Fax: (011) 81-78-929-2380
Characteristics and prognosis of patients after resection of nonsmall cell lung carcinoma measuring 2 cm or less in greatest dimension
Article first published online: 5 JUN 2003
Copyright © 2003 American Cancer Society
Volume 98, Issue 3, pages 535–541, 1 August 2003
How to Cite
Okada, M., Sakamoto, T., Nishio, W., Uchino, K. and Tsubota, N. (2003), Characteristics and prognosis of patients after resection of nonsmall cell lung carcinoma measuring 2 cm or less in greatest dimension. Cancer, 98: 535–541. doi: 10.1002/cncr.11530
- Issue published online: 18 JUL 2003
- Article first published online: 5 JUN 2003
- Manuscript Accepted: 16 APR 2003
- Manuscript Revised: 14 APR 2003
- Manuscript Received: 25 FEB 2003
- nonsmall cell lung carcinoma (NSCLC);
- small-sized tumor;
There remains ongoing controversy with regard to the optimal management strategy and the prognostic significance of small-sized nonsmall cell lung carcinoma. Therefore, in the current study, the authors analyzed the clinical characteristics of patients who underwent complete resection of these lung tumors, the follow-up data, and the significant prognostic factors.
Of 1726 consecutive patients surgically treated for proven primary lung carcinoma, 265 patients underwent complete removal of a nonsmall cell lung carcinoma in which the greatest dimension of the resected specimen was ≤ 2 cm.
The cancer-specific 5-year and 10-year survival rates were 86% and 83%, respectively. Univariate analyses revealed that advanced pathologic stage, a tumor size of 16–20 mm, lymphatic vessel invasion, vascular vessel invasion, a high serum level of carcinoembryonic antigen (CEA), and extended resection were significantly unfavorable prognostic factors. Among these factors, multivariate analyses demonstrated that pathologic stage (P < 0.0001), vascular vessel invasion (P = 0.0040), and CEA level (P = 0.0291) were significant, independent determinants of survival. None of the patients with pathologic Stage I disease, no vascular vessel invasion, and a low serum CEA level died of their disease after undergoing complete resection.
The preoperative level of serum CEA and vascular vessel invasion by tumor cells were found to be independent prognostic factors that were as significant as the well established determinant of pathologic stage for patients with a nonsmall cell lung carcinoma measuring ≤ 2 cm in greatest dimension. These data may contribute to the explanation of the lower-than-expected survival of patients after complete surgical resection of such a small-sized tumor. Cancer 2003;98:535–41. © 2003 American Cancer Society.
Recently, as a result of the development of high-resolution computed tomography (CT) and the advent of spiral CT to screen patients for lung carcinoma, thoracic surgeons have become able to detect small-sized lung tumors and the frequency of such detection has increased rapidly. Some studies have shown that the surgical outcome of patients with a nonsmall cell lung carcinoma (NSCLC) measuring ≤ 2 cm in greatest dimension is significantly better than that of patients with a tumor measuring ≥ 2.1 cm.1–3 At the study institute, patients with lung carcinoma used to undergo lobectomy, but since 1992 as a prospective trial we have attempted segmentectomy with lymph node dissection, even in noncompromised patients with Stage IA NSCLC measuring ≤ 2 cm.1, 4
Because the number of the small-sized lung carcinomas being detected is dramatically increasing, it becomes vital to consider the clinicopathologic behavior of these tumors and to design an optimal management strategy. To our knowledge, there is little information concerning the long-term outcome of patients who undergo complete resection of small-sized NSCLC. In addition, there remains ongoing controversy regarding the prognostic significance of these tumors. The objectives of the current study were to analyze the clinical characteristics of patients who underwent complete resection of NSCLCs measuring ≤ 2 cm in greatest dimension and the follow-up data regarding cancer-related death and to determine those factors believed to be significant for prognosis.
MATERIALS AND METHODS
Between June 1984 and May 2002, 1726 consecutive patients underwent surgery for primary lung carcinoma by the same surgical team. Among these patients, 265 with proven NSCLC in which the greatest dimensions of the resected specimens were ≤ 2 cm underwent complete removal of the primary tumor together with the lymph nodes.5 We retrospectively reviewed the characteristics of the patients, conducted a follow-up survey, and examined noteworthy prognostic factors. All resected specimens were fixed in formalin and sliced at 10-mm intervals. Patients who had evidence of residual tumor at the surgical margin, malignant effusion, or N3 disease as verified by intraoperative findings or postoperative pathologic examination were considered to have undergone an incomplete surgery and were excluded from the current study.6 The following criteria, based on and modified from those of Martini and Melamed,7 were used for the designation of multiple primary lung tumors8: the tumors were anatomically separate (different lobes or lungs) and they were histologically different. If the tumors were histologically the same, there could be no systemic metastases or mediastinal spread. When all the lesions were determined to be squamous cell carcinoma, a classification was accepted if an origin separate from an area of carcinoma in situ could be identified for each lesion. In the case of metachronous tumors with the same histology, the tumor-free interval had to be at least 2 years. Tumors with minute satellite nodules that were found incidentally within the same lobe of the resected specimen were not excluded from the current study because we were not certain whether these lesions should be considered to be local tumor spread. Tumors with satellite lesions in another lobe were excluded. Patients whose tumors subsequently were classified as small cell carcinoma or a low-grade malignant tumor also were excluded. Surgical-pathologic staging was performed according to the New International Staging System for Lung Cancer.9
The medical record of each patient was reviewed for age, gender, histology, pathologic stage, maximum tumor dimension of the resected specimen, lymphatic vessel invasion, vascular vessel invasion, serum carcinoembryonic antigen (CEA) level, and surgical procedure performed. The specimens were systematically rereviewed for all these pathological factors. Lymphatic invasion and vascular vessel invasion indicated tumor cells identifiable in the lymphatic vessel lumen and vascular vessel lumen, respectively. The normal limit for the serum CEA level as assessed by immunoenzymometric assay was 5 ng/mL.
Survival was calculated by the Kaplan–Meier method, and differences in survival were determined using the log-rank analysis. A multivariate analysis of several independent prognostic factors was performed using the Cox proportional hazards regression model. Zero time was the date of pulmonary resection, and the terminal event was death attributable to cancer (i.e., all other deaths resulting from noncancer or unknown causes were treated as withdrawals in all survival analyses but one, in which overall survival curves were examined rather than disease-related survival curves). Differences were considered to be statistically significant when the P value was < 0.05.
The current study population was comprised of 163 men and 102 women with a mean age of 63 years (range, 38–83 years). Overall follow-up ranged from 9–181 months, with a median of 43 months. Histologically, adenocarcinoma was by far the most common occurrence (199 patients; 75%) (Table 1). Squamous cell carcinoma was diagnosed in 62 patients (23%), adenosquamous carcinoma was diagnosed in 2 patients (1%), and large cell carcinoma was diagnosed in 2 patients (1%). Lobectomy was performed in 132 patients (50%), including 11 combined sleeve resections of the bronchus (Table 2). One hundred six segmentectomies (40%), including 6 bronchoplasties, 3 pneumonectomies (1%), and 24 wedge resections (9%), were performed. Not surprisingly, on the basis of the New International Staging System for Lung Cancer,9 the majority of patients (208 patients; 78%) were found to have Stage IA disease (Table 3). Unexpectedly, 22 patients (8%) were determined to have Stage II disease, 22 patients (8%) had Stage IIIA disease, and 7 patients (3%) had Stage IIIB disease, indicating that even among patients with a tumor measuring ≤2 cm in greatest dimension, approximately 20% already had developed advanced disease.
|Cell type||No. (%)|
|Squamous cell carcinoma||62 (23)|
|Adenosquamous carcinoma||2 (1)|
|Large cell carcinoma||2 (1)|
|Lobe only||121 (46)|
|+ sleeve||11 (4)|
|Segment only||100 (38)|
|+ sleeve||6 (2)|
|Wedge resection||24 (9)|
|Stage||No. (%)||Cancer-related deaths (no.)|
|IA T1N0||208 (78)||7|
|IB T2N0 (p2)||6 (2)||0|
|IIA T1N1||20 (8)||6|
|IIB T3N0 (p3)||2 (1)||0|
|IIIA T1N2||19 (7)||5|
|T2N2 (p2)||1 (0.4)||1|
|T3N2 (p3)||2 (1)||1|
|IIIB T4N0 (pm+)||1 (0.4)||1|
|T4N1 (pm+)||1 (0.4)||1|
|T4N2 (pm+)||5 (2)||5|
The disease-specific 5-year and 10-year survival rates of the patients who underwent complete resection for NSCLCs measuring ≤ 2 cm were 86% and 83%, respectively (Fig. 1A). There were no significant differences in survival with regard to age (P = 0.4359), gender (P = 0.7365), or histologic type (P = 0.1110). The outcome of patients with pathologic Stage I disease was excellent (5-year survival rate of 96% and a 10-year survival rate of 94%), and was significantly better than that of patients with pathologic Stage II or Stage III disease (P < 0.0001) (Fig. 1B). The cancer-related survival of patients with a tumor measuring ≤ 15 mm in maximum dimension was significantly better that that of patients with tumors measuring 16–20 mm in greatest dimension (10-year survival rate of 90% vs. 77%; P = 0.0157) (Fig. 1C). The 10-year cancer-specific survival rate of patients without lymphatic vessel invasion or vascular vessel invasion was 86% and 91%, respectively (Figs. 1D, 1E). There were significant differences with regard to prognosis between those patients with and those without lymphatic vessel invasion (P = 0.0008), and between those with and those without vascular vessel invasion (P < 0.0001). The cancer-related prognosis of patients with low preoperative serum levels of CEA was 88% at 10 years and was significantly better than that of patients with a high level of CEA (P = 0.0261) (Fig. 1F). It is interesting to note that the outcome apparently was affected considerably by the extent of the surgical procedure. The 10-year cancer-related survival after lobectomy or pneumonectomy was 75% compared with 91% after lesser resection, such as segmentectomy or wedge resection (P = 0.0181) (Fig. 1G). This result could not be supported by a subsequent multivariate analysis (Table 4) and appeared most likely to be influenced by the selection bias that may have occurred when we determined the procedure according to surgical findings and rapid intraoperative pathologic examination. Next, we also analyzed our data in terms of overall survival rather than cancer-specific survival (Fig. 2) because the former is reliable as an endpoint, especially in a retrospective study. The overall 5-year and 10-year survival rates of those patients who underwent complete resection for NSCLCs measuring ≤ 2 cmr were 80% and 69%, respectively (Fig. 2A). As a result, there were significant differences in the overall survival with regard to pathologic stage (P < 0.0001), lymphatic vessel invasion (P = 0.0011), vascular vessel invasion (P = 0.0001), and preoperative serum level of CEA (P = 0.0100).
|Factors||Unfavorable||Favorable||Risk ratio||95% CI||P value|
|Pathologic stage||II + III||I||7.674||2.881–20.443||< 0.0001|
|Ly factor||ly (+)||ly (−)||1.220||0.461–3.226||0.6887|
|V factor||v (+)||v (−)||3.747||1.524–9.212||0.0040|
|CEA||CEA (+)||CEA (−)||2.356||1.091–5.086||0.0291|
|Extent of procedure||Pn + Lo||Se + We||2.022||0.572–7.141||0.2742|
Multivariate Cox analysis of patients who underwent complete resection for NSCLC measuring ≤ 2 cm in greatest dimension demonstrated that pathologic stage (P < 0.0001), vascular vessel invasion (P = 0.0040), and CEA level (P = 0.0291) were independent, significant prognostic determinants whereas maximum tumor dimension, lymphatic vessel invasion, and the extent of the procedure, which were judged as having prognostic significance by univariate analysis, were not found to be significant factors (P = 0.3351, P = 0.6887, and P = 0.2742, respectively) (Table 4). Patients with all these favorable factors who had pathologic Stage I tumors, no apparent vascular vessel invasion, and a low serum CEA level did not die of lung carcinoma after undergoing complete resection of NSCLCs measuring ≤ 2 cm.
During the course of follow-up, there were 27 cancer-related deaths among the 265 patients (10%) (Table 3). Of these, 9 patients (33%) died within the first 2 years after surgery. Although the incidence of death did not appear to diminish with time during the 5 years after the surgery, no patients died of their disease ≥ 6 years after the initial surgical treatment. The first sites of recurrence were mostly distant. The most common sites were the brain and lungs and no local recurrence was reported to occur at the bronchial stump. Unpredictably, seven patients who had pathologic Stage IA tumor died of their disease despite undergoing complete resection of the tumor (Table 3).
The univariate analysis of the current study data demonstrated that advanced pathologic stage, larger tumor size, lymphatic vessel invasion, vascular vessel invasion, high preoperative serum CEA level, and extent of surgical resection were unfavorable prognostic determinants for patients with NSCLC measuring ≤ 2 cm. However, it is of great interest to note that among these possible determinants, a multivariate analysis demonstrated that the preoperative serum level of CEA and vascular vessel invasion were independent prognostic factors that were as significant as the pathologic stage. Tumor size was not the only factor found to have an impact on survival after complete resection of a small-sized lung tumor. Macchiarini et al,10 also were unable to correlate tumor size with survival in a multivariate analysis of completely resected T1N0M0 tumors. Vascular or lymphatic vessel invasion as a predictor of survival has been controversial. The Veterans Administration Surgical Oncology Study Group, which reviewed 865 cases of resection for lung carcinoma, found lymphatic but not vascular vessel invasion to have a predictive value in patients with squamous cell carcinoma.11 Conversely, vascular vessel invasion was reported to be a main predictor of prognosis in patients with early NSCLC in a study by Macchiarini et al.,10 a finding that was supported by the larger amount of data available from the current study. Recently, Sawabata et al.12 reported that among patients with Stage I NSCLC, those with a high preoperative serum CEA level had a 5-year survival rate of 49% compared with 72% for those with a low serum CEA level (P < 0.0001), a finding that was compatible with the current study data.
A similar controversy concerning surgical intervention for NSCLCs measuring ≤ 2 cm exists between advocates of lesser (sublobar) resection and those who recommend a lobectomy as the procedure of choice. In general, lesser resection, which can provide many benefits to patients such as preservation of vital lung tissue and the chance for further resections if a second primary lung carcinoma develops,13, 14 has been accepted for small-sized lesions in compromised patients with an impaired cardiopulmonary reserve who are not candidates for lobectomy. However, a deep-rooted argument continues over the application of this procedure in patients who might otherwise tolerate a lobectomy. We have considerable doubts as to whether lobectomy, which has been commonly established as a standard to remove the local disease, is necessary to treat such small-sized lesions. The univariate analysis in the current series showed that patients who underwent a lesser resection had a better survival than those treated by lobectomy or pneumonectomy because the size of the tumor for which the lesser resection was performed generally was smaller (so-called selection bias), whereas the multivariate analyses demonstrated that lesser resection had no benefits with regard to survival. However, we should direct our attention to the finding that patients who underwent a lesser resection did not have a worse prognosis. Considering the fact that some patients still experienced recurrence (mainly distant metastases) despite complete resection of tumors measuring ≤ 2 cm, discrimination against those patients at a high risk of developing a systemic disease also appears to be an important issue for the appropriate selection of patients for lesser surgical intervention. Remembering the risk of local recurrence, some surgeons, including us, have advocated the usefulness of lesser surgical intervention with minimal lung parenchymal resection.1, 3, 4, 15, 16 Although the final surgical results should be assessed in the long term, in the proper selection of patients, segmentectomy with intraoperative exploration of lymph nodes, in which resection line is placed beyond the anatomic segmental border, has been reported to achieve comparable prognostic outcomes.
Although the results of the current study are limited by its retrospective nature, the large patient population and long temporal perspective of the data are essential for supporting evidence-based treatments and future directions. These data should help guide other clinicians and surgeons who treat patients with NSCLC measuring ≤ 2 cm. Ideally, randomized trials will need to be performed to identify the optimal management strategy for patients with small-sized NSCLC.