A micromorphometry-based concept for routine classification of sentinel lymph node metastases and its clinical relevance for patients with melanoma

BACKGROUND. The sentinel lymph nodes (SLNs) as the primary targets for lymp hatic metastases can be removed selectively by gamma probe-guided sentinel lymph nodectomy (SLNE) in nearly all patients with cutaneous melanoma. Corr espondingly high standards in terms of specificity, sensitivity, and micros taging are required for the evaluation of SLNs. METHODS. Since 1995, the authors have performed SLNE in 389 lymph node regi ons (LNRs) on 342 patients with melanoma. The harvested 636 SLNs and a furt her 1394 nonsentinel lymph nodes (non-SLNs) were evaluated by standardized, semiserial section histology, including immunohistochemistry. For each LNR , this technique permitted routine S classification using two simple morpho metric parameters: the number of tumor-involved 1-mm slices of the SLNs (n) and the centripetal depth of metastatic cell invasion (d). S1 was defined by 1 less than or equal to n less than or equal to 2 and d less than or equ al to 1 mm, equivalent to localized peripheral tumor cell deposits; S2 was defined by n > 2 and d less than or equal to 1 mm, indicating more extended peripheral metastases; S3 was defined by d > 1 mm in SNLs with deeper meta static infiltration; and S0 meant no detectable tumor cells (n = 0). RESULTS. The authors diagnosed 325 SLNs as S0, 24 SLNs as S1, 22 SLNs as S2 , and 18 SLNs as S3. The occurrence of at least one melanoma-positive non-S LN subsequent regional completion lymph node dissection (RCLND) significant ly increased from 0 of 12 in S1 SLNs to 2 of 13 in S2 SLNs and 9 of 15 in S 3 SLNs (P = 0.001; chi-square test). Like the T classification of the prima ry melanoma, the S classification proved to be a highly significant predict or for distant metastasis (P < 0.001). It turned out to be an independent f actor of influence on distant metastasis and survival in multivariate Cox a nalyses, which included tumor thickness, primary tumor site, patient gender , and patient age as covariates. CONCLUSIONS. The data presented recommend the S-staging concept as a promis ing option to fill a gap between the T and conventional N component of the pTNM classification. If its predictive capacity can be confirmed in multice nter studies, then the S classification may become the decisive criterion f or or against RCLND, and a combined T plus S staging system will help to im prove prognostically relevant stratification of melanoma patients in adjuva nt therapy trials. Cancer 2001;91:2110-20. (C) 2001 American Cancer Society .