P. Buttner et al., PRIMARY CUTANEOUS MELANOMA - OPTIMIZED CUTOFF POINTS OF TUMOR THICKNESS AND IMPORTANCE OF CLARKS LEVEL FOR PROGNOSTIC CLASSIFICATION, Cancer, 75(10), 1995, pp. 2499-2506
Background. Maximum tumor thickness and level of invasion are known to
be the most important prognostic factors for patients with primary cu
taneous melanoma. However, the classification of tumor thickness and t
he question of whether the combination of tumor thickness and level of
invasion provides a better prognostic classification than tumor thick
ness alone are still matters of debate. The present study examined the
relationship between tumor thickness and survival probability to defi
ne cutoff points of tumor thickness. Secondly, it investigated the pro
gnostic value of the combination of tumor thickness and level of invas
ion as proposed in the current TNM classification system. Methods. A s
eries of 5093 patients with invasive primary cutaneous melanoma follow
ed from 1970 to 1988 at four University centers in Germany (Department
s of Dermatology in Tubingen, Wurzburg, Berlin-Steglitz, and at the Fa
chklinik) were analyzed by multivariate Cox models. Results. The relat
ionship between tumor thickness and relative risk of death caused by m
elanoma was found to be almost linear to a tumor thickness of 6 mm. Fo
r tumors greater than 6 mm, no further marked increase in relative ris
k was observed. The stratification of tumor thickness with endpoints a
t 1, 2, and 4 mm resulted in the best fit to the authors' data among a
ll classifications with three endpoints, but differences were only sli
ght. By multivariate analysis, the combination of tumor thickness and
level of invasion as proposed by the current TNM classification were f
ound to be prognostically less significant than tumor thickness alone.
The prognostic influence of level of invasion was proved statisticall
y only for tumor thickness less than or equal to 1 mm. Conclusions. Th
e proposed stratification of tumor thickness with cutoff points at 1,
2, and 4 mm was supported by multivariate statistical analysis. The an
alysis of the current TNM staging system indicates the precedence of t
umor thickness for the staging of patients with primary cutaneous mela
noma in the case of discordance between tumor thickness and level of i
nvasion.