Relation between limb toxicity and treatment outcomes after isolated limb perfusion for recurrent melanoma

Background: The optimal toxic reaction of the normal tissues in perfused limbs after isolated limb perfusion (ILP) is unknown. Theoretically, more severe limb toxicity could reflect a concomitant increased toxic effect to the tumor and improved outcomes. We determined whether there is a relation between limb toxicity and treatment outcomes after ILP for recurrent limb melanoma. Study Design: Among 252 patients with recurrent melanoma of the limbs, treatment outcomes in 192 patients (76%) with no or mild acute limb toxicity were compared with those in 60 (24%) with more severe reactions. Multivariate analysis was used to identify prognostic factors for complete response, limb recurrence–free interval, and survival. Results: Among 112 patients with measurable disease, 65 patients (58%) had a complete response and 27 (42%) experienced a relapse in the perfused limb. For complete response, uninvolved regional lymph nodes (p = 0.0025) and ILP using tumor necrosis factor-α (p = 0.0076) appeared to be favorable prognostic factors in multivariate analysis. There was no evidence of a relation between limb toxicity and complete response either in univariate (p = 0.16) or multivariate analysis (p = 0.46). For limb recurrent–free interval, only the number of lesions was a significant prognostic factor (p = 0.047); limb toxicity was not (p = 0.095). In 140 patients with recurrent melanoma excised before or at the moment of ILP, independent prognostic factors for survival were gender, the number of positive nodes, and stage of disease. There was no relation between limb toxicity and survival in either univariate (p = 0.53) or multivariate analysis (p = 0.94). Forty-eight (34%) of the 140 patients had a relapse in the perfused limb. No prognostic factors for limb recurrent–free interval could be identified; limb toxicity was not related to relapse time in univariate or multivariate analyses (p = 0.16 and p = 0.14, respectively). Conclusions: More severe acute limb toxicity is not associated with improved outcomes. One should aim at grade II toxicity (slight erythema or edema, compatible with complete recovery) at the most to increase the therapeutic ratio of ILP.