Functional and Oncologic Outcomes of Primary versus Salvage Transoral Laser Microsurgery for Supraglottic Carcinoma

Abstract

Objective

To evaluate the functional and oncologic outcomes of transoral laser microsurgery (TLM) in patients with previously untreated supraglottic carcinoma compared with salvage cases after radiation-based treatment.

Methods

We conducted a retrospective case-control study at a single academic tertiary care institution. Functional outcomes were stratified by prior radiation, and assessed at baseline, <1 week postoperatively, and last follow-up.

Results

Five patients underwent TLM for previously untreated disease and five previously radiated patients underwent salvage TLM for local failure. No patient required tracheostomy. There was no local recurrence after TLM as primary therapy and none required radiotherapy. One salvage patient developed local recurrence. Length of feeding tube dependence (p=0.049) and rates of chronic aspiration (>1 month postoperatively, p=0.048) were significantly higher in salvage TLM cases compared with previously untreated cases. Median PSS-HN Understandability of Speech scores were 75 (“usually understandable”) in the salvage group compared with 100 (“always understandable”) in the previously untreated group.

Conclusions

Both local control and function were superior in previously untreated patients compared with salvage patients. Our findings provide support for the use of TLM as a primary treatment modality for selected supraglottic carcinomas, but also suggest a potential for functional recovery in both previously untreated and salvage cases.

INTRODUCTION

Surgery and radiotherapy are the primary treatment options for definitive management of supraglottic carcinoma. Historically, conventional open surgery has provided comparable oncologic outcomes and laryngeal preservation rates to radiotherapy-based regimens.^1–3 Conventional open supraglottic laryngectomy, however, is associated with high rates of aspiration in the acute postoperative setting and requires temporary tracheostomy in most cases.^4,5 Adjuvant radiotherapy after open supraglottic laryngectomy has been shown to further compound postoperative functional problems.⁶ In addition, nonsurgical laryngeal preservation regimens combining radiation and chemotherapy commonly result in dysphagia; current evidence suggests that chronic aspiration is an unfortunate late consequence of these treatments in up to 40% of patients.^7–9

Transoral laser microsurgery (TLM) has become widely accepted as a minimally invasive surgical alternative for management of supraglottic carcinoma. TLM demonstrates equivalent oncologic outcomes comparable to radiotherapy-based regimens and conventional open techniques.^10–12 Moreover, several advantages of TLM relative to open supraglottic laryngectomy appear in the literature including faster recovery of swallowing function, lower incidence of aspiration pneumonia, lower rates of tracheostomy, and shorter hospital stays.^10,13–17 TLM also plays a role in the salvage setting, as an alternative to total laryngectomy, with satisfactory oncologic results reported for select patients after previous irradiation.¹⁸ Although previous authors have demonstrated broad applications of TLM with functional advantages over conventional open surgery, comparative studies of functional outcomes in primary and salvage settings are lacking. We conducted a retrospective case-control review for the purpose of comparing functional and oncologic outcomes of TLM in the primary and salvage setting.

MATERIALS AND METHODS

We conducted a retrospective case-control review at a single academic tertiary care institution. Ten patients who underwent TLM for supraglottic laryngeal carcinoma between January, 2004 and September, 2009 were included. All patients with histologically proven carcinoma of the supraglottic larynx treated with TLM in the primary or salvage setting were eligible. Patients with tumor involvement of the vocal cords and those with hypopharyngeal carcinoma that extended to the supraglottis were excluded. Outcomes were stratified by previous radiation exposure. Disease characteristics, rates of disease control, and final level of tube dependence were evaluated by medical chart review. Follow-up data were available for all patients.

Preoperative assessment

All patients underwent staging endoscopy and imaging of the pharynx, larynx, esophagus, oral cavity, and cervical lymph nodes to permit detailed assessment of the tumor and to rule out synchronous primaries. Tumor classification was conducted according to the American Joint Committee for Cancer staging, 2002 version.¹⁹

Operative technique

The operative procedures were performed under general anesthesia with orotracheal intubation. A Lindholm or distending Steiner laryngoscope was used to visualize the supraglottic larynx. All patients had TLM tumor resection comparable to standard open supraglottic resection, including the epiglottis, the false vocal folds (FVF), and the preepiglottic fat, but sparing the thyroid cartilage. A carbon dioxide laser was coupled with an operating microscope allowing precise coaxial delivery of both the aiming beam and the carbon dioxide cutting beam to the operative field. The initial incision was made using the carbon dioxide laser along the median glossoepiglottal fold to split the epiglottis in the sagittal plane. The hyoid bone and the preepiglottic space were visualized carefully dissecting layer by layer. The preepiglottic space was first surveyed. Gross infiltration of the preepiglottic space can be seen under the microscope whereas microscopic invasion is detectable only by histopathologic evaluation of the resected specimen. The extent of preepiglottic excision was, therefore, considered based on the frozen section results. In case of tumor extension to the paraglottic space, the tumor was followed and parts of the vocalis muscle were resected. The resected specimens were histopathologically examined. If histopathologic analysis showed a positive resection margin, additional resection was carried out to obtain clear margins.

Elective neck dissection was recommended for all previously untreated patients with clinically negative nodal (N0) disease. Possibilities of occult nodal metastasis including risks, benefits, and alternatives of surgical management were thoroughly discussed with all patients. If elective neck dissection was declined, the patient’s neck was followed closely using a “watch and wait” policy. This included careful physical examination as well as routine Doppler ultrasonography or CT scan every 8–12 weeks. Fine needle aspiration was performed in case of a suspected node. A priori criteria for adjuvant radiotherapy or chemoradiation included the presence of at least one of following features: positive tumor margin, perineural invasion, tumor embolus, extracapsular extension, or more than one positive lymph node.

Functional Assessment

Modified barium swallow (MBS) study results and the Performance Status Scale-Head and Neck²⁰ (PSS-HN) were uniformly collected in institutional databases. MBS studies were conducted using standard methods described previously.²¹ Varibar® thin liquid barium was administered to all subjects; additional thickened liquid consistencies and solids were administered at the discretion of the treating clinician. Compensatory swallowing strategies were selected and attempted at the discretion of the treating clinician based upon swallowing deficits observed during the study. MBS results included the presence or absence of aspiration (sensate or silent), silent laryngeal penetration, effectiveness of compensatory swallowing strategies, and recommendations of the MBS study. Tracheostomy and feeding tube status were recorded at the time of the MBS study. The PSS-HN was administered by the speech pathologist at the time of MBS. All patients had at least one postoperative MBS study. Results of fiberoptic endoscopic evaluation of swallowing (FEES) and/or clinical swallow evaluations were recorded using medical chart review for subjects who did not receive an MBS study during an assessment period. Diet levels were recorded using the PSS-HN Normalcy of Diet subscale, in which 0 indicates non-oral nutrition and 100 represents full oral nutrition without restrictions. Speech intelligibility was measured using the PSS-HN Understandability of Speech subscale, in which 0 indicates “never understandable” speech and 100 represents “always understandable” speech. The PSS-HN scores for these two subscales are described in Table 1.²⁰

Table 1.

Performance Status Scale-Head and Neck Cancer (PSS-HN), Normalcy of Diet and Understandability of Speech Subscales

Score	Description
Normalcy of diet
100	Full diet (no restriction)
90	Full diet (liquid assist)
80	All meat
70	Raw carrots, celery
60	Dry bread and crackers
50	Soft chewable foods (e.g., macaroni, canned/soft fruits, cooked vegetables, fish, hamburger, small pieces of meat)
40	Soft foods requiring no chewing (e.g., mashed potatoes, applesauce, pudding)
30	Pureed foods (in blender)
20	Warm liquids
10	Cold liquids
0	Non-oral feeding (tube fed)
Understandability of speech
100	Always understandable
75	Understandable most of the time; occasional repetition necessary
50	Usually understandable; face-to-face contact necessary
25	Difficult to understand
0	Never understandable; may use written communication

Speech and swallowing outcomes were analyzed at three assessment periods: 1) preoperatively, 2) in the immediate postoperative period (<1 week postsurgery), and 3) final (≥1 month postsurgery) postoperative period. The study was approved by the Institutional Review Board at the University of Texas MD Anderson Cancer Center. A waiver of informed consent was obtained.

Statistical Methods

Descriptive statistics were calculated to summarize sample characteristics, and functional and oncologic outcomes. Statistical associations between groups for categorical variables were analyzed using Fisher’s exact test. Statistical differences between groups for continuous variables were analyzed using Wilcoxon’s rank sum test. Statistical significance was considered α-level 0.05. Statistical analyses were performed using the STATA data analysis statistical software, version 10.0 (StataCorp LP, College Station, TX).

RESULTS

Population and Oncologic Outcomes

Ten patients (8 males, 2 females) who underwent TLM for supraglottic carcinoma were included. Five patients underwent TLM for stage I-III (T1-3N0M0) previously untreated disease and five previously radiated patients underwent salvage TLM for local failure . Of the salvage cases, two had prior chemoradiation and three had radiation therapy alone. Nine of 10 patients were diagnosed with squamous cell carcinoma, and one patient with T3N0 disease was diagnosed with adenoid cystic carcinoma of the epiglottis.

All patients had adequate tumor visualization and satisfactory resections without positive tumor margins. No patient required tracheostomy. No surgical complications occurred. There was no recurrence (local or regional) for any patient undergoing TLM as primary therapy (local control rate, 100%); furthermore, none required adjuvant radiotherapy (mean length of follow-up: 24.0 months [sd: 15.2]). One of five salvage patients developed recurrence at the primary site and was treated with repeated TLM. The local control rate for previously radiated patients was 80% (mean length of follow-up: 22.6 months [sd: 12.9]). Overall survival and laryngeal preservation rates were 100% in both groups. Population characteristics and oncologic outcomes are summarized in Table 2.

Table 2.

Population and Oncologic Outcomes

	Previously Untreated	Salvage
Sex
Male	3	5
Female	2	0
Pathologic Diagnosis
SCCA	4	5
Adenoid cystic	1	0
TNM classification
T1N0	1	0
T2N0	3	0
T3N0	1	0
Recurrent	0	5
Site
False Cord	2	2
False Cord/AE fold	0	1
Epiglottis	3	2
Age, Median (Range)	76 (65–77)	60 (47–78)
Hospital stay, Median (Range)	4 (2–7)	3 (2–4)
Local control	5	4
Overall survival	5	5
Months of follow-up, Mean (SD)	24.0 (15.2)	22.6 (12.9)

Abbreviations: SCCA, squamous cell carcinoma; AE, aryepiglottic; SD, standard deviation

During the follow-up period, one patient in the previously untreated group developed an ipsilateral level IIa neck node confirmed by ultrasound and CT scan 15 weeks after TLM. Fine needle aspiration cytologically confirmed metastatic disease. Subsequent neck dissection, levels II-IV, revealed one of 20 positive lymph nodes without extracapsular spread or lymphovascular invasion on pathologic examination.

Baseline Functional Status

Baseline functional outcomes for both groups are tabulated in Table 3. All patients received a baseline MBS study. Neither aspiration nor silent laryngeal penetration was identified at baseline in the previously untreated patient group. In contrast, baseline MBS studies documented impaired airway protection in 4/5 salvage patients; three of these patients demonstrated trace, silent aspiration responsive to swallowing strategies and one demonstrated silent laryngeal penetration. No patient in either group was gastrostomy or tracheostomy dependent at baseline. All patients ate a soft or regular diet without liquid restrictions at baseline, and 4/5 patients in each group had an unrestricted diet per PSSHN scores (Normalcy of Diet = 100) at baseline.

Table 3.

Baseline and Final Functional Outcomes

	Previously Untreated		Salvage
	Baseline	Final*	Baseline	Final*
Feeding Tube	0	0	0	1
Trach	0	0	0	0
Penetration/Aspiration**	0	1	4	5
Diet
Regular/Soft Unrestricted	5	4	5	0
Regular/Soft restricted***	0	1	0	4
Tube + PO	0	0	0	0
NPO	0	0	0	1
PSS-HN Speech Score
100 (“Always understandable”)	5	5	4	2
75 (“Usually understandable”)	0	0	1	3

Abbreviations: PO, per oral; NPO, nothing per oral; PSS-HN, Performance Status Scale Head and Neck Cancer

^*Aspiration status and PSS-HN scores determined at time of final swallowing study; diet, feeding tube, and tracheostomy status at last follow-up.

^**Includes silent laryngeal penetration, sensate aspiration, and silent aspiration.

^***Dietary restrictions included the need for swallowing strategies or thickened liquids to control aspiration.

All previously untreated patients were rated as “always understandable” (PSS-HN Understandability of Speech score = 100) at baseline. Four of five salvage patients were rated as “always understandable” (PSS-HN Understandability of Speech score = 100) and one as “usually understandable” (PSS-HN Understandability of Speech score = 75) at baseline.

Immediate Postoperative Swallowing Outcomes

Functional status and diet recommendations in the first postoperative week are outlined in Table 4. MBS studies were conducted in 4/5 and FEES in 1/5 previously untreated patients in the immediate postoperative period (median: 2 days postoperatively, range: 1–6). Sensate aspiration was documented in 3/5 (60%) previously untreated patients, while two patients who had false cord primaries did not aspirate in the first postoperative week. Three patients began a regular diet with thin liquid (with or without strategies) and NPO status was recommended for only one previously untreated patient in the first postoperative week. Although 3/5 (60%) previously untreated patients required a feeding tube in the first postoperative week, one tube was removed after two days after the MBS study found mild aspiration responsive to swallowing strategies.

Table 4.

Immediate Postoperative (<1 week) Functional Status

	Previously Untreated	Salvage
Feeding Tube	3*	3
Trach	0	0
Penetration/Aspiration**	3	5
Diet recommendation from swallow evaluation
Regular/Soft Unrestricted	1	0
Regular/Soft restricted***	2	2
Tube + PO	1	1
NPO	1	2

Abbreviations: PO, per oral; NPO, nothing per oral

^*One feeding tube removed 2 days postoperatively.

^**Includes silent laryngeal penetration, sensate aspiration, and silent aspiration.

^***Dietary restrictions included the need for swallowing strategies or thickened liquids to control aspiration.

In the immediate postoperative period (<1 week), 3/5 salvage patients underwent an MBS study, and the remaining two were assessed with a clinical swallow evaluation only (median: 3 days postoperatively, range: 1–7). Instrumental evaluation was postponed in these two patients because of the severity of aspiration symptoms on bedside examination in one and the ability to alleviate aspiration symptoms at bedside in the other. All salvage patients (5/5) aspirated in the immediate postoperative period; silent aspiration was documented in two (40%) of these patients. Two of five (40%) remained NPO in the immediate postoperative period due to significant aspiration. A total of 3/5 salvage patients were feeding tube dependent in the first postoperative week.

Final Postoperative Swallowing Evaluation

In the final assessment period (>1 month), aspiration was significantly more common in salvage patients compared with previously untreated patients (p = 0.048). Four of five salvage patients underwent MBS in this assessment period, and the remaining patient who demonstrated mild aspiration responsive to strategies on MBS in the immediate postoperative period was assessed with a clinical swallow evaluation only (median: 7 weeks, range: 4 - 59). All salvage patients (5/5) continued to aspirate during swallowing evaluations in the final postoperative period; silent aspiration was documented in 2 (40%) of these patients. A majority of salvage patients 4/5 tolerated oral intake despite aspiration, but all four of these patients required the use of compensatory swallowing strategies to prevent aspiration. Only 1/5 (20%) remained NPO in the final postoperative period.

Four of five patients in the previously untreated group underwent MBS in the final postoperative period (median: 17 weeks, range: 9 – 206), and one patient whose MBS in the immediate postoperative period was normal did not return for repeat MBS in the final period. Silent aspiration, responsive to swallowing strategies, was documented in only 1/5 (20%) previously untreated patients at last assessment who subsequently reported tolerating a regular, unrestricted diet at 6 months follow-up. At the time of last assessment, all previously untreated patients ate a regular oral diet without restrictions.

Postoperative Speech Outcomes

PSS-HN Understandability of Speech scores were captured for all patients in the postoperative setting. Median postoperative PSS-HN Understandability of Speech scores were 75 (“usually understandable”) in the salvage group compared with 100 (“always understandable”) in the previously untreated group.

Feeding tube dependency

Overall, 60% (6/10) of patients required a feeding tube postoperatively, 3/5 in the salvage group and 3/5 in the previously untreated group. All patients with epiglottic tumors (5/5) required a feeding tube, whereas one of five patients with FVF tumors required a feeding tube (p = 0.048). Length of feeding tube dependence was significantly longer in salvage patients compared with previously untreated patients (p = 0.049). Among the three salvage patients who required gastrostomy, all remained tube dependent more than one year postsurgery. Two of these three salvage patients slowly weaned off tube feedings (413 and 724 days) and one remained tube dependent at last follow-up (1,077 days). Both salvage patients with epiglottic tumors were tube dependent two years or more. All previously untreated patients with epiglottic tumors (3/5) required a feeding tube postoperatively, which were subsequently removed 2, 24, and 181 days postoperatively. Neither of the previously untreated patients with false cord primaries (2/5) required a feeding tube and no previously untreated patient remained tube dependent at last follow-up. Overall functional outcomes (baseline and final) are summarized in Table 3.

DISCUSSION

TLM has evolved as an important minimally invasive surgical option for management of supraglottic carcinoma. Satisfactory oncologic results have been achieved in patients with early T1 and T2 disease and selected T3 cases, but also in cases of salvage resection after radiation failure.^{10–12,18,22} Local control and laryngeal preservation rates above 85% are commonly reported in patients with previously untreated supraglottic disease,^10,22 and 88% 2-year locoregional control was achieved in patients with recurrent supraglottic carcinoma treated with salvage TLM in a single multicenter case series.¹⁸ Our outcomes compare favorably with these reports as the local control rate was 90% overall, and none of our patients required total laryngectomy for local recurrence or laryngeal dysfunction. We observed only one local failure; this patient was in the previously radiated group and was treated with repeat TLM 15 months after his initial TLM.

Radiotherapy is known to impart a number of normal tissue effects that commonly manifest as dysphagia and aspiration in individuals treated for supraglottic carcinomas.^7,9 The potential to avoid radiotherapy is a particularly attractive feature for the use of TLM for primary treatment of supraglottic carcinoma. In our study, all previously untreated patients who underwent supraglottic TLM were treated with a single modality and avoided adjuvant radiotherapy. It is, therefore, not surprising that both early and final functional outcomes were clearly superior in previously untreated patients. In fact, all five salvage TLM cases continued to aspirate on final swallow studies as compared to only one of five previously untreated patients. These data suggest that chronic aspiration is to be expected after salvage supraglottic TLM. It is noteworthy that, despite chronic aspiration, most salvage cases slowly resumed oral intake postoperatively and only one remained gastrostomy dependent at last follow-up. These data suggest a potential for recovery of functional swallowing abilities after salvage TLM with proper training on modifications and compensations that minimize aspiration. Further study in a larger population is needed to substantiate this finding.

Although all patients communicated functionally with natural laryngeal voice after TLM, we observed below normal postoperative speech scores (PSSHN Understandability of Speech scores) in three of five salvage cases. This was somewhat unexpected because supraglottic laryngectomy spares the true vocal folds and is not traditionally associated with voice changes¹³ It is important to consider that the rating we used (PSS-HN subscale) provides a global assessment of speech intelligibility and may be impacted by changes in voice, articulation, and/or resonance. As such, lower speech ratings in salvage cases likely resulted from factors beyond true vocal fold functioning that are commonly associated with radiotherapy, such as laryngeal edema, mucosal dryness, or decreased secretion management.

Rehabilitation efforts after salvage supraglottic TLM are prolonged by the concurrent effects of prior radiotherapy, as evidenced by the fact that the majority of our salvage cases were tube dependent well beyond one year. The delayed functional recovery we observed is supported by Motamed’s systematic review on the topic of salvage conservation laryngeal surgery.²³ It is, therefore, imperative that individuals considering salvage supraglottic TLM be counseled appropriately to expect lengthy dysphagia rehabilitation, particularly when considering salvage resection of epiglottic tumors. Both of our salvage TLM patients with epiglottic tumors were tube dependent two years or more. This may be attributed to the larger defect that results from resection of an epiglottic tumor (compared with a FVF tumor) coupled with baseline dysfunction from prior treatment. Despite a prolonged swallowing recovery, intelligible laryngeal voicing was preserved for all salvage TLM patients in this series. Thus, salvage supraglottic TLM may be an appropriate alternative for those individuals who prioritize more normal verbal communication over the faster swallowing recovery that is offered with salvage total laryngectomy.

In contrast, chronic dysphagia was not present in any of our patients undergoing TLM as primary treatment. Sensate aspiration was observed in only three of five cases in the first postoperative week with resolution of aspiration documented in all but one patient who was unable to return for swallowing assessment after rehabilitation. Tube dependency was also significantly shorter in previously untreated cases; only one patient in this group (72 year-old white male with mild COPD) required enteral feedings longer than 1 month. These findings are consistent with historical institutional data that found prolonged gastrostomy dependence in patients who required adjuvant radiotherapy (55–63 Gy) after open supraglottic laryngectomy.⁶ Higher radiation doses and concurrent chemotherapy delivered during definitive radiotherapy likely further exacerbated postsurgical dysphagia observed in salvage TLM cases in this study.

To date, comparisons of functional outcomes after TLM supraglottic laryngectomy have focused on outcomes relative to conventional open transcervical supraglottic laryngectomy. Although we acknowledge the limitations imposed by retrospective observational methods and a small sample size, we present a novel comparison of functional outcomes between frontline and salvage TLM in patients with supraglottic carcinoma. Statistically significant distinctions in functional outcomes (higher aspiration rates and longer feeding tube dependence after salvage TLM) were apparent between groups despite a limited sample size.

In addition, complete baseline data strengthen our ability to discern the discrete impact of prior cancer treatment on swallowing function. Baseline MBS studies confirmed that airway protection was substantially impaired in four of five salvage cases prior to TLM. Similar tumor burden was observed in previously untreated and salvage cases, but only salvage cases demonstrated impaired airway protection on baseline MBS studies. It is, therefore, our contention that these baseline impairments were likely related to effects of prior radiotherapy and chemoradiotherapy. A recent review of three prospective clinical trials from the Radiation Oncology Treatment Group has underscored the long-term functional impairment of patients undergoing combined treatment.²⁴ Machtay found that 43% of patients had some form of severe late toxicity, especially laryngopharyngeal dysfunction requiring gastrostomy in 13% and less often tracheotomy. This trend was particularly significant in patients with a primary tumor arising in the larynx or hypopharynx, as opposed to the oral cavity or oropharynx (odds ratio, 4.17; p = .0041). When considered against the backdrop of modern multidisciplinary care of head and neck cancer,^24,25 which often supports first treatment with radiation therapy, our experience and that of others suggest that TLM might provide patients added options for function-sparing treatment. Further research with a larger sample size is needed to confirm these preliminary observations.

CONCLUSION

Both local control and postoperative speech and swallow function after supraglottic TLM were superior in previously untreated patients compared with salvage TLM after radiation or chemoradiation failure. Specifically, the prevalence of chronic aspiration and length of gastrostomy tube dependence were significantly higher after salvage TLM. Within groups, those treated for false cord primaries also demonstrated superior functional outcomes compared with those who had epiglottic primaries. These findings provide preliminary evidence that support the use of TLM as a primary treatment modality for selected patients with supraglottic carcinoma, but also suggest a potential for functional recovery in both previously untreated and salvage cases.