Reconstructive Airway Operation After Irradiation

Transcription

1 Reconstructive Airway Operation After Irradiation Derek D. Muehrcke, MD, Hermes C. Grillo, MD, and Douglas J. Mathisen, MD General Thoracic Surgical Unit, Massachusetts General Hospital and the Harvard Medical School, Boston, Massachusetts After a patient died of anastomotic necrosis following a tracheal resection for the management of recurrent thyroid cancer invading the trachea, which had been treated 6 years previously by thyroid lobectomy and 4,800 cgy of radiation to control known residual disease, we explored methods to promote the healing of tissues damaged by irradiation. Between 1979 and 1992, 22 patients underwent major airway resection and reconstruction after receiving large doses of radiation. The average dose was 4, ,113 cgy (range, 3,150 to 6,840 cgy); the number of fractions, 20 to 38; and the average dose per fraction, 180 cgy (range, 150 to 200 cgy). The interval between irradiation and surgical treatment was months (range, 1 to 480 months). Seven cervical, eight midtracheal, and five carinal resections were performed, as well as two mainstem sleeve resections. Omentum was used to protect the anastomosis in 15 patients (68%), a pericardial fat pad was used in 2, and pleura was used in 2. In 3 patients, sternohyoid muscle was placed between the anastomosis and a major vascular structure, but without a tissue wrap. Two patients (9.0%) died postoperatively. Anastomotic dehiscence was the cause of death in a patient treated for lymphoma, and adult respiratory distress syndrome was the cause in the other patient; this patient had undergone carinal pneumonectomy. Complications developed in 8 patients (36%). Two cervical dehiscences were treated by T-tube placement, 2 patients suffered wound infection, and 1 patient each suffered a myocardial infarction, dysphagia, hemoptysis, and bronchitis. Major airway surgical procedures can be performed despite prior irradiation given remote in time, but the likelihood of complication is increased in this setting. The use of vascularized tissue flaps, preferably omentum, to enhance the blood supply and promote fibroplasia seems beneficial. (Ann Thorac Surg 1995;59:14-8) he fact that the healing potential of tissues is impaired T when large doses of radiation are administered has long been recognized by surgeons. Major factors responsible for this are the dose of radiation received and the time interval between irradiation and a subsequent surgical procedure [1, 2]. Traditionally, surgeons are most concerned about those patients whose dose has exceeded 4,500 cgy and who have undergone radiation more than a year before their operation [3]. Although an excisional procedure, such as neck dissection, may entail only a moderate increase in the complications of wound healing, anastomoses may dehisce with disastrous consequences. The ability of irradiated tissues to initiate capillary budding and fibroblastic proliferation in response to injury is minimized or lost [1, 4, 5]. Because of hyalinizing changes in small vessels, necrosis may occur once the integrity of the organ is disturbed during operation [5]. In 1968, we encountered a patient who suffered anastomotic dehiscence after undergoing tracheal resection for the management of invasive recurrent carcinoma of the thyroid. Six years before, the patient had undergone thyroid lobectomy followed by the administration of 4,800 cgy of radiation to control known residual disease in the tracheal wall. Although the integrity of the tissue Presented at the Poster Session of the Thirtieth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 31-Feb 2, Dr Muehrcke's current address is Department of Thoracic and Cardiovascular Surgery, The Cleveland Clinic Foundation, Cleveland, OH Address reprint requests to Dr Grillo, Department of Thoracic Surgery, Massachusetts General Hospital, Boston, MA had been maintained up to that point, transection of the irradiated trachea led to nonhealing, necrosis, dehiscence, and ultimately death. Thereafter, reconstruction of the airways was avoided insofar as possible. However, we continued to encounter patients who had received similar radiation treatment and who later required reconstructive airway procedures to relieve airway obstruction. In most patients, suture lines were buttressed with unirradiated tissue that might also serve as a source of reparative tissue elements. The omentum in particular was used because of its favorable anatomic and biologic characteristics [3, 6-11]. From 1979 to 1992, 22 consecutive patients underwent airway reconstruction after receiving large doses of radiation administered at widely ranging periods before operation. Heterogeneous as this group is, the information yielded by an analysis of their characteristics may provide some guidance to the resolution of this major problem. Only patients who underwent airway anastomoses are included. We have excluded those patients receiving irradiation who underwent pulmonary resection involving bronchial closures, secondary closure of bronchopleural fistulas, and exenterations with end tracheostomies. In such patients, tissue buttresses, generally using omentum, have been successful in promoting wound healing [12, 13]. Material and Methods Twenty-two patients who had previously undergone radiation therapy administered to the thorax or neck for various indications subsequently underwent tracheal, 1995 by The Society of Thoracic Surgeons /95/$ (94)00751-R

2 Ann Thorac Surg MUEHRCKE ET AL ;59:14-8 AIRWAY RECONSTRUCTION AFTER IRRADIATION Table 1. Preoperative in the 22 Patients Undergoing Variable a Age (y)b 49.5 ± 16.2 (9--71) Radiation dose (cgy) b 4, ,112.5 (3,150-6,840) Fractions b 27.6 ± 5.7 (20-38) Dose per fraction (cgy) b ( ) Time between irradiation and _ (1-480) operation (mo) b Sex 16 male/6 female Use of steroids 2 yes/20 no Current chemotherapy 3 yes/19 no Pathologic status Primary tracheal cancer 10 Squamous cell 8 Adenoid cystic 1 Lymphoepithelial 1 Benign stricture 6 Prior tracheal anastomosis 4 Due to irradiation 2 Secondary tracheal tumor 3 Rhabdomyosarcoma 1 Laryngeal squamous cell 1 Squarnous cell lung cancer 1 Stage IV lung cancer: 3 squamous cell lung cancer (invading carina) a on radiation therapy only available for 21 patients, b Value is the mean -+ the standard deviation with the range in parentheses. carinal, or main bronchial sleeve resection and reconstruction at the Massachusetts General Hospital. The preoperative variables for each patient were retrospectively reviewed; these included age, sex, the total radiation dose, the time between irradiation and operation, the dose per fraction of radiation, the administration of steroids or chemotherapy at the time of resection, and the current pathologic status (Table 1). The data for the intraoperative variables are summarized in Table 2. Release techniques were used to prevent excessive anastomotic tension. The postoperative data are given in Table 3. Early mortality includes death during hospitalization or within 30 days of the operation. Follow-up information was obtained by review of charts and from telephone interviews with the patients and referring physicians. The length of follow-up in hospital survivors ranged from I to 144 months and was complete for all 20 patients. Radiation Therapy on the exact dosage, portal size, time period, and number of radiation treatment fractions were available for 21 of the 22 patients (see Table 1). These data could not be obtained in I patient who had received irradiation for the management of cervical tuberculosis 40 years before undergoing cervical tracheal reconstruction for the relief of stenosis that was sufficient to cause dyspnea at rest. Patients fell into four categories: (1) those with Table 2. Intraoperative in the 22 Patients Undergoing Variable Reoperation 16/22 (72.7%) Location of lesion Cervical 7 Midtracheal 8 Distal trachea/carina 5 Main bronchus 2 Length of trachea resected (cm) a 3.27 ± 1.28 (1-5) Release procedures 7/22 (31.8%) Hilar 5 Laryngeal 2 Suture material Vicryl 18 (82%) PDS 2 Tevdek 1 Nylon 1 Anastomotic wrap Omental 15 (68.2%) Pericardial 2 Pleural 2 None 3 a Value is the mean -+ the standard deviation with the range in parentheses. symptomatic postirradiation tracheal stenosis; (2) those with recurrent tracheal tumor after irradiation, either primary or adjunctive; (3) those with recurrent extratracheal tumor invading the trachea; and (4) those who had been given preoperative radiotherapy. Six patients were referred because of postirradiation Table 3. Postoperative in the 22 Patients Undergoing Variable Hospital stay (days) a _+ 6.8 (8-33) Length of follow-up (mo) a 21.5 ± (0-144) Positive resection margins 2/22 (9%) Mortality (30-day) 2/22 (9%) Dehiscence 1 Adult respiratory distress 1 syndrome Complications 8/22 (36.4%) Dehiscence 2 Wound infection 2 Myocardial infarction 1 Dysphagia 1 Hemoptysis 1 Bronchitis 1 Status at follow-up Alive and well 11 Mive with T-tube 2 Died from recurrent cancer 6 Died from myocardial infarction 1 a Value is the mean +- the standard deviation with the range in parentheses.

3 16 MUEHRCKE ET AL Ann Thorac Surg AIRWAY RECONSTRUCTION AFTER IRRADIATION 1995;59:14-8 tracheal stenosis. All were suffering from dyspnea at rest. Three had undergone irradiation for the control of positive tumor margins after resection of adenoid cystic carcinoma of the trachea. Another had received mediastinal irradiation 13 years before for the treatment of stage Ia Hodgkin's lymphoma. A fifth had undergone irradiation for the control of cervical tuberculosis 40 years earlier. The sixth patient had received radiation treatment for the management of recurrent thyroid cancer. The stenosis was located at the site of tracheostomy, which had been included in the field of irradiation. Ten patients presented with recurrent primary tracheal tumors after undergoing radiation therapy. In 5, the radiation therapy had failed as the primary method of treatment. In 3 others, treated by tracheal resection for the removal of primary squamous cell carcinoma, irradiation had been given postoperatively because the resection margins were found to be positive for tumor. All suffered recurrence and were referred for surgical therapy. In another patient who received irradiation for the management of lingual cancer, a primary tracheal tumor developed in the previously irradiated field. The tenth patient in this group received irradiation after undergoing two tracheal resections and reconstructions for the removal of separate primary tracheal squamous cell cancers. He presented with a third primary tracheal tumor in the irradiated area. Three patients received irradiation when a previously resected extratracheal tumor recurred and invaded the trachea. One patient had undergone two prior courses of cervical irradiation for the control of recurrent rhabdomyosarcoma; the doses were 3,728 cgy and 5,040 cgy and the treatment had been carried out 93 and 67 months before, respectively. A second patient had a laryngeal squamous cell cancer that had initially been treated primarily with irradiation. Locally recurrent lung cancer had developed in the final patient after upper lobectomy and radiation therapy. Three patients underwent preoperative irradiation and chemotherapy shortly before undergoing carinal or mainstem bronchial resection as part of a cancer and leukemia group B (CALGB) protocol for the treatment of stage IIIa squamous cell bronchogenic carcinoma, which was extended to include carinal involvement (stage IIIb). Three patients who had received prior radiation therapy also underwent postoperative irradiation after a definitive operation. One patient suffered a fourth recurrence of rhabdomyosarcoma in his neck, for which he received an additional 2,000 cgy of radiation delivered to the cervical area several months after he had undergone tracheal resection. The airway remained free of tumor. A second patient, who had undergone preoperative irradiation elsewhere for the management of a large right lung cancer and a subsequent right pneumonectomy, suffered recurrence at the bronchial stump. Carinal resection involving an end-to-end anastomosis of the trachea to the left main bronchus was performed. Even though 3 cm of trachea was resected and the intraoperative frozen section margins were read as normal, tumor was found upon final study. He received another 1,500 cgy postop- eratively. In the third patient, the tracheobronchial lymph nodes were found to be positive after he had been given 4,200 cgy of radiation as well as chemotherapy preoperatively for the control of stage IIIb, right upper lobe carcinoma. Postoperatively, he was treated with an additional 1,500 cgy delivered to a mediastinal portal, which included the carinal pneumonectomy site and lymph node-bearing area. Anesthesia The anesthetic technique has been described elsewhere [14l. A deep Ethrane (Anaquest, Madison, WI) technique was used that allowed spontaneous ventilation during operation and prompt extubation at the completion of the procedure. The airway was managed intraoperatively through the insertion of a flexible armored tube into the distal airway that was attached to sterile connecting tubing placed across the operative field. Operative Procedures After complete preoperative radiographic examination of the airways had been carried out, each patient underwent rigid bronchoscopy with Hopkins telescopes and, when indicated, esophagoscopy to identify the nature, location, and extent of the lesion. Well-defined principles of tracheal surgical procedures were observed [15, 16] (see Table 2). Most cases were associated with fibrosis and inflammation that stemmed from prior irradiation and surgical treatment. Tissue buttresses were used to provide a healthy source of regenerating tissue to help promote healing of the anastomosis; a sealing layer and tissue interposition between the suture line and major blood vessels were also used for this purpose. A tissue wrap was not routinely employed if a patient had received less than 4,500 cgy of radiation earlier than 12 months before the airway operation. Vascularized tissue was sutured around the anastomosis using 4-0 silk taking care to ensure that the tissue was flush against the airway. Tissue pedicles were prevented from kinking so that the blood supply was not impeded, and these were attached without tension. When omentum was used, it was usually delivered into the chest through the substernal space. If a cervical or cervicomediastinal exposure to the airway was used, the omentum was brought up into the anterior mediastinal space. It was frequently necessary to remove one of the clavicular heads or a portion of the upper manubrium. When a right thoracotomy was used, the omentum was brought up substernally or through the diaphragm anterior to the hilum of the lung. In 19 patients, the anastomoses were wrapped with a vascularized tissue flap. In 15, the airway was buttressed using a pedicled omental flap [3]. In 2, the omentum was unavailable due to prior gastrectomy; therefore, a pericardial fat pad flap [5] was pedicled in 1 and an intercostal muscle flap was formed in the other, in whom the pericardial fat pad was insubstantial. A pleural wrap was used in 2 patients. In 3 patients, the airway anastomosis was not protected with a vascularized tissue flap. In 1, the sternohyoid muscle was interposed between the tracheal anastomosis

4 Ann Thorac Surg MUEHRCKE ET AL ;59:14-8 AIRWAY RECONSTRUCTION AFTER IRRADIATION and the innominate artery; in the second patient, the sternohyoid muscle was placed between the tracheal anastomosis and the esophagus. The intent of these procedures was to prevent fistula formation. In I patient who had received only 4,200 cgy just 12 months earlier, the airway was not protected and no complications ensued. The interrupted 4-0 Vicryl sutures (Ethicon, Somerville, NJ) used for tracheal reconstruction appear to all but eliminate the formation of granulation tissue at the suture line [17]. No superiority has been shown for monofilament sutures. In 3 earlier patients, nonabsorbable sutures were used for reconstruction. Results Two patients died as the result of operation, for a mortality rate of 9.0% (see Table 3). One suffered adult respiratory distress syndrome after undergoing right carinal pneumonectomy. Another with late tracheal stenosis due to treatment for Hodgkin's lymphoma died after the tracheal anastomosis separated. He had received 4,000 cgy of radiation 150 months before. Although no active lymphoma was detected, the trachea was encased in massive scar tissue presumed to stem from his original disease. The quality of the tracheal wall used for anastomosis was therefore very unsatisfactory in this patient, and an omental wrap was used. Complications developed in 8 patients (36%). One patient with tracheal dehiscence has already been noted. In another, a paratracheal abscess was encountered at the resection of a necrotic tumor; this required placement of a T tube to maintain the airway. He has since died of recurrent squamous cell carcinoma. Two patients suffered wound infections. Hemoptysis developed in 1 patient as the result of granulation tissue that had formed at the anastomotic site, 1 had a myocardial infarction, and 1 had transient dysphagia after undergoing suprahyoid laryngeal release. All of these complications resolved. In another, stenosis developed at the tracheal anastomosis, and repeated dilations were required to relieve the resulting constriction. One patient who suffered recalcitrant anastomotic stenosis postoperatively had undergone irradiation 480 months before. Her primary disorder was irradiation-induced tracheal stenosis. Seventeen patients experienced excellent results after operation, with no evidence of exertional dyspnea, and 2 patients experienced good results, but suffered dyspnea with moderate exercise. Good healing was confirmed bronchoscopically. Six died of recurrent cancer or sarcoma from 3 to 45 months after resection (mean, 17 months; median, 10 months), most often without airway obstruction. Comment Surgeons have learned to be wary of performing closures and anastomoses in operative fields where high doses of radiation have been administered, particularly when this has been done remotely in time. Such hazards have been particularly recognized in the setting of intestinal surgical procedures [2]. Documented guidelines for the management of such patients in terms of dose and time do not exist with regard to tissues in general or to airways in particular. Irradiated tissue is more susceptible to injury during surgical manipulation and is less likely to heal satisfactorily [1, 5]. Deficits in tracheobronchial anastomotic healing are not common in healthy tissue, assuming a competent surgical technique has been used, unless there is excessive tension on the anastomosis or the blood supply is interrupted. In 365 consecutive patients undergoing tracheal resection and reconstruction for the relief of postintubation stenosis or removal of primary tumors, separation developed in 2.7% and partial or complete stenosis arose in 6.5% [17]. Late stenoses also occurred after resections involving a considerable length of the trachea in 2 patients on high doses of prednisone, a circumstance now avoided. Necrosis, such as that noted in our initial experience with late resection after high-dose radiation therapy, was not seen. Even less common are defects in healing after bronchial sleeve resection, in which situation tension is rarely a factor [18, 19]. Experimental studies conducted in dogs have demonstrated the detrimental effects of irradiation on tracheal anastomoses [5, 20]. When more than 3,500 cgy of radiation was delivered over a 3-week period before resection and reconstruction, the attendant incidence of subsequent tracheal stenosis increased proportionally with increasing doses of radiation. Histologically, this was associated with more edema, inflammation, and fibrosis at the site of the anastomosis than that seen in nonirradiated control animals [20]. Irradiation inhibits capillary proliferation, mucosal blood flow [5], and the appearance of local fibroblast populations [4], key components in the reparative response that finally unites the anastomosis. A late effect of irradiation includes hyalinization of arterioles, which then do not respond with a reparative response when injured [1]. Irradiation-induced damage appeared to be the major source of dehiscence in our early experience. Others have noted similar problems with wound healing, which in turn has led thoracic surgeons to be reluctant to perform these procedures after irradiation [20, 21]. Tsubota and associates [20] described a patient in whom anastomotic necrosis developed following tracheal resection performed 3 months after receiving 3,500 cgy of radiation for the management of a primary tumor. Abbey-Smith [21] noted healing problems in patients undergoing sleeve lobectomy and bronchoplasty after they had received high doses of radiation preoperatively, although Jensik [22] disclaimed this effect. It seemed likely, therefore, that irradiation-impaired tissue might heal better if an unirradiated, vascularized tissue flap was applied to the anastomotic site. This would provide a better blood supply and promote fibroplasia. Therefore, in our patients who had received either high-dose irradiation (more than 4,500 cgy) or who had undergone irradiation more than 12 months before resection, the anastomosis was wrapped with vascularized tissue.

5 18 MUEHRCK ET AL Ann Thorac Surg AIRWAY RECONSTRUCTION AFTER IRRADIATION 1995;59:14-8 Our experience indicates that airway reconstruction can be performed fairly safely in patients undergoing irradiation despite the potentially harmful effects of ionizing irradiation on wound healing. The incidence of complications and the mortality rates were, however, higher than those observed in patients who had undergone major airway reconstructions but not received radiation therapy. In our series, the hospital mortality rate of 9.0% was higher than the 1.8% that was noted for patients with postintubation lesions who had not received radiation therapy [23] or the 5% in patients with primary tumors who have undergone only resection [24], including carinal resections. In our 22 patients, only one of two deaths resulted directly from failure of the anastomosis to heal, although even in this patient special factors were present. Subsequent patients with a similar pathologic condition have been electively managed with T tubes. The complication rate in patients who underwent irradiation was 36% versus a 25.8% incidence in patients who did not receive radiation therapy [23]. Although tumor recurrence limited the palliative effects of the operation for some patients in terms of time (mean, 17 months), the relief of the dreadful symptoms of airway obstruction was definitive. The only other palliative alternatives available are repetitive mechanical removal of the tumor (core out or laser), brachytherapy, or mechanical intubation (tracheostomy, T tube, or stent)--all generally associated with early recurrent airway obstruction. The omentum is preferred for augmenting anastomotic healing. It appears to elaborate an angiogenic factor that stimulates the development of a new blood supply in avascular areas [6, 11], as well as provides new fibroblasts to enhance healing [7]. As observed in experiments studying devascularized canine bronchi and in the setting of human single-lung transplantations, a network of fine vessels originating from omentum wrapped around anastomoses forms within 4 days of the operation [8, 9]. The same phenomenon probably occurs in irradiated tissues, thus initiating the healing process. The use of omentum has a long surgical history. It has been used to cover intrathoracic esophageal anastomoses to lessen the risk of anastomotic leakage [10]. The omentum will reach any part of the airway with ease. It has been used in the treatment of difficult thoracic problems, such as in patients undergoing complex chest wall reconstruction, and in those with mediastinal infection after cardiac procedures, vascular infection, esophageal perforation [3], and a bronchopleural fistula [13]. The omentum also appears to function well in the presence of infection. As it is not yet clear whether a pedicled pericardial fat pad confers similar biologic advantages, this remains a second choice in those instances when omentum is not available. Pedicled muscle flaps are a less attractive option. Major airway reconstructive surgical procedures can be performed in patients who have undergone irradiation remote in time, but the likelihood of complications is increased. The use of vascularized tissue flaps, preferably omentum, to enhance blood supply and promote fibroplasia seems beneficial. References 1. Cox JD, Byhardt RW, Wilson JF. Complications of radiation therapy and factors in their prevention. World J Surg 1986; 10: DeCosse JJ, Rhode RS, Wentz WB, et al. The natural history and management of radiation-induced injury of the gastrointestinal tract. Ann Surg 1969;170: Mathisen DJ, Grillo HC, Vlahakes GJ, Daggett WM. The omentum in the management of complicated cardiothoracic problems. J Thorac Cardiovasc Surg 1988;95: Grillo HC, Potsaid MS. Studies in wound healing: IV. Retardation of contraction by local x-irradiation and observations relating to the origin of fibroblasts in repair. Ann Surg 1961;154: Inni K, Takahashi Y, Hasegawa S. Effect of preoperative irradiation on wound healing after bronchial anastomosis in mongrel dogs. J Thorac Cardiovasc Surg 1993;106: Silverman KJ, Lung DP, Zetter BR, et al. Fat antiogenesis: a possible link to coronary atherosclerosis and thrombosis. Circulation 1985;72(Suppl 3): Lieberman-Meffert D, White H. The greater omentum. Berlin: Springer-Verlag 1983;63-6,187-9, Morgan E, Lima O, Goldberg M, et al. Successful revascularization of totally ischemic bronchial autografts with omental pedicle flaps in dogs. J Thorac Cardiovasc Surg 1982;84: Lima O, Goldberg M, Peters WJ, et al. Bronchial omentopexy in canine lung transplantation. J Thorac Cardiovasc Surg 1982;83: Fekete F, Ronsse BH, Tossen JC, Langonnet F. EEA stapler and omentum graft in esophagogastrectomy: experience with 30 intrathoracic anastomoses for cancer. Ann Surg 1981;193: Goldsmith HS, Griffith AL, Kupferman A, Catsimpools N. Lipid angiogenic factor from omentum. JAMA 1984;252: Grillo HC, Mathisen DJ. Cervical exenteration. Ann Thorac Surg 1990;49: Puskas JD, Mathisen DJ, Grillo HC, Wain JC, Wright CD, Moncure AC. Treatment strategies for bronchopleural fistula. J Thorac Cardiovasc Surg (in press). 14. Wilson RS. Tracheal resections. In: Marshall BE, Longnecker DE, Fairly HB, eds. Anesthesia for thoracic procedures. Boston: Blackwell, 1988: Grillo HC. Surgery of the trachea. In: Keen G, Farndon JR, eds. Operative surgery and management, 3rd ed. Bristol: Butterworth, Heinemann 1994: Grillo HC. Surgery of the trachea. In: Ravitch M, Steichen F, eds. Atlas of general thoracic surgery, Philadelphia: Saunders, 1988: Grillo HC, Zannini P, Michelassi F. Complications of tracheal reconstruction: incidence, treatment and prevention. J Thorac Cardiovasc Surg 1986;91: Frist WH, Mathisen DJ, Hilgenberg AD, Grillo HC. Bronchial sleeve resection with and without pulmonary resection. J Thorac Cardiovasc Surg 1991;93: Newton JR, Grillo HC, Mathisen DJ. Main bronchial sleeve resection with pulmonary conservation. Ann Thorac Surg 1991;52: Tsubota N, Simpson WJ, VanNostrand AWP, Pearson FG. The effect of preoperative irradiation on primary tracheal anastomosis. Ann Thorac Surg 1975;20: Abbey-Smith R. Personal communication. In Tsubota et al [20]. 22. Jensik RJ. Discussion of Tsubota et al [20]. 23. Grillo HC. Surgical treatment of post-intubation tracheal injuries. J Thorac Cardiovasc Surg 1979;78: Grillo HC, Mathisen DJ. Primary tracheal tumors. Ann Thorac Surg 1990;49:69-77.