From the publishers of The New England Journal of Medicine

Save time and stay informed. Our physician-editors offer you clinical perspectives on key research and news.

  1. Home>
  2. Specialties>
  3. Dermatology>
  4. Summary and Comment

The Lethal Thin Melanoma — Wolf in Sheep’s Clothing

A new decision tree adds previously unavailable parameters to the prediction of risk in thin melanomas.

Although the survival rate associated with thin melanomas exceeds 90%, about 15% of annual reported deaths in the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database involve thin melanomas. This mortality rate is largely explained by the high number of thin melanomas currently diagnosed in the U.S. and highlights the need for effective identification of high-risk thin melanomas, especially if adjuvant therapy is to be considered.

To create a classification system that predicts thin melanoma risk better than the current American Joint Commission on Cancer (AJCC) staging system, researchers at the University of Pennsylvania examined records on 26,291 melanomas in the SEER database and proposed new parameters that they validated in 2389 melanomas seen by their Pigmented Lesion Group. The AJCC parameters — age, sex, anatomic site, and tumor characteristics (level, thickness, and ulceration) — were available for all patients. The researchers looked at four additional candidate prognostic factors: radical (RGP) or vertical (VGP) growth phase; presence of RGP regression; presence of VGP mitogenicity; and presence of tumor-infiltrating lymphocytes (TILs). They developed a statistical model called a prognostic tree to sequentially divide patients into classes, with each tree node representing an "if-then" stratification decision (e.g., presence or absence of RGP).

Receiver-operating characteristic (ROC) curve analysis showed that the new model performed significantly better than AJCC staging (mean difference in area under the curve, 0.06). For example, in the new system, a man presenting with a 0.81-mm nonulcerated mitogenic melanoma would be assigned to a class with an estimated 10-year survival rate of 90.6% (comparable to stage 1B in AJCC). Under the AJCC system, the same patient would be classified as stage IA or T1a, with an expected 10-year survival rate of 97.4%.

Comment: If the number of melanomas continues to increase while median thickness remains under 1 mm, deaths attributable to thin melanomas will also increase. Identifying the patients with thin melanomas at greatest risk for death is particularly important. The University of Pennsylvania Pigmented Lesion Group has been devising elegant models to predict outcome since the days of Wallace Clark. According to this new prognostic tree, it appears that older men with thin, nonulcerated, mitogenic lesions have a worse prognosis than the present system suggests. Although this system has implications for sentinel lymph node sampling, other clinical practice changes should wait for further validation in additional sets of patients. The ultimate goal, of course, is the identification of molecular markers that will replace even the most sophisticated mathematical models.

— Hensin Tsao, MD, PhD

Published in Journal Watch Dermatology April 20, 2007

Citation(s):

Gimotty PA et al. Identification of high-risk patients among those diagnosed with thin cutaneous melanomas. J Clin Oncol 2007 Mar 20; 25:1129-34.

Your Remark:

Reader Remarks are intended to encourage lively discussion of clinical topics with your peers in the medical community. Please consider this when composing your remark.

Fields marked with an * are required.

Name as you'd like it to appear:

Submitting a comment indicates you have read and agreed to the remark guidelines and declare:*

PRIVACY: We will not use your email address, submitted for a comment, for any other purpose nor sell, rent, or share your e-mail address with any third parties. Please see our Privacy Policy.

 

CLEAR erases anything you've added in any part of the form. CONTINUE allows you to check your entire post (and edit it if necessary) before submitting.

To ensure that your Reader Remark is not formatted as one long paragraph, precede new paragraphs with either a blank line or an indentation.

Search

Advanced

Article Tools

Reader Remarks

Sign-In

Forgot your password?

New to Journal Watch?

E-mail Alerts

Delivered to your inbox.
Tailored to your interests. Free.

Sign Up Now!

Journal Watch Newsletters

Available in 13 specialties with convenient delivery and 10 free online CME exams.

Subscribe Now!

Copyright © 2007. Massachusetts Medical Society. All rights reserved.