015. Written informed consent was provided for prospective data collection. Deidentified data were subsequently analyzed under approval of the MDACC Institutional Review Board with a consent waiver.

The study sample comprised those enrolled in the registry from March 1, 2015, to February 28, 2018, who according to the AJCC Staging Manual, 7th edition TNM classification had low- to intermediate-risk human papillomavirus–related OPSCC¹⁶ possibly resectable by TORS. Eligibility criteria were as follows: (1) pathologically confirmed T1-2N0-2b or T3N0 OPSCC, (2) TORS or radiotherapy as primary treatment modality at the MDACC, and (3) met the 3- to 6-month posttreatment time point by September 2018. Of the 533 patients enrolled, 257 (75 undergoing primary TORS and 182 receiving primary radiotherapy) were eligible and included in this analysis (Figure 1). Primary treatment was determined by multidisciplinary tumor board presentation of each case with review by medical oncology, head and neck surgery, radiation oncology, radiology, pathology, and speech language pathology specialists (including K.A.H., C.M.K.L.Y., B.E.E., R.G., A.C.H., M.E.K., S.Y.L., C.D.F., G.B.G., A.S.G., F.J., R.F., J.S.L., and N.D.G.). In general, patients were considered for TORS if they had low-volume disease at the primary site and neck (T0-2N0-1, AJCC Cancer Staging Manual, 8th edition) and were deemed to be surgically fit.^17-20

Dysphagia was graded from MBS studies according to DIGEST criteria.²¹ The MBS study was scheduled routinely before treatment, postoperatively, and 3 to 6 months after the end of locoregional treatment regardless of symptoms. The target for the postoperative MBS study was 4 weeks to establish new baseline swallow function before adjuvant therapy and/or need for more intensive dysphagia intervention after initial postsurgical recovery. The MBS studies followed a previously reported standard protocol.²²

DIGEST is a validated method to grade the severity of pharyngeal-phase dysphagia from videofluoroscopy based on the safety (using a penetration-aspiration scale²³) and efficiency (using estimation of the percentage of pharyngeal residue) of bolus clearance. DIGEST grades align to the National Cancer Institute’s Common Terminology Criteria for Adverse Events framework (grade 1, mild; 2, moderate; 3, severe; and 4, life threatening).²² DIGEST was derived by independent review of MBS DICOM files by a trained, blinded speech pathologist who met reliability standards. DIGEST was dichotomized with grade 2 or higher as moderate-severe dysphagia based on published data, suggesting that this is a meaningful split associated with quality of life, diet, and swallow muscle radiotherapy dose.²⁰ DIGEST grade 2 or higher dysphagia indicates at least intermittent high-grade penetration or aspiration (penetration-aspiration scale score ≥5) and/or postswallow residue of a 50% or greater bolus.

Patients completed the MDADI and the MD Anderson Symptom Inventory–Head and Neck Module (MDASI-HN) at the MBS time points and MDASI-HN also during radiotherapy. The MDADI is a 20-item measure of swallowing-related quality of life with a range of scores from 20 (worst) to 100 (best).⁹ The MDASI-HN is a validated multisymptom inventory that includes a swallowing-related item (“your difficulty swallowing/chewing”) rated from 0 (not present) to 10 (as bad as you can imagine).²⁴ Patient-reported outcome assessments allow frequent administration to ascertain evolution of symptoms in periods during which status changes frequently (and frequent MBS studies are not feasible to follow this change), such as the radiotherapy interval. Thus, responses to weekly MDASI-HNs during radiotherapy assessed differences in symptom trajectories at the start and end of radiotherapy when MBS studies were not routine.

Clinicopathologic data queried from the Oropharynx Cancer Registry are detailed in Table 1. Critical clinical variables, including staging, tumor site, and treatment, were double verified. Primary tumor volume was delineated on the pretreatment diagnostic contrast-enhanced computed tomography scan in a commercially available radiation oncology contouring software (VelocityAI software, version 3.0.1; Varian Medical Solutions). Nonvisible primary tumors on computed tomography were excluded from this portion of the analysis.

To estimate prevalence rates and severity of acute postoperative dysphagia after TORS and at 3 to 6 months after treatment, DIGEST grades before and after TORS and at 3 to 6 months after treatment in the primary TORS group were compared pairwise using a nonparametric signed rank test, with the effect size r estimated with bootstrap 95% CIs and interpreted as follows: r = 0.10, small; r = 0.30, medium; and r = 0.50, large.²⁵ Univariable and multivariable logistic regression models were used to assess moderate-severe acute post-TORS dysphagia (DIGEST grade ≥2) by pretreatment factors. To compare dysphagia grades at 3 to 6 months in patients treated with primary TORS vs primary radiotherapy, prevalence of moderate-severe (DIGEST grade ≥2) dysphagia at 3 to 6 months was estimated with exact (Clopper-Pearson) 95% CIs and compared between the primary TORS and radiotherapy groups using the Fisher exact test, with Cramer V estimated for magnitude of difference with bootstrap bias-corrected 95% CIs among all 257 patients.

To compare symptom trajectories during and after radiotherapy, longitudinal analysis of the MDASI swallowing symptom severity item across 3 time points (start of radiotherapy, end of radiotherapy, and at 3 to 6 months) used generalized estimating equations among 219 patients treated with radiotherapy. Patients were grouped based on their treatment status at the start of radiotherapy as follows: (1) treatment naive, (2) after TORS, and (3) after induction. The MDASI scores skewed right, with a high frequency of zeros. Poisson distribution with log link function and unstructured correlation matrix (according to quasi-information criteria) corrected for within-patient correlations. Backward selection (α < .05 retention criteria) was used for model building by retaining the baseline MDASI scores. Statistically significant group × time interactions were explored comparing least-squared means.

For all prevalence estimates, we provided 95% exact binomial CIs and included missing MBS studies in the denominator to avoid overestimating dysphagia. We reported effect size estimates and 95% CIs rather than P values, in keeping with journal standards to improve transparency and quality of reporting.²⁶ The magnitude of effect sizes was interpreted using Cohen benchmarks.²⁵ All analyses were conducted using SAS statistical software for Windows, version 9.4, TS Level 1M5 (SAS Institute Inc) and RStudio, version 1.2.1335 (RStudio Inc). Additional details of statistical methods are summarized in eTable 1 in the Supplement.

A total of 257 patients (mean [SD] age, 59.54 [9.07] years; 222 [86.4%] male) were included in this study (75 undergoing primary TORS and 182 receiving primary radiotherapy) as detailed in Table 1. The primary TORS group had significantly smaller primary tumor volume (median, 5.15 vs 6.95 cm³), lower N classification (25 of 75 [33.3%] vs 128 of 182 [70.3%] with N2b; difference in proportions, −37.0%; 95% CI, −50.5% to 23.5%), lower radiotherapy dose (median, 60 vs 70 Gy), lower likelihood of bilateral radiotherapy (20 of 75 [26.7%] vs 145 of 182 [79.7%]; difference in proportions, −53.0%; 95% CI, −65.5% to −40.5%), higher likelihood of neck dissection (75 of 75 [100%] vs 18 of 182 [9.9%]; difference in proportions, 90.1%; 95% CI, 84.8%-95.4%), and worse baseline dysphagia (3 of 75 [4.0%] vs 1 of 182 [0.6%] with DIGEST grade 2; difference in proportions, 3.5%; 95% CI, −2.1% to 9.0%) compared with the primary radiotherapy group. Baseline differences between groups for primary tumor volume and DIGEST grade were small, whereas the difference in N classification was large. Differences in radiotherapy dose, bilateral radiotherapy fields, and neck dissection were large. Baseline MBS studies found no or mild dysphagia in 244 patients (98.4%) (DIGEST grade 0 or 1 in 244 of 246 with baseline MBS outcomes). Baseline MDADI scores were optimal (composite score ≥80) in 223 of 235 patients (94.9%) with complete MDADI data.

Among 75 patients in the primary TORS group, 49 returned for postoperative MBS study a median of 3.1 weeks (interquartile range, 1.86 to 4.14 weeks) after surgery. The DIGEST grade significantly worsened after TORS (Wilcoxon r = −0.63; 95% CI, −0.78 to −0.44), indicating a large effect size for the association of surgery with dysphagia severity. After TORS, 11 of 75 patients (14.7%; 95% CI, 7.56%-24.73%) had moderate (DIGEST grade 2) and 6 of 75 patients (8.0%; 95% CI, 2.99%-16.60%) had severe (DIGEST grade 3) acute post-TORS dysphagia before adjuvant treatment (Table 2 and Figure 2A). In 14 of 75 patients (18.7%; 95% CI, 10.6%-29.3%), a DIGEST grade of 0 or 1 before TORS increased to a DIGEST grade of 2 (9 of 75 [12.0%]; 95% CI, 5.6%-21.6%) or 3 (5 of 75 [6.7%]; 95% CI, 2.2%-14.9%) after TORS.

In univariable regression, moderate-severe dysphagia (prevalence, 22.7%; 95% CI, 13.8%-33.8%) with a post-TORS DIGEST grade of 2 or higher was significantly associated with T stage and baseline primary tumor volume (Table 3). Age, N classification, tumor site, and baseline DIGEST grade were not significantly associated with a DIGEST grade of 2 or higher, although a DIGEST grade of 2 or higher after TORS was more prevalent with tonsil primary tumors and increased as baseline DIGEST grade increased. Controlling for baseline DIGEST grade, primary tumor volume was significantly associated with post-TORS dysphagia (odds ratio, 1.43 per cubic centimeter increase; 95% CI, 1.11-1.84). Patients with a DIGEST grade of 2 or higher after TORS had a median tumor volume of 9.35 cm³ (range, 0.7-21.9 cm³) vs 2.25 cm³ (range, 0.3-11.9 cm³) among those with no or mild dysphagia.

After the acute post-TORS MBS study, among 38 complete cases with all 3 MBS time points, dysphagia grades (per DIGEST) improved by 3 to 6 months (Wilcoxon r = 0.28; 95% CI , −0.03 to 0.53) but remained worse than baseline (Wilcoxon r = −0.41; 95% CI, −0.63 to −0.13). Compared with baseline DIGEST grade, at 3 to 6 months, 17 of 38 patients (44.7%) had the same DIGEST grade and 5 of 38 (13.2%) had an improved DIGEST grade, whereas 16 of 38 (42.1%) had 1 or 2 DIGEST grades worse (most with worse grades representing a DIGEST grade of 0 at baseline converted to a DIGEST grade of 1 at 3-6 months, representing mild subacute dysphagia in patients who had normal baseline MBS study results).

Among all 75 patients in the primary TORS group, moderate-severe dysphagia (DIGEST grade ≥2) was observed in 0 treated with surgery alone compared with 3 of 22 (13.6%; 95% CI, 2.9%-34.9%) treated with TORS and radiotherapy and 2 of 15 (13.3%; 95% CI, 1.7%-40.5%) treated with TORS and chemoradiotherapy (Figure 2D). Among the 38 complete cases in the primary TORS group, 3- to 6-month DIGEST grades differed significantly by adjuvant therapy received, with 15 of 24 patients (62.5%; 95% CI, 40.6%-81.2%) who received adjuvant radiotherapy and chemoradiotherapy vs 1 of 14 (7.1%; 95% CI, 0.2%-33.9%) who received TORS alone having poorer function than baseline (difference in proportions, 55.4%; 95% CI, 26.1%-84.6%; V = 0.57; 95% CI, 0.29-0.75), indicating a large effect size of adjuvant therapy.

Among all 257 patients, 207 (80.5%) returned for MBS studies at 3 to 6 months. Five of the 75 patients (6.7%; 95% CI, 2.2%-14.9%) in the primary TORS group had dysphagia of DIGEST grades of 2 or higher compared with 29 of the 182 patients (15.9%; 95% CI, 10.9%-22.1%) in the primary radiotherapy group (difference in proportions, 9.3%; 95% CI, 0.6%-18.0%; V = 0.12; 95% CI, 0.02-0.20), indicating a small effect size of the primary treatment group (Table 2 and Figure 2C).

Considering subgroups by TNM classification (AJCC Cancer Staging Manual, 7th edition), there were no cases of dysphagia with a DIGEST grade of 2 or higher at 3 to 6 months among the 47 patients treated for T1N0-2a (AJCC Cancer Staging Manual, 7th edition) disease in either treatment group and similar proportions of DIGEST grade of 2 or higher among 153 of the total cohort of 257 (59.5%) treated for T1-2N2b disease. The greatest differences in dysphagia prevalence between groups was among those with T2N0-2a disease, with the prevalence of dysphagia among patients receiving primary TORS being lower. At the start of radiotherapy, MDASI swallowing symptom severity item scores were significantly worse in the post-TORS group compared with postinduction (mean [SD] change, 2.6 [1.1]) and treatment-naive (mean [SD] change, 1.7 [0.3]) patients (eTable 2 in the Supplement).

Among 219 patients who received radiotherapy, 165 (75.3%) were treatment naive, whereas 37 (16.9%) started radiotherapy after TORS and 17 (7.8%) after induction chemotherapy. Multivariable longitudinal generalized estimating equations controlled for baseline MDASI responses found significant group × time interaction for MDASI swallowing symptom severity item based on treatment status at radiotherapy; age and concurrent chemotherapy were also significant in the model (eTable 3 in the Supplement). At the start of radiotherapy, MDASI-HN swallow symptoms were significantly worse in the post-TORS group compared with the postinduction and treatment-naive groups. This trend inverted at the end of radiotherapy. Differences at the end of radiotherapy and after radiotherapy were not significant after controlling for concurrent chemotherapy. Swallowing symptoms in all groups converged by 3 to 6 months (Figure 2B).

Swallowing outcomes are a key factor in clinical decision-making for low- to intermediate-risk OPSCC.²⁷ An increasing body of literature suggests favorable swallowing outcomes after a primary TORS strategy largely based on low gastrostomy use and early return to oral intake.^8,11,17 Using prospectively acquired patient-reported outcomes and MBS studies, we found a 22.7% prevalence of moderate-severe acute post-TORS pharyngeal dysphagia (according to MBS DIGEST grade ≥2) that was significantly associated with primary tumor volume. Dysphagia symptoms (according to patient-reported outcome MDASI-HN swallowing symptom severity item) remained significantly elevated at the start of radiotherapy among patients undergoing TORS who required adjuvant therapy compared with their treatment-naive or postinduction counterparts. By 3 to 6 months, however, both DIGEST and patient-reported MDASI swallowing symptom severity item scores partially recovered and did not significantly differ from those in the primary radiotherapy group.

Moderate to severe dysphagia with a DIGEST grade of 2 or higher a median of 3 weeks after TORS reflects clinically significant pharyngeal dysfunction at least transiently in the acute postsurgical period, a scenario rarely reported previously. The observed prevalence of post-TORS dysphagia was higher than that reported by Lazarus et al² in the only published longitudinal TORS MBS data set to our knowledge. Differences in baseline functioning may, in part, explain these differences because all patients in the Mount Sinai Beth Israel series had normal baseline MBS study results (DIGEST grade 0) compared with only 77.8% in the cohort in our study. Primary tumor volume was also significantly associated with post-TORS dysphagia in our study but was not measured comparably in the Mount Sinai Beth Israel series to enable comparison. Other baseline factors, such as depth of invasion and tumor proximity to swallowing structures, may prove to be important variables associated with post-TORS dysphagia specifically but were not available in either cohort to consider.

Given that most findings suggest superior swallowing outcomes after primary TORS^2,17,28,29 with the exception of recently presented ORATOR (A Phase II Randomized Trial for Early-Stage Squamous Cell Carcinoma of the Oropharynx: Radiotherapy vs Transoral Robotic Surgery) trial data,³⁰ it was somewhat unexpected to find small, nonsignificant differences in 3- to 6-month swallowing outcomes by primary treatment group. The low number of high-grade dysphagia events at 3 to 6 months after primary TORS (n = 5) precluded multivariable adjustment for potential sources of selection bias. The primary radiotherapy group, for instance, had larger primary tumor volume and were twice as likely to have N2b disease. These features likely result in larger target volumes, causing intermediate- to high-dose radiotherapy fields to overlap more of the swallowing-critical muscles. The primary TORS group, however, had worse baseline dysphagia on MBS studies; baseline dysphagia was previously reported to be associated with double the odds of moderate-severe dysphagia after treatment.²⁰ Effect size estimates suggested small differences in primary tumor volume and baseline DIGEST grades between primary treatment groups, which may minimize the consequences of these imbalances. Nevertheless, imbalance in tumor burden inherent in TORS-eligible patients in addition to differences in baseline function is a critical limitation to consider when interpreting these results but also the larger body of published, nonrandomized comparisons of TORS and radiotherapy outcomes.

The observed 16% prevalence of postradiotherapy moderate-severe (DIGEST grade ≥2) dysphagia in the primary radiotherapy comparator group was better than previously reported in unselected OPSCC cohorts. A 31% prevalence of dysphagia with a DIGEST grade of 2 or higher at 3 to 6 months after intensity-modulated radiotherapy among all stages of OPSCC in patients from the same registry (ie, the sample included a larger proportion of patients with T3 and T4 or N2c and N3 disease) was recently reported.²⁰ This meaningful difference in dysphagia prevalence when sampling all stages vs low- to intermediate-risk OPSCC in these 2 reports highlights the need to consider the comparator group carefully when appraising outcomes of primary TORS and provides evidence against the straightforward comparison of primary surgery to primary chemoradiotherapy that exists in many publications. Outcomes of patients with low- to intermediate-risk disease treated with primary radiotherapy were better than those from historic data of patients with all stages of OPSCC. This finding has been demonstrated in both patient-reported outcome assessments that showed favorable symptom profiles after primary radiotherapy for early-stage tonsil disease³¹ and multisite estimates of gastrostomy use that showed low rates of long-term dependence (<3%) after oropharyngeal intensity-modulated radiotherapy for likely TORS-eligible disease per TNM classification³² and in the present work for MBS outcomes. Likewise, unilateral radiotherapy for early-stage tonsil cancer reportedly requires extremely low gastrostomy use (8% placement during radiotherapy).¹⁸

Details of surgical and nonsurgical therapy that influence swallowing outcomes are nuanced and many. Key factors to consider include laterality of radiotherapy fields, target volumes, radiotherapy technique, radiotherapy dose, neck dissection, and concurrent chemotherapy, many of which differed largely between primary treatment groups in this cohort. The primary TORS group had lower-volume radiotherapy and lower-dose radiotherapy but a higher likelihood of neck dissection, and 20.0% had trimodality therapy. Despite 51.7% avoiding radiotherapy or receiving more favorable radiotherapy, when required, we failed to detect significantly better swallowing outcomes after primary TORS by patient report or MBS study 3 to 6 months after treatment. Trade-offs between surgical pharyngeal injury and radiotherapy dose and volume reduction in the full cohort might have attenuated differences in certain subgroups, in which, for instance, surgery alone was highly probable and primary TORS offered a swallowing advantage by sparing radiotherapy injury altogether. Exploratory subgroup comparisons by TNM stage suggested similarly low dysphagia burden among patients with T1N0-2a disease and higher burden among patients with T1-2N2b disease regardless of primary treatment; the most likely differences in subacute dysphagia by primary TORS vs radiotherapy were in the small subgroup treated for T2N0-2a disease. These trends might be considered for hypothesis-generating purposes, but subgroup comparisons require numbers beyond those available currently in the registry and likely should also account for site of disease.

This prospective registry has matured only to the point to evaluate subacute outcomes in the window of 3 to 6 months after treatment. Overall, moderate-severe dysphagia (by MBS study or patient-reported outcome) was prevalent in 6.7% of patients after primary TORS and 15.9% of patients after primary radiotherapy at this time point, suggesting reasonable functional recovery for most patients with low- to intermediate-risk OPSCC regardless of treatment modality. This early functional recovery, however, does not guarantee long-term function. Previous reports^13,14,33 characterized late radiotherapy-associated dysphagia as a disabling late effect in this population, with global paresis of oropharyngeal swallow causing severe to profound dysphagia, recurrent pneumonia, and lifelong gastrostomy dependence. Late radiotherapy-associated dysphagia is commonly preceded by de novo lower cranial neuropathy in patients many years after cancer cure. A 4% to 5% risk of lower cranial neuropathy has been reported after intensity-modulated radiotherapy for OPSCC at a dose of 66 to 72 Gy^15,34 and primary surgery with adjuvant radiotherapy has been reported to be an independent risk factor that accelerates time to neuropathy.³⁵ Long-term follow-up of modern TORS and primary radiotherapy cohorts is required to examine late outcomes, a particularly critical outcome for human papillomavirus–associated OPSCC.

This work is the first, to our knowledge, to prospectively compare MBS study and patient-reported swallowing outcomes of primary TORS and radiotherapy jointly to describe the postsurgical effect and subacute recovery of swallowing function by modality. Strengths include use of validated, multidimensional swallowing outcomes and a relatively large, consecutively enrolled cohort. Missing data are a limitation. Missing acute post-TORS MBS studies in part reflect the evolution of the early registry time points and attrition after treatment that is common to clinical registries. Baseline DIGEST grades did not significantly differ among patients with missing posttreatment MBS data, and missingness at 3 to 6 months was similar by treatment groups and subgroups. For these analyses, we chose to focus on dysphagia severity as measured by the MBS DIGEST and MDASI-HN swallowing symptom severity item. Clinically, however, differences were noted in the dysphagia profiles between primary treatment groups. For instance, impaired velopharyngeal seal and unilateral pharyngeal weakness were often noted in dysphagia cases after TORS. Future work will be critical to detail the pathophysiologic findings of post-TORS dysphagia.

Although most patients had a functional swallow after primary TORS, 22.7% developed moderate-severe pharyngeal dysphagia in the acute postsurgical period that improved but did not recover to baseline by 3 to 6 months. Posttreatment dysphagia grades according to the MBS DIGEST did not significantly differ between patients treated with primary TORS and those treated with primary radiotherapy. Symptom trajectories suggest that the trade-off for potentially favorable late swallowing outcomes after primary TORS may be higher swallowing symptom burden during early weeks of postoperative radiotherapy. These data suggest important implications on supportive care and preoperative counseling.

Accepted for Publication: July 28, 2019.

Corresponding Author: Katherine A. Hutcheson, PhD, Department of Head and Neck Surgery, The University of Texas MD Anderson Cancer Center, PO Box 301402, Unit 1445, Houston, TX 77030 (karnold@mdanderson.org).

Published Online: September 26, 2019. doi:10.1001/jamaoto.2019.2725

Author Contributions: Dr Hutcheson and Ms Warneke had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hutcheson, Elgohari, Goepfert, Hessel, Kupferman, Fuller, Lewin.

Acquisition, analysis, or interpretation of data: Hutcheson, Warneke, Yao, Zaveri, Elgohari, Goepfert, Lai, Fuller, Gunn, Garden, Johnson, Ferrarotto, Gross.

Drafting of the manuscript: Hutcheson, Zaveri, Lewin.

Critical revision of the manuscript for important intellectual content: Hutcheson, Warneke, Yao, Elgohari, Goepfert, Hessel, Kupferman, Lai, Fuller, Gunn, Garden, Johnson, Ferrarotto, Gross.

Statistical analysis: Hutcheson, Warneke.

Obtained funding: Hutcheson.

Administrative, technical, or material support: Yao, Elgohari, Kupferman, Gunn, Ferrarotto, Gross.

Supervision: Hutcheson, Lai, Fuller, Gunn.

Conflict of Interest Disclosures: Dr Hutcheson reported receiving research grants and contracts from the National Institutes of Health and the Patient Centered Outcomes Research Institute. Dr Lai reported receiving personal fees from Cardinal Health outside the submitted work. Dr Fuller reported receiving grants and personal fees from the National Institutes of Health, grants and personal fees from Elekta AB, and personal fees from Greater Baltimore Medical Center, Tianjin Memorial Hospital, and the American Association of Physicists in Medicine outside the submitted work. Dr Johnson reported receiving grants from Trovagene and PIQUR outside the submitted work. Dr Ferrarotto reported receiving consulting fees from Ayala Pharmaceuticals, Regeneron Sanofi, and Klus. Dr Gross reported receiving personal fees from Intuitive Surgical and nonfinancial support from MedRobotics outside the submitted work. No other disclosures were reported.

Funding/Support: This work was directly supported by the Charles and Daneen Stiefel Oropharynx Fund.

Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Members: MD Anderson Head and Neck Cancer Symptom Working Group (The University of Texas MD Anderson Cancer Center, Houston): Richard Cardoso, DDS, MS, FACP; Mark S. Chambers, DMD, MS; Renata Ferrarotto, MD; Steven J Frank, MD; Clifton D. Fuller, MD, PhD; Adam S. Garden, MD; Ann Gillenwater, MD; Maura Gillison, MD, PhD; Neil D. Gross, MD; G. Brandon Gunn, MD; Amy C. Hessel, MD; Kate A. Hutcheson PhD; Faye Johnson, MD, PhD; Jason Johnson, MD; Jan S. Lewin, PhD; Carol M. Lewis, MD, MPH; Stephen Y. Lai, MD, PhD; Charles Lu, PhD; Tito R. Mendoza, PhD, MS, MEd; William H. Morrison, MD; Abdallah S. R. Mohamed, MD, MSc; Jack Phan, MD, PhD; David I. Rosenthal, MD; Erich M Sturgis, MD, PhD; Shirley Su, MD; Jihong Wang, PhD; and Randal S. Weber, MD.

Meeting Presentation: This paper was presented at the Combined Otolaryngology Spring Meetings of the American Head and Neck Society Section; May 1, 2019; Austin, Texas; and portions of this work were presented at the Seventh European Swallowing and Swallowing Disorders Congress and World Dysphagia Summit; September 22, 2017; Barcelona, Spain.

Additional Contributions: The MD Anderson Oropharynx Program team, particularly the members of the Patient-Reported Outcomes/Function (PROF) Core, contributed to the data collection efforts, Clare Alvarez and Casey Richardson provided modified barium swallow study analysis, and Angela Kurtz provided administrative support. All were compensated for their work with their salaries.