- Email: [email protected]
Prognosis of patients with resected non-small cell lung cancer: Impact of clinical and pathologic variables
Lung Cancer (2006) 52, 207—212
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/lungcan
Prognosis of patients with resected non-small cell lung cancer: Impact of clinical and pathologic variables Peter M. Ravdin a,∗, Gregory Davis b a b
Division of Medical Oncology, University of Texas Health Sciences Center at San Antonio, United States Center for Epidemiology and Biostatistics, University of Texas Health Science Center at San Antonio, United States
Received 26 October 2005 ; received in revised form 18 January 2006; accepted 23 January 2006
KEYWORDS Lung cancer; Non-small cell lung cancer; Staging; Prognosis; Bronchoalveolar cell; Adjuvant therapy; SEER
Summary The objective of this study was to determine whether AJCC staging, supplemented by additional clinical and pathologic variables could identify a subpopulation of pathologically staged NSCLC patients who had been resected for cure with a low risk (<20%) of 5-year disease specific mortality. The Surveillance, Epidemiology, and End Results (SEER) registry was analyzed to identify and evaluate the disease specific 5-year mortality of 17,130 patients with resected and pathologically defined Stage I—IIIA disease initially diagnosed from 1988 to 1997. The analyses showed that American Joint Commission on Cancer (AJCC) T and N defined stage Stage I—IIIA subgroups had approximately a 30, 60, and 75% 5-year disease specific mortality, respectively. The 5366 Stage IA patients were identified as having a 5-year disease specific mortality of 25%. Further multivariate analyses of Stage IA cases showed predictors of favorable outcome to be tumor size (<10 mm), histologic grade (well differentiated), and histologic subtype (bronchoalveolar cell). Subgroups identified with two or all three of these additional features had a ∼10% 5-year lung cancer specific mortality. Although nearly all patient subgroups with Stages I—IIIA resected and pathologically staged NSCLC have substantial (>20%) risk of death from their cancer within 5 years of diagnosis, for Stage IA patients additional information (tumor size, histologic grade, and histologic subtype) allows additional refinement in prognostic estimates and identification of some low risk subgroups. © 2006 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Abbreviations: SEER, Surveillance, Epidemiology, and End Results; AJCC, American Joint Commission on Cancer; NSCLC, nonsmall cell lung cancer; S.E., standard error ∗ Corresponding author at: 19931 Encino Royale, San Antonio, TX 78259, United States. Tel.: +1 210 497 7965. E-mail address: [email protected] (P.M. Ravdin).
Lung cancer is the leading cause of cancer death in the United States [1]. Most patients with this malignancy are diagnosed at a point when the disease is unresectable, although a substantial number of patients with non-small cell lung cancer have their disease resected and become candidates for possible systemic adjuvant therapy. Over the
0169-5002/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2006.01.016
208 last decade a series of meta-analyses [2—6] and individual trials [7—10] have made it clear that systemic adjuvant therapy can improve the outcome of patients with resected NSCLC. The decision about whether to receive systemic adjuvant therapy always involves a balancing of risk and benefit. Current adjuvant chemotherapy regimens with platinum compounds have considerable toxicity with many patients being unable to finish their adjuvant regimens because of toxicity, and a 1—2% treatment related mortality being reported for some regimens [8—10]. The benefit is also far from ideal with most meta-analyses suggesting that cis-platinum based regimens reduce the hazard rate of relapse and death by only about 20%. Thus adjuvant therapy decision making for NSCLC, like that for breast and colon cancer involves a careful weighing of risk and benefit. For patients at low risk of negative outcome the decision of whether to use adjuvant chemotherapy is always more difficult because these patients will experience the same toxicity but gain less benefit than high-risk patients. For patients at very low risk the risk of toxicity may outweigh the benefit. Thus, it is important to be able to identify low risk patients. Classical pathologic staging such as the AJCC staging can provide some guidance [11], but for NSCLC such staging is not ideal in that even the most favorable group (Stage 1) has a 30% risk of disease specific death within 5 years [12]. Stage 1 NSCLC patients have a heterogeneous set of prognoses. To be able to identify Stage I patients with a substantially lower risk might help distinguish a subset of NSCLC patients who might avoid adjuvant chemotherapy from those patients for whom it might be considered. This paper explores whether additional information in the SEER databases (tumor size, histologic grade, and histologic subtype) can further refine prognostic estimates.
2. Materials and methods The Surveillance, Epidemiology, and End Results (SEER) program Public Use Database (SEER 13 registry 2004) was used for these analyses [13]. The database included data from patients from 1973 to 2002. The analyses were done on patients diagnosed starting in 1988 because detailed extent of disease information was available starting in that year, and with diagnoses until 1997, allowing 5 years of follow-up for all patients. Analyses were restricted to invasive non-small cell lung cancer, and excluded small cell lung cancer cases, carcinoids, and cancers with ambiguous or unrecorded histologic subtypes. Morphology codes (from International Classification of Diseases for Oncology-2 morphology codes) included 8010—8040, 8050—8076, 8140, 8143, 8250—8260, 8310, 8320, 8323, 8470—8490, and 8550—8573 that were diagnosed prior to autopsy. SEER extent of disease codes and staging manuals were used to identify patients in American Joint Committee on Cancer T and N subsets. T1 corresponded to a SEER extent of disease code of 10 with the additional criteria that the tumors must have been ≤ 30 mm in size. T2 corresponded to an extent of disease code of 10 if the tumor was >30 mm in size, or codes 20 or 40. T3 corresponded to tumors with extent of disease codes of 50, 60, or 73.
P.M. Ravdin, G. Davis Table 1 5-Year disease specific survival (in %) as defined by AJCC T and N stage [9] for Stages I—IIIA NSCLC
Number in parentheses is the standard error. Categories surrounded by bold line are Stage I, dashed line Stage II, and thin line Stage III.
Analyses were restricted to patients who had not had preoperative radiation therapy. Patients must have undergone surgery with curative intent (surgery codes of 30—70; at least a lobectomy or pneumonectomy). Survival analyses were done using SEER*Stat (http:// www.seer.cancer.gov/seerstat) statistical software version 6.1.4. This software used an actuarial method of estimate of cause specific survival based on the cause of death provided in the SEER registry. Cox regression analyses were done with NCSS statistical software of Kayden, Utah.
3. Results Table 1 shows that disease specific survival stratified by AJCC T and N stage at 5 years of follow-up for the 17,310 patients who underwent surgery with curative intent, had pathologically staged Stage I—IIIA disease, and were thus included in this analysis. Stage I patients (T1N0 and T2N0) have the best survival with an average of a 71% 5-year disease specific survival. The majority (on average 60%) of Stage II patients (T1N1, T2N1, and T3N0) will die during this period. Stage IIIA patients (T1N2, T2N2, T3N2, and T3N1) patients had, on average, a still worse prognosis with an average 5-year specific survival of only 24%. Further analyses focused on the 5366 patients with Stage IA disease (T1N0). The first of these analyses (see Table 2) examined the impact of number of nodes reported as pathologically examined. Patients with 0 nodes pathologically examined had a poorer outcome than patients with 1—5, 6—10, >10, or an unknown number of nodes examined. Because the completeness of staging of these patients might be questioned, these patients were excluded from subsequent analyses. Table 3 shows the outcome of the 4866 patients with a known number of nodes examined with T1NO disease stratified by sex, age, tumor size, histologic grade, and histologic
Table 2 Effect of number of nodes examined on 5-year disease specific survival of T1N0 cases All cases 0 1—5 6—10 >10 ?
76.6 67.4 77.3 77.5 78.9 76.5
(0.6), (2.2), (1.0), (1.3), (1.5), (1.2),
n = 5366 n = 500 n = 1881 n = 1074 n = 793 n = 1262
Prognosis of patients with resected non-small cell lung cancer Table 3
209
Outcome in subgroups n
5-Year lung cancer specific survival % (S.E.)
Overall
4866
77.5 (0.6)
Sex Male Female
2533 2333
76.0 (0.9) 79.1 (0.9)
1.15 (0.02) —
Age 21—50 50—59 60—69 70—79
293 964 1933 1634
77.7 81.9 77.9 74.4
0.80 (0.11) 0.64 (< 0.001) 0.82 (0.01) —
Tumor size <10 mm 11—20 21—30
369 2268 2229
85.3 (1.9) 80.0 (0.9) 73.6 (1.0)
0.53 (<0.001) 0.73 (<0.001) —
Histologic grade Well diff. Moderately diff. Poorly diff. Undiff./anaplastic Unknown
640 1769 1567 223 667
84.9 79.0 71.8 71.4 81.8
(1.5) (1.0) (1.2) (3.2) (1.6)
0.51 (<0.001) 0.72 (0.02) 1.01 (ns) —
Histologic type Adenocarcinoma Squamous Large cell Bronchoalveolar Other
2151 1164 268 694 589
76.6 76.2 76.0 84.8 75.4
(0.9) (1.3) (2.7) (1.4) (1.0)
0.95 0.98 0.90 0.63 —
(2.5) (1.3) (1.0) (1.1)
subtype. All the parameters in univariate analysis were predictive of prognosis with the relative risk (RR) conferred being generally modest (between 0.51 and 1.15). Univariate analyses (see Table 3) showed all the tumor and patient related variables tested to be related to prognosis. Tumor size and histologic grade were statistically significant predictors on both univariate and multivariate analyses, with the most favorable of their categories (<10 mm, and well differentiated) having half the risk of their least favorable categories. Sex was weakly predictive in univariate analysis. Patients 70—79 years of age did more poorly than the younger subsets of patients. Patients with bronchoalveolar subtype (BAC) tumors had a more favorable prognosis. Multivariate analysis (see Table 4) identified tumor size, histologic grade, patient age, and histologic subtype as independent predictors of outcome. Patients with combinations of favorable features had excellent outcomes. As can be seen in Table 3 patients with favorable tumor size, histology grade, or bronchoalveolar
Table 4
Results of multivariate analysis
Overall
RR
p
Age (70—79 vs. <70) Grade (>1 vs. 1) Size (11—30 mm vs. 1—10 mm) BAC vs. other types
1.31 1.56 1.52 0.74
<0.001 <0.001 0.003 0.02
Univariate RR (p value)
(ns) (ns) (ns) (0.002)
subtype had a 5-year disease specific survival of 85.3% (S.E. 1.9), 84.9% (S.E. 1.5), and 84.8% (S.E. 1.4), respectively. For patients with combinations of two such favorable features, size/grade, size/BAC, and grade/BAC, the 5-year disease specific survivals were 91.6% (S.E. 3.3), 93.9% (S.E. 4.2), and 87.3% (S.E. 2.5), respectively. For the small subset of these patients (n = 23) with small well-differentiated bronchoalveolar cancers, the 5-year lung cancer specific survival was 95.5% (S.E. 4.4%).
4. Discussion This analysis was done to better define the prognosis of patients with Stage IA NSCLC. Whether these patients should receive adjuvant chemotherapy hinges in large part on the answers to two questions: (1) is their prognosis poor enough to justify the risks and costs of such additional therapy and (2) is adjuvant chemotherapy effective in this patient population? At the present time different national guidelines list adjuvant therapy for Stage I disease as recommended for some patients, or as optional, or not recommended [14—16]. None of these guidelines discusses in quantitative terms the risks faced by the patients or the relative benefit that might be expected. The rationale for not recommending therapy is: (1) that compared to other NSCLC patients, Stage IA patients have a relatively favorable prognosis and (2) that the meta-analyses of clinical trials of adjuvant chemotherapy tends to show only modest efficacy [2—6] and that for
210 the subset of patients with Stage IA disease whether adjuvant therapy reduces the hazard for relapse and death is uncertain. Is the prognosis of Stage IA disease well defined and is it favorable? Published estimates for the 5-year OS of Stage IA patients vary widely, from as high as 90% [17] to as low as 63% [18]. A widely quoted North American analysis of MD Anderson Center experience with 511 Stage IA pathologically staged and resected NSCLC patients found them to have 5year disease specific survival of only 67% [12]. This estimate is slightly more pessimistic than the 76% estimate based on SEER, but seems roughly comparable. The MD Anderson Center estimate for outcome at 5 years of pathologically staged Stage IB patients of 57% is comparable, but again slightly more pessimistic than the SEER based estimate of 62% presented in this paper. Given a number of differences between the two databases some differences in the estimates might be expected. The MD Anderson Center patients were from an earlier era from 1973 to 1988 (in SEER the prognosis of Stage IA patients was no different in years 1988—1992 and 1993—1997; data not shown) and were referred to a tertiary cancer center. In addition, none of the MD Anderson Center patients received adjuvant therapy whereas some of the patients in the SEER database had received adjuvant therapy, even though the impact of this on prognostic estimates derived using the SEER registry would have been expected to be minor (see below). The prognostic estimates for Stage I NSCLC patients based on the National Cancer Data Base (NCDB) are much more pessimistic, and although quoted in the AJCC staging manuals, these estimates are methodologically flawed [19]. The analysis includes both pathologically and clinically staged patients. Its estimate of 45% for the 5-year disease specific survival of Stage I patients is much more pessimistic than other published estimates. One especially valuable piece of data from the NCDB is information about how commonly adjuvant chemotherapy was being given to patients in the United States. This database shows that <1% of patients with surgically resected Stage I NSCLC in 1990 received any adjuvant chemotherapy, and that in 1995 the number was still low, with only about 5% of Stage I patients receiving such therapy. Given that the adjuvant chemotherapy of this era had only a very modest effect (<20% proportional risk reduction of mortality), the impact of treating so few patients with therapy of such modest effectiveness would have no detectable effect on the SEER patients’ outcome presented in this analysis. This is an important point because the SEER registry does not include information on whether a patient received system adjuvant therapy. Because this analysis was done on a large set of 4866 Stage IA patients, the results of this study have more statistical power than the results of many of the studies that analyzed the data of only a at most few hundred cases. The analyses show that in this low risk subgroup, that gender has no independent effect on outcome, but that age, tumor size, histologic grade, and histologic subtype have prognostic significance. The effects of tumor size in Stage I patients have been previously reported by Wisnivesky et al. using data from SEER, but this paper analyzed data from all Stage I patients and did not examine the impact of variables other than size [20].
P.M. Ravdin, G. Davis Other published analyses of tumor size in Stage I patients have also reported it to be an independent predictor of outcome [21—23]. It has been suggested than bronchoalveolar tumors have a better prognosis that the other main NSCLC histologic subtypes. The literature on this topic for Stage I NSCLC has been inconclusive, with the weakness of being based on small data sets in which bronchoalveolar cases represent only about 10—20% of the cases [22,23]. In the analysis presented here based on 694 bronchoalveolar cases, it is clear that this is a favorable histologic subtype (HR of 0.74). Although this is statistically significant, the impact is small in magnitude, and identifies a relatively small cohort of patients. Finding the histologic grade is a predictor of outcome has been reported in the past [21]. An additional analysis was done to assess the impact of histologic grade with the bronchoalveolar cases omitted. The 5-year disease specific survivals were 70.1% (2.5), 64.4% (1.2), 59.6% (1.0), and 56.5% (2.3) for well, moderately, poorly, and undifferentiated cases, respectively, which shows that the effect of grade is not driven by the inclusion of bronchoalveolar cases. The literature about the histologic grading of NSCLC is sparse and questions such as the reproducibility of the interpretation histologic grade apparently have not been addressed. Although much of this analysis has focused on identification of low risk Stage IA patients, another way of looking at the results is that a 22% 5-year disease specific mortality can only be interpreted as a favorable prognosis relative to the prognosis of other patients with lung cancer. This level of risk would be viewed as a reason for giving adjuvant chemotherapy for breast cancer. The inclusion of Stage IA patients in clinical trials would seem to be justified because the majority of these patients are not really at low risk for lung cancer related mortality. The current literature about the efficacy of adjuvant chemotherapy for Stage I (and Stage IA) disease is weak, and to some degree there is uncertainty about the effectiveness of adjuvant chemotherapy in Stage I patients. This uncertainty is reflected in recent trial results. In the IALT trial which showed an overall favorable mortality hazard ratio reduction for adjuvant chemotherapy (HR = 0.86, p = 0.03) there was a trend for Stage I—III subsets of patients all to benefit from adjuvant therapy [10]. In the IALT trial the greatest impact on hazard rate was in patients with Stage III disease, and only in this patient subset did the impact on hazard rate reach statistical significance. Although there was no statistical heterogeneity by stage in benefit, it might still be argued that it had not been proven that Stage I patients benefited. In the JBR-10 trial, which allowed Stage IB and Stage II patients, only the Stage II patients had a survival advantage, and no benefit was seen for the Stage IB patients [8]. In the report of the JBR-10 trial an underpowered statistical test for interaction between treatment and stage failed to show an interaction, so again, although Stage I patients might be presumed to be benefiting this was not rigorously demonstrated. In the ANITA trial Stage I and Stage IIIA patients clearly benefited from therapy, but no benefit was seen in Stage IB patients. In the ANITA trial no formal statistical test was done to assess the possible interaction between treatment effect and stage of disease [9]. Thus the evidence that adjuvant chemotherapy is effective in Stage I patients is not as complete as would be ideal.
Prognosis of patients with resected non-small cell lung cancer The two meta-analyses that have addressed whether there was an interaction between stage and treatment effectiveness did not find such an interaction [5,6], so the lack of statistically significant benefit in the Stage I patients from individual trials is presumably due to a lack of statistical power to demonstrate those advantages. There are some limitations to the analyses presented here because they are based on information from a large registry based on a patient population for which there were no absolute guidelines directing patient evaluation and treatment. Thus there is the possibility of patient selection as to which patients were resected with curative intent. Although this is a particular issue for Stage IIIA patients, for whom there is substantial controversy as to whether an initial surgical approach is the best approach [24,25], this is less likely to be a major issue with Stage I disease. Patients older that 79 were excluded from this analysis because of the concern that the high rates of comorbidity in these older patients would be more likely to cause patient selection, and would also be associated with higher levels of uncertainty about the actual cause of death. Another issue is that there is no verification that histologic grading and histologic subtyping was done uniformly across centers. It should be noted that the major classification system used in the United States is that of the World Health Organization [26], and that there were revisions to this classification system in 1999. Although these are weaknesses, these drawbacks also apply to much of the information derived from clinical trials, as well as patient series at single institutions. The results of these analyses are of potential utility because they show that patients with an approximate 15% risk of 5-year disease specific mortality can be identified in the Stage IA population. These are patients with tumors that are <10 mm in size, are well differentiated, or are of the bronchoalveolar histology. The results are consistent with the interpretation that patients with two favorable features have an approximate 10% 5-year lung cancer specific mortality. How much benefit these low risk subsets of patients might receive from adjuvant chemotherapy depends on estimates of the proportional risk reduction afforded by adjuvant chemotherapy, and assumptions about whether these estimates of efficacy can be applied to these populations of patients. If one uses a meta-analysis derived estimate of a 20% proportional risk reduction, the projected benefit from adjuvant chemotherapy for some low risk patients would be small and approximately equal to treatment related mortality for low risk Stage IA patients. Conversely, on average, Stage IA patients have a risk of NSCLC related death at 5 years of 25% and a possible absolute survival benefit of about 5%. Guidelines suggesting that such patients might consider adjuvant therapy seem justified [14], but further clinical trials in this patient population would be of value.
References [1] Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55(1):10—30. [2] Non-small Cell Lung Cancer Collaborative Group. Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomised clinical trials. BMJ 1995;311:899—909.
211 [3] Hotta K, Matsuo K, Ueoka H, Kiura K, Tabata M, Tanimoto M. Role of adjuvant chemotherapy in patients with resected non-small-cell lung cancer: reappraisal with a meta-analysis of randomized controlled trials. J Clin Oncol 2004;22:3860—7. [4] Sedrakyan A, van der Meulen J, O’Byrne K, Prendiville J, Hill J, Treasure T. Postoperative chemotherapy for non-small cell lung cancer: a systematic review and meta-analysis. J Thorac Cardiovasc Surg 2004;128:414—9. [5] Hamada C, Tanaka F, Ohta M, Fujimura S, Kodama K, Imaizumi M, et al. Meta-analysis of postoperative adjuvant chemotherapy with Tegafur—Uracil in non-small-cell lung cancer. J Clin Oncol 2005;23:4999—5006. [6] Berghmans T, Paesmans M, Meerta AP, Branle F, Lafitte JJ, Lemaitre F, et al. Survival improvement in resectable non-small cell lung cancer with (neo)adjuvant chemotherapy: results of a meta-analysis of the literature. Lung Cancer 2005;49:13—23. [7] Strauss GM, Herndon J, Maddaus MA, Johnstone DW, Johnson EA, Watson DM, et al. Randomized clinical trial of adjuvant chemotherapy with paclitaxel and carboplatin following resection in Stage IB non-small cell lung cancer (NSCLC): Report of Cancer and Leukemia Group B (CALGB) Protocol 9633. Proceedings of the 2004 ASCO Annual Meeting. J Clin Oncol 2004;22(14S) [Abstract 7019]. [8] Winton TL, Livingston R, Johnson D, Rigas J, Johnston M, Butts C, et al. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005;352: 2589—97. [9] Douillard J, Rosell R, Delena M, Legroumellec A, Torres A, Carpagnano F. ANITA: phase III adjuvant vinorelbine (N) and cisplatin (P) versus observation (OBS) in completely resected (stage I—III) non-small-cell lung cancer (NSCLC) patients (pts): final results after 70-month median follow-up. Proceedings of the 2005 ASCO Annual Meeting. J Clin Oncol 2004;23 [Abstract 7013]. [10] Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J. International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N EngI J Med 2004;350(4):351—60. [11] Greene FL, Page DL, Fleming ID, et al., editors. AJCC cancer staging manual. New York: Springer-Verlag; 2002. [12] Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710—7. [13] Surveillance, Epidemiology, and End Results (SEER) Program (http://www.seer.cancer.gov) SEER*Stat Database Incidence— –SEER 13 Regs Public-Use, November 2004. National Cancer Institute, DCCPS, Surveillance Research Program. [14] National Comprehensive Cancer Network. Clinical practice guidelines in oncology, v. 2; 2005 (http://www.nccn.org). [15] http://cancer.gov/cancerinfo/pdg/cancerdatabase. [16] National Institute for Clinical Excellence. Lung cancer: the diagnosis and treatment of lung cancer. http://www.nice. org.uk/CG024NICEguideline. [17] Kato H, Ichinose Y, Ohta M, Hata E, Tsubota N, Tada H, et al. Japan Lung Cancer Research Group on Postsurgical Adjuvant Chemotherapy. A randomized trial of adjuvant chemotherapy with uracil—tegafur for adenocarcinoma of the lung. N EngI J Med 2004;350(17):1713—21. [18] van Rens MTM, de la Riviere, Elbers HR, van Den Bosch JM. Prognostic assessment of 2,361 patients who underwent pulmonary resection for non-small cell lung cancer, stage I, II, and IIIA. Chest 2000;117:374—9. [19] Fry WA, Phillips JL, Menck HR. Ten-year survey of lung cancer treatment and survival in hospitals in the United States. A National Cancer Data Base Report. Cancer 1999;86:1867—76. [20] Wisnivesky JP, Yankelevitz D, Henschke CI. The effect of tumor size on curability of stage I non-small cell lung cancers. Chest 2004;126:761—5.
212 [21] Gajra A, Newman N, Gamble GP, Abraham NZ, Kohman LJ, Graziano SL. Impact of tumor size on survival in stage IA nonsmall cell lung cancer: a case for subdividing stage IA disease. Lung Cancer 2003;42:51—7. [22] Birim O, Kappetein AP, Takkenberg JJM, van Klaveren RJ, Bogers AJ. Survival after pathological stage IA nonsmall cell lung cancer: tumor size matters. Ann Thorac Surg 2005;79:1137—41. [23] Port JL, Kent MS, Korst RJ. Tumor size predicts survival within stage IA non-small cell lung cancer. Chest 2003;124: 1828—33.
P.M. Ravdin, G. Davis [24] van Meerbeeck JP, Kramer G, Van Schil PE, Legrand C, Smit EF, Schramel FM, et al. A randomized trial of radical surgery (S) versus thoracic radiotherapy (TRT) in patients (pts) with stage IIIA-N2 non-small cell lung cancer (NSCLC) after response to induction chemotherapy (ICT) (EORTC 08941). Proceedings of the 2005 ASCO Annual Meeting. J Clin Oncol 2005;23 [Abstract 7015]. [25] Turrisi AT. A case against surgery for most IIIa non-small cell lung cancer. Semin Oncol 2005;32(Suppl 3):S6—8. [26] Travis WD, Sobin HL. Histological typing of lung and pleural tumours. 3rd ed. Berlin: Springer-Verlag; 1999.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/lungcan
Prognosis of patients with resected non-small cell lung cancer: Impact of clinical and pathologic variables Peter M. Ravdin a,∗, Gregory Davis b a b
Division of Medical Oncology, University of Texas Health Sciences Center at San Antonio, United States Center for Epidemiology and Biostatistics, University of Texas Health Science Center at San Antonio, United States
Received 26 October 2005 ; received in revised form 18 January 2006; accepted 23 January 2006
KEYWORDS Lung cancer; Non-small cell lung cancer; Staging; Prognosis; Bronchoalveolar cell; Adjuvant therapy; SEER
Summary The objective of this study was to determine whether AJCC staging, supplemented by additional clinical and pathologic variables could identify a subpopulation of pathologically staged NSCLC patients who had been resected for cure with a low risk (<20%) of 5-year disease specific mortality. The Surveillance, Epidemiology, and End Results (SEER) registry was analyzed to identify and evaluate the disease specific 5-year mortality of 17,130 patients with resected and pathologically defined Stage I—IIIA disease initially diagnosed from 1988 to 1997. The analyses showed that American Joint Commission on Cancer (AJCC) T and N defined stage Stage I—IIIA subgroups had approximately a 30, 60, and 75% 5-year disease specific mortality, respectively. The 5366 Stage IA patients were identified as having a 5-year disease specific mortality of 25%. Further multivariate analyses of Stage IA cases showed predictors of favorable outcome to be tumor size (<10 mm), histologic grade (well differentiated), and histologic subtype (bronchoalveolar cell). Subgroups identified with two or all three of these additional features had a ∼10% 5-year lung cancer specific mortality. Although nearly all patient subgroups with Stages I—IIIA resected and pathologically staged NSCLC have substantial (>20%) risk of death from their cancer within 5 years of diagnosis, for Stage IA patients additional information (tumor size, histologic grade, and histologic subtype) allows additional refinement in prognostic estimates and identification of some low risk subgroups. © 2006 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Abbreviations: SEER, Surveillance, Epidemiology, and End Results; AJCC, American Joint Commission on Cancer; NSCLC, nonsmall cell lung cancer; S.E., standard error ∗ Corresponding author at: 19931 Encino Royale, San Antonio, TX 78259, United States. Tel.: +1 210 497 7965. E-mail address: [email protected] (P.M. Ravdin).
Lung cancer is the leading cause of cancer death in the United States [1]. Most patients with this malignancy are diagnosed at a point when the disease is unresectable, although a substantial number of patients with non-small cell lung cancer have their disease resected and become candidates for possible systemic adjuvant therapy. Over the
0169-5002/$ — see front matter © 2006 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.lungcan.2006.01.016
208 last decade a series of meta-analyses [2—6] and individual trials [7—10] have made it clear that systemic adjuvant therapy can improve the outcome of patients with resected NSCLC. The decision about whether to receive systemic adjuvant therapy always involves a balancing of risk and benefit. Current adjuvant chemotherapy regimens with platinum compounds have considerable toxicity with many patients being unable to finish their adjuvant regimens because of toxicity, and a 1—2% treatment related mortality being reported for some regimens [8—10]. The benefit is also far from ideal with most meta-analyses suggesting that cis-platinum based regimens reduce the hazard rate of relapse and death by only about 20%. Thus adjuvant therapy decision making for NSCLC, like that for breast and colon cancer involves a careful weighing of risk and benefit. For patients at low risk of negative outcome the decision of whether to use adjuvant chemotherapy is always more difficult because these patients will experience the same toxicity but gain less benefit than high-risk patients. For patients at very low risk the risk of toxicity may outweigh the benefit. Thus, it is important to be able to identify low risk patients. Classical pathologic staging such as the AJCC staging can provide some guidance [11], but for NSCLC such staging is not ideal in that even the most favorable group (Stage 1) has a 30% risk of disease specific death within 5 years [12]. Stage 1 NSCLC patients have a heterogeneous set of prognoses. To be able to identify Stage I patients with a substantially lower risk might help distinguish a subset of NSCLC patients who might avoid adjuvant chemotherapy from those patients for whom it might be considered. This paper explores whether additional information in the SEER databases (tumor size, histologic grade, and histologic subtype) can further refine prognostic estimates.
2. Materials and methods The Surveillance, Epidemiology, and End Results (SEER) program Public Use Database (SEER 13 registry 2004) was used for these analyses [13]. The database included data from patients from 1973 to 2002. The analyses were done on patients diagnosed starting in 1988 because detailed extent of disease information was available starting in that year, and with diagnoses until 1997, allowing 5 years of follow-up for all patients. Analyses were restricted to invasive non-small cell lung cancer, and excluded small cell lung cancer cases, carcinoids, and cancers with ambiguous or unrecorded histologic subtypes. Morphology codes (from International Classification of Diseases for Oncology-2 morphology codes) included 8010—8040, 8050—8076, 8140, 8143, 8250—8260, 8310, 8320, 8323, 8470—8490, and 8550—8573 that were diagnosed prior to autopsy. SEER extent of disease codes and staging manuals were used to identify patients in American Joint Committee on Cancer T and N subsets. T1 corresponded to a SEER extent of disease code of 10 with the additional criteria that the tumors must have been ≤ 30 mm in size. T2 corresponded to an extent of disease code of 10 if the tumor was >30 mm in size, or codes 20 or 40. T3 corresponded to tumors with extent of disease codes of 50, 60, or 73.
P.M. Ravdin, G. Davis Table 1 5-Year disease specific survival (in %) as defined by AJCC T and N stage [9] for Stages I—IIIA NSCLC
Number in parentheses is the standard error. Categories surrounded by bold line are Stage I, dashed line Stage II, and thin line Stage III.
Analyses were restricted to patients who had not had preoperative radiation therapy. Patients must have undergone surgery with curative intent (surgery codes of 30—70; at least a lobectomy or pneumonectomy). Survival analyses were done using SEER*Stat (http:// www.seer.cancer.gov/seerstat) statistical software version 6.1.4. This software used an actuarial method of estimate of cause specific survival based on the cause of death provided in the SEER registry. Cox regression analyses were done with NCSS statistical software of Kayden, Utah.
3. Results Table 1 shows that disease specific survival stratified by AJCC T and N stage at 5 years of follow-up for the 17,310 patients who underwent surgery with curative intent, had pathologically staged Stage I—IIIA disease, and were thus included in this analysis. Stage I patients (T1N0 and T2N0) have the best survival with an average of a 71% 5-year disease specific survival. The majority (on average 60%) of Stage II patients (T1N1, T2N1, and T3N0) will die during this period. Stage IIIA patients (T1N2, T2N2, T3N2, and T3N1) patients had, on average, a still worse prognosis with an average 5-year specific survival of only 24%. Further analyses focused on the 5366 patients with Stage IA disease (T1N0). The first of these analyses (see Table 2) examined the impact of number of nodes reported as pathologically examined. Patients with 0 nodes pathologically examined had a poorer outcome than patients with 1—5, 6—10, >10, or an unknown number of nodes examined. Because the completeness of staging of these patients might be questioned, these patients were excluded from subsequent analyses. Table 3 shows the outcome of the 4866 patients with a known number of nodes examined with T1NO disease stratified by sex, age, tumor size, histologic grade, and histologic
Table 2 Effect of number of nodes examined on 5-year disease specific survival of T1N0 cases All cases 0 1—5 6—10 >10 ?
76.6 67.4 77.3 77.5 78.9 76.5
(0.6), (2.2), (1.0), (1.3), (1.5), (1.2),
n = 5366 n = 500 n = 1881 n = 1074 n = 793 n = 1262
Prognosis of patients with resected non-small cell lung cancer Table 3
209
Outcome in subgroups n
5-Year lung cancer specific survival % (S.E.)
Overall
4866
77.5 (0.6)
Sex Male Female
2533 2333
76.0 (0.9) 79.1 (0.9)
1.15 (0.02) —
Age 21—50 50—59 60—69 70—79
293 964 1933 1634
77.7 81.9 77.9 74.4
0.80 (0.11) 0.64 (< 0.001) 0.82 (0.01) —
Tumor size <10 mm 11—20 21—30
369 2268 2229
85.3 (1.9) 80.0 (0.9) 73.6 (1.0)
0.53 (<0.001) 0.73 (<0.001) —
Histologic grade Well diff. Moderately diff. Poorly diff. Undiff./anaplastic Unknown
640 1769 1567 223 667
84.9 79.0 71.8 71.4 81.8
(1.5) (1.0) (1.2) (3.2) (1.6)
0.51 (<0.001) 0.72 (0.02) 1.01 (ns) —
Histologic type Adenocarcinoma Squamous Large cell Bronchoalveolar Other
2151 1164 268 694 589
76.6 76.2 76.0 84.8 75.4
(0.9) (1.3) (2.7) (1.4) (1.0)
0.95 0.98 0.90 0.63 —
(2.5) (1.3) (1.0) (1.1)
subtype. All the parameters in univariate analysis were predictive of prognosis with the relative risk (RR) conferred being generally modest (between 0.51 and 1.15). Univariate analyses (see Table 3) showed all the tumor and patient related variables tested to be related to prognosis. Tumor size and histologic grade were statistically significant predictors on both univariate and multivariate analyses, with the most favorable of their categories (<10 mm, and well differentiated) having half the risk of their least favorable categories. Sex was weakly predictive in univariate analysis. Patients 70—79 years of age did more poorly than the younger subsets of patients. Patients with bronchoalveolar subtype (BAC) tumors had a more favorable prognosis. Multivariate analysis (see Table 4) identified tumor size, histologic grade, patient age, and histologic subtype as independent predictors of outcome. Patients with combinations of favorable features had excellent outcomes. As can be seen in Table 3 patients with favorable tumor size, histology grade, or bronchoalveolar
Table 4
Results of multivariate analysis
Overall
RR
p
Age (70—79 vs. <70) Grade (>1 vs. 1) Size (11—30 mm vs. 1—10 mm) BAC vs. other types
1.31 1.56 1.52 0.74
<0.001 <0.001 0.003 0.02
Univariate RR (p value)
(ns) (ns) (ns) (0.002)
subtype had a 5-year disease specific survival of 85.3% (S.E. 1.9), 84.9% (S.E. 1.5), and 84.8% (S.E. 1.4), respectively. For patients with combinations of two such favorable features, size/grade, size/BAC, and grade/BAC, the 5-year disease specific survivals were 91.6% (S.E. 3.3), 93.9% (S.E. 4.2), and 87.3% (S.E. 2.5), respectively. For the small subset of these patients (n = 23) with small well-differentiated bronchoalveolar cancers, the 5-year lung cancer specific survival was 95.5% (S.E. 4.4%).
4. Discussion This analysis was done to better define the prognosis of patients with Stage IA NSCLC. Whether these patients should receive adjuvant chemotherapy hinges in large part on the answers to two questions: (1) is their prognosis poor enough to justify the risks and costs of such additional therapy and (2) is adjuvant chemotherapy effective in this patient population? At the present time different national guidelines list adjuvant therapy for Stage I disease as recommended for some patients, or as optional, or not recommended [14—16]. None of these guidelines discusses in quantitative terms the risks faced by the patients or the relative benefit that might be expected. The rationale for not recommending therapy is: (1) that compared to other NSCLC patients, Stage IA patients have a relatively favorable prognosis and (2) that the meta-analyses of clinical trials of adjuvant chemotherapy tends to show only modest efficacy [2—6] and that for
210 the subset of patients with Stage IA disease whether adjuvant therapy reduces the hazard for relapse and death is uncertain. Is the prognosis of Stage IA disease well defined and is it favorable? Published estimates for the 5-year OS of Stage IA patients vary widely, from as high as 90% [17] to as low as 63% [18]. A widely quoted North American analysis of MD Anderson Center experience with 511 Stage IA pathologically staged and resected NSCLC patients found them to have 5year disease specific survival of only 67% [12]. This estimate is slightly more pessimistic than the 76% estimate based on SEER, but seems roughly comparable. The MD Anderson Center estimate for outcome at 5 years of pathologically staged Stage IB patients of 57% is comparable, but again slightly more pessimistic than the SEER based estimate of 62% presented in this paper. Given a number of differences between the two databases some differences in the estimates might be expected. The MD Anderson Center patients were from an earlier era from 1973 to 1988 (in SEER the prognosis of Stage IA patients was no different in years 1988—1992 and 1993—1997; data not shown) and were referred to a tertiary cancer center. In addition, none of the MD Anderson Center patients received adjuvant therapy whereas some of the patients in the SEER database had received adjuvant therapy, even though the impact of this on prognostic estimates derived using the SEER registry would have been expected to be minor (see below). The prognostic estimates for Stage I NSCLC patients based on the National Cancer Data Base (NCDB) are much more pessimistic, and although quoted in the AJCC staging manuals, these estimates are methodologically flawed [19]. The analysis includes both pathologically and clinically staged patients. Its estimate of 45% for the 5-year disease specific survival of Stage I patients is much more pessimistic than other published estimates. One especially valuable piece of data from the NCDB is information about how commonly adjuvant chemotherapy was being given to patients in the United States. This database shows that <1% of patients with surgically resected Stage I NSCLC in 1990 received any adjuvant chemotherapy, and that in 1995 the number was still low, with only about 5% of Stage I patients receiving such therapy. Given that the adjuvant chemotherapy of this era had only a very modest effect (<20% proportional risk reduction of mortality), the impact of treating so few patients with therapy of such modest effectiveness would have no detectable effect on the SEER patients’ outcome presented in this analysis. This is an important point because the SEER registry does not include information on whether a patient received system adjuvant therapy. Because this analysis was done on a large set of 4866 Stage IA patients, the results of this study have more statistical power than the results of many of the studies that analyzed the data of only a at most few hundred cases. The analyses show that in this low risk subgroup, that gender has no independent effect on outcome, but that age, tumor size, histologic grade, and histologic subtype have prognostic significance. The effects of tumor size in Stage I patients have been previously reported by Wisnivesky et al. using data from SEER, but this paper analyzed data from all Stage I patients and did not examine the impact of variables other than size [20].
P.M. Ravdin, G. Davis Other published analyses of tumor size in Stage I patients have also reported it to be an independent predictor of outcome [21—23]. It has been suggested than bronchoalveolar tumors have a better prognosis that the other main NSCLC histologic subtypes. The literature on this topic for Stage I NSCLC has been inconclusive, with the weakness of being based on small data sets in which bronchoalveolar cases represent only about 10—20% of the cases [22,23]. In the analysis presented here based on 694 bronchoalveolar cases, it is clear that this is a favorable histologic subtype (HR of 0.74). Although this is statistically significant, the impact is small in magnitude, and identifies a relatively small cohort of patients. Finding the histologic grade is a predictor of outcome has been reported in the past [21]. An additional analysis was done to assess the impact of histologic grade with the bronchoalveolar cases omitted. The 5-year disease specific survivals were 70.1% (2.5), 64.4% (1.2), 59.6% (1.0), and 56.5% (2.3) for well, moderately, poorly, and undifferentiated cases, respectively, which shows that the effect of grade is not driven by the inclusion of bronchoalveolar cases. The literature about the histologic grading of NSCLC is sparse and questions such as the reproducibility of the interpretation histologic grade apparently have not been addressed. Although much of this analysis has focused on identification of low risk Stage IA patients, another way of looking at the results is that a 22% 5-year disease specific mortality can only be interpreted as a favorable prognosis relative to the prognosis of other patients with lung cancer. This level of risk would be viewed as a reason for giving adjuvant chemotherapy for breast cancer. The inclusion of Stage IA patients in clinical trials would seem to be justified because the majority of these patients are not really at low risk for lung cancer related mortality. The current literature about the efficacy of adjuvant chemotherapy for Stage I (and Stage IA) disease is weak, and to some degree there is uncertainty about the effectiveness of adjuvant chemotherapy in Stage I patients. This uncertainty is reflected in recent trial results. In the IALT trial which showed an overall favorable mortality hazard ratio reduction for adjuvant chemotherapy (HR = 0.86, p = 0.03) there was a trend for Stage I—III subsets of patients all to benefit from adjuvant therapy [10]. In the IALT trial the greatest impact on hazard rate was in patients with Stage III disease, and only in this patient subset did the impact on hazard rate reach statistical significance. Although there was no statistical heterogeneity by stage in benefit, it might still be argued that it had not been proven that Stage I patients benefited. In the JBR-10 trial, which allowed Stage IB and Stage II patients, only the Stage II patients had a survival advantage, and no benefit was seen for the Stage IB patients [8]. In the report of the JBR-10 trial an underpowered statistical test for interaction between treatment and stage failed to show an interaction, so again, although Stage I patients might be presumed to be benefiting this was not rigorously demonstrated. In the ANITA trial Stage I and Stage IIIA patients clearly benefited from therapy, but no benefit was seen in Stage IB patients. In the ANITA trial no formal statistical test was done to assess the possible interaction between treatment effect and stage of disease [9]. Thus the evidence that adjuvant chemotherapy is effective in Stage I patients is not as complete as would be ideal.
Prognosis of patients with resected non-small cell lung cancer The two meta-analyses that have addressed whether there was an interaction between stage and treatment effectiveness did not find such an interaction [5,6], so the lack of statistically significant benefit in the Stage I patients from individual trials is presumably due to a lack of statistical power to demonstrate those advantages. There are some limitations to the analyses presented here because they are based on information from a large registry based on a patient population for which there were no absolute guidelines directing patient evaluation and treatment. Thus there is the possibility of patient selection as to which patients were resected with curative intent. Although this is a particular issue for Stage IIIA patients, for whom there is substantial controversy as to whether an initial surgical approach is the best approach [24,25], this is less likely to be a major issue with Stage I disease. Patients older that 79 were excluded from this analysis because of the concern that the high rates of comorbidity in these older patients would be more likely to cause patient selection, and would also be associated with higher levels of uncertainty about the actual cause of death. Another issue is that there is no verification that histologic grading and histologic subtyping was done uniformly across centers. It should be noted that the major classification system used in the United States is that of the World Health Organization [26], and that there were revisions to this classification system in 1999. Although these are weaknesses, these drawbacks also apply to much of the information derived from clinical trials, as well as patient series at single institutions. The results of these analyses are of potential utility because they show that patients with an approximate 15% risk of 5-year disease specific mortality can be identified in the Stage IA population. These are patients with tumors that are <10 mm in size, are well differentiated, or are of the bronchoalveolar histology. The results are consistent with the interpretation that patients with two favorable features have an approximate 10% 5-year lung cancer specific mortality. How much benefit these low risk subsets of patients might receive from adjuvant chemotherapy depends on estimates of the proportional risk reduction afforded by adjuvant chemotherapy, and assumptions about whether these estimates of efficacy can be applied to these populations of patients. If one uses a meta-analysis derived estimate of a 20% proportional risk reduction, the projected benefit from adjuvant chemotherapy for some low risk patients would be small and approximately equal to treatment related mortality for low risk Stage IA patients. Conversely, on average, Stage IA patients have a risk of NSCLC related death at 5 years of 25% and a possible absolute survival benefit of about 5%. Guidelines suggesting that such patients might consider adjuvant therapy seem justified [14], but further clinical trials in this patient population would be of value.
References [1] Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55(1):10—30. [2] Non-small Cell Lung Cancer Collaborative Group. Chemotherapy in non-small cell lung cancer: a meta-analysis using updated data on individual patients from 52 randomised clinical trials. BMJ 1995;311:899—909.
211 [3] Hotta K, Matsuo K, Ueoka H, Kiura K, Tabata M, Tanimoto M. Role of adjuvant chemotherapy in patients with resected non-small-cell lung cancer: reappraisal with a meta-analysis of randomized controlled trials. J Clin Oncol 2004;22:3860—7. [4] Sedrakyan A, van der Meulen J, O’Byrne K, Prendiville J, Hill J, Treasure T. Postoperative chemotherapy for non-small cell lung cancer: a systematic review and meta-analysis. J Thorac Cardiovasc Surg 2004;128:414—9. [5] Hamada C, Tanaka F, Ohta M, Fujimura S, Kodama K, Imaizumi M, et al. Meta-analysis of postoperative adjuvant chemotherapy with Tegafur—Uracil in non-small-cell lung cancer. J Clin Oncol 2005;23:4999—5006. [6] Berghmans T, Paesmans M, Meerta AP, Branle F, Lafitte JJ, Lemaitre F, et al. Survival improvement in resectable non-small cell lung cancer with (neo)adjuvant chemotherapy: results of a meta-analysis of the literature. Lung Cancer 2005;49:13—23. [7] Strauss GM, Herndon J, Maddaus MA, Johnstone DW, Johnson EA, Watson DM, et al. Randomized clinical trial of adjuvant chemotherapy with paclitaxel and carboplatin following resection in Stage IB non-small cell lung cancer (NSCLC): Report of Cancer and Leukemia Group B (CALGB) Protocol 9633. Proceedings of the 2004 ASCO Annual Meeting. J Clin Oncol 2004;22(14S) [Abstract 7019]. [8] Winton TL, Livingston R, Johnson D, Rigas J, Johnston M, Butts C, et al. Vinorelbine plus cisplatin vs. observation in resected non-small-cell lung cancer. N Engl J Med 2005;352: 2589—97. [9] Douillard J, Rosell R, Delena M, Legroumellec A, Torres A, Carpagnano F. ANITA: phase III adjuvant vinorelbine (N) and cisplatin (P) versus observation (OBS) in completely resected (stage I—III) non-small-cell lung cancer (NSCLC) patients (pts): final results after 70-month median follow-up. Proceedings of the 2005 ASCO Annual Meeting. J Clin Oncol 2004;23 [Abstract 7013]. [10] Arriagada R, Bergman B, Dunant A, Le Chevalier T, Pignon JP, Vansteenkiste J. International Adjuvant Lung Cancer Trial Collaborative Group. Cisplatin-based adjuvant chemotherapy in patients with completely resected non-small-cell lung cancer. N EngI J Med 2004;350(4):351—60. [11] Greene FL, Page DL, Fleming ID, et al., editors. AJCC cancer staging manual. New York: Springer-Verlag; 2002. [12] Mountain CF. Revisions in the International System for Staging Lung Cancer. Chest 1997;111:1710—7. [13] Surveillance, Epidemiology, and End Results (SEER) Program (http://www.seer.cancer.gov) SEER*Stat Database Incidence— –SEER 13 Regs Public-Use, November 2004. National Cancer Institute, DCCPS, Surveillance Research Program. [14] National Comprehensive Cancer Network. Clinical practice guidelines in oncology, v. 2; 2005 (http://www.nccn.org). [15] http://cancer.gov/cancerinfo/pdg/cancerdatabase. [16] National Institute for Clinical Excellence. Lung cancer: the diagnosis and treatment of lung cancer. http://www.nice. org.uk/CG024NICEguideline. [17] Kato H, Ichinose Y, Ohta M, Hata E, Tsubota N, Tada H, et al. Japan Lung Cancer Research Group on Postsurgical Adjuvant Chemotherapy. A randomized trial of adjuvant chemotherapy with uracil—tegafur for adenocarcinoma of the lung. N EngI J Med 2004;350(17):1713—21. [18] van Rens MTM, de la Riviere, Elbers HR, van Den Bosch JM. Prognostic assessment of 2,361 patients who underwent pulmonary resection for non-small cell lung cancer, stage I, II, and IIIA. Chest 2000;117:374—9. [19] Fry WA, Phillips JL, Menck HR. Ten-year survey of lung cancer treatment and survival in hospitals in the United States. A National Cancer Data Base Report. Cancer 1999;86:1867—76. [20] Wisnivesky JP, Yankelevitz D, Henschke CI. The effect of tumor size on curability of stage I non-small cell lung cancers. Chest 2004;126:761—5.
212 [21] Gajra A, Newman N, Gamble GP, Abraham NZ, Kohman LJ, Graziano SL. Impact of tumor size on survival in stage IA nonsmall cell lung cancer: a case for subdividing stage IA disease. Lung Cancer 2003;42:51—7. [22] Birim O, Kappetein AP, Takkenberg JJM, van Klaveren RJ, Bogers AJ. Survival after pathological stage IA nonsmall cell lung cancer: tumor size matters. Ann Thorac Surg 2005;79:1137—41. [23] Port JL, Kent MS, Korst RJ. Tumor size predicts survival within stage IA non-small cell lung cancer. Chest 2003;124: 1828—33.
P.M. Ravdin, G. Davis [24] van Meerbeeck JP, Kramer G, Van Schil PE, Legrand C, Smit EF, Schramel FM, et al. A randomized trial of radical surgery (S) versus thoracic radiotherapy (TRT) in patients (pts) with stage IIIA-N2 non-small cell lung cancer (NSCLC) after response to induction chemotherapy (ICT) (EORTC 08941). Proceedings of the 2005 ASCO Annual Meeting. J Clin Oncol 2005;23 [Abstract 7015]. [25] Turrisi AT. A case against surgery for most IIIa non-small cell lung cancer. Semin Oncol 2005;32(Suppl 3):S6—8. [26] Travis WD, Sobin HL. Histological typing of lung and pleural tumours. 3rd ed. Berlin: Springer-Verlag; 1999.