Long-Term Survival After Resection of Primary Adenoid Cystic and Squamous Cell Carcinoma of the Trachea and Carina

Transcription

1 ORIGINAL ARTICLES: Long-Term Survival After Resection of Primary Adenoid Cystic and Squamous Cell Carcinoma of the Trachea and Carina Henning A. Gaissert, MD, Hermes C. Grillo, MD, M. Behgam Shadmehr, MD, Cameron D. Wright, MD, Manjusha Gokhale, MA, John C. Wain, MD, and Douglas J. Mathisen, MD Division of Thoracic Surgery and Center for Clinical Effectiveness in Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, and Policy Analysis Inc, Brookline, Massachusetts Background. Tracheal resection for primary carcinoma may extend survival. We evaluated survival after surgical resection or palliative therapy to identify prognostic factors. Methods. We conducted a retrospective study of patients diagnosed with primary adenoid cystic carcinoma () or squamous cell carcinoma () of the trachea between 1962 and Laryngotracheal, tracheal, or carinal resection was performed when distant metastasis and invasion of adjacent mediastinal structures were absent and tumor length permitted. Radiotherapy was administered after operation (54 Gy), except in superficial tumors, or as palliation (60 Gy). Results. Of 270 patients with or (135 each), 191 (71%) were resected. Seventy-nine were not resected due to tumor length (67%), regional extent (24%), distant metastasis (7%), or other reasons (2%). Overall operative mortality was 7.3% (14/191) and improved each decade from 21% to 3%. Tumor in airway margins was present in 40% (17/191) of resected patients ( 59% versus 18%) and lymph node metastasis in 19.4% (37/191). Overall 5- and 10-year survival in resected was 52% and 29% (unresectable 33% and 10%) and in resected 39% and 18% (unresectable 7.3% and 4.9%). Multivariate analysis of long-term survival found statistically significant associations with complete resection (p < 0.05), negative airway margins (p < 0.05), and adenoid cystic histology (p < 0.001), but not with tumor length, lymph node status, or type of resection. Conclusions. Locoregional, not distant, disease determines resectability in primary tracheal carcinoma. Resection of trachea or carina is associated with long-term survival superior to palliative therapy, particularly for patients with complete resection, negative airway margins, and. (Ann Thorac Surg 2004;78: ) 2004 by The Society of Thoracic Surgeons The surgical treatment of primary tracheal carcinoma was conceived with new surgical techniques and an adaptation of oncologic principles. Complete, en-bloc resection as applied in other solid tumors was modified to account for adjacent mediastinal structures, finite tracheal length, and segmental tracheal blood supply. Surgical advances consisted of improving techniques of tracheal mobilization and anastomosis, gaining judgment to assess maximal resectable airway length, and adding adjuvant therapy. Tracheal carcinomas are considered for resection when localized, but selection criteria are not uniform. Their low incidence has obviated prospective studies to evaluate and compare therapy. Alternatives, including primary radiotherapy, photodynamic therapy, Accepted for publication May 20, Presented at the Fortieth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 26 28, Address reprint requests to Dr Gaissert, Division of Thoracic Surgery, Massachusetts General Hospital, Blake 1570, 55 Fruit St, Boston, MA 02114; hgaissert@partners.org. and stenting, have not improved survival and remain of palliative value. Long-term survival after surgical resection of malignant tracheal tumors has been previously demonstrated [1 4]. The benefit of surgical therapy is less clear when resection is incomplete or lymph node metastasis present. In a preliminary analysis of tracheal squamous cell carcinoma (), Grillo and Mathisen [5] showed a negative impact on survival when tracheal margins or lymph nodes were involved. For adenoid cystic carcinoma (), a French cooperative study [6] demonstrated higher survival after complete resection, whereas Maziak and colleagues [7] did not find a significant difference in survival between complete and incomplete resection. A study of the two most common primary tracheal tumors, and, was conducted to examine presentation and resection rates, to further elucidate factors predictive of long-term survival, and to observe the role of incomplete resection and lymph node metastasis by The Society of Thoracic Surgeons /04/$30.00 Published by Elsevier Inc doi: /j.athoracsur

2 1890 GAISSERT ET AL Ann Thorac Surg SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA 2004;78: Abbreviations and Acronyms adenoid cystic carcinoma CR carinal resection CT computed tomography Gy gray LTR laryngotracheal resection MGH Massachusetts General Hospital squamous cell carcinoma TR tracheal resection Patients and Methods Patient Selection A retrospective analysis was conducted at Massachusetts General Hospital (MGH) of all patients evaluated for primary and of the trachea since Medical records and Department of Pathology databases were searched to identify tracheal tumors. Patients who underwent their first resection elsewhere were included if the purpose of the second procedure at MGH was to complete tumor excision. Pathology reports, operative reports, and hospital charts were reviewed to exclude laryngeal cancer and tumors of the lung extending to the carina. The survival period began on the day of operation in resectable patients and the day of bronchoscopy in unresectable patients. The MGH institutional review board approved this protocol in October Consent was received from patients before obtaining follow-up information. portion of infraglottic larynx while preserving at least one recurrent laryngeal nerve not involved with tumor. Reconstruction was achieved by shaping the remaining trachea to conform to the laryngeal defect. Carinal resection was performed with or without concomitant lung resection. Carcinomas involving the carina by extension from the lung or a mainstem bronchus were excluded from analysis. Systematic lymph node dissection was assumed to interfere with the tracheal blood supply and was therefore not performed. Regional lymph nodes were often not included in the specimen if not grossly enlarged. Absence of tumor at the airway margins was confirmed by frozen section unless the limits of resection had been reached and no additional trachea could be removed. The resection was judged to be complete when airway margins were found to be disease free and no statement in the operative note indicated that the resection was incomplete. Tumor Data Tumor dimension in the long axis of the airway, depth of invasion, and the presence of tumor at airway or radial margins were recorded in resected patients from a review of pathology reports. Positive airway margins indicated true residual, most often microscopic, tumor. In contrast, the radial margins were interpreted as positive when tumor was seen within 1 mm of the margin, usually indicating that only a thin layer of connective tissue covered the outside of the tumor. Tumor involvement of excised lymph nodes was noted. Determination of Resectability Patients underwent an assessment of locoregional and distant metastatic disease before resection. Radiographic evaluation included conventional tomography of trachea and carina [8] and computed tomography once available [9]. Every patient underwent bronchoscopy. The pathologic diagnosis was made from bronchoscopic biopsies or the resected specimen. Tracheal resection was considered only when complete resection seemed feasible. Microscopic involvement of airway margins was accepted if the airway was normal on gross inspection and no further length of airway could be resected. Resection was not performed when the procedure would have resulted in grossly positive peritracheal margins. A tumor was also considered unresectable when metastatic disease was present, the length of involved airway at bronchoscopy was judged to preclude safe reconstruction, or when invasion of certain adjacent organs was found during operative exploration. Esophageal invasion was resected in some patients. Types of Resection Resections were grouped by anatomic region and extent. For standard tracheal resection, a sleeve of trachea was removed with end-to-end reconstruction. When tracheal resection was combined with laryngectomy, cervical or mediastinal end-tracheostomy was required. For laryngotracheal resection, the tumor was removed with a Adjuvant and Primary Radiotherapy Adjuvant postoperative radiotherapy was introduced during the second decade of this experience for close or tumor-bearing margins. After bronchoscopic assessment of anastomotic healing, a recommended dose of 54 Gy was administered 6 weeks after resection. Primary radiotherapy with 60 Gy was recommended for patients who had unresectable tumors, but could not be confirmed in all patients. Most patients received treatment outside MGH; a report from our hospital [10] found a median postoperative dose of 60 Gy. Follow-Up Patients and their physicians were contacted for follow-up information. The social security death index was searched. Patients were determined to have died if name, date of birth, and social security number matched. Follow-up information on treatment and tumor status in patients who did not undergo resection was often limited to the date of death. Statistical Analysis Patient averages for dichotomous and continuous variables were calculated overall and for tumor type ( and ). Where appropriate, statistically significant differences were between and were tested using the 2 statistic (dichotomous variables) and the t-statistic (continuous variables). Finally, to better under-

3 Ann Thorac Surg GAISSERT ET AL 2004;78: SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA Table 1. Characteristics of 270 Patients With Primary Tracheal Carcinoma stand factors associated with long-term survival, we calculated univariate logistic analyses, multivariate logistic analyses, and survival analyses. The multivariate analyses predicted the likelihood of survival at 5 and 10 years based on clinical relevant factors (for tumor type, tumor length, type of resection, age, completeness of resection, positive lymph nodes, smoking status, and airway margin). The survival analysis depicted predicted survival based on patients with complete follow-up. All analyses were calculated using SAS version 8.0 (SAS Institute, Carey, NC). Results 2 p Value Sex Female 53% 32% Male 46% 68% Mean age (median), years 49 (47) 61 (62) Comorbidities (%) Smoker Hypertension EtOH use Past MI Diabetes mellitus Steroid use Angina Arrhythmia Prior stroke Prior cancers (%) Lung Larynx Head and neck Colon Prostate Oropharynx Other Symptoms (%) Dyspnea Cough Hemoptysis Wheeze Stridor Hoarseness Dysphagia Fever Other adenoid cystic carcinoma; EtOH alcohol; MI myocardial infarction; squamous cell carcinoma. A total of 270 patients (135 patients with and 135 with ) with primary tracheal carcinoma were evaluated between 1962 and Ten patients with nonsquamous bronchogenic carcinoma were excluded from analysis. As shown in Table 1, patients with were more likely to be male, smokers, have prior cancers of the respiratory tract, and present with hemoptysis. Patients with presented more often with dyspnea and wheezing. The mean duration of symptoms was 12.2 months, with longer durations in and in tumors later found to be unresectable (resected 18.3 months, resected 4.54 months; unresectable 23.7 months, unresectable 7.58 months; p 0.001). Thirty-four patients had prior thoracic surgical procedures including nine tracheal resections and 12 lung resections. Twenty-three patients (8.5%) received thoracic radiation after diagnosis and before presentation; after presentation, 14 underwent resection at MGH. Unresectable Tumors Contraindications precluded resection in 34 (25.2%) patients with and 45 (33.3%) patients with. Unresectable tumors were found during thoracotomy in 11 patients and cervical or cervicomediastinal exploration in 6. The leading contraindication was locoregional disease in more than 90% of unresectable patients, whereas distant metastasis was present in less than 7% (Table 2). Immediate management in selected patients included endoscopic resection before radiation. Five patients underwent T-tube placement, 2 tracheostomy, and 2 placement of self-expanding airway stents. Resected Tumors One hundred one patients with (75%) and 90 patients with (67%) underwent resection. There were 103 tracheal resections, 41 parenchyma-sparing carinal resections, 20 combined carinal and pulmonary resections, 15 laryngotracheal resections, and 12 tracheal resections with laryngectomy (7 patients) or surgical separation of larynx and trachea to transfer tracheal length for intrathoracic reconstruction (tracheal devolvement, 5 patients). Table 3 lists the type of resection for each tumor and hospital mortality. One hundred eightyfour patients underwent single-stage reconstruction, 5 had multiple stages, and 2 underwent resection for residual disease after previous resection elsewhere. With increasing experience and improved surgical technique over four decades, the resection rates for both tumors increased, the total number of resections rose, hospital mortality decreased, and certain extreme reconstructive techniques to bridge long defects (skin tube reconstruction and tracheal devolvement) were largely abandoned (Table 4). Table 2. Contraindications to Resection in Unresected Patients Reason for Contraindication (%) (n 34) (n 45) Extent of airway involvement Extent of regional disease Distant disease 6 7 Medical contraindication 0 2 Patient choice 3 0 adenoid cystic carcinoma; 1891 squamous cell carcinoma.

4 1892 GAISSERT ET AL Ann Thorac Surg SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA 2004;78: Table 3. Types of Resection for Primary Tracheal Carcinoma Type n % n % Total % Laryngotracheal Tracheal Tracheal with permanent tracheostomy Carinal without pulmonary resection Carinal with pulmonary resection Total adenoid cystic carcinoma; squamous cell carcinoma. The hospital mortality was 10% (10/101) in and 4.4% (4/90) in. Two thirds of carinal resections, the procedure with the highest mortality (laryngotracheal resection 0; tracheal resection 3.9%; carinal resection 16%) were performed for. The cause of death was related to the anastomosis in 8 patients and to respiratory failure in 6. The early high mortality was in part attributable to complex staged procedures, which were later abandoned. Anastomotic complications, fatal and nonfatal, occurred in 28 patients (14.6%; separation 7.3%, stenosis 3.1%, granulations 2.1%, tracheoarterial fistula 1.0%, necrosis 0.5%, and other 0.5%). Postoperative tracheostomy was required in 17 patients and remained temporary in 13. Length and depth of resected tumors and status of resection margins are listed in Table 5. Adenoid cystic carcinomas were longer and extended more often into peritracheal soft tissues or to adjacent organs, most commonly the esophageal muscle coat. Tumor was therefore more often present at or close to the radial resection margin, and tracheal margins frequently contained microscopic tumor. Fifty-nine percent of patients with had positive airway margins, in contrast to only 18% with. The lymph node status after 191 resections is listed in Table 6. Selective sampling established lymph node metastasis in 19.4% (37/191) of patients, most commonly from peritracheal and subcarinal stations; nodal biopsies were not obtained in 35%. Postoperative radiation could be confirmed in 70% of patients (71/101) with and 46% of patients (41/90) with. Long-Term Follow-Up Follow-up of survival was complete in 88.5% (239/270) of patients. The overall survival for all patients with primary tracheal carcinoma was 84% at 1 year, 45% at 5 years, and 25% at 10 years. Mean survival time was 38 months with resected, 8.8 months with unresectable, 69 months with resected, and 41 months with unresectable. Figure 1 depicts the survival by tumor type and resection status. At 5 years, 39.1% of patients with resected and 7.3% of patients with unresectable survived. Five-year survival was 52.4% for resected and 33.3% for unresectable. A comparison of survival after laryngotracheal, tracheal, and carinal resections is shown in Figure 2. Despite higher hospital mortality after carinal resection, differences in long-term survival between these types of resection were not significant. Survival according to airway margin status is depicted in Figures 3 and 4. Complete resection with negative airway margins resulted in higher survival compared with incomplete resection or unresectable tumors for both and, although in there were few patients with tumor-bearing airway margins and the difference from tumor-free margins was not significant. In, the survival after incomplete resection separates from unresectable tumors after 10 years; we know of no survivor beyond 13 years in the unresectable group, whereas survival after incomplete resection at 15 years was 14.5%. Figure 5 shows the survival according to lymph node status. Lymph node-positive had a markedly lower survival than tumors without nodal metastasis. No cor- Table 4. MGH Experience With Primary Tracheal Carcinoma Over 4 Decades Decade Total Patients (n) Patients (n) Patients (n) Resections Performed (n) Resection Rate (%) Hospital Mortality (%) Total Two patients operated in 2002 were excluded from the Table. adenoid cystic carcinoma; MGH Massachusetts General Hospital; squamous cell carcinoma.

5 Ann Thorac Surg GAISSERT ET AL 2004;78: SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA Table 5. Length and Depth of Resected Tumors, and Status of Resection Margins Descriptor Tumor length (cm), mean 3.2 ( ) 2.4 ( ) (95% confidence interval) Tumor depth (%) Mucosa and submucosa Limited to airway wall Peritracheal soft tissue Adjacent organ Depth not known Margin status (%) Only tracheal margin positive 8 10 Only radial margin positive Tracheal and radial margin positive 51 8 Any tracheal margin positive Total positive margins Any tracheal margin was positive if either cephalad or caudad tracheal margins, or mainstem bronchial margin contained tumor. Total positive margins denote tumor involving or approaching within 1 mm either the radial or tracheal surgical margin. adenoid cystic carcinoma; squamous cell carcinoma Fig 1. Actuarial overall survival by resection status and tumor type. ( adenoid cystic carcinoma; squamous cell carcinoma.) prostate, and other tumor in 1 patient each; in, lung and breast cancer developed in 1 patient each. relation was noted between lymph node status and survival in. Seven airway tubes were placed late after resection: 4 for local recurrence, 2 for aspiration, and 2 for a radiation-induced stricture. The logistic multivariate analysis of resected patients (Table 7) identified, complete resection, and negative resection margins to be associated with 5-year survival and, complete resection, and age to be associated with 10-year survival. Sensitivity analyses showed that either negative pathologic airway margins or complete resection as determined by the surgeon were associated with improved survival when run in separate models. Recurrence Table 8 lists vital status and recurrence in 177 operative survivors. Of 55 patients with recurrence, the site was locoregional only in 26, distant in 16, combined in 7, and unstated in 7 patients. New primary malignancy was observed in 14 patients. In, lung cancer occurred in 6, head and neck carcinoma in 2, and leukemia, colon, Comment This study describes the outcome in a consecutive series of patients undergoing evaluation and treatment for primary tracheal carcinoma. More than two thirds of the study group underwent extirpation of their malignancy, surpassing resection rates of 10% reported in epidemiologic studies. The wide gap between these rates lends credence to the reproach by Licht and associates [11] who perceived a nihilistic attitude based on ignorance toward tracheal resection in one country s experience, although selection bias probably lowered referral to us of patients with distant disease. The typical patient with tracheal cancer in this study had tumor-related symptoms for 12 months, resected patients had a mean tumor Table 6. Lymph Node Status in 191 Patients Undergoing Tracheal Resection for Primary Carcinoma (n) (n) Total (%) No lymph node biopsy Tumor-free lymph nodes One lymph node positive More than one lymph node positive Total adenoid cystic carcinoma; squamous cell carcinoma. Fig 2. Actuarial survival by type of resection. (CR carinal resection; LTR laryngotracheal resection; TR tracheal resection.)

6 1894 GAISSERT ET AL Ann Thorac Surg SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA 2004;78: Fig 3. Actuarial survival in adenoid cystic carcinoma by airway margin. Note 15-year observation interval. ( adenoid cystic carcinoma.) length of 2.8 cm, and only 13 of 79 unresectable patients had distant metastasis. The sum of these observations suggests that most patients with tracheal cancer, in contrast to peripheral lung cancer, have limited disease when symptoms occur and are therefore good candidates for surgical therapy. Nonspecific symptoms, however, delay diagnosis and this delay may affect therapy as suggested by the difference in symptom duration between resected and unresectable patients. We found that complete resection in and was associated with improved survival. A retrospective French multicenter study of 208 patients had also demonstrated the value of complete resection in [6], but not so a study describing the experience with 38 patients Fig 4. Actuarial survival in squamous cell carcinoma by airway margin. ( squamous cell carcinoma.) Fig 5. Actuarial survival of primary tracheal carcinoma by lymph node status. ( adenoid cystic carcinoma; squamous cell carcinoma.) at the University of Toronto [7]. The French report did not find a decrease of survival for patients with positive lymph nodes, which we observed in. However, we did not find lymph node involvement predictive of survival in the multivariate analysis. These seemingly conflicting data were based on small numbers of lymph node-positive patients, and the multivariate analysis may have attributed any effect on outcome to squamous histology. The multivariate analysis did not establish other variables of disease extent as predictors of survival in either tumor, such as tumor depth or tumor length. These findings indicate the problems in characterizing basic oncologic relationships in primary tracheal carcinoma due to its low incidence and suggest a current lack of data to formulate a precise staging system. Current staging methods have limited accuracy in defining the locoregional extent of disease. Computed tomography does not predict mediastinal organ involvement and underestimates the longitudinal reach of tumor in the airway [9]. Endoscopy supplemented by mucosal biopsies provides at present the best assessment of luminal tumor length. Although endoluminal ultrasound may distinguish reliably between compression and infiltration of the trachea by an extrinsic tumor [12], detection of submucosal infiltration, by far the most common cause of positive tracheal resection margins, has not been reported. Despite the shortcomings of present locoregional staging, only 17 of 208 patients (8.2%) underwent operative exploration without resection. A comparison of and of the trachea highlights their commonalities and distinctions. Both tumors were referred with equal frequency and together constituted more than 95% of primary tracheal carcinomas. In contrast, epidemiologic studies, without review of pathology or radiographs, consistently find other bronchogenic carcinomas in 15% to 20% and adenoid cystic

7 Ann Thorac Surg GAISSERT ET AL 2004;78: SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA Table 7. Influence of Perioperative Variables on Survival at 5 and 10 Years No. Overall Survival at 5 Years Survival at 10 Years Average/ Percent Bivariate OR (CI) Multivariate OR (CI) Bivariate OR (CI) Multivariate OR (CI) 1895 Histology % 2.59 ( ) 8.76 ( ) 3.19 ( ) 7.58 ( ) (REF) 90 47% p Tumor length ( 1.4) 0.84 ( ) 0.76 ( ) 0.85 ( ) 0.75 ( ) p Type of resection CR % 3.59 ( ) 8.10 ( ) 2.29 ( ) ( ) LTR/TR (REF) 16 8% p Age ( 15.4) 0.98 ( ) 0.98 ( ) 0.96 ( ) 0.96 ( ) p Resection Complete 76 41% 2.69 ( ) ( ) 3.04 ( ) 1.55 ( ) Other (REF) % p Lymph node status Not positive % 0.27 ( ) 0.60 ( ) 0.26 ( ) 0.55 ( ) Positive (REF) 42 22% p Smoker No 67 35% Yes % 1.06 ( ) 1.35 ( ) 1.04 ( ) 1.56 ( ) p Airway margin Negative % 0.71 ( ) 7.13 ( ) 0.64 ( ) 9.55 ( ) Other (REF) 76 40% p adenoid cystic carcinoma; CI confidence interval; CR carinal resection; LTR laryngotracheal resection; OR odds ratio; REF reference; squamous cell carcinoma; TR tracheal resection. tumors in fewer than 15% of patients [11, 13]. A difference in growth rate may introduce lead-time bias but offers no satisfactory explanation for this disparity. Both tumors have high resection rates and share locoregional disease as the most common contraindication to resection. Effects of smoking, a higher age at presentation, and growth rates combine so that 43% of patients with but only 10% of patients with survive beyond 5 years. We did Table 8. Vital Status and Recurrent Disease in 177 Patients Surviving Tracheal Resection for Primary Carcinoma (n) (n) Total (%) Died of disease Died without evidence of disease Died, unknown status Alive with disease Alive without evidence of disease Incomplete follow-up Total adenoid cystic carcinoma; squamous cell carcinoma. not determine whether incomplete resection is of curative or merely palliative value; the number of patients was too small. Although usually discovered when already committed to resection and thus of limited usefulness in determining resectability, we assume that incomplete resection due to positive tracheal margins continues to be justified in selected patients with in the hands of experienced surgeons. The operative risk is low, a small although statistically not significant survival benefit is observed, and it is our impression that death occurs from metastatic disease rather than airway obstruction. Because locoregional tumor is the most important determinant of resectability and current resection methods rely on primary reconstruction with trachea, any further increase of the resection rate will depend on the development of a safe tracheal replacement. References 1. Perelman M, Korolyova N. Surgery of tumours in the thoracic portion of the trachea. Thorax 1968;23: Eschapasse H. Les tumeurs trachéales primitives: traitement chirurgical. Rev Fr Mal Respir 1974;2:

8 1896 GAISSERT ET AL Ann Thorac Surg SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA 2004;78: Pearson FG, Thompson DW, Weissberg D, Simpson WJ, Kergin FG. Adenoid cystic carcinoma of the trachea. Experience with 16 patients managed by tracheal resection. Ann Thorac Surg 1974;18: Grillo HC. Tracheal tumors: surgical management. Ann Thorac Surg 1978;26: Grillo HC, Mathisen DJ. Primary tracheal tumors: treatment and results. Ann Thorac Surg 1990;49: Regnard JF, Fourquier P, Levasseur P. Results and prognostic factors in resections of primary tracheal tumors: a multicenter retrospective study. The French Society of Cardiovascular Surgery. J Thorac Cardiovasc Surg 1996;111: Maziak DE, Todd TR, Keshavjee SH, Winton TL, Van Nostrand P, Pearson FG. Adenoid cystic carcinoma of the airway: thirty-two-year experience. J Thorac Cardiovasc Surg 1996;112: Weber AL, Grillo HC. Tracheal tumor: radiological, clinical and pathological evaluation. Adv Otorhinolaryngol 1978;24: Spizarny DL, Shepard JA, McLoud TC, Grillo HC, Dedrick CG. CT of adenoid cystic carcinoma of the trachea. AJR Am J Roentgenol 1986;146: McGrath JJ, Choi NC, Mathisen DJ, et al. Role of radiation therapy in the management of tracheal tumors [Abstract]. Int J Radiat Oncol Biol Phys 2001;51: Licht PB, Friis S, Pettersson G. Tracheal cancer in Denmark: a nationwide study. Eur J Cardiothorac Surg 2001;19: Herth F, Ernst A, Schulz M, Becker H. Endobronchial ultrasound reliably differentiates between airway infiltration and compression by tumor. Chest 2003;123: Manninen MP, Antila PJ, Pukander JS, Karma PH. Occurrence of tracheal carcinoma in Finland. Acta Otolaryngol 1991;111: DISCUSSION DR DONNA E. MAZIAK (Ottawa, Ontario, Canada): I would like to commend you on this landmark report in which you have presented the largest series and the longest follow-up in this relatively rare disease. I have a few questions. First, looking at the lymph node status, you reported, and maybe I misunderstood, that patients with adenoid cystic who were lymph-node-negative did worse in terms of survival than people who were lymph-node-positive. Did you find most of your lymph nodes that were positive were more from direct extension rather than from a separate site away from the tumor? And did you look at perineural spread in patients with adenoid cystic carcinoma and the effect on survival? Quite often perineural spread is more common than lymphatic. With regard to your cut-off for length of tumor that was resected, I think the longest was 3.5 cm. Sometimes longer resections are possible, and in our series with Dr Pearson we resected even up to 8 cm. I was wondering if you did no resections longer than 3.5 cm by choice or if that length was the standard cut-off at your institution. Also, did you notice if any patients who received radiation went from being unresectable to resectable? With regard to radiation, only 70% of your patients received postoperative radiation. With your findings and your conclusions, do you think now that most likely everyone should get radiation postoperatively? Again, I would like to thank you for your report and for your presentation of these patients. DR GAISSERT: Doctor Maziak, thank you very much for your questions. With regard to the apparent better survival for lymph-nodepositive patients in adenoid cystic carcinoma, there were so few patients and they did not have an early decline in their survival as did the lymph-node-negative patients, so we suspect that this particular selection of patients underwent resection at a later time after an initial high mortality. I doubt that this survival difference can be attributed to lymph node disease. It was our impression that most lymph nodes were not involved by direct extension but that they had tumor deposits by lymphatic spread. We did not examine the specific item of perineural spread in a separate manner. We reported tumor length as a mean with standard deviation. Some patients had resections of up to 7 or 8 cm, but most of those underwent a laryngectomy and a mediastinal tracheostomy at the same time. Finally, with regard to preoperative radiation, we accepted patients with preoperative radiation and examined whether they were candidates and then treated them differently from other patients in that we wrapped the anastomosis in vascularized tissue, such as omentum, but because of our own bias, we did not treat patients with preoperative radiotherapy. DR ROBERT J. KORST (New York, NY): I enjoyed the presentation immensely. What are the patients dying of? Specifically what are the recurrence patterns? Are you seeing a lot of local recurrence? Are there distant failures? DR GAISSERT: I would like to show the next slide, if you will allow me. [Shows slide.] To answer your question, Dr Korst, this slide shows the vital status and the recurrence pattern as best as we can determine, and this is the weakness of a retrospective study that looks at a long time period during which patients have been referred mostly from other regions and we sometimes have difficulty getting the data. You see from this slide that a fairly large number of patients died with an unknown status or had social security numbers that were not deactivated and they were lost to follow-up. Most of the 23% of patients who died of disease had distant metastatic disease; in the case of adenoid cystic carcinoma, most metastases were to lung, bone, and brain. Regarding the frequency of local recurrence, we had 18 patients with squamous cell carcinoma who had a local recurrence compared with 5 patients with adenoid cystic carcinoma. We did not find, by the way, a correlation with postoperative radiotherapy, so we could not identify an effect, positive or negative, of postoperative radiotherapy. DR GIUSEPPE CARDILLO (Rome, Italy): Some endoscopists perform an endoscopic resection of the tumor before sending patients to the surgeon. We believe that this procedure is dangerous. What is your opinion about an endoscopic resection before definitive surgery? DR GAISSERT: We have performed endoscopic resection in some instances in which we wanted to delay treatment because the patient was still on steroids. At that time we did the endoscopic resection with bronchoscopic methods and without the use of laser. We have a great concern about using laser for

9 Ann Thorac Surg GAISSERT ET AL 2004;78: SURVIVAL AFTER RESECTION OF THE TRACHEA AND CARINA endoscopic resection because the laser may extend the damage and thus convert a potentially resectable lesion to an unresectable lesion. So we discourage the use of laser. DR JOSEPH B. SHRAGER (Philadelphia, PA): This is a beautiful study, and, as everyone knows, it could not come from anywhere but Massachusetts General Hospital given the volume of patients undergoing tracheal surgery up there. For the adenoid cystic group, something that really struck me is that the survival curves for those who had a positive airway margin and the nonsurgical cases were basically overlapping. Apparently operation confers no survival benefit on patients who have a positive airway margin, thus perhaps you could opt to not operate on those patients. So the question I have and I m not aware of any data on the issue so maybe you can shed light on this is whether there are any data that tell you the mean microscopic extension of disease beyond the gross tumor. If you knew what the extension beyond the gross disease is likely to be, then you could choose which patients might be reasonably approached surgically and which would be better served by radiation alone. DR GAISSERT: Thank you, Dr Shrager. That is an excellent question and one that is important to us: Can you predict whether a resection is going to be complete or incomplete on the airway? We do not think you can. You may say that a length of tumor likely will result in a positive airway margin, but you cannot predict with certainty whether eventually that is going to be the case. And you find out when all the bridges are burned and you are essentially committed to reconstruction. The next question is whether an incomplete resection has a palliative value. I have not shown this here and I do not think that we have all the data to show that. Nevertheless, with an operative mortality of 3%, I think one can make a case for palliative resection, even though only a small group of patients who undergo an incomplete resection reach the 15-year mark. Some do, but we do not know who does and who does not DR SHRAGER: I believe I see the back of Dr Grillo s head a few rows in front of me. If Dr Grillo is here, I would love to hear him comment. DR GAISSERT: Doctor Grillo, would you like to comment on this question? DR HERMES C. GRILLO (Boston, MA): I do not know that I have that much to add to the excellent presentation Dr Gaissert made, but I think this study does show some of the problems of the rarity of the disease. It has taken 40 years to make this accumulation, and, as you point out, there are problems inherent in follow-up, especially when the patients have come from basically all over the world, and it is difficult to extract some of the refined questions that have been asked from these data or to set up anything prospectively that is ever going to answer some of those questions. So I regret that we cannot answer them all. Insofar as the problem of whether it is worth doing a resection if there is a positive margin, first of all, as Dr Gaissert pointed out, you have no option. In that sense you have reached that point, and at that point you have to stop. Early on we did not stop, and you can see what the mortality was. However, we do know that death from airway obstruction is a wretched and miserable death, and most of these people, when they did die of their disease, died from distant metastases, lung in particular, along with some bone and some other, which is not pleasant but is different; this has been actually one of the rationales for why we have gone ahead. In, say, thyroid cancer invading the airway, even in the presence of slow-moving pulmonary metastases, as are some of these adenoid cystic pulmonary metastases, we have considered it worthwhile to go ahead and remove the airway obstruction for the patient s benefit. DR RICHARD FEINS (Rochester, NY): Thank you very much, Dr Gaissert and Dr Grillo. We are very fortunate in general thoracic surgery this year to have the publication of Dr Grillo s book on tracheal surgery and I am sure it will stand as the sentinel work in that area. So for those of you who are interested in great books, that book, as a lifelong work, has now been published. Thank you, Dr Grillo, for that.