See related article on page 834. See related article on page 834. In 1960, during a traveling fellowship in England, I met Jack Leigh Collis in Birmingham. There I was introduced to his original concept of adding a “gastroplasty” to the Collis repair for patients with extreme degrees of peptic esophagitis and stricture. Such advanced cases are rarely encountered today, but in those earlier times, there was no doubt whatsoever that such patients had a significant extent of acquired “short esophagus.” Collis used a thoracoabdominal exposure and fashioned a long, 9-cm gastric tube, or gastroplasty, over a 29F bougie. The outcomes though imperfect, were often improved, and this operation avoided the forbidding operative mortality usual in that era after resection and bowel interposition. I did my first lengthening gastroplasty in 1963, on a frail 70-year-old woman with a long and extreme peptic stricture caused by peptic esophagitis after prolonged postoperative ileus and nasogastric suction. I used a purely transthoracic exposure, fashioned a 5-cm gastric tube over a 50F bougie, and added a Belsey mark IV repair, rather than a Collis repair. The patient survived, and her condition remained much improved for more than a decade thereafter. I was convinced that short esophagus was a reality, and with subsequent experience recognized lesser and more subtle degrees of short esophagus. These cases were usually associated with some gross endoscopic evidence of peptic esophagitis (past or present) and were seen in patients who also had Barrett's esophagus, in those with massive hiatus hernia, and in a significant number of instances after previous, failed antireflux operations. As Mattioli and colleagues note in their article, the controversy concerning the incidence (or even the existence) of short esophagus has raged through the subsequent decades. In their prospective, multicenter report, Mattioli and colleagues document the incidence of short esophagus in a consecutive series of surgically treated patients with gastroesophageal reflux disease managed in 7 different centers. The methodology includes a meticulous and detailed technique of preoperative contrast radiographs, with which one of the authors (Mattioli) has had long and well-documented experience. Subsequent intraoperative measurements are obtained with precise and controlled laparoscopic or thoracoscopic measurements. These measurements define the distance between the esophagogastric junction and the apex of the diaphragmatic hiatus (the length of intra-abdominal esophagus in a resting state). The level of the esophagogastric junction is located with intraoperative esophagoscopy. I find this study the most critical evaluation for acquired short esophagus in any series of surgical patients as yet reported. Despite the extremely effective medication available today for the relief and resolution of peptic esophagitis, there remains a significant incidence of acquired short esophagus among patients selected for operative correction of gastroesophageal reflux disease. The short esophagus: Intraoperative assessment of esophageal lengthThe Journal of Thoracic and Cardiovascular SurgeryVol. 136Issue 4PreviewTo define the frequency and predictors of short esophagus in a case series of patients undergoing antireflux surgery. Full-Text PDF

PURPOSE:This is a phase II study to assess the role of induction chemotherapy in the management of stage IIIA non-small-cell lung cancer (NSCLC). We are now reporting the long-term follow-up of the Toronto phase II trial.METHODS:Sixty five patients with mediastinoscopy proven stage IIIA NSCLC received two cycles of preoperative MVP or VLB/P followed by thoracotomy followed by two further courses of chemotherapy.RESULTS:The overall response rate was 67.7% with three complete and 41 partial responders. Forty seven patients went on to thoracotomy with 35 complete resections. Pathologically 4.6% of patients had no tumour remaining. There were three postop deaths as well as five chemotherapy related deaths. Of the 35 patients completely resected 19 have recurred including eight in brain. The median survival for the entire 65 patients is 18.6 months with a 1 year survival of 66%, 5 year survival of 29% and a 10 year survival of 22%.CONCLUSIONS:The long-term survival of induction chemotherapy is maintained. The high incidence of brain recurrences warrants assessment of the role of prophylactic cranial radiation. The role of surgery for stage IIIA NSCLC following induction chemotherapy awaits further study.

During the past 40 years, Dr Hermes Grillo and his colleagues have established a center for upper airway surgery that is unique in terms of the volume of clinical experience that has been attracted, managed, and clearly reported on. Their referral cases include every significant condition in this relatively esoteric field. Idiopathic laryngotracheal stenosis (ILTS) is infrequently reported and is one of the least understood upper airway pathologies. There are even fewer published reports of successful management by means of resection and primary reconstruction in a single stage. In this issue Ashiku and colleagues1Ashiku SK, Kuzucu A, Grillo HC, Wright CD, Wain JC, Lo B, et al. Idiopathic laryngeal stenosis: effective definitive treatment with laryngotracheal resection. J Thorac Cardiovasc Surg. 2004;127:99-107Google Scholar describe an extensive and detailed review of 73 patients with ILTS, all managed by means of a 1-stage resection and primary anastomosis between 1971 and 2002. The classic features of ILTS warrant emphasis. ILTS is a relatively uncommon condition of unknown cause characterized by a nonspecific inflammation of the mucus membrane of the upper airway, which can progress to cicatricial and circumferential stenosis of variable severity. The condition is almost exclusively confined to women (71/73 patients) between 20 and 60 years of age. The authors provide a clear, concise description of the clinical presentation (most commonly dyspnea and wheezing respiration), natural history, differential diagnosis, and surgical pathology. The inflammatory lesion and subsequent cicatricial stenosis is found in the upper airway at the subglottic and upper tracheal levels. On occasion, the inflammation extends to the inferior margin of the vocal cords. Spontaneous resolution was not observed, although progression of subglottic stenosis can progress very slowly over periods of many years. All 73 patients were operated on at a time in the course of their disease when the acute inflammation had subsided and when symptomatic stenosis was not adequately manageable by means of dilation alone. A single-stage operation consisted of circumferential resection of the upper cervical trachea, which included all or part of the anterior cricoid arch in 63 patients. In 36 of these 63 subglottic resections, the cicatricial process extended high in the subglottis and required a resection close to or abutting the inferior margin of the vocal cords. A flap of posterior membranous trachea was fashioned from the distal tracheal resection margin and used to cover the denuded surface of the posterior cricoid plate, a technique described by Grillo and colleagues in 1982. There were no operative deaths, and good (64%) to excellent (26%) results were obtained in 90% of cases. Importantly, considering the idiopathic cause of this inflammatory process, these favorable results were well maintained in long-term follow-up. Our Toronto group has had a small experience with the 1-stage surgical management of ILTS. Our observations concerning clinical presentation, natural history, and surgical pathology mirror those reported by Ashiku and colleagues.1Ashiku SK, Kuzucu A, Grillo HC, Wright CD, Wain JC, Lo B, et al. Idiopathic laryngeal stenosis: effective definitive treatment with laryngotracheal resection. J Thorac Cardiovasc Surg. 2004;127:99-107Google Scholar A few variations in management are noted. We reported initial experience with 2 cases of ILTS in 19862Pearson F.G. Brito-Filomeno L. Cooper J.D. Experience with partial cricoid resection and thyrotracheal anastomosis.Ann Otol Rhinol Laryngol. 1986; 95: 582-585PubMed Google Scholar and with 9 cases of ILTS among a group of 53 patients undergoing a 1-stage cricotracheal resection for subglottic stenosis in 1992.3Maddaus M.A. Toth J.L.R. Gullane P.J. Pearson F.G. Subglottic tracheal resection and synchronous laryngeal reconstruction.J Thorac Cardiovasc Surg. 1992; 104: 1443-1450PubMed Google Scholar Our original technique of partial cricoid resection, described in 1975,4Pearson F.G. Cooper J.D. Nelems J.M. Van Nostrand A.W.P. Primary tracheal anastomosis after resection of the cricoid cartilage with preservation of recurrent laryngeal nerves.J Thorac Cardiovasc Surg. 1975; 70: 806-816PubMed Google Scholar differed from that of Grillo and colleagues in that the subglottic mucosa was divided transversely at the upper line of transection and anastomosed to the distal tracheal margin. Some of the anterior aspect of the posterior cricoid plate was usually rongeured away to widen and facilitate the end-to-end anastomosis. A posterior tracheal mucosal flap was used only in those cases in which the upper resection margin included an interarytenoid scar. Interarytenoid scarring was present in only 1 of our 9 reported cases. Good-to-excellent results were achieved and maintained in all 9 patients. In patients requiring temporary postoperative airway support, we have used a Montgomery T tube as an alternative to a distal tracheostomy tube. The upper tracheal arm of the T tube is placed about 1 cm above the vocal cords because the anastomosis lies within millimeters of the inferior margin of the vocal folds. A T tube has the advantage over an open tracheostomy of providing a closed and well-humidified airway. We have also had some experience with cricotracheal resection and primary reconstruction in 5 patients with cicatricial subglottic stenosis caused by Wegener granulomatosis.5Herridge M.S. Pearson F.G. Doowney G.P.P. Subglottic stenosis complicating Wegener's granulomatosis surgical repair as a viable treatment option.J Thorac Cardiovasc Surg. 1996; 111: 961-966Abstract Full Text Full Text PDF PubMed Scopus (37) Google Scholar In all 5 instances, the acute inflammatory process had subsided, and the collagen disorder appeared to be well controlled with immunosuppressive medication other than high-dose steroids. In contrast to the unfavorable results in 6 patients reported by Akishu and colleagues,1Ashiku SK, Kuzucu A, Grillo HC, Wright CD, Wain JC, Lo B, et al. Idiopathic laryngeal stenosis: effective definitive treatment with laryngotracheal resection. J Thorac Cardiovasc Surg. 2004;127:99-107Google Scholar all 5 of our patients were significantly benefited over long-term follow-up.5Herridge M.S. Pearson F.G. Doowney G.P.P. Subglottic stenosis complicating Wegener's granulomatosis surgical repair as a viable treatment option.J Thorac Cardiovasc Surg. 1996; 111: 961-966Abstract Full Text Full Text PDF PubMed Scopus (37) Google Scholar

Survival following surgical resection of non-small cell lung cancer (NSCLC) has improved since the 1960s, although the 5-year survival rate remains low. This article provides an overview of the role of surgery for NSCLC stages I-III, with a focus on optimizing long-term survival in those patients with resectable disease. Topics explored include diagnosis and staging, indications for resection, types of resection, and indications for adjuvant therapy. A review of the literature indicates a clear survival advantage for complete resection, and is suggestive of an advantage for mediastinal lymph node dissection (vs lymph node sampling) and neoadjuvant therapy (vs adjuvant therapy).

Objective: Paraesophageal hernias represent advanced degrees of sliding hiatus hernia with intrathoracic displacement of the intraesophageal junction. Gastroesophageal reflux disease occurs in most cases, resulting in acquired short esophagus, which should influence the type of repair selected. Methods: Between 1960 and 1996, 94 patients with massive, incarcerated paraesophageal hiatus hernia were operated on at the Toronto General Hospital. The mean age was 64 years (39 to 85 years), with a female to male ratio of 1.8:1. Organoaxial volvulus was present in 50% of cases. Clinical presentation in these patients included postprandial pain in 56%, dysphagia in 48%, chronic iron deficiency anemia in 38%, and aspiration in 29%. Symptomatic reflux, either present or remote, was recorded in 83% of cases. All patients underwent endoscopy by the operating surgeon. In 91 of 94 patients, the esophagogastric junction was found to be above the diaphragmatic hiatus, denoting a sliding type of hiatus hernia. Gross, endoscopic peptic esophagitis was observed in 36% of patients: ulcerative esophagitis in 22% and peptic esophagitis with stricture in 14%. A complete preoperative esophageal motility study was obtained for 41 patients. The lower sphincter was hypotensive in 21 patients (51%), and the amplitude of peristalsis in the distal esophagus was diminished in 24 patients (59%). These abnormalities are both features of significant gastroesophageal reflux disease. In 13 recent, consecutive patients with paraesophageal hernia, the distance between the upper and lower esophageal sphincters was measured during manometry. The average distance was 15.4 +/- 2.33 cm (11 to 20 cm), which is consistent with acquired short esophagus. The normal distance is 20.4 cm +/- 1.9 (p < 0.0001). Results: All 94 patients were treated surgically: 97% had a transthoracic repair with fundoplication. A gastroplasty was added in 75 cases (80%) because of clearly defined or presumed short esophagus. There were two operative deaths, and two patients were never followed up. Among the 90 available patients, the mean follow-up was 94 months; median follow-up aas 72 months. Seventy-two patients (80%) are free of symptoms (excellent result); 13 (13%) have inconsequential symptoms requiring no therapy (good result); and three patients (4%) are improved but have symptoms requiring medical therapy or interval dilatation (fair result). Two patients had poor results because of recurrent hernia and severe reflux. Both were successfully treated by reoperation with the addition of gastroplasty because of acquired shortening, which was not recognized at the first operation. Conclusions: Most of these 94 patients had symptoms or endoscopic, manometric, and operative findings that were consistent with a sliding hiatus hernia. There was a high incidence of endoscopic reflux esophagitis and of acquired short esophagus. True paraesophageal hernia, with the esophagogastric junction in a normal abdominal location, appears rare. Our observations were supported by measurements obtained at preoperative endoscopy and manometry, and by findings at the time of surgical repair. These observations support the choice of a transthoracic approach for repair in most patients.

It is generally agreed that no single surgical technique of repair provides adequate results under all circumstances. It is equally important to recognize that almost all conditions which complicate hiatus hernia and gastroesophageal reflux are of a benign order, and the decision for surgery is an option for the adequately informed patient. This article presents the indications for operation and the criteria for selection of open repairs. It is not a review article, but rather expresses the experience and opinions of the author.

Background. Mediastinal parathyroid cysts are a relatively rare clinical entity. The clinical presentation can be quite varied, although most are found incidentally during investigations for M esophageal or respiratory symptoms.Methods. We present a review of the literature and describe two instructive cases showing specific clinical findings. The clinical presentation, radiologic and pathologic findings, and treatment of mediastinal parathyroid cysts are discussed.Results. In the first patient, the presenting symptom was increasing hoarseness resulting from paresis of the right recurrent laryngeal nerve. This case illustrates the rare association of a beign mediastinal parathyroid cyst with unilateral vocal cord palsy. The second patient presented with the more classic findings of progressive dyspnea and strider related to tracheal compression.Conclusions. Although mediastinal parathyroid cysts are rare and can have varied presentations, thorough investigation can reveal the underlying cyst. Surgical excision is the treatment of choice and can be expected to produce excellent results. (C) 1997 by The Society of Thoracic Surgeons.

Background. Pulmonary arteriography has been reported to be useful in the preoperative assessment of patients with lung cancer to determine the technical resectability and feasibility of pneumonectomy by imaging the main right and left pulmonary arteries. In this report, we describe the use of selective pulmonary arteriography in the assessment of lobar resectability. Methods. Selective pulmonary arteriography provides a detailed anatomic view of the lobar branches and has been used at our institution for the past 30 years to preoperatively investigate patients who are candidates for a sleeve lobectomy.Results. Three cases are described that demonstrate the usefulness of selective pulmonary arteriography in the assessment of the technical feasibility of sleeve resection in patients with lung cancer.Conclusions. Arteriographic findings may accurately show whether a sleeve lobectomy is technically possible, that only a pneumonectomy is possible, or that the only safe way to ensure clearance of the pulmonary artery is to perform arterioplasty. This information may obviate an unnecessary thoracotomy in patients who are judged on the basis of a physiologic assessment to be unable to tolerate a pneumonectomy. (C) 1997 by The Society of Thoracic Surgeons.

This article reviews the histories and techniques of the Belsey and Collis-Belsey antireflux repair operations and discusses the complications and pitfalls associated with these two procedures.

Exposure and mobilization of the esophagus, cardia and proximal stomach is identical to that used for a Belsey Mark 4 repair.

Objective: Surgery of pulmonary aspergillosis followed by higher incidence of post-operative complications, This was the purpose ro evaluate our material. Methods: Between January 1983 and December 1995, the operation was carried out on a total of 84 patients for pulmonary aspergillosis. The patients were comprised of 71 males and 13 females. with a mean age of 49 years (range, 24-71). Previous lung disorders were observed in about half of the cases (most frequently tuberculosis), while in the other half aspergillosis was developed on the basis of (sub)-acute infections. Haemoptysis was present in 48% of patients. The diagnosis was suspected in 47 cases by chest X-ray. Aspergilloma was diagnosed in 50 patients pre-operatively (excluding 12 typical aspergillomas for cavernostomies). with;he other pre-operative diagnoses being tuberculosis, lung cancer, pyoscelrosis, etc. Results: In 71 cases pulmonary resection was carried out (52 lobectomies, 13 wedge resections and six pulmonectomies). A total of 12 cavities were opened by cavernostomy and one lung biopsy was performed for disseminated lung disease. The post-operative mortality rate was 9.5%. The most common complications were bleeding, empyema, bronchial fistula and wound infection. In 23 patients with developed prolonged air leak and/or residual air space, complications were observed more frequently in patients with greater cavitation near the chest wall. Conclusions: In most cases of pulmonary aspergilloma surgical intervention remains the only effective therapy. The operation has a lower risk factor in asymptomatic patients and in patients without pleural or chest wall involvement. In some cases, cavernostomy may be the only remaining surgical choice. (C) 1997 Elsevier Science B.V.

There have been innumerable approaches to the management of subglottic stenosis, which bear testimony to the difficulty in obtaining predictably satisfactory results. Management includes techniques of endoscopic dilation, laser resection, laryngofissure and stenting, and an ingenious array of plastic reconstructions with or without postoperative stenting. On occasion, permanent intubation with a conventional distal tracheotomy or a silicone rubber T tube may be used.

The long-term clinical results of surgical treatment for esophageal achalasia were reviewed in 35 patients having a minimum follow-up of 10 years. Group A (n = 22) are those patients whose first procedure (myotomy and Belsey partial fundoplication) was done at our hospital. Group B (n = 13) are those who had undergone one or more previous operations elsewhere. In group A good to excellent results occurred in 21/22 patients (95%) at 1 year, 17/22 (77%) at 5 years, 15/22 (68%) at 10 years, 11/16 (69%) at 15 years, and 6/9 (67%) at 20 or more years. Two patients underwent early reoperation (2 and 5 years) for dysphagia due to incomplete myotomy. Three patients underwent esophagectomy (7, 19, and 23 years) and one patient underwent an antrectomy and Roux-en-Y diversion (23 years) for late-onset complications of reflux. Three of 13 group B patients had had multiple prior operations and had severe reflux damage at presentation and underwent immediate esophagectomy. Ten patients had one or more conservative operations in our hospital and 4 of these eventually required esophagectomy for disabling reflux. Therefore, there were 10 patients (groups A + B) who underwent esophageal resection, all but 1 of whom had endoscopically documented reflux and 5 of whom had peptic strictures. Six of the 10 esophagectomies were performed more than 10 years (13 to 23 years) after the first operation. These data show that there is a deterioration of initially good results after surgical myotomy and hiatal repair for achalasia due to the late development of disabling reflux disease.

Spontaneous pneumomediastinum is a rare clinical entity; when diagnosis is certain, no treatment is required and symptoms rarely recur. The clinical presentation is usually diagnostic; however, atypical symptoms may mandate further investigation before diagnosis can be established. We describe 2 patients with spontaneous pneumomediastinum who presented with dominant esophageal symptoms (odynophagia and dysphagia) suggestive of esophageal perforation. Investigation and management are discussed.

There have been no major breakthroughs in surgical management for primary lung cancer during the past 40 years. Improved 5-year survival relates primarily to improved preoperative staging and appropriate selection of patients for resection. Perioperative morbidity and mortality, however, has been significantly reduced. Certain principles pertain to current surgical management: resection remains the best treatment for patients with localized, non-small cell primary lung cancer. Accurate preoperative diagnosis and staging: whenever possible, it is desirable to establish the diagnosis and cell type before operation. Accurate evaluation of the N status warrants wide application of invasive staging with mediastinoscopy or a variant. Indications for resection: only patients in whom a complete resection is anticipated should be selected for surgery. Such cases included T1 to T4 stages, N0 and N1 tumors, and selected N2 cases. The indication for resection in patients with hematogenous metastases are anecdotal. Intraoperative staging: accurate and deliberate intraoperative staging with evaluation of nodes using the American Thoracic Society map is highly desirable. The nature of nodal metastases exerts a critical influence on prognosis and in the selection of patients for surgical resection. At present, there is no clear indication for adjuvant therapy in surgically resected cases other than for evaluation and clinical trials.

During the 1940s and 1950s, as many as 50% of thoracotomies identified nonresectable tumors. At present, better than 90% of patients undergoing thoracotomy for presumably resectable lung cancer are found to have operable tumors. This improvement is the result of major advances in the preoperative staging of this disease. Mediastinoscopy and computed tomography (CT) are the most valuable techniques for evaluating the mediastinum in patients with primary cancer of the lung. For each modality, the primary objective is to define the presence or absence of spread to mediastinal lymph nodes. In patients with non-small-cell lung cancer, surgical resection remains the treatment of choice so long as all recognizable tumor can be removed at operation. Both mediastinoscopy and CT provide critical information concerning the potential for a complete resection. Computed tomography remains the most effective noninvasive technique for the evaluation of mediastinal nodes.

PURPOSE The 5-year survival rates with surgical resection for preoperatively identified stage IIIA N2 non-small-cell lung cancer (NSCLC) are less than 10%. A pilot study of mitomycin, vindesine, and cisplatin (MVP) induction chemotherapy was undertaken in an attempt to improve the curative potential of surgery in this group of patients. PATIENTS AND METHODS Thirty-nine patients with mediastinoscopy stage IIIA N2 NSCLC received two cycles of MVP. Responding patients underwent thoracotomy for resection and two further courses of MVP. RESULTS The overall response rate was 64% (25 of 39) with three complete and 22 partial responses. Twenty-two patients were resected, which included a radical mediastinal node dissection. Eighteen resections were complete and four were incomplete. Pathologically, three patients (7.7%) had no tumor remaining. Toxicity included two postoperative deaths secondary to a bronchopleural (BP) fistula, mitomycin pulmonary toxicity in two patients, and septic deaths in four patients. Twenty-eight patients have died; 20 have recurrent or progressive disease. Eight of the 18 patients completely resected have recurred, with a median time to recurrence of 20.6 months. Sites of recurrence include two locoregional, five distant (two in brain), and one in both. Median survival of all 39 patients is 18.6 months, with a 3-year survival of 26%. The median survival for those patients completely resected was 29.7 months with a 3-year survival of 40%. CONCLUSIONS We conclude (1) that MVP is an effective but toxic chemotherapeutic regimen for limited NSCLC; (2) the median survival seems to be prolonged; and (3) the role of induction chemotherapy followed by surgery in stage IIIA N2 NSCLC requires a phase III randomized trial to compare it with other treatment modalities.

Postintubation injury of the upper airway commonly results in stenotic lesions of the larynx, subglottis, and adjacent trachea. The traditional approach to surgical correction is laryngofissure for the laryngeal component and staged plastic reconstruction of the subglottic stenosis. Reported results are variable and unpredictable, and permanent extubation is impossible in a significant number of patients. We report experience with 15 patients with combined laryngeal, subglottic, and tracheal stenosis who were managed by a one-stage operation: circumferential resection of the subglottis and trachea with primary thyrotracheal anastomosis, combined with laryngofissure and laryngeal reconstruction. These procedures required the collaboration of the Departments of Otolaryngology and Thoracic Surgery of the Toronto General Hospital. Between 1972 and 1991, our thoracic surgical division did 53 circumferential subglottic tracheal resections with primary thyrotracheal anastomosis for benign disease. There were no operative deaths and 51 of 53 patients were successfully extubated. In 15 of these patients, a concomitant laryngofissure for laryngeal reconstruction was required. Laryngeal repair included excision or incision of interarytenoid scar (n = 13), interarytenoid mucosal graft (n = 6), or mobilization of cricoarytenoid joint (n = 3). A temporary laryngotracheal stent (usually a Montgomery T tube) was maintained after the operation in all cases (duration 3 to 42 months). Thirteen of these 15 patients are now permanently extubated and none has functionally significant restenosis. Vocal function is satisfactory to good in these patients. The approach described for these combined laryngotracheal lesions provides better results than those reported with traditional staged and plastic techniques of reconstruction. The collaboration of the departments of otolaryngology and thoracic surgery was essential to achieve these results.

Combined modality treatment with chemotherapy and radiation produces tumor regression in most patients with small-cell lung cancer, but the impact on survival has been small, and less than 20% of patients with limited disease survive 2 years. Survival time is extremely short after failure to respond or relapse after treatment. Local control remains a problem, with one third of patients having recurrence only at the primary site. In an attempt to prolong survival and perhaps achieve cure, we undertook surgical resection in 28 patients with limited small-cell lung cancer who did not have complete remission with standard treatment or who had only local recurrence after treatment. There were 28 patients, 22 male and six female, median age 61 years (range 41 to 76). All patients had been treated with chemotherapy and 13 had received preoperative radiotherapy to the primary site and mediastinum. Eight patients underwent an operation for relapse after complete remission. Five patients had had no response to treatment, three had had a slight response followed by progression during chemotherapy, and 12 had achieved partial response but had greater than 3 cm residual masses. Twelve patients required pneumonectomy, 15 lobectomy, one patient had unresectable disease, and two had bulky residual masses after the operation. Three others had microscopic residual disease. Pathologic examination showed only small-cell lung cancer in 18 patients, mixed small-cell and non-small-cell in four, and only non-small-cell lung cancer in six. There were only four patients with stage I disease, 10 with stage II, and 14 with stage III. The median survival from the date of diagnosis for the entire group is 105 weeks and from the date of operation, 74 weeks. The projected 5-year survival rate is 23%. The two patients with residual masses died with local progression, and distant metastatic disease developed in 17 others. One patient died at 6 years without recurrent disease. Eight patients are alive 2 to 5 years after diagnosis. Seven of these patients required only a lobectomy, four had stage I disease, two had stage II, and two had stage III disease. Five had pure small-cell lung cancer and three had mixed small-cell and non-small-cell tumors. All of the patients with pathologic stage I disease remain alive compared with one of 10 with stage II disease and two of 14 with stage III. In summary, relapse or failure to respond to chemotherapy may be due to non-small-cell lung cancer or a mixed tumor.(ABSTRACT TRUNCATED AT 400 WORDS)