Surgical resection of multifocal non-Small cell lung cancer is associated with prolonged survival

ORIGINAL ARTICLES: GENERAL THORACIC

Surgical Resection of Multifocal Non-Small Cell Lung Cancer Is Associated With Prolonged Survival Richard J. Battafarano, MD, PhD, Bryan F. Meyers, MD, Tracey J. Guthrie, BSN, Joel D. Cooper, MD, and G. Alexander Patterson, MD Department of Surgery, Division of Cardiothoracic Surgery, Washington University School of Medicine, St. Louis, Missouri

Background. Revisions in the international system for the staging of lung cancer, adopted in 1997, assigned the T4 descriptor to separate tumor nodules in the same lobe and the M1 descriptor to tumor nodules in a different lobe. Consequently, these changes shifted the stage of patients with these lesions to stage IIIB or stage IV. The goal of this review was to determine the impact of multifocal non-small cell lung cancer on survival. Methods. A database analysis of our cardiothoracic surgery tumor registry was performed to identify all patients who underwent surgical resection of non-small cell lung cancer (NSCLC), who were ultimately determined to have pathologically node-negative disease from 1994 to 1999. All pathology reports were individually reviewed. Survival data were collected on each patient from the date of surgery with a mean duration of follow-up of 46.3 months. Kaplan Meier actuarial survival was determined for all patients. Results. Forty-four patients were identified who underwent complete resection of multiple NSCLC tumors. During this same period, 504 patients underwent com-

plete resection of stage I NSCLC tumors. The 3-year actuarial survival for patients with T1/N0/M0 tumors was 79.6%. In comparison with patients with T1/N0/M0 tumors, the 3-year actuarial survival rates of patients with T2/N0/M0 tumors (72.3%, p ⴝ 0.056), T4/N0/M0 tumors (66.5%, p ⴝ 0.058), and T1 to T2/N0/M1 tumors (63.6%, p ⴝ 0.201) were lower. However, these differences did not achieve statistical significance. Conclusions. Although there was a trend toward decreased survival in patients with multifocal NSCLC compared with patients with stage I NSCLC, this did not achieve statistical significance. Importantly, survival in these subgroups of patients with stage IIIB or stage IV disease (stage determined solely on the basis of multifocal NSCLC) is better than the survival reported in the series that formed the foundation for these staging changes. These data support complete surgical resection of multifocal lung tumors in patients with node-negative NSCLC. (Ann Thorac Surg 2002;74:988 –94) © 2002 by The Society of Thoracic Surgeons

T

patients with stage IIIB or stage IV NSCLC are considered to have a poor prognosis, they are often treated with chemotherapy or chemoradiation therapy and are not offered surgical resection. Accurate differentiation of synchronous primary lung cancers and from satellite pulmonary metastases in the preoperative clinical and radiographic staging of patients is not possible. Therefore, the purpose of this study was to determine the impact of completely resected multifocal NSCLC on patient survival.

he American Cancer Society estimates that there will be approximately 169,000 patients who will be diagnosed with non-small cell lung cancer (NSCLC) in 2001 [1]. The majority of patients present with locally advanced or metastatic NSCLC and ultimately succumb to their disease. However, patients with NSCLC that is amenable to complete surgical resection have a more favorable prognosis. Patients with completely resected stage I and stage II NSCLC have actuarial 5-year survival rates that range from 55% to 75% and 35% to 45%, respectively [2–5]. Revisions in the international system for the staging of lung cancer, adopted in 1997, assigned the T4 descriptor to separate tumor nodules in the same lobe and the M1 descriptor to tumor nodules in a different lobe [5]. Consequently, these changes shifted the stage of patients with these lesions to stage IIIB or stage IV. Because

Presented at the Thirty-eighth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –30, 2002. Address reprint requests to Dr Battafarano, Department of Surgery, Division of Cardiothoracic Surgery, Washington University School of Medicine, One Barnes-Jewish Plaza, 3107 Queeny Tower, St. Louis, MO 63110-1013; e-mail: [email protected].

© 2002 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

Material and Methods Population Under Study A database analysis of our cardiothoracic surgery tumor registry was performed. Five hundred forty-eight patients underwent surgical resection of NSCLC with pathologically node-negative disease from January 1, 1994 to December 31, 1999. The surgical pathology reports of all patients identified by this database query were individually reviewed and the final pathologic staging was assigned using the 1997 revisions in the International System for Staging Lung Cancer. Patients with pathologically staged T1/N0/M0 (Stage IA), T2/N0/M0 0003-4975/02/$22.00 PII S0003-4975(02)03878-X

Ann Thorac Surg 2002;74:988 –94

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

Table 1. Demographic Features and Treatment Information About the 548 Patients With Stage I or Multifocal NSCLC Undergoing Resection Stage I NSCLC n ⫽ 504 Gender Male Female Age (years) Comorbidity severity None Mild Moderate Severe Type of resection Lobectomy Wedge(s) Segmentectomy Bilobectomy Lobectomy and Wedge Sleeve lobectomy Pneumonectomy Largest tumor size ⬍ 3 cm ⬎ 3 cm

Multifocal NSCLC n ⫽ 44

273 (54.2%) 231 (45.8%) 66.6 ⫾ 9.9

15 (34.1%) 29 (65.9%) 68.5 ⫾ 8.7

159 (31.5%) 167 (33.2%) 129 (25.6%) 49 (9.7%)

9 (20.5%) 16 (36.4%) 13 (29.5%) 6 (13.6%)

412 (81.7%) 28 (5.5%) 16 (3.2%) 12 (2.4%) 7 (1.4%)

19 (43.2%) 6 (13.6%) 2 (4.6%) 3 (6.8%) 8 (18.2%)

11 (2.2%) 18 (3.6%)

0 (0%) 6 (13.6%)

273 (54.2%) 231 (45.8%)

26 (59.1%) 18 (40.9%)

NSCLC ⫽ non-small cell lung cancer.

(Stage IB), T4(multifocal)/N0/M0 (Stage IIIB), and T1 to 2/N0/M1(multifocal) NSCLC (stage IV) tumors were included in the analysis. Patients with carcinoid tumors, patients who had been previously treated for lung cancer, patients with contralateral disease, and patients who received preoperative chemotherapy or radiation therapy were excluded from this analysis. Tumor recurrence, patient survival, and cause of death were determined for each patient. Follow-up information on all patients was acquired within the last 6 months through clinic follow-up notes, direct patient or family contact, contact with the patient’s primary care physician, and review of all death certificates. This study represents a secondary data analysis of a prospective cohort study. Approval for this study was granted from the Washington University School of Medicine Human Studies Committee.

Comorbidity Severity Since 1994, specifically trained cancer registrars have prospectively coded overall severity of comorbidity on cancer patients treated at Barnes-Jewish Hospital, St. Louis, MO, using a modified version of the KaplanFeinstein Index (KFI). Minor modifications to the KFI were made to include diabetes mellitus and other comorbid conditions not present on the original KFI. The KFI classifies the pathophysiologic derangement of each comorbid ailment based on a 4-category system [0, 1, 2, 3]. An overall comorbidity score (none, mild, moderate, or severe) is determined based on the number of ailments

989

and their individual degree of decompensation. The cancer registrars determined the comorbidity severity in their usual abstraction of the medical record [6].

Statistical Analysis Descriptive statistics were used to describe the patients’ characteristics and outcomes. Normally distributed continuous data are expressed as mean ⫾ standard deviation throughout. Medians with ranges are used when continuous data are not normally distributed. Categorical data are expressed as counts and proportions. ␹2 or Fisher’s exact tests were used to analyze the categorical data. Kaplan-Meier (product-limit) graphs were used to demonstrate survival over time and freedom from recurrence of disease. Survival and event-free survival comparison between groups of patients was completed using the Mantel-Haenszel log rank test. Cox multivariate proportional hazards regression model was used to identify independent risk factors for death following surgical resection for stage I NSCLC. Time to death following surgery was selected as the primary outcome. The likelihood ratio method was used to determine hazard ratios, and the hazard ratio was used to approximate the relative risk (RR). All data analysis was performed using Systat (Systat 10.0 for Windows; SPSS Inc., Chicago, IL). All p values less than 0.05 were considered to be statistically significant.

Results Patient Demographics Five hundred forty-eight patients underwent complete surgical resection of NSCLC who were ultimately determined to have pathologically node-negative disease between January 1, 1994 and December 31, 1999 (Table 1). Two hundred eighty-eight patients were male (53%) and 260 patients (47%) were female. The mean age was 66.7 years old (⫾ 9.8 years). The distribution of patients by pathologic stage was stage IA 273 (49.8%), stage IB 231 (42.2%), stage IIIA 27 (4.9%), and stage IV 17 (3.1%). Table 2. Histology of Resected Lesions From 44 Patients Same histology Adenocarcinoma Squamous Total Different histology Adenocarcinoma ⫹ squamous Adenocarcinoma ⫹ large cell Adenocarcinoma ⫹ large cell ⫹ squamous Squamous ⫹ large cell Total Bronchioalveolar histology Pneumonic form of bronchioalveolar Multifocal bronchioalveolar Adenocarcinoma ⫹ bronchioalveolar Squamous ⫹ bronchioalveolar Total

18 6 24 4 3 1 1 9 4 3 2 2 11

990

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

Table 3. Forty-Four Patients With Multifocal NSCLC Ipsilateral distribution Same lobe: 27 Different lobes: 17

44

NSCLC ⫽ non-small cell lung cancer.

Forty-four patients had multiple foci of NSCLC resected. Twenty-seven patients had multiple lesions in a single lobe (T4/N0/M0) and 17 had lesions in a separate lobe (T1 to 2/N0/M1). Thirty-eight patients (86%) underwent anatomic resection. The majority underwent lobectomy (43%) or lobectomy and wedge excision (18%). However, 9 patients (20.5%) required a more extensive operation to assure complete resection. Eight patients with small lesions underwent limited pulmonary resection by segmentectomy (4.5%) or multiple wedge excisions (14%). In comparison with patients with stage I NSCLC, patients with multifocal NSCLC had a similar distribution of comorbidity severity, underwent anatomic resection at an equivalent rate, and had a comparable proportion of patients whose largest tumor was greater than 3 cm. In patients with multifocal NSCLC (Table 2), 24 patients had multiple tumors with the same histology, 9 patients had tumors with different histology, and 11 patients had multifocal disease that included bronchoalveolar carcinoma. Twenty-eight patients (64%) had a least one tumor that was an adenocarcinoma. Seven other patients (16%) had multifocal bronchoalveolar carcinoma and 2 patients (4.5%) had a bronchoalveolar carcinoma in association with a separate squamous cell carcinoma (Table 3).

Survival Survival data were collected on each patient from the date of surgery. Ten patients (1.8%) died in the immediate postoperative period. The 3-year actuarial survival for Fig 1. Survival stratified by pathologic stage.

Ann Thorac Surg 2002;74:988 –94

patients with T1/N0/M0 tumors was 79.6% (Fig 1). In comparison with patients with T1/N0/M0 tumors, the 3-year actuarial survival rates of patients with T2/N0/M0 tumors (72.3%, p ⫽ 0.056), T4/N0/M0 tumors (66.5%, p ⫽ 0.058), and T1 to T2/N0/M1 tumors (63.6%, p ⫽ 0.201) were lower. However these differences were not statistically significant. The association of tumor stage and survival was analyzed using a multivariate Cox regression analysis. The additional variables of age, gender, and comorbidity severity were included in the model (Table 4). There was no significant association between tumor stage and survival in this analysis. Compared with patients with T1/ N0/M0 tumors, patients with T2/N0/M0 tumors, T4/ N0/M0 tumors, and T1 to T2/N0/M1 tumors had no increased risk of death in the follow-up period. Only comorbidity severity and age had a significant impact on survival.

Incidence of Recurrent Lung Cancer One hundred five patients (19.2%) developed recurrent lung cancer in the follow-up period. The rate of recurrence as a function of stage was T1/N0/M0 (38 of 273, 13.9%), T2/N0/M0 (54 of 231, 23.4%), T4/N0/M0 (7 of 27, 25.9%), T1 to 2/N0/M1 (6 of 17, 35%). The recurrence-free 3-year survival for patients with T1/N0/M0 tumors was 88.0%. The recurrence-free 3-year survival for patients with T2/N0/M0 (78.1%, p ⫽ 0.003), T4/N0/M0 (72.5%, p ⫽ 0.027), and T1 to T2/N0/M1 (61.3%, p ⫽ 0.002) was significantly lower than the survival for patients with T1/N0/M0. However, there was no significant difference in recurrence-free survival between patients with T2/N0/ M0, T4/N0/M0, and T1 to 2/N0/M1 tumors (Fig 2).

Effect of Tumor Histology on Survival in Patients With Multifocal NSCLC In order to determine if specific subgroups of multifocal NSCLC had a more favorable prognosis, we examined

Ann Thorac Surg 2002;74:988 –94

991

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

Table 4. Results of Multivariate Analysis of Prognostic Factors Influencing Survival for the Entire Group of Node Negative Patients Variable Comorbidity severity None Mild Moderate Severe Age (years) Gender Female Male Stage T1N0M0 T2N0M0 T4N0M0 T1-2N0M1

Number

Relative Risk

95% CI

p Value

168 183 142 55 551

1.000 1.406 1.919 1.770 1.025

... 0.965–2.052 1.320 –2.790 1.088 –2.883 1.009 –1.041

... 0.076 ⬍ 0.001 0.022 0.002

260 288

1.000 0.934

... 0.705–1.239

... 0.601

273 231 27 17

1.000 0.857 1.575 1.197

... 0.325–2.264 0.874 –2.838 0.452–3.177

... 0.938 0.131 0.688

survival in patients with multiple tumors with the same histology (which would include patients with intrapulmonary metastases), different histology (separate primary tumors), and where at least one of the tumors was bronchoalveolar carcinoma. There was no difference in survival between these groups (Fig 3).

Comment The optimal management of patients with multifocal non-small cell carcinoma remains controversial. Preoperatively, it is often difficult to determine if multiple lesions represent separate primary lung cancers or intrapulmonary metastases. This distinction can remain difficult even after pathologic examination of the surgical specimens when the histology of the lesions is the same. Martini and Melamed [7] described their experience with multifocal lung cancer and outlined their recommendations differentiating separate primary lung cancers and intrapulmonary metastases. They defined lesions as mul-

tiple synchronous lung cancers if the tumors were physically distinct and separate, and if the histology of the tumors was different. In cases where the histology of the lesions was the same, lesions without carcinoma in the lymphatics common to both lesions and without evidence of extrapulmonary metastases at the time of diagnosis were considered synchronous lung cancers. The original report describing the revisions in the international system for staging lung cancer, adopted in 1997, assigned the T4 descriptor to a tumor of any size with satellite tumor nodule(s) within the ipsilateral primary-tumor lobe of the lung and the M1 descriptor to separate metastatic tumor nodules in the ipsilateral nonprimary-tumor lobe of the lung. However, the American Joint Committee on Cancer (AJCC) staging manual assigns the T4 descriptor to a tumor of any size with separate tumor nodules in the same lobe and the M1 descriptor to separate tumor nodule(s) in a different lobe (ipsilateral or contralateral) [8]. The AJCC descriptors

Fig 2. Freedom from disease recurrence stratified by pathologic stage.

992

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

Ann Thorac Surg 2002;74:988 –94

Fig 3. Survival stratified by pathology of multifocal disease. (BAC ⫽ bronchioloalveolar cancer.)

include both synchronous primary lung cancers and intrapulmonary metastases. The importance of distinguishing synchronous primary lung cancers and intrapulmonary metastases has been debated. Okada and associates [9, 10] described their experience with surgical resection of multiple primary lung cancers and intrapulmonary metastases. Using the criteria of Martini and Melamed [7], they found that the 5-year actuarial survival of patients with completely resected synchronous primary lung cancers was 70%. In a separate analysis, these authors found that node-negative patients with intrapulmonary metastases had a 5-year actuarial survival rate of 45% [10]. Recently, Okumura and colleagues [11] reported their experience with intrapulmonary metastases. In their series, nodenegative patients with pulmonary metastases had a 5-year actuarial survival rate of 37%. In this analysis, patients who underwent resection of multifocal NSCLC had a favorable 3-year actuarial survival. Patients with T4/N0/M0 tumors and T1 to T2/N0/M1 tumors had survival rates of 66.5% and 63.6%, respectively. Although this was lower than the survival rates of patients with T1/N0/M0 lesions (79.6%) and T2/N0/M0 lesions (72.3%), this difference did not achieve statistical significance. There was a significant difference in recurrence-free survival at 3 years between patients with T1/N0/M0 tumors and patients with T2/N0/M0, T4/N0/M0, or T1 to T2/N0/M1 tumors. Overall, patients with multifocal tumors (T4/ N0/M0 or T1 to T2/N0/M1) had a 3-year survival and recurrence-free survival that was similar to the survival of patients with T2/N0/M0 tumors. We did not attempt to differentiate between synchronous primary lung cancers and intrapulmonary metastases in this analysis. However, the incidence of multifocal lung cancer in this series (44 of 551, 8.0%) was similar to the incidence reported by other investigators [10 –12]. Importantly, there was no difference in survival between patients with multiple tumors of the same histology (which would include patients with intrapulmonary metastases) and patients with multiple tumors of different

histology (synchronous primary tumors). Because patients with bronchoalveolar carcinoma often have a favorable prognosis, we specifically analyzed the survival of patients where at least one of the tumors was bronchoalveolar carcinoma. Again, there was no difference in survival in this subset of patients with multifocal NSCLC. In approaching patients with multifocal NSCLC, it is critical to accurately stage each patient before proceeding with surgical resection. In addition to a chest CT that includes the liver and the adrenal glands, all patients have a CT scan or MRI of the brain, and a bone scan (or more recently a PET scan) to rule out metastatic disease. We routinely perform a cervical mediastinoscopy with biopsy of the paratracheal and subcarinal lymph nodes before proceeding with thoracotomy. Only patients with no evidence of disease after this thorough evaluation proceed to surgical resection. The optimal management of patients with multifocal NSCLC remains controversial. However, patients with node-negative disease that undergo resection have a favorable survival that is comparable to patients with stage IB tumors. Therefore we continue to recommend complete surgical resection of multifocal NSCLC in nodenegative patients with adequate pulmonary reserve.

References 1. Greenlee RT, Hill-Harmon ME, Murray T, Thun M. Cancer statistics, 2001. CA Cancer J Clin 2001;51:15–36. 2. Martini N, Burt ME, Bains MS, McCormack PM, Rusch VW, Ginsberg RJ. Survival after resection of stage II non-small cell lung cancer. Ann Thorac Surg 1992;54:460–5. 3. Ginsberg RJ, Rubinstein LV. Randomized trial of lobectomy versus limited resection for T1 N0 non- small cell lung cancer. Lung Cancer Study Group. Ann Thorac Surg 1995;60: 615–22. 4. Martini N, Bains MS, Burt ME, et al. Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Cardiovasc Surg 1995;109:120–9. 5. Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710–7. 6. Battafarano RJ, Piccirillo JF, Meyers BF, et al. Impact of comorbidity on survival after surgical resection in patients

Ann Thorac Surg 2002;74:988 –94

7. 8. 9. 10.

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

with stage I non-small cell lung cancer. J Thorac Cardiovasc Surg 2002;123:280 –7. Martini N, Melamed MR. Multiple primary lung cancers. J Thorac Cardiovasc Surg 1975;70:606–12. American Joint Committee on Cancer. Cancer Staging Manual, 5th ed. Philadelphia: Lippincott Williams & Wilkins, 1997. Okada M, Tsubota N, Yoshimura M, Miyamoto Y. Operative approach for multiple primary lung carcinomas. J Thorac Cardiovasc Surg 1998;115:836– 40. Okada M, Tsubota N, Yoshimura M, Miyamoto Y, Nakai R.

993

Evaluation of TMN classification for lung carcinoma with ipsilateral intrapulmonary metastasis. Ann Thorac Surg 1999;68:326–30. 11. Okumura T, Asamura H, Suzuki K, Kondo H, Tsuchiya R. Intrapulmonary metastasis of non-small cell lung cancer: a prognostic assessment. J Thorac Cardiovasc Surg 2001;122: 24– 8. 12. Deslauriers J, Brisson J, Cartier R, et al. Carcinoma of the lung. Evaluation of satellite nodules as a factor influencing prognosis after resection. J Thorac Cardiovasc Surg 1989;97: 504–12.

DISCUSSION DR JOE B. PUTNAM (Houston, TX): Thank you Dr Battafarano, for your excellent presentation and outstanding results in this challenging lung cancer patient population: pathologically staged T4 tumors/satellite nodules, IIIB; or nodules in separate lobes, M1. I was delighted to review the manuscript in preparation for this discussion. These excellent results reflect both the authors’ careful care of patients in the clinical environment and their careful review of their outcomes after such treatment. The staging of lung cancer assists the clinician in evaluation of effects of therapy, allows the exchange of information, and guides the choice of the most appropriate treatment. In 1997, Clifton F. Mountain, MD, presented revisions in the international staging system for lung cancer. Specific changes were made in classifying multiple lung nodules: satellite tumor nodules were designated as T4 and separate metastatic tumor nodules in different lobes as M1. Survival at 4 years was 8% for all T4 tumors (including all nodal stages), and 2% for M1 disease. An earlier review by Dr Naruke and colleagues showed that overall survival for pathologic T4 lesions was about 20%. Dr Naruke also showed that separate tumor nodules in the same lobe approximates all T4 survival and that separate tumor nodules in different lobes approximates all metastatic disease survival. With separate nodules in the same lobe, 177 patients had 22% five-year survival, and in different lobes, 7%. The study presented here is a retrospective review of pathologic data. A numerator is presented, [for example, T4 tumors resected]; however, the denominator [T4 resected ⫹ T4 not resected] is missing. How many patients underwent pulmonary resection for lung cancer during this period? How many patients had clinical T4/satellite nodules, or M1/nodules in separate lobes, who were not selected for pulmonary resection? In this article the selection of patients for resection based upon clinical staging criteria was not clearly described, nor was it compared to the final pathologic stage. Would you describe the clinical staging process which was undertaken on these patients and how these patients were selected for resection? At The University of Texas M. D. Anderson Cancer Center, the guidelines for resection of T4 satellite nodules are well established. Patients with resectable T4 satellite nodules are clinically staged radiographically and with mediastinoscopy. If mediastinoscopy is negative, resection (typically lobectomy) is undertaken to obtain local control. A mediastinal lymph node dissection is performed to enhance the surgical staging. If the mediastinoscopy is positive, alternatives to surgery are considered. What criteria would you use to exclude patients from resection and consider alternatives to surgery? I applaud the authors’ efforts in describing their excellent results after resection in patients with pathologically defined multifocal lung cancer. However, in the absence of clinical staging information, we miss the opportunity to establish criteria for selecting patients with T4/Satellite nodules or M1/nodules in

separate lobes, who will optimally benefit prior to the initiation of definitive therapy. These excellent results reflect outcome in resected patients. We must select our patients on clinical criteria prior to resection to optimize survival and local control in these clinically challenging but resectable advanced tumors. Thank you. DR BATTAFARANO: Thank you, Dr Putnam. In addressing your first question, which is what is the denominator, it is oftentimes difficult in a retrospective study to determine that exactly. For instance, in the patients with T4 tumors, or multifocal lung cancer who have N1 disease, I do not know the exact number. We are going through our database now to determine that exactly. Any patients with N2 disease would not undergo resection, especially if this was determined at the time of mediastinoscopy, which we also use routinely, as you do, in your preoperative evaluation of patients. The hard part with determining the denominator is that we don’t know how many patients are seen by an internist or a medical oncologist who see two lesions on the CAT scan and never send those patients to us. But I do know that in our practice, if a patient is sent to us with multifocal tumor, that in and of itself does not dissuade us necessarily from resection. As long as they have adequate pulmonary reserve, as long as they have no distant metastatic disease, and if at the time of surgery we performed mediastinoscopy the nodes are negative, then we will proceed with exploration and resection. In the preoperative preparation of these patients, approximately two-thirds of the tumors were identified either radiographically preoperatively or were identified by the surgeon at the time of surgery. So prior to the definitive resection, in two-thirds of the patients we knew that they had multifocal disease. In the other third, these were tumors when the lobectomy specimen was sent to the pathologists, they identified the multifocality in that specimen. To address your second question with the clinical staging, we use almost the exact same algorithm that you use. We get routine MRI, brain scans, and we also obtain bone scans, and now, more recently, PET scanning. But I think routine use of mediastinoscopy is an important aspect, because if the mediastinoscopy is positive, then we do not proceed with thoracotomy or resection. I think one of the things that would dissuade us from taking these patients to the operating room would be if their lymph nodes were positive at the time of mediastinoscopy. We have not utilized PET to exclude patients unless they have distant metastatic disease. If a patient has a positive PET in the mediastinum, we would do mediastinoscopy, and if that were positive, we would stop. If that were a false positive and truly negative at the time of mediastinoscopy, we would proceed with resection.

994

BATTAFARANO ET AL PROLONGED SURVIVAL WITH SURGICAL RESECTION OF NSCLC

DR JOHN BENFIELD (Los Angeles, CA): Doctor Peter Roberts and I, also wished to evaluate the impact of multifocal lung cancer upon the therapeutic effectiveness of resection. Our review differed from yours somewhat in that we focused upon bronchoalveolar carcinoma, or BAC, because of its propensity to be multifocal. Among 73 BAC patients for whom we did complete pulmonary resection, there were 14 patients with a mean age of 65 years, ranging from 51 to 87, who had multifocal lesions without lymph node metastases, N0. Follow-up was 100%, for a mean of 4.9 years, ranging from 2.6 to 8.5. The tumor distribution was unilateral in 9 and bilateral in 5 of the patients with multifocal disease. Nine of the patients had 2 lesions, 4 patients had 3 lesions, and one patient had innumerable diffuse or discrete foci in a single lobe. The multifocal nature of the disease was discovered intraoperatively in four patients. Operative mortality was zero. Postoperatively 9 patients were staged pIIIB or pIV on the basis of multiple foci of similar morphology. Two patients had clearly different cell types, implying multiple stage I primaries, and 3 patients had a combination of adenocarcinoma and BAC. The overall five-year survival after resection of multifocal cancers that included BAC was 64%. Unilateral or bilateral distribution had no impact upon survival. Therefore, in our experience, the current staging system has not been prognostic for N0 multifocal BAC, and, like you, we recommend complete resection of multifocal BAC. DR DOUGLAS E. WOOD (Seattle, WA): I have a question about your definition of multifocality. Multifocality obviously includes synchronous primary tumors as well as satellite lesions, and multifocality is a convenient way of merging them. I am not sure that histology alone is adequate differentiation of those two very different tumor characteristics that may determine behavior. I am wondering if you are using immunocytochemistry, a further examination not just of histology type but of histology appearance, as well as flow cytometry to differentiate between lesions that may have the same histology but actually may be synchronous primary tumors. I think all of us would be far more

Ann Thorac Surg 2002;74:988 –94

comfortable recommending aggressive surgical resection in patients with synchronous primary tumors, and I think those patients are really biologically different than the T4 or M1 disease as satellite nodules have been characterized in the most recent staging classification. DR BATTAFARANO: Well, it is very difficult to be sure, even from the pathology standpoint, whether or not these represent satellite nodules or intrapulmonary metastases or whether or not they are synchronous tumors. One issue that our pathologists rely heavily on, they have not done specific immunocytochemistry stains, but they look at the parenchyma around the lesions, especially the second lesion, to give an idea. Primary lung cancers tend to have a very vigorous inflammatory reaction with the lung, and that is why you see the spiculated lesions on CAT scan, whereas intrapulmonary metastases tend to be tumor cells that are either well circumscribed or have very little desmoplastic reaction. So that is one aspect of it. But we have not specifically had our pathologists using immunocytochemistry or flow cytometry. DR WOOD: I guess I would say you have got this great cohort of patients, and that is one thing that can still be done retrospectively. It might be interesting to look at that to see with ICC and, if necessary, with flow cytometry to differentiate true apparent satellite lesions from what appears to be synchronous primary tumors and see if that differentiates the outcomes in your group of patients, since you have got such a great cohort. Dr Battafarano, I think it is important from a retrospective analysis and pathology analysis, but as clinicians, we are still faced with the problem of patients with multiple nodules either radiographically identified preoperatively or at the time of surgery when you identify them by palpation. I think that all of those other analyses have to be done after the resection is performed, and that is why we grouped them together for this analysis. I think that this gives us more clinical relevance for how to approach the next patient. Thank you.