Relevance of an Intensive Postoperative Follow-up After Surgery for Non Small Cell Lung Cancer

Transcription

1 Relevance of an Intensive Postoperative Follow-up After Surgery for Non Small Cell Lung Cancer Virginie Westeel, MD, Didier Choma, MD, François Clément, MD, Marie-Christine Woronoff-Lemsi, PhD, Jean-François Pugin, MD, André Dubiez, MD, and Alain Depierre, MD Chest Disease Department and Departments of Thoracic Surgery and Pharmacy, University Hospital, Besançon, France Background. Although a minimal follow-up with periodic clinic visits and chest radiographs is usually recommended after complete operation for non small cell lung cancer, the ideal follow-up has not been defined yet. Objectives of this prospective study were to determine the feasibility of an intensive surveillance program and to analyze its influence on patient survival. Methods. Follow-up consisted of physical examination and chest roentgenogram every 3 months and fiberoptic bronchoscopy and thoracic computed tomographic scan with sections of the liver and adrenal glands every 6 months. Influence of patient and recurrence characteristics on survival from recurrence was successively analyzed using the log-rank test and a Cox model adjusted for treatment. Results. Among the 192 eligible patients, recurrence developed in 136 patients (71%) and was asymptomatic in 36 patients (26%). In 35 patients, recurrence was asymptomatic and detected by a scheduled procedure: thoracic computed tomographic scan in 10 (28%) patients and fiberoptic bronchoscopy in 10. Fifteen patients (43%) had a thoracic recurrence treated with curative intent. From the date of recurrence, 3-year survival was 13% in all patients and 31% in asymptomatic patients whose recurrence was detected by a scheduled procedure. Asymptomatic recurrences (p < 0.001), female sex (p < 0.001), performance status 2 or less (p 0.01), and age 61 years or younger (p 0.01) were shown to be significantly favorable prognostic factors. Conclusions. This intensive follow-up is feasible and may improve survival by detecting recurrences after surgery for non small cell lung cancer at an asymptomatic stage. (Ann Thorac Surg 2000;70: ) 2000 by The Society of Thoracic Surgeons Surgery remains the best chance of cure in non small cell lung cancer, but the risk of recurrence or of new lung cancer is high. The 5-year survival rate observed in patients who have been operated on varies by stage, from 20% to 65%. There is no consensus for the best follow-up of patients after complete resection for non small cell lung cancer. A survey among the members of The Society of Thoracic Surgeons showed a marked variation in surveillance practice in the United States [1, 2]. Moreover, significant differences in follow-up practice patterns can exist within single health-care facilities as shown in the St. Louis Department of Veterans Affairs Medical Center [3]. Wide variations in surveillance testing are also observed in Europe. On the basis of retrospective studies [3 7], it is usually considered that there is no justification for a follow-up surveillance beyond periodic physical examination and chest roentgenogram in operated lung cancer. To clearly evaluate the interest of a more intensive surveillance after lung cancer operation, a randomized trial should be undertaken as it was in breast cancer patients who were operated on [8, 9]. Before initiation of a randomized trial Accepted for publication April 1, Address reprint requests to Dr Westeel, Service de pneumologie, CHU, Hôpital J Minjoz, Blvd Fleming, Besançon, France; virginie.westeel@ufc-chu.univ-fcomte.fr. comparing a strategy including physical examination and chest roentgenogram with a more intensive program, relevant follow-up strategies need to be defined. A prospective study including a cost analysis was performed to evaluate the feasibility and the impact of a standardized follow-up program consisting of physical examination, chest roentgenogram, thoracic computed tomography (CT) scan with sections of the liver and adrenal glands, and fiberoptic bronchoscopy on patient survival. Material and Methods Eligibility Criteria All patients of the Respiratory Medicine Department of Besançon who underwent resection for non small cell lung cancer between January 1980 and December 1993 and whose postoperative care was performed by this department were considered for this study. Patients were eligible if the cancer had been completely resected. Incomplete resections included positive bronchial or pulmonary parenchymal margins and extranodal cancer invasion of mediastinal lymph nodes. Patients who had had a previous malignancy in the 5 years before the operation (except for carcinoma in situ of the uterine cervix and basal cell carcinoma of the skin), patients who died within 30 days of operation, and patients with a 2000 by The Society of Thoracic Surgeons /00/$20.00 Published by Elsevier Science Inc PII S (00)

2 1186 WESTEEL ET AL Ann Thorac Surg INTENSIVE FOLLOW-UP IN RESECTED NSCLC 2000;70: mixed histologic profile that included small cell carcinoma or neuroendocrine tumor cells could not be included. Patients who had undergone primary resection of synchronous brain metastasis were eligible. Informed consent was obtained from all patients, and all procedures were conducted in accord with French ethical guidelines. Preoperative Evaluation Preoperative radiologic evaluation included chest roentgenogram, thoracic CT scan including liver and adrenals, and brain CT scan. Mediastinoscopy was not routinely required but was left to the discretion of investigators. Follow-up Procedures During the first 3 years after the operation, follow-up procedures included: physical examination and chest roentgenogram 1 month after operation and every 3 months thereafter and thoracic CT scan with sections of the liver and adrenal glands and fiberoptic bronchoscopy every 6 months. Between 1980 and 1984, fiberoptic bronchoscopy was planned every 3 months during the first year after operation. Because it was frequently omitted, fiberoptic bronchoscopy was scheduled every 6 months from the first year of surveillance, at the same time as thoracic CT scan, between 1985 and From the fourth to the seventh year after operation, chest roentgenogram was performed every 6 months and thoracic CT scan and fiberoptic bronchoscopy once a year. From the eighth year after operation, surveillance testing only consisted of a yearly chest roentgenogram. Compliance With Follow-up Procedures Compliance with the surveillance procedures was studied in the patients included during the last 4 years of the study. In these patients, the number of chest roentgenograms, thoracic CT scans, and fiberoptic bronchoscopies actually realized until recurrence, death, or June 15, 1997, was calculated and compared with the number of scheduled procedures. The frequency of physical examinations was the same as the frequency of chest roentgenograms, as they were realized at the same visit. Recurrences The criteria reported by Martini and colleagues [10] were used to differentiate second primary lung cancers and lung recurrences. Recurrences were recorded as detected during a scheduled or an unscheduled procedure and as localized within the chest or distant. Unscheduled procedures consisted of additional visits or procedures because of new symptoms. Patients whose tumor recurred were recorded as being either symptomatic or asymptomatic at the time of recurrence. The method by which the recurrence was diagnosed was noted; these included physical examination, chest roentgenogram, CT scan of the chest, CT sections of the liver and adrenal glands, fiberoptic bronchoscopy, and other procedures. Other procedures consisted of procedures different from those included in the program and performed because of new symptoms either detected during a scheduled visit or during an additional visit. Treatment was considered to be curative in intent if the patient had cryotherapy for a carcinoma in situ, underwent complete surgical resection of a chest or brain lesion, or had thoracic radiotherapy or chemoradiation for a local or locoregional recurrence (at most limited to the lung and ipsilateral mediastinum). Palliative therapy was defined as thoracic radiotherapy or chemotherapy for advanced disease. Best supportive care was defined as comfort-oriented care. Survival and Statistical Analysis Potential follow-up was defined as the time elapsed between the date of operation and June 15, Survival was calculated from the date of operation to the date of death or last follow-up visit. In patients whose cancer recurred, survival was also calculated from the date of recurrence to the date of death or last follow-up visit. Disease-free survival was calculated from the date of operation to the date of recurrence or censured at the date of death or last follow-up visit in patients whose disease did not recur. Survival curves were constructed using the Kaplan-Meier method [11]. Survival curves from the date of recurrence were compared by the log-rank test for the following variables: sex, age at operation ( median age versus median age), pathology (adenocarcinoma and large cell carcinoma versus squamous cell carcinoma), performance status ( 2 versus 3 to 4) and symptoms (asymptomatic versus symptomatic) at the time of recurrence, mode of detection of recurrence (scheduled versus unscheduled procedure), diagnostic procedure (physical examination or chest roentgenogram versus fiberoptic bronchoscopy or thoracic CT scan or CT sections of the liver and adrenal glands versus others), site of recurrence (thorax only versus extrathoracic or both intrathoracic and extrathoracic), and disease-free interval ( 1 year versus 1 year). Variables with a p value less than 0.20 were included in a Cox model adjusted for treatment (curative versus palliative versus best supportive care). Qualitative variables of more than two categories were introduced in the model as dummy variables. Second cancers were introduced in the model as time-dependent variables [12]. All statistical analyses were performed using the BMDP software [13]. Cost Analysis The cost was calculated per patient for the median disease-free survival duration. Reimbursement prices are determined for each procedure by the French health-care system. French health-care system 1998 repayment tariffs and 1998 exchange rate ($ French francs) were used. They were $21.43 for a visit, $19.89 for a chest roentgenogram, $ to $ for a thoracic CT scan with sections of the liver and adrenals depending on the age of the scan system, and $ for a fiberoptic bronchoscopy. For one patient, the cost of the follow-up program was the total of the costs for each procedure. For each procedure, the cost was the cost for one procedure multiplied by the number of times this procedure was theoretically performed during the studied period.

3 Ann Thorac Surg WESTEEL ET AL 2000;70: INTENSIVE FOLLOW-UP IN RESECTED NSCLC 1187 Table 1. Characteristics of the 192 Eligible Patients at Operation and Treatments Characteristics n 192 (%) Males/females 177 (92)/15 (8) Pathology Squamous cell a 146 (76) Adenocarcinoma a 43 (22) Large cell carcinoma 4 (2) Pathologic stage I/II 86 (45)/36 (18) IIIa T3/IIIa N2 19 (10)/38 (20) IIIb/IV b 9 (5)/4 (2) Resections Pneumonectomy 113 (59) Lobectomy/bilobectomy c 71 (37)/6 (3) Limited 3 (1) Adjuvant therapy Operation only 124 (65) Operation chemotherapy 33 (17) Operation radiotherapy 29 (15) Operation chemo-radiotherapy 6 (3) a One patient had synchronous squamous cell carcinoma and adenocarcinoma. b Metastatic sites were lung (3 patients) and brain (2 patients). c One patient had a bilobectomy plus a contralateral lobectomy. Compliance With Follow-up Procedures Thirty-eight patients underwent operation between January 1990 and December In these 38 patients, 223 chest roentgenograms were performed for 243 that were scheduled (91.8%), 90 (82.6%) thoracic CT scans with sections of the liver and adrenals and 101 (92.7%) fiberoptic bronchoscopies were performed for a scheduled number of 109 for both. Recurrences One hundred thirty-six patients (71%) had recurrences during the studied period. Median age at the date of recurrence was 61 years (range, 36 to 81 years). Modes of detection, sites and treatments of recurrence, and survival are summarized in Figure 1. Recurrence was detected by a scheduled procedure in 85 patients (63%) and by nonscheduled visits because of new symptoms in 50 patients (37%). In 1 asymptomatic patient, a brain CT scan was systematically performed and detected an asymptomatic brain metastasis. Among the 85 patients whose recurrence was detected by one of the scheduled procedures, 35 (26%) were asymptomatic. Physical examination or chest roentgenogram detected recurrences in 49 patients (36%), thoracic CT scan, in 18 (13%), abdominal CT sections, in 12 (9%), and fiberoptic The cost per years of life potentially gained by thoracic CT scan or fiberoptic bronchoscopy was estimated by dividing the cost because of thoracic CT scan with abdominal sections and fiberoptic bronchoscopy during the median disease-free survival duration for the whole population by the number of years gained. The number of years of life gained by thoracic CT scan or fiberoptic bronchoscopy was the number of asymptomatic patients alive 3 years after recurrence multiplied by 3 among asymptomatic patients who had a thoracic recurrence detected by thoracic CT scan or fiberoptic bronchoscopy and treated with curative intent. Results Patient Characteristics Between January 1980 and December 1993, 298 patients who underwent operations for non small cell lung cancer were monitored by the Chest Disease Department of Besançon. One hundred six patients were not eligible for the following reasons: incomplete resection (76 patients), previous malignancy within the 5 years before operation (16 patients), and death within the 30 days after operation (14 patients). The remaining 192 patients were eligible for the follow-up study. Characteristics of patients at the time of operation and treatments are detailed in Table 1. Median age was 60 years (range, 33 to 81 years). A majority of patients were men and had squamous cell carcinoma. The median potential follow-up was 131 months (range, 46 to 207 months). Fig 1. Recurrences: modes of detection, sites, treatments, and survival. (CI confidence interval; CT computed tomography.)

4 1188 WESTEEL ET AL Ann Thorac Surg INTENSIVE FOLLOW-UP IN RESECTED NSCLC 2000;70: bronchoscopy, in 23 patients (17%). In 9 symptomatic patients (26%) whose recurrence was suspected by a new symptom during a scheduled visit, recurrence was confirmed by other procedures. In symptomatic patients requiring an unscheduled visit for new symptoms, recurrence was confirmed by other procedures in 24 cases (47%). These other diagnostic procedures included brain CT scan (19 patients) and bone scan or radiographs (15 patients). Of the 35 asymptomatic patients whose recurrence was diagnosed by a scheduled procedure, recurrence was diagnosed by thoracic CT scan in 10 patients (29%) and fiberoptic bronchoscopy in 10 patients (29%). Recurrence developed in the chest in 71 patients (52%), outside the chest in 39 patients (29%), and both within and outside the chest in 26 patients (19%). In the 71 patients whose recurrence was diagnosed within the chest only, median time for recurrence was 17 months (range, 1 to 163 months). In the 65 patients whose recurrence occurred outside the chest or both within and outside the chest, median time for recurrence was 9 months (range, 1 to 148 months). Single metastatic sites included brain (18 patients), bone (13 patients), liver (5 patients), and others (3 patients). Recurrence developed in the chest only in 29 (80%) of the 35 asymptomatic patients whose recurrence was detected by a scheduled procedure. Of the 136 recurrences, 35 patients (26%) were treated with curative intent (operation in 22 patients, radiotherapy in 7 patients, chemoradiation in 5 patients, cryotherapy in 1 patient), 71 (52%) received palliative therapy, and 30 (22%) received best supportive care. Treatment was delivered with curative intent for a thoracic recurrence in 15 (43%) of the 36 asymptomatic patients, in 7 (14%) of the 50 symptomatic patients whose recurrence was detected by a scheduled procedure, and in 6 of the 51 patients (12%) whose recurrence was diagnosed by an unscheduled procedure. Recurrence was asymptomatic, treated with curative intent and detected by chest CT scan in 5 of the 136 (4%) patients or by fiberoptic bronchoscopy in 5 (4%) patients. Palliative therapy was given to 17 (47%) of asymptomatic patients, to 35 (70%) of symptomatic patients whose recurrence was detected by one of the follow-up procedures, and to 19 (38%) of symptomatic patients whose recurrence was detected outside the surveillance program. Survival One patient was lost to follow-up 15 days after operation. Median survival was 24 months (range, 1 to 198 months) from the date of operation and 7 months (range, 0 to 164 months) from the date of recurrence. From the date of recurrence, 3-year survival was 13% in the whole population, 31% in asymptomatic patients, 10% in symptomatic patients whose recurrence was detected by one of the follow-up procedures, and 4% in patients whose recurrence was diagnosed by an unscheduled procedure. Among asymptomatic patients, 3-year survival was 40% when the recurrence was exclusively thoracic, 45% when this thoracic recurrence was diagnosed by chest CT scan or fiberoptic bronchoscopy, and 70% when treatment was Table 2. Influence of Patient and Recurrence Characteristics on Survival From Recurrence: Univariate Analysis Variables Median Overall Survival From Recurrence (months) p Value Scheduled/unscheduled 11/ diagnostic procedure Asymptomatic/symptomatic 19/ Thorax only/extrathoracic 12/ or both Disease-free survival from 11/ operation 1/ 1 year Female/male 11/ Adenocarcinoma and large cell/squamous 11/ Diagnostic procedure Physical examination chest x-ray Bronchoscopy CT scan 12 Others 4 Performance status 2/ 2 15/ / 61 years 9/ CT computed tomographic. delivered with curative intent (Fig 1). For the whole population, median disease-free survival was 19 months (range, 1 to 198 months) and the 1-year and 3-year disease-free survival rates were 59% (95% confidence interval, 52% to 66%) and 24% (95% confidence interval, 18% to 30%), respectively. The results of univariate analysis are detailed in Table 2. The detection of recurrence by a scheduled procedure, the absence of symptoms at the time of recurrence, the thoracic site, and a disease-free survival from operation greater than 1 year were found to significantly enhance overall survival from recurrence. No difference in disease-free survival from operation was shown between asymptomatic and symptomatic patients (p 0.25). The results of the Cox model adjusted for treatment are shown in Table 3. Absence of symptoms, female sex, performance status of 2 or less, and age 61 years or younger were found to be significant favorable prognostic factors at the time of recurrence. Occurrence of a second cancer did not enter the model. Second Primary Tumors During this follow-up period, 22 new malignancies developed in the 192 patients (11%): gastrointestinal (5 patients), head and neck (5 patients), lung (3 patients), leukemia (2 patients), esophagus (2 patients), prostate (2 patients), hepatocarcinoma (1 patient), breast (1 patient), and bladder (1 patient). Of 11 (50%) of the 22 second primary cancers located in the surveillance area, 8 (73%) were diagnosed during the 3 postoperative years. Cost Analysis Calculated for a follow-up period of 19 months, which was the median disease-free survival, the cost per patient

5 Ann Thorac Surg WESTEEL ET AL 2000;70: INTENSIVE FOLLOW-UP IN RESECTED NSCLC 1189 Table 3. Cox Model Adjusted for Treatment Variables varied between $1,777 and $1,833, depending on the age of the scan system. For the median of occurrence of a thoracic recurrence of 17 months, the cost per patient varied between $1,231 and $1,291. For the median of occurrence of an extrathoracic or both thoracic and extrathoracic recurrence of 9 months, the cost per patient varied between $679 and $698. Seven patients with an asymptomatic thoracic recurrence detected by CT scan or fiberoptic bronchoscopy and treated with curative intent were alive 3 years after recurrence. The obtained number of years of life gained was 21. The cost because of thoracic CT scan with sections on the liver and adrenal glands and fiberoptic bronchoscopy varied between $270,945 and $281,710. The cost per year of life gained varied between $12,902 and $13,415. Comment Median Overall Survival From Recurrence (months) RR p (95% CI) a Value CU/PA/BSC 26/7/ ( ) 10 4 Asymptomatic/symptomatic 19/ ( ) 10 3 Female/male 11/ ( ) 10 3 Performance status 2/3 4 15/ ( ) 0.01 Age 61/ 61 years 9/ ( ) 0.01 a RR (95% CI) relative risk (95% confidence interval). BSC best supportive care; CU curative; PA palliative. Results of the present study demonstrate that this intensive follow-up is feasible, because 83% to 93% of the scheduled procedures were effectively performed. Even fiberoptic bronchoscopy, considered as the most invasive procedure, was accepted by most patients. Patients who were asymptomatic at the time of recurrence were shown to have a significantly longer survival from recurrence than symptomatic patients. Asymptomatic recurrences were more frequently diagnosed by thoracic CT scan and fiberoptic bronchoscopy than by physical examination and chest roentgenogram. Diversity of follow-up after complete resection of lung cancer shows that the ideal surveillance strategy has not been defined yet [1 3]. The most frequent surveillance includes periodic clinic visits and chest roentgenograms [1 3]. In a retrospective analysis of 182 patients, Virgo and colleagues [3] compared nonintensive with intensive follow-up. Patients were considered to have an intensive follow-up if they had 4 or more clinic visits, blood tests, or chest roentgenograms in any 12-month period, or at least 1 CT scan of the chest, bronchoscopy, or sputum cytology screening. No significant survival advantage was shown for intensive surveillance, but there was a tendency for a longer survival, which suggests that some benefit may exist. In a paper by Walsh and associates [5], the records of 358 lung cancer patients who were operated on and whose follow-up testing procedures and intervals varied were retrospectively reviewed. Patients whose recurrence was asymptomatic were shown to have a significantly longer survival. Survival might benefit from a surveillance strategy that allows the detection of more asymptomatic recurrences than periodic physical examinations and chest roentgenograms would. There is no doubt that only a randomized study can answer the question of whether a more intensive follow-up could enhance survival. We sought to define a follow-up program that would be relevant for a phase III study. The studied surveillance strategy followed principles considered as essential for an effective postoperative testing [7]: The shorter interval between examinations during the first 3 postoperative years, as well as the performance of a surveillance beyond 3 years, was consistent with the time of maximal risk of recurrence and the natural history of the tumor. The tests were directed at the thorax, which was a site of recurrence in 80% of asymptomatic patients. Therapy was available that could result in cure, significant prolongation of life, or palliation of symptoms. Rates of potentially curative and palliative treatments were increased in asymptomatic and symptomatic recurrences detected by one of the follow-up procedures. The initiation of earlier therapy may have improved the outcome compared with therapy given when the patient becomes symptomatic from the tumor. There is an increased risk of second malignancies, especially lung and head and neck cancers, which can be detected by thoracic CT scan or fiberoptic bronchoscopy. Our results are consistent with those of Walsh and coworkers [5]. Indeed, asymptomatic recurrences on which advantages of surveillance testing are focused have been shown to be of better prognosis. As no difference in disease-free interval was shown between asymptomatic and symptomatic patients, lead time bias cannot explain the survival advantage observed in patients who were asymptomatic at the diagnosis of recurrence. The diagnostic procedure (scheduled versus unscheduled) did not enter the multivariate model. However, a majority of recurrences that could be treated with curative intent in asymptomatic patients were diagnosed by chest CT scan or fiberoptic bronchoscopy. A survival advantage might be found with chest CT scan and fiberoptic bronchoscopy in a larger subgroup of patients. A small but significant advantage would be acceptable knowing the number of patients undergoing operation for non small cell lung cancer worldwide. A limitation of this study is the population. Only patients whose postoperative care was performed by our Chest Disease Department were included. They represent only part of the population that has undergone this operation, and this may explain the differences in resection types, sites of recurrences, and survival observed between our series and most published series. In our series, the rates of pneumonectomies of 59% and of

6 1190 WESTEEL ET AL Ann Thorac Surg INTENSIVE FOLLOW-UP IN RESECTED NSCLC 2000;70: thoracic recurrences of 52% were higher than in published studies [3, 5, 6, 10]. Rates of pneumonectomies varied between 17% and 27% in studies including stage I to IIIb tumors [5, 6]. Rates of thoracic recurrences were approximately 20% in most studies [3, 5, 10], with a particularly low rate (7%) in the study reported by Younes and associates [6]. These studies had a surgical recruitment. Patients analyzed in the present study had been referred to the Chest Disease Department rather than directly to the Thoracic Surgery Department, often because their disease presentation was responsible for difficult therapeutic decision. In determining appropriate follow-up practices, besides quantity of life, other important factors to be considered are quality of life and financial aspects. Intensive follow-up might improve quality of life by a better diagnosis and treatment of tobacco- related comorbidities and of postoperative symptoms such as pain or dyspnea. It might also be deleterious on quality of life because of potential anxiety generated by repeated procedures. Younes and colleagues [6] retrospectively evaluated two follow-up strategies. Sixty-seven patients had a follow-up consisting of physical examination every 3 months, chest roentgenogram alternating with thoracic CT scan every 6 months, and liver function tests every year for 2 years. Sixty-three patients were seen on a symptom-oriented basis. The number of health problems other than recurrence requiring emergency room admissions or in-hospital stays was significantly reduced in the strict follow-up group. There is no published study evaluating follow-up strategies after lung cancer resection that includes a quality-of-life questionnaire. The cost for a follow-up period of 19 months was among the lowest when compared with the costs estimated by Virgo and associates [4] from $1,533 to $9,145 for eight published follow-up strategies of increasing intensities for the 5-year period between 1990 and The cost because of thoracic CT scan and fiberoptic bronchoscopy for 1 year of life gained was acceptable. In the study by Younes and coworkers [6], costs of the strict group were significantly higher than those of the symptom group, with no difference in survival after recurrence between the two groups. In the present study, a significant survival advantage was shown in asymptomatic patients whose recurrence was diagnosed during the follow-up strategy, and this seems mostly related to chest CT can and fiberoptic bronchoscopy. This testing strategy is worth being studied within a large randomized study including an indepth cost analysis and evaluation of quality of life and of potential other medical benefits. To detect a 5% survival advantage for the studied follow-up strategy, a total of 2,973 patients who have been operated on should be included. We thank Doctors Claude Prenat, Didier Pernet, Marie-Lise Paquin, Jean Lahourcade, Fabrice Chaussade, and Jean-Louis Cusenier for their cooperation, and Doctor Willard A Fry for having reviewed this paper. References 1. Johnson FE, Naunheim KS, Coplin MA, Virgo KS. Geographic variation in the conduct of patient surveillance after lung cancer surgery. J Clin Oncol 1996;14: Naunheim KS, Virgo KS, Coplin MA, Johnson FE. Clinical surveillance testing after lung cancer operations. Ann Thorac Surg 1995;60: Virgo KS, McKirgan LW, Caputo MCA, et al. Post-treatment management options for patients with lung cancer. Ann Surg 1995;222: Virgo KS, Nauheim KS, McKirgan LW, et al. Cost of patient follow-up after potentially curative lung cancer treatment. J Thorac Cardiovasc Surg 1996;112: Walsh GL, O Connor M, Willis KM, et al. Is follow-up of lung cancer patients after resection medically indicated and cost effective? Ann Thorac Surg 1995;60: Younes RN, Gross JL, Deheinzelin D. Follow-up in lung cancer. How often and for what purpose? Chest 1999;115: Edelman MJ, Meyers FJ, Siegel D. The utility of follow-up testing after curative cancer therapy. A critical review and economic analysis. J Gen Intern Med 1997;12: The GIVIO Investigators. Impact of follow-up testing on survival and health-related quality of life in breast cancer. A multicenter randomized controlled trial. JAMA 1994;271: Rosselli Del Turco MR, Palli D, Carridi A, et al. Intensive diagnostic follow-up after treatment of primary breast cancer: a randomized trial. JAMA 1994;271: Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumor in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995;109: Kaplan E, Meier P. Nonparametric estimation from incomplete observation. J Am Stat Assoc 1958;53: Cox DR, Oakes D. Time-dependent covariates. In: Cox DR, Oakes D, eds. Analysis of survival data. London: Chapman & Hall Ltd, 1984: Dixon WJ. BMDP statistical software. Berkeley, CA: University of California, 1992.