Cancer staging systems continuously evolve, aiming to minimize prognostic overlap between stages and prognostic heterogeneity within stages. This provides more accurate and predictive survival statistics and may enhance treatment planning for patients at any given stage of disease.
The new AJCC melanoma staging system, instituted this year, is no different; it too attempts to define more homogeneous “at risk” patient populations. It goes without saying that any staging system must rely on previously described prognostic factors proven to be statistically robust in multivariate analysis.
In this issue of The Melanoma Letter, Drs. Balch, Mihm, Gershenwald, and Soong provide a detailed evaluation of the revised melanoma staging system. The system remains largely unaltered, but a few important changes have occurred. Although tumor thickness remains the most vital prognostic factor in stage I and II patients, two other factors have proven important in defining more prognostically homogeneous patient populations with stage I or II disease; namely, ulceration and mitotic rate. Ulceration remains from previous editions, as studies continue to show it is a key discriminating factor, while mitotic rate is newly added, essentially replacing the long-established Clark’s level. The rationales for this quite revolutionary step were multivariate analyses revealing that, once mitotic rate was included in the statistical model, Clark’s level became an insignificant prognostic variable.
While the presence or absence of ulceration and mitotic rate helps minimize prognostic heterogeneity for stage I and II patients, sentinel lymph node status helps minimize overlap between stages. For patients with stage III disease, the tumor burden in the regional nodes and the number of nodes involved remain important prognostic variables. However, based on improved prognosis for patients presenting with satellite or in-transit metastasis compared to other stage IV patients, the former are now classified as having stage III disease.
Finally, stage IV disease with a normal LDH has been regrouped, highlighting the observation that patients presenting with distant skin, subcutaneous, or nodal metastases have a better prognosis than those with lung or other visceral metastases. However, once LDH is elevated, prognosis is poor regardless of metastatic site.
Categorizing patients into groups with similar survival probabilities can help researchers evaluate the efficacy of melanoma treatments. More importantly, groupings provide valuable information to help guide management. One day, molecular and genetic variables will likely be included in cancer staging systems, creating even more homogeneous “at risk” populations. However, no matter how versed we become at predicting outcomes, it is important to remember that no individual patient is a statistic. Regardless of their ‘probability’ of survival, individual patients experience dichotomous outcomes—either surviving or succumbing to their disease. Since the outcome for each melanoma patient, no matter what stage, remains unknown, each should be treated like one whose recovery will tip the survival statistics in a more favorable direction.
Allan C. Halpern, MD
Ashfaq A. Marghoob, MD