Recurrence-free survival of 141 patients with supraglottic cancers treated with laser surgery.
Recurrence-free survival of 141 patients with supraglottic cancers treated with laser surgery, stratified by neck dissection status.
Recurrence-free survival of 141 patients with supraglottic cancers treated with laser surgery, stratified by R classification of the primary tumor (in accordance with Union Internationale Contre le Cancer definitions16).
Iro H, Waldfahrer F, Altendorf-Hofmann A, Weidenbecher M, Sauer R, Steiner W. Transoral Laser Surgery of Supraglottic CancerFollow-up of 141 Patients. Arch Otolaryngol Head Neck Surg. 1998;124(11):1245-1250. doi:10.1001/archotol.124.11.1245
To determine the role of transoral laser resection of supraglottic carcinomas.
Retrospective unicenter study of the oncologic results of transoral carbon dioxide laser microsurgery for supraglottic carcinomas performed between February 1979 and December 1993. Median follow-up was 37 months.
University hospital academic tertiary referral center.
We reviewed the medical records of 141 patients (a consecutive sample of 131 men and 10 women; mean age, 60 years) with histologically proven supraglottic carcinomas undergoing transoral laser surgery, possibly in combination with neck dissection or radiotherapy. Stage distribution of patients was as follows: stage I, 23.4%; stage II, 25.5%; stage III, 16.3%; and stage IV, 34.8% (according to the Union Internationale Contre le Cancer staging system).
Main Outcome Measures
Recurrence-free survival rates and local and regional recurrence rates.
Five-year recurrence-free survival rates were as follows: the whole case load, 65.7%; stage I, 85.0%; stage II, 62.6%; stage III, 74.2%; and stage IV, 45.3%, according to the Union Internationale Contre le Cancer staging system. The local and regional recurrence rates were 16.3% and 9.9%, respectively.
The oncologic results of transoral carbon dioxide laser surgery are satisfying if clean surgical margins (R0 resection) can be reached. In patients in whom tumor-free margins are not achieved (R1 and R2 resection) and transoral revision is not possible, transcervical procedures (partial or total laryngectomy) should be performed. The indication for transoral supraglottic laryngectomy in T3 lesions should be considered with restraint.
SINCE THE introduction of the carbon dioxide (CO2) laser into clinical otolaryngological procedures, laser surgery for malignant tumors in the upper aerodigestive tract has been gaining increasing importance.1 However, most publications on this topic are devoted exclusively to laser surgery of glottic cancer, the indications and contraindications of which remain controversial. To date, the value of CO2 laser resection of supraglottic cancer has been discussed only insufficiently in the literature. Except for individual case reports2,3 and publications on therapeutic results with comparatively small case loads,4- 6 there has been only 1 major report7 of laser surgery in the upper aerodigestive tract presented within the framework of a general survey. Currently, transcervical supraglottic partial resection is, in the opinion of many authors—at least as far as early supraglottic cancer is concerned—the surgical therapy of choice8- 10; for more extensive tumors (T3 and T4), total laryngectomy is usually recommended.8,10 The oncologic results of these therapy methods have been presented in numerous studies, some with large case loads (Table 1). When performing a transcervical supraglottic partial resection, a transient or permanent tracheotomy is often unavoidable; postoperative alteration of the swallowing function caused by a lesion of the superior laryngeal nerve(s) with susceptibility to aspiration is a typical aftereffect of surgery.9 Especially in view of conservation of the organ and associated aspects of functional maintenance, exclusive radiotherapy—if necessary, with consecutive dissection of the cervical lymph nodes—is considered a therapeutic alternative because simultaneous chemoradiotherapy has not yet yielded sufficient results. Results of some studies provide a comparison between surgery and irradiation. When oncologic criteria are applied, surgery is usually preferred.9,17 The present study describes long-term follow-up of 141 patients whose supraglottic carcinomas were treated with laser surgery.
From February 1979 to December 1993, 141 of 387 patients with supraglottic cancer were treated with transoral laser surgery at the Department of Otorhinolaryngology–Head and Neck Surgery of the University of Erlangen-Nuremberg, Erlangen, Germany. The indication for transoral laser surgery was fulfilled if the tumor was completely exposable during preceding diagnostic microlaryngoendoscopy with multiple biopsies. In 34 additional patients, laser surgery was attempted but could not be completed because of the extent of tumor expansion or the assumed lack of oncologic radicality. These patients reached transcervical horizontal resection or laryngectomy in the same session and are excluded from further considerations. The decision for or against performing transoral laser surgery was based mainly on the results of endoscopy; radiographic studies were not performed as a rule, but in the last 4 years magnetic resonance tomography was performed in doubtful cases to assess if an R0 resection would be possible. Patients with additional malignant tumors in their history, patients with cancers treated elsewhere, and patients with recurrent tumors or distant metastases were excluded from the analysis.
The medical records of the patients were evaluated in a retrospective analysis, and a statistical assessment of the data was performed. If necessary, the postoperative course was clarified by physicians in charge of continued treatment, patients or their relatives, and registration office personnel. Follow-up was conducted until December 31, 1995.
At the time of diagnosis, the 131 male and 10 female patients had a mean age of 60 years (range, 36-90 years). Stage distribution of patients was as follows: stage I, 23.4%; stage II, 25.5%; stage III, 16.3%; and stage IV, 34.8% (according to the Union Internationale Contre le Cancer [UICC] staging system16). Table 2 correlates T and N categories. Thus, more than 50% of patients were assigned to the group of advanced supraglottic cancers (stages III and IV). All patients underwent transoral microsurgical CO2 laser resection of the primary tumors. Even larger tumors could, if necessary, be removed piece by piece intraoperatively, according to the recommendations of one of the authors (W.S.), who introduced this technique in the Ear, Nose, and Throat Department of the University of Erlangen-Nuremberg in 1979. Residual tumors were classified according to UICC definitions: RX, the presence of residual tumor could not be assessed; R0, no residual tumor; R1, microscopic residual tumor; and R2, macroscopic residual tumor. An R0 resection was the general objective. All cancers were histologically determined as squamous cell carcinoma. Seventy-one patients (50.4%) were treated by unilateral or bilateral neck dissection. In 30 of these patients with clinical N0 status (33% of the patients with N0 status), an elective neck dissection was indicated (Table 3). Nine patients with clinically suspected metastases in the cervical lymph nodes were nevertheless not subjected to a neck dissection. Three of these patients received irradiation instead; the remaining 6 patients were merely kept under regulated surveillance. Sixty-four patients (45.4%) were treated by postoperative radiotherapy, including cervical lymph drainage. Nine of these patients were in stage I, 10 were in stage II, 9 were in stage III, and 36 were in stage IV. Sixty-nine patients (48.9%) underwent no further postoperative therapy (Table 3). The remaining 8 patients received postoperative concomitant chemoradiotherapy or were only treated with cytostatic agents. Table 3 summarizes the treatment modalities according to UICC stage. Median follow-up was 37 months. The statistical evaluation of the data was performed with a software program package (StatView 4.5 for Macintosh, Abacus Concepts Inc, Berkeley, Calif). Survival analyses were performed by the Kaplan-Meier method and the log-rank test.
A successful R0 resection was achieved in 110 (78.0%) of 141 patients treated with transoral microscopic laser surgery. The duration of the laser procedure ranged from 30 minutes to 3 hours depending on the extent of the tumor and the need to wait for intraoperative frozen sections. In the final pathohistologic workup of the resected specimens, tumor ramifications were microscopically detected (R1) in 14 patients (9.9%). In 9 patients, tumor had to be retained macroscopically (R2); and in 8 patients, no definite R classification could be established retrospectively (RX). The group of R1 and R2 resections will be discussed in detail below. In 80.1% of patients, the T category corresponded with the pT category; in 12.8% of patients, the pT category was overestimated in clinical staging. In 23.3% of patients, occult lymph node metastases were detected in the N0 neck, and clinical neck staging was false-positive in 23.9% of patients. Eighteen patients (12.8%) underwent tracheotomy in the course of tumor resection or during subsequent radiotherapy; in 6 patients (4.2%), the tracheostoma could be resealed after completion of therapy. None of the patients died of treatment-related causes. Six patients developed postoperative aspiration pneumonias, which in all cases could be cured by antibiotic therapy. A secondary laryngectomy for functional reasons was not necessary in any patients. Figure 1 depicts the recurrence-free survival graph according to Kaplan-Meier for the complete case load. The result is a recurrence-free 5-year survival rate of 65.7% and a recurrence-free 10-year survival rate of 56.8%. The recurrence-free 2- and 3-year survival rates are 81.6% and 76.1%, respectively. The corresponding stage-dependent 5-year local control rates are 86.1% for stage I, 74.6% for stage II, 75.4% for stage III, and 78.4% for stage IV. If R1 and R2 resections are excluded, the respective recurrence-free 5-year survival rates are 79.1% for stages I and II and 72.5% for stages III and IV. Patients whose therapy included a neck dissection had a recurrence-free 5-year survival rate of 72.9%, whereas patients who were not operated on at the neck reached a recurrence-free 5-year survival rate of 58.6% (Figure 2). In a subgroup analysis of survival after elective neck dissection, an advantage in survival for the patients who were operated on emerged as a trend (recurrence-free survival rates of 85.3% vs 65.6%); however, the significance level of 5% was reached in neither group.
All patients with microscopic (R1) and macroscopic (R2) residual tumor received postoperative irradiation because revision surgery either had been refused by the patient or had not been performed for other reasons. However, irradiation did not have a significant positive effect on the prognosis of patients with R1 and R2 resections. With a value of P<.001, the difference in survival between the group of patients with R0 resection and the group with microscopic (R1) or macroscopic (R2) residual tumor reached a high significance level in the log-rank test (Figure 3). Nine of 23 patients in whom laser resection did not induce an R0 condition never reached complete remission and ultimately died of tumor progression. Seven of the remaining 14 patients who reached complete remission because of the consecutive therapy afterward developed a local recurrence. Furthermore, 2 of 8 patients with RX surgery had renewed local tumor growth; regional recurrences did not occur in this group.
During follow-up, 23 patients (16.3%) developed local recurrences. In relation to the group of patients whose tumors could be R0 resected, this amounts to a local recurrence rate of 14.5% (16/110). Table 4 shows the rates for local and regional recurrences according to UICC stages and the rates for local recurrences according to T categories. Seven of 23 patients with local recurrences and 7 of 14 patients with regional recurrences had been treated by postoperative irradiation. None of the 6 patients with N+ neck but not treated at the neck site had a regional recurrence. Therefore, it must be assumed that the clinical staging of the cervical lymph nodes was false-positive.
Table 5 provides information on the recurrence-free 5-year survival rates according to treatment modality. In this comparison, 65% of the group with postoperative irradiation belonged to stage IV, whereas only 14% of the group not treated by postoperative irradiation belonged to this stage.
Except for the report by Steiner and Ambrosch,7 our analysis is the first publication worldwide of the long-term results of laser surgery for supraglottic cancer based on a major case load. The literature is predominantly devoted to laser surgery of glottic cancer; few reports have been published on laser treatment of supraglottic cancer. In the history of ear, nose, and throat laser surgery, the treatment of supraglottic cancer is a comparatively recent development. In 1978, Vaughan3 first pointed out the possibility of using this laser for the resection of early supraglottic cancer. In 1983, Davis et al2 described the first experience with the CO2 laser in 10 patients with supraglottic cancer, but laser resection was applied in 6 patients only for staging purposes. Similarly, in a subsequent report18 comprising 24 patients, the laser was used primarily for staging before definitive radiotherapy. In 1990, Zeitels et al6 explicitly described the technique of laser epiglottectomy and reported their experience with 51 patients. However, in this group, only 27 patients had malignant tumors of the supraglottis, and the laser procedure was applied exclusively for diagnostic purposes, serving for excisional biopsy in 20 patients. In 7 patients with T1 or T2 tumors, excisional biopsy was able to achieve R0 resection. The authors considered the presence of T1 and T2 tumors to be an indication for curative laser resection of supraglottic cancer; in the case of invasion of the preepiglottic space (according to American Joint Committee on Cancer classification T3), they recognized the limitations of this method and preferred an external approach. In this study, no long-term results were given. By 1994, Zeitels et al5 could look back on 45 patients treated with laser surgery, 36 of whom had supraglottic cancer (T1-T3 and N0). In 22 patients (T1, 16 patients; T2, 6 patients; 19 patients had supraglottic localization of the primary tumor), a local en bloc resection was carried out without further surgical treatment or radiotherapy. No local recurrences were observed in this group. Twenty-three patients with more extensive tumors (T2-T3; 18 patients had supraglottic tumor localization) who, according to the authors, had not come into consideration for a classic transcervical supraglottic partial resection were subjected to excisional biopsy of the primary tumor with the CO2 laser, followed by radiotherapy with the inclusion of the primary tumor and cervical lymph drainage. In 16 of these 23 patients, R0 resection was achieved; in this group, there was no local recurrence. Five of 7 patients with R1 resection had a local or regional recurrence during follow-up. Therefore, as the authors concede, this patient group represents a highly selective sample of small, respectively "early," supraglottic cancers without formation of cervical metastases and thus substantially differs from our case load exhibiting an N+ condition in 36% of patients and UICC stage IV in 35% of patients. Rudert and Werner4 described their experience with CO2 laser resection in 30 patients with supraglottic cancers; a further 17 patients had received palliative treatment in the form of tumor debulking. Although Zeitels et al5 advocated transcervical tumor resection in the case of tumor invasion of the preepiglottic space, Rudert and Werner4 did not uphold this general demand but rather refrained from transoral laser surgery if the tongue base was infiltrated. Fifteen of 30 tumors treated by Rudert and Werner4 using laser surgery could be assigned to categories T3 and T4. Eleven patients were subjected to neck dissections, and 21 patients received postoperative irradiation. In this context, the authors emphasized the need to include the neck in therapeutic considerations. During mean follow-up of 27 months, 4 of 30 patients died of the tumor. Beyond this, no survival analysis was carried out. Köllisch et al19 compared functional results between transoral laser surgery and supraglottic partial resection with an external approach. Their study yielded the result that transoral laser surgery was superior to transcervical supraglottic partial resection in terms of the time required to restore the patient's swallowing capacity, in terms of tracheotomy rate, and in terms of incidence of aspiration pneumonias. In our case load, only 13% of patients were tracheotomized; and in 4%, the tracheostoma could be resealed after appropriate deglutition training. Twelve patients, however, had to remain tracheotomized because of postoperative aspiration problems; of these, 11 had pT4 cancer with extensive resections in the area of the tongue base. No patient had to be laryngectomized because of persisting aspiration, which—according to Suárez et al20—was necessary in 10% of their patients after transcervical supraglottic partial resection. In a further 24% of their patients, the tracheal cannula could not be removed.
In 1996, Steiner and Ambrosch7 published the results of the Goettingen University group after laser surgery of supraglottic cancer in 99 patients. In 43 patients with stage I and stage II supraglottic cancer, the local recurrence rate was 9.5% and the 5-year overall survival rate was 72.8%; in 56 patients with stage III and stage IV cancer, the respective percentages were 19.5% and 48.8%.
In the patients from Erlangen, discussed in our study, overall 5-year survival rates of 75.4% for stages I and II and 56.0% for stages III and IV were determined.
Further surgical interventions after R1 and R2 resections were not performed, especially at the beginning of the trial; after recognizing the worse prognosis of R1 and R2 resections despite postoperative radiotherapy, transcervical surgical procedures (partial or total laryngectomy) were strongly recommended to the patients affected. As a conclusion of these results, laser surgery for supraglottic carcinoma seems only appropriate in cases of histopathologically proven R0 resections. However, this is not a peculiar characteristic of laser surgery but of oncologic surgery in general. If an R0 resection cannot be achieved by means of transoral surgery, either transoral revision surgery or transcervical procedures (partial or total laryngectomy) must be carried out because postoperative irradiation of patients with R1 and R2 status failed to demonstrate oncologic benefit.
The functional advantages of transoral laser surgery of supraglottic cancer, such as possible avoidance of temporary or permanent tracheotomy, lower incidence of pharyngocutaneous fistulas, and faster rehabilitation of swallowing, are obvious,21 even if neck dissection is performed from a transcervical approach. Nevertheless, comparison of the oncologic results of the different therapies is essential as well. Table 1 shows the survival rates achieved in various treatments of supraglottic cancer reported in several studies. A problem inherent in this comparison is the fact that several different target criteria, such as overall survival, disease-free survival, or recurrence-free survival, are used. Although the rates of survival and local recurrences reported in the literature vary considerably and the case loads are in part composed heterogeneously, it can be inferred that the oncologic results of transoral laser surgery are comparable to those of classic transcervical partial resection if an R0 resection can be achieved by means of laser surgery. This applies to the survival and the local recurrence rates, the latter being the decisive criterion for comparing these methods. Considering, moreover, the functional aspects of transoral laser resection, the advantages of this operative technique predominate. Therefore, transoral laser resection of supraglottic cancer should, together with classic transcervical supraglottic partial resection, be considered an established therapeutic modality. Of course, it would be wrong to conclude that in the future all supraglottic cancers can, without exception, be managed by laser surgery. Rather, the role of classic supraglottic partial resection and, if indicated, laryngectomy should not be underestimated, and these methods will continue to rank high in the management of supraglottic cancers.
However, in a stage-related analysis of our patient population, the comparatively disappointing therapy results in stage II are striking. Of 36 patients, only 8 had been subjected to a neck dissection; 10 patients (including 4 patients with R1 resection who refused a further operation) had received postoperative irradiation. Apart from local recurrences, the causes of death were regional recurrences and distant metastases; thus, the recurrence patterns are inhomogeneous. In view of the fact that in stage III— which is associated with a better prognosis—a neck dissection and postoperative radiotherapy were performed in most patients, it follows that surgical treatment of the primary tumor as a sole therapeutic measure has to be rated as inadequate for managing stage II supraglottic cancer. As a consequence of these results, all patients with T2 lesions are now undergoing neck dissection or postoperative irradiation. Also, patients with stage I disease are referred to postoperative radiotherapy (or neck dissection) because of the high local recurrence rate of 26% in patients only undergoing surgery although R0 resection was obtained.
Local recurrence rates were not strongly correlated to T category, but recurrences occurred, especially in T3 cancer (Table 4). Conversely, the lower incidence of local recurrences in T4 lesions might be partially explained by problems within the UICC classification of laryngeal cancer. In T3 lesions, the indication for larynx-conserving surgery (transoral laser surgery and transcervical partial laryngectomy) and total laryngectomy must be weighed with concern for the possibility of involvement of the cricoarytenoid joint, deep muscular infiltration, or neural invasion with consecutive worse prognosis.
Independent of the controversy about the right choice of therapy for the primary tumor, no definitive consensus regarding the treatment of the neck in the case of supraglottic cancer has been reached. On the basis of their investigations, Gudziol and Beleites,22 Lutz et al,23 Weber et al,24 and Myers and Alvi9 demand routine bilateral neck dissection. In this context, Cummings25 gives reference to a proportion of 22% occult metastases in the cervical lymph nodes in T2 cancers and a simultaneous false-positive clinical status of the cervical lymph nodes in 55% of patients. Our case load exhibits a certain tendency for patients to have a better prognosis after elective neck dissection, although the observed differences did not reach the level of significance. Furthermore, the position of postoperative radiotherapy cannot be clearly defined on the basis of the available data. Also, Suárez et al26 were able to show in this context that postoperative radiotherapy was likely to significantly reduce the incidence of contralateral regional recurrences and of distant metastases. In our study, postoperative radiotherapy seemed to reduce the incidence of local, but not regional, recurrences (Table 5). It, therefore, remains for further studies to examine the effect of postoperative radiotherapy on local recurrences after supraglottic laser resection.
Supraglottic carcinomas could be treated by transoral laser surgery with satisfactory oncologic results if R0 resection is reached. Diagnostic microlaryngoendoscopy is necessary to decide if a transoral laser procedure is possible; magnetic resonance tomography will provide additional information on tumor expansion in some patients, and demonstrated to be helpful to assess if an R0 resection with transoral laser surgery would be possible. However, submucosal tumor growth will not always be detected in magnetic resonance tomography, and multiple intraoperative frozen sections are always necessary to state clean surgical margins. Patients with R1 or R2 resections failed subsequent irradiation; therefore, for these patients, revision surgery up to total laryngectomy must be strongly recommended; every patient undergoing transoral laser surgery should be informed about this aspect. If a sure R0 resection is achievable, transoral laser surgery can be performed in all tumors exposable by means of direct laryngoscopy. T3 supraglottic tumors deserve particular consideration in view of their worse prognosis; this statement is valid not only for transoral laser surgery but also for larynx-conserving surgery in general. Surgical treatment of the cervical lymph nodes and postoperative radiotherapy remain important, prognostically relevant elements of therapy also in the context of laser surgery. Elective neck dissection and postoperative irradiation even in stage I and stage II disease should be performed with reference to the data obtained in this study.
Accepted for publication May 21, 1998.
This work was supported by the Johannes and Frieda Marohn Foundation, Erlangen, Germany.
Reprints: Heinrich Iro, MD, Department of Otorhinolaryngology–Head and Neck Surgery, Saarland University, D-66421 Homburg, Germany (e-mail: HNOHIRO@MED-RZ.UNI-SB.DE).