Ann Thorac Surg 2006;82:220-226
© 2006 The Society of Thoracic Surgeons
Original article: General thoracic
Lung Cancer Staging: A Case for a New T Definition
Major Charles R. Mulligan, MC, USA
a
,
*
,
Amir D. Meram, MD
d
,
c
,
Courtney D. Proctor
d
,
Hongyu Wu, MD
c
,
Kangmin Zhu, PhD
c
,
Aizen J. Marrogi, LTC, MC, USA
d
,
b
a Division of Cardiothoracic Surgery, Departments of Surgery and Pathology and Area Laboratory Services, Walter Reed Army Medical Center, Washington, DC
b Division of Anatomic Pathology, Departments of Surgery and Pathology and Area Laboratory Services, Walter Reed Army Medical Center, Washington, DC
c United States Military Cancer Institute, Walter Reed Army Medical Center, Washington, DC
d Laboratory of Biomarkers and Carcinogenesis, CBCP-IRC, Department of Surgery, Uniformed Services University for Health Sciences, Bethesda, Maryland
Accepted for publication February 13, 2006.
* Address correspondence to Major Mulligan, Cardiothoracic Surgery, Rm 4655, Bldg 2, Walter Reed Army Medical Center, 6900 Georgia Ave NW, Washington, DC 20307-5001 (Email: charles.mulligan{at}na.amedd.army.mil).
Presented at the Poster Session of the Fifty-second Annual Meeting of the Southern Thoracic Surgical Association, Orlando, FL, Nov 10–12, 2005.
 |
Abstract
|
|---|
BACKGROUND: The purpose of this study is to determine a more refined T definition for lung cancer staging on the basis of clinical outcomes.
METHODS: The Walter Reed Army Medical Center Tumor Registry and the Thoracic Surgery Tumor Clinic files were queried for lung cancers diagnosed from 1990 to 2000. Cox regression analysis and Kaplan–Meier survival curves for tumor size were used to analyze the impact of size on survival and relative risk, and then used to redefine T. Using the new T definition, the cohort was restaged, and the two staging system survivals were compared using Cox regression analysis.
RESULTS: Tumor size was documented in 439 males and 226 females. Forty-two tumors were 1.0 cm or less, 133 were between 1.01 and 2.0 cm, 133 were between 2.01 and 3.0 cm, 91 were between 3.01 and 4.0, 96 were between 4.01 and 5.0, and 166 were greater than 5.0 cm. A survival advantage was noted for smaller lesions, with 5-year survivals of 48.6%, 45.9%, 26.6%, 27.0%, 14.4%, and 11.6%, respectively. Cox regression analysis revealed increased risk at 2.0 cm (hazards ratio, 2.014; 95% confidence interval, 1.24 to 3.26), 4.0 cm (hazards ratio, 2.51; 95% confidence interval, 1.53 to 4.09), and 5.0 cm (hazards ratio, 3.14; 95% confidence interval, 1.96 to 5.02). After redefining T, the new staging system showed a better 5-year survival in each stage.
CONCLUSIONS: Lung cancer tumor size criteria should be changed to include T1 tumors 2.0 cm and less; T2 tumors between 2.0 and 4.0 cm or pleural invasion of T1 tumor; T3 tumors greater than 4.0 cm or pleural invasion of T2 tumors.
|
Introduction
|
|---|
Lung cancer is the leading cause of cancer death in both men and women in the United States. An excess of 170,000 cases of new lung cancer are expected to be diagnosed in 2005 with an estimated 160,000 deaths related to lung cancer [1]. The lung cancer cure rates continue to be an abysmal 15%. Overall poor outcomes in lung cancer are likely to be multifactorial and may be related to molecular biologic and physiologic factors. A more accurate staging system, reflective of clinical features, may allow better prediction of outcomes and lead to more appropriate treatment, resulting in better overall survival.
The staging system is the means by which we communicate, prognosticate, and treat malignancies. During the last 30 years, the lung cancer staging system has been revised multiple times, most recently in 1997. Tumor size is an important predictor of survival and part of all staging classifications. Tumor size of 3.0 cm as the cutoff between T1 and T2 tumors was first incorporated in the 1973 staging scheme, and has not been changed since. Advances in imaging, surgery, perioperative management, chemotherapy, and radiation therapy allow us to detect and treat smaller tumors with more active agents. With improving technology each aspect of the staging system must be critically reviewed on a periodic basis.
The Military Health Care System offers a unique opportunity to evaluate outcomes of patients with unlimited access to care, treatment, and follow-up. We have recently examined the role of unlimited access to the health-care system and correlated that with patient outcomes. Our data suggested such a model of health-care delivery offered an ideal mechanism for evaluating clinical outcomes in patients of different sexes and ethnic backgrounds [2]. Consequently, we undertook this study to analyze lung cancer patients' clinical outcomes and to determine the optimum tumor size criteria for lung cancer staging.
|
Material and Methods
|
|---|
After institutional review board approval, the Walter Reed Army Medical Center Tumor Registry was queried to find all lung cancers diagnosed from 1990 to 2000 (tumor site 340 to 349) and cross-referenced to the convenience files of the Thoracic Surgery Tumor Clinic. Of the 907 non–small cell lung cancers (NSCLC) identified, 665 patients had documented tumor size (pathologic or radiographic) and hence were available for analysis in this study. Mean follow-up was 2.5 years, with a range of 1 to 12 years.
To determine the optimum T definition, the original cohort tumor size was evaluated with both Kaplan–Meier survival curve and Cox regression analysis. The tumors were evaluated at 1.0-cm intervals beginning at 0 up to 5.0 cm and then greater than 5.0 cm. The hazards ratio and the 5-year Kaplan–Meier survival percentage for each tumor size was plotted linearly, and transition points of increasing risk and decreasing survival were identified. From these inflection points, a new T definition was defined.
Using this new T definition and the current stage groupings, the cohort of known tumor size was then restaged and reevaluated for survival using the Kaplan–Meier method. The new staging system survival was then compared with the current system survival using Cox regression analysis controlling for possible confounding variables of sex, race, completeness of resection, other treatments, and histology to determine the significance of survival differences between the current and new staging systems. To examine whether the differences in the two staging systems influenced estimation of survival, we created a variable indicating concordance between the systems and reanalyzed each stage to see the impact of the changes using Cox regression analysis on lung cancer survival.
|
Results
|
|---|
Defining Tumor Size
Six hundred sixty-five (73%) patients had documented tumor size, including 439 men and 226 women, a ratio of 2:1. Tumor size was evaluated at 1.0-cm increments from 0 to 5 cm and then for all tumors greater than 5.0 cm. There were 43 tumors of 1.0 cm or less, 136 between 1.0 and 2.0 cm, 133 between 2.0 and 3.0 cm, 91 between 3.0 and 4.0 cm, 96 between 4.0 and 5.0 cm, and 166 greater than 5 cm. A survival advantage regardless of stage was noted for smaller lesions, with 5-year survivals of 48.6%, 45.9%, 26.6%, 27.0%, 14.4%, and 11.6%, respectively (Table 1). On further analysis, a staging bias was discerned. The size of the tumor correlated proportionally with disease stage, with tumors 2.0 cm or less presenting with early stage disease (stage I and II) in 68% of the cases, compared with 49% and 27% for tumors between 2.0 and 4.0 cm and greater than 4.0 cm, respectively (Fig 1).
View larger version (62K):
[in this window]
[in a new window]
|
Fig 1. The graph depicts the dispersion of tumor size and pathologic stage. Note the increasing incidence of advanced stage cancer with increasing size.
|
|
Cox regression analysis demonstrated an incremental increase in hazards ratio from 1.07 (95% confidence interval [CI], 0.65 to 1.75) for 1.0- to 2.0-cm tumors to 3.14 (95% CI, 1.96 to 5.02) for tumors greater than 5.0 cm (Table 1). A closer look at the Cox regression analysis demonstrates a significant relative risk with tumors at 2.0, 4.0, and 5.0 cm. Analysis of 5-year survival on the basis of tumor size reveals a similar correlation. The transition points graphically for both variables occur at 2.0 cm and at 4.0 cm (Fig 2).
View larger version (74K):
[in this window]
[in a new window]
|
Fig 2. Tumor size and relative risk compared with overall cure rates for each size tumor. Note the transitions occurring at 2.0 cm and 4.0 cm for both curves.
|
|
Pleural invasion also was a marker for worse outcome, exhibiting behavior similar to those displayed by the next larger size tumor. The median survival for stage IA was 1,595 days compared with stage IB without pleural involvement of 1,161 day and 1,097 days with pleural involvement. Looking at the smaller subset of patients with pleural involvement, those with stage IB tumors less than 2.0 cm had median survival of 1,826 days, whereas those with stage IB greater than 2.0 cm was 1,328 days. Stage IIB and IIIA median survival of pleural-positive tumors was 1,089 and 794 days, respectively.
On the basis of these data, we defined T1lesion as a tumor 2.0 cm or less; a T2lesion would be between 2.0 and 4.0 cm, 2 cm or more distal to the carina, or a T1lesion with pleural invasion; a T3lesion would be greater than 4.0 cm or a T2lesion with pleural involvement or a tumor within 2.0 cm of the carina not involving the carina or invasion of a structure that could be surgically resected (chest wall, diaphragm, mediastinal pleura, parietal pericardium); and a T4lesion defined by invasion of the mediastinum, heart, great vessels, vertebral body, carina, or satellite nodules within the same lobe or a malignant pleural effusion.
New T Definition: Stage Survival and Analysis
Using the current stage groupings and our new definition for tumor size, our cohort was restaged and reanalyzed using Kaplan–Meier survival curves and Cox regression analysis (Fig 3). There were 93 (14.0%) stage IA, 91 (13.7%) stage IB, 7 (1.1%) stage IIA, 101 (15.2%) stage IIB, 76 (11.4%) stage IIIA, 116 (17.4%) stage IIIB, and 181 (27.2%) stage IV tumors. Table 2
shows the demographic breakdown for our study cohort and their treatments, and Table 3
shows the breakdown of surgical procedures. A complete resection was considered an anatomic resection with lymph node sampling or dissection. Kaplan–Meier 5-year survival for respective stages was 64.5%, 54.0%, 47.6%, 39.4%, 15.5%, 6.0%, and 1.6%. Compared with the survival curves of the original cohort the 5-year adjusted mortality is better in all stages (Fig 4). Using Cox regression analysis and controlling for confounding variables, the survivals in stage IA (p = 0.05), stage IB (p = 0.01), and stage IIB (p = 0.01) were statistically better than the current staging system. The multivariate analysis of both staging systems concurred, with advanced stage disease, lack of surgical intervention, older age, less-differentiated tumors, and need for radiation therapy conferring increased risk of poor outcome (Table 4). The only major difference noted was an increased hazards ratio for stage IIB disease in the new system.
View larger version (109K):
[in this window]
[in a new window]
|
Fig 3. Stage groupings and their changes based upon new T definition. The arrows reflect the stage migration of the new T definition and the number which changed stage.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
Fig 4. Kaplan–Meier survival curves comparing current and new staging systems, showing the difference in stage I cancers (A); stage II cancers (B); stage III cancers (C); and stage IV cancer (D).
|
|
Cox regression analysis comparing the concordance of the our new system versus the current staging system was then performed, adjusting for confounding variables to determine whether these changes better predicted survival. A hazards ratio greater than 1 reflected that the differences in T definition better predicted mortality, whereas a ratio less than 1 meant that the impact on survival was better and therefore the result of upstage migration, not a better-defined T. Table 5
shows that our new staging system better stratifies mortality risk in stage IA and IB, whereas the differences noted in stage IIB are more likely related to upstage migration. Stage IIA was too small to analyze; and stage IIIA and stage IIIB were equivalent in our system versus the current staging system.
|
Comment
|
|---|
The current study shows an increased relative risk of lung cancer mortality that correlates with tumor size. The Kaplan–Meier survival curves and the Cox regression analysis reflect similar conclusions (Fig 2). Our T definition is based on the transition points, which occurred at 2.0 and 4.0 cm. One could argue that the higher relative risk occurs above 5 cm (hazards ratio, 2.5 versus 3.1), making our definition too narrow. A narrow definition is also supported by the Cox regression analysis of congruence, which shows that the improved survival in our new stage IIB is more likely related to upstage bias; however, in the overall cohort evaluation the new definition did demonstrate an increased hazards ratio for stage IIB disease not seen in the current staging system, implying a better definition. Overall, the improved survival for each stage most likely reflects a better stage stratification and does demonstrate that there is an inflection point where tumor size does impart survival similar to a T3 lesion.
Mountain's article [3] on the staging revision in 1997 accepted the T definition of tumor size assigned in 1974 by the Task Force on Lung Cancer of American Joint Committee on Cancer. In his 2003 publication, Watanabe [4] discussed the history and controversy of the TNM staging of lung cancer and pointed to the existence of more than 50 articles in the English literature that challenged Mountain's system. Since this publication, numerous current reviews have appeared in the surgical literature questioning the validity of the current definition of tumor size (Table 6). Some focused on only early stage cancer and others on all stages, but most were surgical reviews, which did not address medically treated patients, unlike the current study.
As expected with increasing tumor size, a stage migration occurred. Larger tumors tend to present at a more advanced stage. In this current study, the percentage of early stage tumors decreased from 68% to 27% with increasing tumor size. Wisnivesky and associates [5] reported a similar trend during SEER database analysis for tumor size, with 58% early stage presentation for tumors less than 1.5 cm compared with 20% for tumors greater than 4.5 cm. The correlation of these variables may account for our multivariate analysis where stage III and IV diseases were significant predictors of mortality but actual size was not. Both Doddoli and colleagues [6] and Wu and coworkers [7] looked at stage I cancers using the current 3.0-cm cutoff and found that large tumors, both in univariate and multivariate analysis, had significant impact on lung cancer mortality. Wu and associates [7] reported a significantly improved 5-year survival of 59.7% versus 40.1%. On the other hand, Gajra and colleagues [8] looked at a further subdivided tumor size impact in stage I disease, choosing 1.5 cm as a cutoff on the basis of Cox regression analysis, with a hazards ratio of 1.96 (95% CI, 1.02 to 4.08; p = 0.04) and a 5-year survival of 81.5% versus 70.9%. Using the SEER database to evaluate a surgical cohort of stage I tumors with known tumor size, Wisnivesky and coworkers [9] demonstrated incremental survival at less than 1.5 cm (69%), 1.6 to 2.5 cm (63%). 2.6 to 4.5 cm (53–58%), and greater than 4.5 cm (43%). Further, Kondo and colleagues [10] examined a group of peripherally located 1.0-cm tumors and reported a survival advantage for tumors less than 1.0 cm of 97% compared with 75%; however, this data set was skewed by a larger number of bronchioalveolar tumors in the series.
Okada and associates [11] looked at tumors 2.0 cm in size. Their stage distribution was reported as 78% stage I, 8% stage II and IIIA, and 3% stage IIIB, which is similar to our data. In their univariate analysis, they found tumors less than 1.5 cm did better than larger tumors at 10 years (90% versus 77%), with overall cohort 5- and 10-year survival of 80% and 69%, respectively [11]. In 2005, this group reviewed a larger surgical cohort of 1,272 lung cancer patients, analyzing survival at 1.0-cm intervals up to 3.0 cm and then greater than 3.0 cm, and found significant improvement in 5-year survival of 87.3%, 75.5%, 68%, and 49%, respectively [12]. They did not define an optimal size criteria, however, as we have in the current study. Okada and colleagues [12] did report a similar stage migration, with 4% of stage III tumors being less than 1 cm compared with 31% for tumors greater than 3.0 cm.
The impact of larger tumors in surgical cohorts was investigated by Cangir and associates [13], who reported larger tumors had a worse 5-year survival (31% versus 56%) and a noticeable stage migration with increasing size. Their univariate Cox regression analysis demonstrated an increased risk beginning with 4.0-cm tumors (hazards ratio, 1.7; 95% CI, 1.0 to 2.8; p < 0.05) and increasing further at 5.0 cm (hazards ratio, 2.23; 95% CI, 1.5 to 3.3) similar to our data (hazards ratio, 2.5 and 3.1, respectively) [13]. Unlike our study, their multivariate analysis showed significantly increased risk for tumors greater than 5.0 cm (hazards ratio, 3.6; 95% CI, 1.6 to 7.9; p = 0.001), reflecting perhaps the surgical bias in their cohort compared with our combined medical and surgical group, and thereby bringing into question whether our T definition is too narrow.
Some might also question the definition of tumors greater than 4.0 cm or 5.0 cm as a T3 tumor. However, the survival noted by Cangir and coworkers [13] of 31% falls in the high end of quoted stage IIIA survival. Our study likewise noted a 5-year survival of 14.4% and 11.6% for tumors of greater than 4 cm and 5 cm, respectively, which correlates with the lower end of quoted survival in stage IIIA and the upper end of survival quoted in stage IIIB [14]. The 2.7:1 ratio of advanced stage tumors in the greater than 4.0-cm cohort might argue that this weights the results toward lower survival; however, when the median survival of all early stage tumors of 1,238 days is compared with 954 days for the larger tumors, which by the current system were classified as stage I or II tumors, there is 25% decrement in survival solely on the basis of tumor size.
In our study, pleural involvement demonstrated an increased risk similar to the next larger size tumor. Kang and associates [15] analyzed the outcome of T2 tumors and a 5-year survival of 63% without pleural invasion compared with 44% for tumors, which is similar to the 5-year survival for stage IIB disease. Osaki and colleagues [16] found a similar migration of survival with pleural invasion of 46% compared with 68%. These reports support the case for a stage migration with pleural involvement.
In conclusion, a better-refined T definition may allow better staging and treatment of lung cancer patients. Our data, obtained from a combined medical and surgical cohort, demonstrates a better survival based on a new T definition. Our recommended changes to the current system include T1 tumors 2.0 cm or less; T2 between 2.0 and 4.0 cm, or T1 tumors with pleural invasion; and T3tumors greater than 4.0 cm, or T2 tumors with pleural invasion. Further evaluation of other lung cancer cohorts looking more closely at tumor size is needed to confirm this observation and to assess whether there is a better upper limit of tumor size that can be defined.
|
Footnotes
|
|---|
The views expressed in this article are the views of the authors and should not be construed as official or as reflecting the policies of the Department of the Army or Department of Defense.
|
References
|
|---|
- American Lung Association, Epidemiology and Statistics Unit, Research and Program Services. Trends in lung cancer morbidity and mortality. May 2005. Available at: http://www.lungusa.org/site/pp.asp?c=dvLUK900E&b=33347..
- Mulligan CR, Meram DA, Proctor CD, Wu H, Zhu K, Marrogi AJ. Unlimited access to careeffect on racial disparity and prognostic factors in lung cancer. Cancer Epidemiol Biomark Prev 2006;15:25-31.[Abstract/Free Full Text]
- Mountain CF. Revisions in the International System for staging lung cancer Chest 1997;111:1710-1717.[Abstract/Free Full Text]
- Watanabe Y. TNM classification for lung cancer Ann Thorac Cardiovasc Surg 2003;9:342-350.
- Wisnivesky JP, Yankelevitz D, Hensche CI. Stage of lung cancer in relation to its size Chest 2005;127:1136-1139.[Abstract/Free Full Text]
- Doddoli C, Aragon A, Barlesi F, et al. Does the extent of lymph node dissection influence outcome in patients with stage I non–small cell lung cancer? Eur J Cardiothorac Surg 2005;27:680-685.[Abstract/Free Full Text]
- Wu Y, Lin CJL, Hsu W, Huang B, Huang M, Wang L. Long-term results of pathologic stage I non–small cell lung cancervalidation of using the number of totally removed lymph nodes as a staging control. Eur J Cardiothorac Surg 2003;24:994-1001.[Abstract/Free Full Text]
- Gajra A, Newman N, Gamble GP, Abraham NZ, Kohman LJ, Graziano SL. Impact of tumor size in stage IA non–small cell lung cancera case for subdividing stage IA disease. Lung Cancer 2003;42:51-57.[Medline]
- Wisnivesky JP, Yankelevitz D, Henschke CI. The effect of tumor size on curability of stage I non–small cell lung cancer Chest 2004;126:761-765.[Abstract/Free Full Text]
- Kondo D, Yamada K, Kitayama Y, Hoshi S. Peripheral lung adenocarcinomas10 mm or less in diameter. Ann Thorac Surg 2003;76:350-355.[Abstract/Free Full Text]
- Okada M, Sakamoto T, Nishio W, Unchino K, Tsubota N. Characteristics and prognosis of patients after resection of nonsmall cell lung carcinoma measuring 2 cm or less in greatest dimension Cancer 2003;98:535-541.[Medline]
- Okada M, Nishio W, Sakamoto T, et al. Effect of tumor size on prognosis in patients with non–small cell lung cancerthe role of segmentectomy as a type of lesser resection. J Thorac Cardiovasc Surg 2005;129:87-93.[Abstract/Free Full Text]
- Cangir AK, Kutlay H, Akai M, Gungor A, Ozdemir N, Akay H. Prognostic value of tumor size in non–small cell lung cancer larger than five centimeters in diameter Lung Cancer 2004;46:325-331.[Medline]
- Ponn RB, LoCiecero J, Daly BD. Surgical treatment of non–small cell lung cancer. In: Shields TW, LoCiecero J, Ponn RB, Rusch VW (eds). General Thoracic Surgery, 6th ed. Philadelphia: Lippincott Williams & Wilkins; 1548–87..
- Kang J, Kim KD, Chung KY. Prognostic value of visceral pleural invasion in non–small cell lung cancer Eur J Cardiothorac Surg 2003;23:865-869.[Abstract/Free Full Text]
- Osaki T, Nagashima A, Yoshimatsu T, Yamada S, Yasumoto K. Visceral pleural involvement in non–small cell lung cancerprognostic significance. Ann Thorac Surg 2004;77:1769-1773.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
S.-H. I. Ou, J. A. Zell, A. Ziogas, and H. Anton-Culver
Prognostic Significance of the Non-Size-Based AJCC T2 Descriptors: Visceral Pleura Invasion, Hilar Atelectasis, or Obstructive Pneumonitis in Stage IB Non-small Cell Lung Cancer Is Dependent on Tumor Size
Chest,
March 1, 2008;
133(3):
662 - 669.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
