Sigston E, de Mones E, Babin E, Hans S, Hartl DM, Clement P, Brasnu DF. Early-Stage Glottic CancerOncological Results and Margins in Laser Cordectomy. Arch Otolaryngol Head Neck Surg. 2006;132(2):147-152. doi:10.1001/archotol.132.2.147
Copyright 2006 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2006
To assess local control of early-stage glottic cancer by laser cordectomy in comparison with previously published external partial laryngectomy series and to determine the relevance of histological margins in glottic cancers excised with laser cordectomy.
Retrospective review of laser cordectomy for carcinoma in situ (Tis) and stage T1 glottic cancer from January 1991 to January 2004.
Fifty-two patients with Tis or T1 glottic cancer.
Endoscopic laser cordectomy, classified using the system proposed by the European Laryngeal Society Working Committee.
Main Outcome Measures
Local control after initial surgery and after salvage compared with a published historical control group, according to the type of cordectomy performed and the histological margins of the removed specimen.
Sixteen patients with Tis, 30 with T1a tumors, and 6 with T1b tumors were followed up for an average of 38 months. Type I cordectomy was the most common procedure used to treat Tis, and type II and type III were the most common for treating T1a and T1b tumors. Of 6 recurrences, 4 were treated with laser cordectomy and 2 were treated with external partial laryngectomy. The rate of laryngeal preservation was 100%. There were 3 recurrences despite histologically clear margins. Three (17%) of 18 patients with suspicious margins developed recurrences. The rate of local control with single intervention (46 [89%] of 52) was lower than with partial external laryngectomy. However, 46 (89%) of 52 patients ultimately had less tissue removed by laser than would have been removed by external partial laryngectomy.
Laser cordectomy provides excellent local control and laryngeal preservation. Close follow-up of patients with positive or suspicious margins is an alternative to further routine treatment.
Treatment of patients with early-stage glottic cancer with external partial surgery, radiotherapy, and, more recently, transoral laser surgery, is an area of continual change, complexity, and controversy. Focus is now not just on oncological outcome but also on preservation of vocal function. This results in a continuous conflict between the need to remove tissue for oncological stability and the need to preserve tissue for optimal vocal function. Our objective was to find and define the precarious balance.
Use of transoral laser surgery to treat early-stage glottic cancer, initially described by Strong and Jako in 1972,1 was brought into prominence by the work of Steiner2 and other European pioneers. This approach allows early-stage tumors to be removed with minimal sacrifice of healthy tissue and with retention of good voice quality. Since the early 1990s, our interest has increasingly turned to use of transoral carbon dioxide laser for treatment of carcinoma in situ (Tis) and stage T1 lesions, with external approaches being reserved for larger lesions (T2) and those lesions with adverse characteristics such as gross anterior commissure involvement or poor access transorally or in the case of failed previous treatment.
This article describes our series of 52 patients with Tis or T1 glottic lesions, initially treated with transoral carbon dioxide laser excision. The objectives of this retrospective study were to assess local control relevant to tumor stage, anatomical location, amount of tissue removed, and pathological margins. Oncological outcomes were compared with those of our previously published reports3,4 of a series of patients with similar lesions treated with external partial laryngectomy performed by the same team of surgeons (D.F.B., E.M., and S.H.). We also discuss the role of further routine management for early-stage invasive cancer diagnosed by excisional biopsy findings and the difficulty of interpreting histological margins and their clinical relevance in early-stage glottic cancers excised with transoral laser cordectomy.
Fifty-two patients with clinical Tis or T1N0M0 glottic lesions (using the TNM staging system), determined by the 2002 Union Internationale Contre le Cancer classification,5 were treated between January 1991 and January 2004 with transoral carbon dioxide laser resection as first treatment. All patients who had been previously treated for the same lesion or previous cancer were excluded. There were 46 men and 6 women aged 34 to 83 years (mean age, 63 years). Sixteen patients were classified as having Tis, 30 had involvement of 1 vocal fold only with normal movement and were classified as having T1a tumors, and 6 were classified as having T1b tumors with involvement of both cords without impaired mobility. Six patients had anterior commissure involvement: 2, Tis; 2, T1a tumors; and 2, T1b tumors.
Staging was achieved by clinical examination and laryngoscopic examination using general anesthesia. Routine biopsy specimens were taken from lesions with obvious infiltration for histological confirmation of a diagnosis of squamous cell carcinoma. For superficial lesions, an excisional biopsy with a type I or II cordectomy, using the technique described in the following section, was performed.
All interventions in this series were performed using general anesthesia employing a Mallinckrodt Laser-Flex endotracheal tube (Athlone, Ireland) with a double cuff inflated with isotonic sodium chloride solution for airway maintenance. The larynx was exposed as widely as possible with suspension laryngoscopy and counterpressure on the cricoid cartilage as required for optimal visualization of the anterior commissure region. Examination of the anterior and subglottic regions was also facilitated by using 30° and 70° rigid endoscopes.
Microscopic examination with microinstrumental palpation was undertaken before performing the cordectomy. For superficial lesions, the submucosal space (superficial lamina propria) was infiltrated with isotonic sodium chloride solution following the technique described by Kass et al.6 If the mucosa was completely hydrodissected, a type I cordectomy was performed. If adherences were observed, a type II (subligamental) cordectomy was performed. Type III cordectomy was the minimal procedure performed to treat obviously invasive lesions. Better exposure, if required, was achieved with ipsilateral removal of the false vocal fold.
Cordectomy was performed using a carbon dioxide laser (model 1040; Sharplan, Tel Aviv, Israel), coupled to an operating microscope, on a power setting of 0.5 to 2 W in pulse or superpulse mode. The type of cordectomy was classified according to the system proposed by the European Laryngological Society Working Committee,7 summarized in Table 1>. In all cases, the excision was en bloc. The surgeon made careful, clinical assessment of the margins of resection with the operating microscope both during the resection and immediately on completion. No patient underwent prophylactic neck dissection.
Data relating to the initial lesion, including stage, size, anatomical site, surgeon, and type of cordectomy performed, were documented and reviewed retrospectively with additional information relating to histopathologic findings and follow-up.
Each patient was examined every 1 to 2 months the first year and subsequently every 2 to 4 months depending on progress and patient requirements. Flexible video nasolaryngoscopy was performed at each follow-up visit. Patients in whom invasive carcinoma was found when excisional biopsy was performed or in whom the histological margins were positive, close, or suspicious were not routinely retreated but were strictly and regularly followed up. Routine repeated rigid laryngoscopy using general anesthesia was not performed. The length of follow-up ranged from 6 to 93 months, with an average follow-up period of 38 months. Forty-four patients were followed up for more than 12 months, 29 patients for more than 24 months, and 23 patients for more than 36 months.
A summary of relevant anatomical details is presented in Table 2>. Two patients with Tis had distinct synchronous lesions. Three patients had both vocal cords involved without anterior commissure involvement, and based on the 2002 Union Internationale Centre le Cancer criteria5 and in accordance with the TNM rules, these tumors were classified as T1b rather than as synchronous T1a. All synchronous glottic lesions were treated simultaneously but were counted as a single lesion in the analysis. Six patients had anterior commissure involvement.
Type I cordectomy was the most common type of cordectomy used to remove Tis (13 of 16 lesions). Most T1 lesions were removed with a type II or type III resection (31 of 36). Table 3> summarizes the types of cordectomy used for each stage of tumor. For bilateral vocal fold lesions classified as T1b, only the larger resection is presented in our results.
The length of hospitalization ranged from 2 to 6 days (mean, 2.3 days). All patients were routinely admitted the day before surgery. The 1 patient with a 6-day stay also underwent removal of a benign lesion from the base of the tongue. Two patients experienced minor complications: 1 developed mild subcutaneous emphysema and the other, a subcutaneous hematoma following removal of an anterior glottic lesion with cricoid counterpressure. Both had slightly longer hospital stays than average but with no long-term sequelae.
Definitive histological findings and margins for all patients were obtained. Histological margins were defined as clear, close (tumor present <0.5 mm from the cut edge), suspicious (possibility of tumor at the margin but an artifact secondary to laser usage or tissue preparation prevented definitive assessment), or positive.
Six patients developed local recurrence 3 to 78 months after initial treatment. Four of the recurrences recurred within the first 12 months. The other 2 occurred at 30 months and 78 months, respectively. Two patients had a second recurrence, also treated with transoral laser cordectomy. The details of the recurrences and subsequent salvage therapy are presented in Table 4>. Three patients with recurrence initially had histologically clear margins. The association of recurrences with the histological margins is presented in Table 5>.
The rate of local control with 1 intervention was 91% (40 of 44 patients) at 12 months, 85% (24 of 29 patients) at 24 months, and 79% (18 of 23 patients) at 36 months. With further treatment, the rates of local control and laryngeal preservation were both 100%. No patients developed regional or metastatic disease. Only 1 death occurred in the series, at 86 months after initial treatment as the result of cerebral metastasis from an unrelated cancer. Given the distribution and length of follow-up, we did not perform Kaplan-Meier analysis or other statistical analyses.
This series supports the findings of other published studies8- 13 that have demonstrated the transoral carbon dioxide laser to be an effective treatment method for stage Tis and T1 glottic cancer. Particularly pertinent is the high rate of laryngeal preservation (93%-100%) associated with high rates of ultimate cancer control (96%-100%). Radiotherapy has also proved to be a viable alternative treatment for Tis and T1 glottic tumors because patients retain good voice quality. However, other series14,15 have demonstrated that the ultimate rate of total laryngectomy for patients initially treated with radiotherapy is significantly higher than in patients initially treated with surgery. Furthermore, even using narrow fields to limit radiation exposure, the physician's ability to safely reirradiate is limited. Radiotherapy is therefore a 1-time treatment. About 10% of patients with early-stage glottic cancer will develop a second primary upper aerodigestive cancer, potentially more advanced or aggressive.3,9,16 This second malignancy may be treated by radiotherapy if the patient has not been previously treated by radiotherapy.
Previous laser cordectomy does not preclude the use of external surgery or radiotherapy for recurrence or a second laryngeal primary tumor. This contributes to the high laryngeal preservation rate when transoral laser is used as the initial treatment modality. However, the laryngeal preservation rate for salvage by external or transoral surgery following radiotherapy is lower.9,10,14,15,17
This study can be compared with our previous articles3,4 on partial laryngectomy. The results for laryngofissure and cordectomy in 33 patients with T1 tumors confined to the midcord revealed a 100% initial control rate and 5- and 10-year ultimate control rates of 100%.3 No patient in this series required a tracheostomy or nasogastric tube. This argument is also used in favor of transoral laser treatment.9,11,12 In a second series of 60 patients treated with vertical partial laryngectomy involving a more heterogeneous group of tumors, including 7 cases with anterior commissure involvement, the rate of initial local control was 94.8%.4 Local recurrences were salvaged with supracricoid partial laryngectomy. The recurrences were in patients with anterior commissure involvement. The 5-year overall cancer control rate was 90%, and the laryngeal preservation rate was 100%.
In the study discussed herein, the initial control rate was 88% (46 of 52 patients). This is lower than in both of our external partial laryngectomy series.3,4 A difference in patient selection influenced this result. In the laryngofissure and cordectomy series, all the tumors were T1a and confined to the midcord. In this transoral laser series, 4 of the 6 recurrences had involvement of the anterior third of the vocal cord, a factor that has been associated with an increased recurrence rate.10,13 In this present laser series, however, if just the T1a subgroup is considered, the initial control rate is comparable with the rate of external partial laryngectomy, 97% (29 of 30 patients). The rate of local control with the inclusion of salvage surgery is comparable with our previously reported series.3,4 Ultimate control rates cannot be directly compared because of the differences in the length of follow-up. The rate of laryngeal preservation was 100% in all of these series.
The lower initial control rate in the present series is offset by the advantage that a tumor can be removed with a smaller and more targeted margin than that afforded by an external approach. The amount of tissue removed in the open procedures described earlier in this section (partial laryngectomies, ie, laryngofissure and cordectomy and vertical partial laryngectomy) equates to a type IV or V cordectomy. In the present series, including the salvage surgery, local control was achieved in 46 of 52 patients (89%) with a type I, II, or III cordectomy, including those having tumors with anterior commissure involvement. The amount of tissue removed correlated with vocal outcome. Our institution (Hôpital Européen Georges Pompidou, Paris, France) is currently undertaking a prospective study to evaluate this. Because this is a retrospective study, no systematic assessment of vocal function was undertaken; however, voice results described in the literature for types I and II cordectomy are excellent17- 19 and for type III are good.19 Thus, voice preservation for these early-stage cancers seems to be optimized with laser resection compared with an external approach. There is currently no evidence to demonstrate that the quality of voice after type IV or V cordectomy is any better or worse than after an external procedure with primary reconstruction.4
In the present series, early-stage invasive disease determined by excisional biopsy findings and microscopically positive, close, or suspicious margins were not routinely reexcised but were closely monitored with strict and regular review. This option was chosen for a number of reasons.
The glottic area is a region that, in most patients, is readily observed and reveals earlier symptoms than other areas, such as the supraglottis or the hypopharynx.
There is no compelling evidence that defines an adequate margin for laser excision of early-stage glottic lesions. A range of 2 to 5 mm of healthy tissue seems to be the suggested margin, but this has not been proved.9
No conclusive data exist to suggest that the ultimate outcome is significantly altered by further immediate treatment. Peretti et al13 found that positive margins made no statistical difference in 5-year disease-free survival rates. Only 23 of 45 patients with positive histological margins received complementary therapy, either reexcision or radiotherapy, in their series.
Patient compliance with follow-up was 100%. Noncompliance or inability to adequately visualize the larynx with nasolaryngoscopy would be a contraindication to this plan of management.
The highest recurrence rate was with a type I cordectomy, which was performed on 13 of 16 patients with Tis. Three patients had local recurrence: 1 with clear margins, 1 with close margins, and 1 with a positive margin. Subsequently, the patient with clear margins required an external partial laryngectomy for recurrent disease, whereas the other 2 recurrences were salvaged with repeated transoral laser treatment. Carcinoma in situ can be multifocal, which may explain the higher recurrence rate and why the patient with clear margins required more aggressive intervention for salvage.10 There is an ongoing discussion10,13,18 about the adequate treatment of Tis tumors. Follow-up is of prime necessity. Interestingly, there was only 1 local recurrence in a patient in the T1a group who had undergone excisional biopsy with a type II cordectomy and 1 in the group who underwent a type III cordectomy. This would suggest that in small T1a tumors, a type II cordectomy may be appropriate, provided close follow-up is undertaken. Overall, only 3 (16%) of the 18 patients with positive, close, or suspicious margins developed local recurrence, including just 1 (25%) of 4 with positive margins (Tis). Follow-up in this subset was for a minimum of 18 months, and 13 patients (72%) were followed up for more than 36 months. All 3 of these recurrences were treated with repeated transoral laser administration. Routine reexcision or adjunct therapy would have meant that 15 patients (84%) in this category received potentially unnecessary treatment. Radiotherapy after glottic transoral surgery has been associated with poorer vocal outcome than surgery or radiotherapy alone.20 The 100% rate of laryngeal preservation supports the appropriateness of this approach.
The preconception that suspicious histological margins are associated with an inevitably high rate of recurrence is reasonable. However, in our series and in that by Peretti et al,13 a very high recurrence rate in this group of patients was not observed. To our knowledge, there is no other comment in the literature specifically addressing this phenomenon. Two reasonable explanations can be postulated to explain why lesions with histologically inadequate margins do not have the anticipated higher recurrence rate. First, problems with assessing the margins may result in false inadequate margins. The small size of the excised tissue can make assessment of margins technically difficult. Tissue contraction caused by the elastic fibers of the subepithelial connective tissue, enhanced by the thermal effect of the laser, will result in a margin being smaller in vitro than in vivo.21 Artifacts from laser treatment, especially coagulation, can interfere with the pathologists' ability to comment confidently on the presence or absence of tumor cells at or near the margin.21,22 The use of frozen section margins taken from the patient has been described,21 but this still has the potential technical difficulties associated with dealing with very small pieces of tissue and the possibility of interference from laser artifacts. For these reasons, frozen sections are not used routinely at our institution. Second, an innate property of the tumor may be responsible for this observation. Molecular biology has shown laminin 5, a glycoprotein constituting a major component of the basement membrane with its γ2 chain expression, to be a sensitive marker of invasiveness in various carcinomas. Nordemar et al23 demonstrated a highly statistically significant association between laminin 5 expression and progression of Tis lesions of the larynx to invasiveness. Nakayama et al24 have recently shown that increased expression of laminin 5 in hypopharyngeal cancer correlates with its invasive and migratory properties and that the overexpression is not uniform throughout the tumor. It could be hypothesized that an inadequate margin in a tumor with high expression of laminin 5 at the margin in question is associated with a higher risk of recurrence and may allow the determination of patients who require further treatment. Immunohistological and molecular margins do not seem to be altered by laser effects.25 The possibility of applying the progress of molecular biology to clinical practice is an area for further research.
Surgeons' evaluation of the margins of resection at the time of cordectomy is of utmost importance. Histological margins remain important because our results suggest that the risk of local recurrence with clear margins is low. Positive pathological margins provide a guide as to which patients may require a more intensive follow-up but do not reliably predict local recurrence.
In conclusion, this study supports the role of transoral laser surgery in treating early-stage glottic cancer. It is comparable with the follow-up results of other reported transoral laser series1,2,8,9,12,13,21 and with the results of our previous series3,4 of external partial laryngectomies.
Type II cordectomy may be adequate treatment for Tis or very small T1a tumors diagnosed by excisional biopsy findings. This approach allows for maximal preservation of healthy tissue and, in theory, optimal functional outcome. In 89% of patients, the cancer was ultimately controlled by a type III cordectomy or smaller resection.
Routine reexcision or other additional treatment for inadequate margins may not be necessary in early-stage glottic cancers if certain strict criteria for follow-up are met. Eighty-four percent of patients in this subgroup did not require further treatment after a minimum of 18 months of follow-up. The vocal cords of all 3 patients in this group who developed local recurrence were successfully treated with further transoral laser excision.
Correspondence: Daniel F. Brasnu, MD, Department of Otorhinolaryngology, Head and Neck Surgery, Hôpital Européen Georges Pompidou, Assistance Publique des Hôpitaux de Paris, University Paris V, 20 rue Leblanc, 75015 Paris, France (firstname.lastname@example.org).
Submitted for Publication: May 22, 2005; final revision received August 4, 2005; accepted September 7, 2005.
Financial Disclosure: None.