Continuing Controversies in Staging NSCLC: An Analysis of the Revised 1997 Staging System

ABSTRACT: The 1997 revision of the lung cancer staging system has added very little to disease staging, and many changes have been totally unnecessary. Before the next revision of the staging system, anticipated in the year 2007, a worldwide effort to collect accurate data bases and analyze a variety of T, N, and M prognostic factors as well as the newer biologic factors is required to define more accurately the clinical and pathologic stages of lung cancer according to prognosis. This approach would allow a more rational consideration of changes in TNM staging. As well, a single, universally accepted lymph node map is needed desperately to unify staging concepts worldwide.[ONCOLOGY 12(Suppl 2):51-54, 1998]

The value of clinical and surgical-pathologic staging in managing lung cancer cannot be underestimated. Both noninvasive and invasive staging techniques, short of resection, can be used for clinical staging parameters and, although the accuracy of such staging is less than ideal, newer techniques promise improvements. In comparison, surgical/pathologic staging depends on cooperation between the surgeon and pathologist in identifying local invasion and the location and involvement of lymph nodes in the resected specimen. The importance of routine evaluations of this type has been well established, allowing as it does for accurate final surgical/pathologic staging, which is vitally necessary for developing new treatment strategies.

The 1987 American Joint Committee on Cancer/International Union Against Cancer (AJCC/UICC) TNM Classification of Lung Cancer[1,2] proved to be an excellent tool for developing stage-related treatment strategies and estimates of prognosis. Very recently, the revised 1997 iteration has been published by the UICC and AJCC (Table 1).[3,4] Changes have included further subdivision of lung cancer stages I and II into subsets according to clinical and pathologic prognosis. Unfortunately, the minor modifications made leave much to be desired. For example, recent retrospective and prospective reviews have identified problems in the current T and N definitions and inconsistencies in prognosis within the subsets of the various stages that the 1997 revision does not address. The International Association for the Study of Lung Cancer was not consulted about these changes. Before the next TNM staging revision, anticipated for the year 2007, detailed prospective and retrospective analyses should be made by interested groups throughout the world, then collated and collected by the International Association for the Study of Lung Cancer to provide the best possible staging system for clinical application worldwide.

TNM Classification

T Status

At present, T1 and T2 tumors are considered to be well defined, but recent analyses[5,6] suggest that survival is much worse for patients with tumors larger than 5 cm than for those with tumors smaller than 5 cm, and it appears that this 5-cm parameter has greater prognostic value than the current 3-cm cutoff for T1 tumors. A prospective analysis should therefore address whether a T1 tumor should retain its £ 3-cm definition or should be changed to include a 4- or 5-cm lesion, with tumors larger than 5 cm reclassified as marking the border of the higher T2 status. One report questions whether tumor invasion of the elastic layer of the pleura determines a worse prognosis and, thus, whether such invasion should continue to define a T2 tumor.[7]

The definitions of T3 versus T4 tumors tend to be extremely vague. In the 1997 classification, a T3N0 tumor has been categorized as stage IIB. However, a tumor that invades only the parietal pleura has a much more optimistic prognosis than one invading through the endothoracic fascia, chest wall musculature, or ribs. A retrospective analysis from Memorial Sloan-Kettering Cancer Center suggests that simple parietal pleural invasion has a much better prognosis than does deeper chest wall invasion.[8] However, larger subset analyses are needed to identify the prognostic implications of this type of invasion. Quite possibly, this minimal invasion (pleura only) should be reclassified as T2, since 5-year survival following complete resection of such lesions is 50% or greater. Deeper invasion should retain its T3 classification. Tumors invading the mediastinum and superior sulcus tumors, now classified as T3, do not have the same prognosis as stage II lesions and could thus remain as T3, with these T3 tumors reverted to stage IIIA disease once again.

Although the T4 category has been reasonably well described, tumors invading the phrenic nerve or vagus nerve in the aortic or supra-aortic region, now classified as T3, have extremely poor prognoses and probably should be upgraded to a T4 designation. More precise definitions to identify the T status of the superior sulcus tumor (T3 vs T4) are required. For example, the presence of Horner’s syndrome, motor (vs sensory) dysfunction of the lower brachial plexus (C8 involvement), or involvement of the subclavian vessels in all likelihood should designate this tumor as T4.

All of these proposed changes require confirmation by a worldwide effort, rather than a single institutional data base.

N Status

Although N2 lymph nodes are considered mediastinal, the exact location of the pleural envelope has not been well defined. A general consensus is needed by which, for example, it is agreed that the pleural envelope begins just proximal to an identifiable fixed structure (eg, the upper lobe bronchus) and, therefore, all lymph nodes proximal to this point should be designated as mediastinal. At present, this definition is accepted by the AJCC but not by the UICC. Ultimately, this definition will affect the designation of hilar nodes, the involvement of which appears to signify a poorer prognosis than does the involvement of other N1 nodes.

The definition of N3 disease within the mediastinum is quite vague and depends on the nodal map used. The lymphatic drainage of the various lobes to the superior mediastinum should be taken into account in designating N3 disease (Figures 1 and 2 [9,10]).

Left-sided tumors, for example, drain ipsilaterally in the superior mediastinum, along only the left tracheoesophageal groove. Any lymph node drainage to the right of the left paratracheal border should be considered contralateral or N3 disease for these left-sided tumors.

In contrast, right-sided disease drains to lymph nodes in the superior mediastinum, anterior to the trachea as far as the left tracheal border. These should all be considered ipsilateral (right) paratracheal lymph nodes. Because of these anatomic features, the left anterior border of the trachea and not the midline of the trachea should be designated as the dividing point between ipsilateral and contralateral. By this definition, contralateral N3 nodes from a right-sided tumor would include only those lymph nodes along the left tracheoesophageal groove and along the lateral border of the left mainstem bronchus. According to this new definition, the normal lymphatic drainage of the hemithoraces rather than the midline would define a lymph node as left or right paratracheal (levels 4 and 2).

Controversy persists as to what constitutes subcarinal (N2) versus contralateral (N3) nodes in the subcarinal space. Many Japanese surgeons identify subcarinal nodes along the medial border of the contralateral mainstem bronchus as contralateral N3 lymph nodes.[11] In contrast, most surgeons in North America consider lymph nodes anywhere within the subcarinal space as ipsilateral. A universally acceptable definition is needed.

The number of involved lymph nodes versus stations has never been factored into the assessment of prognosis. Other solid tumors (eg, breast or esophagus) are apparently associated with a much poorer prognosis when more than four lymph nodes are involved. This same prognostic effect is also seen in lung cancer, at least for N1 nodes and likely for N2 nodes as well. A prospective analysis would clarify this and could have an impact on TNM staging in the future.

M Status

In lung cancer, patients with a solitary site of metastasis, especially the brain, can be considered for surgical resection and possible cure.[12] Thus, a designation of M1a disease would better categorize this solitary site subset and would further differentiate such patients from those who have multiple metastatic sites (M1b).

Lymph Node Maps

Currently at least three lymph node maps are being used worldwide (Naruke, the American Thoracic Society, and the Lung Cancer Study Group modification of the American Thoracic Society).[13-15] Another lymph node map recently suggested by Mountain and Dresler[16] is really more of a minor modification of the Lung Cancer Study Group map. A single lymph node map that can be adopted by all TNM authorities and used universally is needed urgently.

Stricter definitions also are needed to differentiate level 1 and level 3 lymph nodes. Posterior level 3 lymph nodes (3p, UICC), for example, would best be considered as level 8 (paraesophageal) nodes, which rarely are involved in tumors of the lung. Anterior level 3 lymph nodes (3a, UICC), however, should include only those prevascular nodes draining the right upper lobe, anterior and lateral to the superior vena cava. These nodes are in reality levels 5 and 6 lymph nodes in the left hemithorax, although the frequency of involvement of these right-sided nodes and their prognosis are unknown.

In North America, level 1 nodes are considered by many as the lowest cervical nodes (N3) since they lie superior to the innominate vein and abut the thoracic inlet. Although the prognosis associated with tumors involving level 1 lymph nodes is not known, it is suspected that these are equivalent in prognosis to tumors with any other N3 involvement and should probably be classified as such.

Staging Classification and Prognosis

The newly revised 1997 TNM staging guidelines provide new subgroups for stages I and II. There certainly is reason to subdivide stage I disease since patients with T1N0 tumors have a much better prognosis than those with T2N0 tumors. Similarly, T3N0 tumors involving only the parietal pleura or by virtue of proximity to the main carina are associated with a prognosis similar to that of other stage II tumors (approximately 40% to 50% 5-year survival after complete resection). It is therefore valid to classify these T3N0 tumors as stage II disease. Unfortunately, all T3N0 tumors do not indicate as positive a prognosis. The value of subdividing T1N1 and T2N1 tumors is questionable and such changes are probably unnecessary.

Clinical vs Pathological Staging

All tumors, including small-cell lung cancer, should be staged clinically by TNM criteria. According to TNM rules, NX (nodes not assessable) is allowed as a valid clinical staging identifier. With the advent of computed tomography scans, however, NX should now be disallowed in clinical staging using chest x-rays and computed tomography scans as the means of assessing nodal disease (N0 vs N1 vs N2 and 3). The NX classification should be allowed only for surgical/pathologic staging when no information has been provided by the surgeon and/or pathologist to determine lymph node involvement.

Other Prognostic Factors

Many new and potentially useful biologic prognostic factors have been reported in the literature.[17,18] At present, weight loss is the only non-TNM prognostic factor considered to be valid.[19] Over the next 10 years, however, it is quite possible that tumor markers and pathologic markers (eg, epithelial markers and angiogenesis) will be identified as having proven prognostic value and as such may be included in the next iteration of TNM staging.

Conclusions

The 1987 TNM classification for lung cancer has served us well. The minor modifications presented in the 1997 iteration added very little to the guidelines set forth in 1987, with many alterations unnecessary or uncalled for. Changes, however, are needed: More precise definitions of the T and N categories are called for. The prognostic value of the number of involved nodes (vs stations) should be assessed. A single worldwide nodal map, with more precise definitions, should be agreed upon. This nodal map should have strict guidelines and compatibility such that it can be used by both clinicians and radiologists. The newer prognostic factors recently identified also should be verified by large prospective studies so that they can be included where indicated in the next TNM staging classification.

Long before the next TNM pronouncement, anticipated for 2007, the International Association of the Study of Lung Cancer should become proactive and strike a representative TNM staging committee to prospectively collect data and advise both the UICC and AJCC concerning any further changes. This process should begin immediately.

References:

1. International Union Against Cancer: TNM Classification of Malignant Tumours, 4th ed, pp 69-73. Berlin, Springer-Verlag, 1987.

2. International Union Against Cancer: TNM Classification of Malignant Tumours, 5th ed, pp 93-100. New York, Wiley-Liss, 1997.

3. American Joint Committee on Cancer (AJCC): Manual for Staging of Cancer, 3rd ed, pp 115-121. Philadelphia, J. B. Lippincott, 1988.

4. American Joint Committee on Cancer (AJCC): Manual for Staging of Cancer, 4th ed, pp 127-137. Philadelphia, J. B.Lippincott, 1997.

5. Martini N, Burt ME, Bains MS, et al: Survival after resection of stage II non-small-cell lung cancer. Ann Thorac Surg 54:460-466, 1992.

6. van Velzen E, Snijder RJ, Brutel de la Riviere A, et al: Lymph node type as a prognostic factor for survival in T2 N1 M0 non-small-cell lung carcinoma. Ann Thorac Surg 63:1436-1440, 1997.

7. Martini N, Bains M, Burt ME, et al: Incidence of local recurrence and second primary tumors in resected stage I lung cancer. J Thorac Surg 109:1-10, 1995.

8. McCaughan BC, Martini N, Bains MS, et al: Chest wall invasion in carcinoma of the lung. Therapeutic and prognostic implications. J Thorac Cardiovasc Surg 89:836-841, 1985.

9. Hata H: Resection of N2/N3 mediastinal disease, in Motta G (ed): Lung Cancer—Frontiers in Science and Treatment, pp 431-444. Genoa, Rapallo, 1993.

10. Lee JD, Ginsberg RJ: Lung cancer staging: The value of ipsilateral scalene lymph node biopsy performed at mediastinoscopy. Ann Thorac Surg 62:338-341, 1996.

11. Naruke T: Mediastinal lymph node dissection, in Pearson FG, Deslauriers J, Ginsberg RJ, et al (eds): Thoracic Surgery, pp 909-917. New York, Churchill Livingstone, 1995.

12. Wronski M, Arbit E, Burt M, et al: Survival after surgical treatment of brain metastases for lung cancer: A follow-up study of 231 patients treated between 1976-1991. J Neurosurg 83:605-616, 1995.

13. Naruke T, Sumasu K, Ishikawa S: Lymph node mapping and curability at various levels of metastasis in resected lung cancer. J Thorac Cardiovasc Surg 76:832-839, 1978.

14. American Thoracic Society: Clinical staging of primary lung cancer. Am Rev Respir Dis 127:1-6, 1983.

15. Rusch VW, Ginsberg RJ, Holmes EC: A Thoracic Surgical Handbook for Clinical Trials. New Jersey, Bristol-Myers Squibb, 1993.

16. Mountain CF, Dresler CM: Regional lymph node classification for lung cancer staging. Chest 111(6):1718-1723, 1997.

17. Kern JA, Slebos RJC, Top B, et al: C-ErbB-2 expression and Codon 12 K-ras mutations both predict shortened survival for patients with pulmonary adenocarcinoma. J Clin Invest 93:516-520, 1994.

18. Minna JD, Sekido Y, Fong KW, et al: Molecular biology of lung cancer, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practices of Oncology, 5th ed, pp 849-857. Philadelphia, Lippincott-Raven, 1996.

19. Ginsberg RJ, Vokes E, Raben A: Non-small-cell lung cancer, in DeVita VT, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology, 5th ed, pp 858-911. Philadelphia, J. B. Lippincott-Raven, 1996.