Primary cutaneous melanoma is often infiltrated by lymphocytes that provide the opportunity to study what may be the local immunologic reaction to the tumor and to correlate the presence of these lymphocytes with overall survival. In an attempt to delineate the histologic diagnostic criteria, to classify different categories of lymphocytic infiltrates, previously described by Elder et al. as brisk, nonbrisk, and absent, and to verify their prognostic significance, we reviewed 285 consecutive cases of primary cutaneous melanomas (American Joint Committee on Cancer Stage I and II).
In addition to clinical variables (age, sex, and location of tumor) and the presence of tumor infiltrating lymphocytes in the vertical growth phase, the histopathologic attributes reviewed included mitotic rate, thickness, and regression. The results were derived from independent histopathologic review by two pathologists (C.G.C., M.C.M., Jr.) on separate occasions. A multivariate analysis of survival was performed with the Cox's regression model.
The 5- and 10-year survival rates for melanoma with a vertical growth phase and a brisk infiltrate were 77% and 55%, respectively. For tumors with a nonbrisk infiltrate, the 5- and 10-year survival rates were 53% and 45%, respectively, and for tumors with absent tumor infiltrating lymphocytes, the 5- and 10-year survival rates were 37% and 27%, respectively. Mitotic index, thickness, and tumor infiltrating lymphocytes were statistically (univariate analysis) significant prognostic factors (P = 0.003, 0.000001, 0.0003, respectively), whereas the presence or absence of regression is not. In the univariate statistical analysis, the sex of patients and site of melanoma also were statistically significant (P = 0.00001 and 0.002 respectively), whereas age (P = 0.98) was not statistically significant. The multivariate analysis of thickness, mitotic rate, and tumor infiltrating lymphocytes showed that thickness and presence of tumor infiltrating lymphocytes were significant and independent histologic prognostic factors. With regard to the clinical factors, sex retained its independent prognostic significance. The histologic characteristics of melanoma with vertical growth phase (brisk, nonbrisk, and absent) are exemplified.
We demonstrated that when categories of tumor infiltrating lymphocytes are strictly defined, they indeed have very strong predictive value for primary cutaneous melanomas with a vertical growth phase. This work confirms the work of Clark et al. and fully illustrates the brisk, nonbrisk, and absent categories of infiltration. Finally, a multivariate analysis comparing thickness, mitotic rate and presence of tumor infiltrating lymphocytes showed that only thickness and presence of tumor infiltrating lymphocytes are significant and independent positive histologic prognostic factors. Cancer 1996;77:1303-10.