by ecancer reporter Clare Sansom

The treatment and prognosis of patients with solid tumours depend crucially on the stage of their disease, and staging a tumour is therefore one of the most important factors in planning the treatment for an individual patient.

For decades, the principal classification system used has been the TNM system maintained by the International Union against Cancer (UICC). In this, T refers to the size and spread of the primary tumour; N to the number of lymph nodes involved; and M to the presence or absence of distant metastases.

Each TNM classification is mapped onto a stage (Stage I -Stage IV) and this is used to predict prognosis and plan therapy. The TNM system has been constantly developing since it was first introduced in the 1970s, with the seventh edition replacing the sixth in 2009. 

A group of clinicians and scientists led by Helmut Friess of Technische Universität München, Munich, Germany, has now investigated whether the more complex seventh edition of the classification scheme for colorectal cancer improves the prediction of prognosis.

The researchers collected clinical and histopathological data from 2229 patients with colorectal cancer, taken from a series of 2364 patients who underwent surgery at the Technische Universität München between 1996 and 2009. 

Unusual tumours and cases with missing pathological data or insufficient follow-up data were excluded from the analysis. Each tumour was assigned to a TNM category and a UICC stage according to both the sixth and the seventh classification systems, and a further independent prognostic indicator was calculated using a scoring system and nomogram based on all available clinical and histopathological information (the “Munich model”). 

Outcome data from a mean follow-up of 86 months was used to compare the predictive value of the three classification systems.

In colorectal cancer, the changes from the sixth to the seventh edition of the TNM classification system have involved the introduction of a number of new subdivisions in the T, N and M groupings, leading to changes in stage classification for some tumours. In particular, the M1 classification, indicating distant metastases, was subdivided into M1a (metastasis into one organ) and M1b (more than one), and the Stage IV classification for metastatic cancer was similarly split into Stage IVa if one organ was involved and Stage IVb if two or more were.

In this series of patients, the assigned UICC tumour stage changed between edition 6 and edition 7 in only 155 (7%) of cases and the division of Stage IV tumours by the extent of metastasis was found to have no influence on prognosis.

In summary, five year survival rates for all patients when tumours were staged using the UICC sixth edition were: stage I 96%; IIA 90%; IIB 86%; IIIA 90%; IIIB 72%; IIIC 48%; and IV, 13%. The equivalent survival rates with staging calculated using the seventh edition were: stage I 96%; IIA 90%; IIB 84%; IIC 87%; IIIA 89%; IIIB 72%; IIIC 36%; IVA 15%; and IVB, 10%. The three systems were evaluated based on one-, two- and five-year cancer-specific survival probability.

The performance of each of the classification systems was assessed by calculating its concordance index, or c-index; this assesses both sensitivity (true positive rate) and specificity (true negative rate), and a c-index of 1 indicates a perfectly predictive model.

The c-value of the UICC sixth edition was found to be 0.829 (CI: 0.815-0.842); of the UICC seventh edition 0.833 (CI: 0.819-0.847); and of the Munich model 0.843 (CI: 0.828-0.857). Any differences between these three systems therefore failed to reach statistical significance. 

However, the specificity increased for the Munich model compared to each of the UICC editions, and the sensitivity of the seventh edition was significantly improved over that of the sixth edition.

The researchers concluded that the more complex, and therefore more difficult to handle, seventh edition of the UICC classification and staging system did not confer greater accuracy in prognosis prediction over its predecessor. 

Reference

Nitsche, U., Maak, M., Schuster, T., PhD, Künzli, B., Langer, R., Slotta-Huspenina, J., Janssen, K-P., Friess, H. and Rosenberg, R. (2011). Prediction of Prognosis Is Not Improved by the Seventh and Latest Edition of the TNM Classification for Colorectal Cancer in a Single-Center Collective. Ann. Surg. 254: 793–801. doi: 10.1097/SLA.0b013e3182369101