Prognostic factors in patients undergoing lymphadenectomy for malignant melanoma.

Review of a 19 year experience in melanoma patients undergoinglymphadenectomy at the National Cancer Institute revealedthat the preoperative assessment of the status of theregional lymph nodes was accurate 91% of the time when thesurgeon felt the nodes were clinically positive, and accurate79% of the time when the nodes were judged clinically negative. The 10-year survival in patients with one to three histologicallypositive nodes or no positive nodes was 50-55%, compared to a25% 8-year survival in patients with four or more histologicallypositive nodes. Stepwise multivariate evaluation of prognosticfactors indicated that the most important factor for predictingprognosis is the number of nodes histologically involved. Nodepalpability was the second most important factor because of itshigh correlation with number of nodes histologically involved. Site of melanoma was the third most important factor, aspatients with extremity (upper or lower) melanoma had a bettersurvival (P = 0.002) than patients with axial melanoma (trunkor head and neck). Five years following lymphadenectomythere appeared to be substantial differences in survivalaccording to differences in the level of invasion of the primarylesion, however, these differences were not nearly aspronounced 10 years following node dissection.B The division of melanoma thicknesses into <1.50 mm and>1.50 mm provided some prognostic discrimination at fiveyears but again the differences were not pronounced 10 yearsfollowing node dissection. The thickness measurements wereeasier to determine than the level of invasion, and more reproduceableon resubmission to the same pathologist. Fourpatients with melanoma less than 0.76 mm had subsequentmetastases, but these may represent inadequate sampling of theprimary melanoma both in our series and in the four similarpatients previously reported with such thin metastasizingmelanomas.