A low protein dose (73 +/- 10 micrograms total) 131I-labeled monoclonal antibody cocktail made of equal microgram quantities of 225.28S (IgG2a) and 763.24T (IgG1) murine monoclonal antibodies, which bind additively to a high molecular weight antigen of melanoma, was evaluated as a lymphoscintigraphic agent in 17 patients with intermediate to thick (mean Breslow depth, 3.39 +/- 0.64 mm) melanomas or clinical Stage II disease scheduled for nodal dissection. Eleven of the patients were clinically Stage I while 6 were clinically Stage II. 131I antibody cocktail, 258 +/- 10 microCi, was administered s.c. at the site of the primary melanoma or its scar following surgical removal. In eight patients, 63 +/- 8 microCi of 125I nonspecific normal sheep IgG was coadministered s.c. Gamma camera imaging was conducted beginning immediately after and continuing for several days following injection. Surgical resection, weighing, and gamma counting of the draining lymph nodes were undertaken in all patients. On gamma scans, early nodal uptake of antibody was most pronounced and of longest duration in the tumor pathologically positive patients (5 of 7 had visible nodal uptake, 4 of 7 visually stable or rising with time), with the t 1/2 of nodal clearance by gamma scan significantly (P less than 0.05) longer than in the negative patients in whom 4 of 10 showed some, although generally transient (0 of 10 stable or rising), nodal uptake. Scans were not easily interpretable when the injection site was very near the draining nodal group, in part due to the detection of scatter activity from the injection site. In several instances the scan was correct and the clinical examination was incorrect as regards nodal disease. Quantitative analysis of the surgically excised draining nodes showed significantly (P less than 0.001) more 131I anti-melanoma antibody uptake in the 21 tumor-involved nodes [0.01217% injected dose (ID)/node median] than in the 512 tumor-negative nodes (0.00051% ID/node median). Median percentage ID/g of anti-melanoma antibody in tumor-involved nodes was significantly greater (P less than 0.01) than in tumor-negative nodes (0.01984 versus 0.003215% ID/g). 125I-labeled nonspecific antibody did not accumulate significantly more in the tumor-involved nodes on a per node or per g basis in the 283 of 533 nodes studied using the dual-label approach (0.0036 versus 0.00092% ID/g). These data demonstrate that by external imaging and by tissue counting that a radiolabeled anti-melanoma monoclonal antibody cocktail can specifically accumulate to melanoma-involved lymph nodes following s.c. administration.(ABSTRACT TRUNCATED AT 400 WORDS)