Prognostic Significance of Periodic Acid-Schiff-positive Patterns in Primary Cutaneous Melanoma

• Published in: Clinical cancer research Vol. 7; no. 3; pp. 473 - 477
• Main Authors: Warso, M A, Maniotis, A J, Chen, X, Majumdar, D, Patel, M K, Shilkaitis, A, Gupta, T K, Folberg, R
• Format: Journal Article
• Language: English
• Published: United States American Association for Cancer Research 01-03-2001
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Biopsy; Disease-Free Survival; Female; Humans; Male; Melanoma - diagnosis; Melanoma - pathology; Middle Aged; Periodic Acid-Schiff Reaction; Prognosis; Proportional Hazards Models; Regression Analysis; Skin Neoplasms - diagnosis; Skin Neoplasms - pathology; Time Factors
• ISSN: 1078-0432; 1557-3265

The patterns of periodic acid-Schiff (PAS) staining of extracellular matrix in histological sections of certain melanomas
may be predictive of outcome. Recent in vitro and molecular genetic data suggest that the appearance of these patterns in both uveal and cutaneous melanoma is a function
of aggressive tumor cells. We studied 96 patients with primary cutaneous melanomas treated at the University of Illinois at
Chicago who were monitored for disease-free survival. Survival probabilities were determined by Kaplan-Meier estimates, and
prognostic factors were evaluated by multivariate analysis. By univariate analysis, there was a significant decrease in disease-free
survival among patients whose tumors contained parallel with cross-linking or network patterns (PXNs; P = 0.0070). Stepwise regression with Cox models that included the combinations of the PAS-positive patterns, tumor thickness,
female gender, ulceration, and age yielded a model with thickness and the PAS-positive parallel with cross-linking or networks.
Despite the relatively small sample size in this study, the detection of the PAS-positive parallel with cross-linking or networking
in cutaneous melanoma was associated with a decrease in disease-free outcome. Additional studies of the prognostic significance
of these patterns is warranted on larger data sets.