Cutaneous leishmaniases present similar clinical appearances, but differing prognosis in the course of infection. Ulcers caused by parasites of the subgenus Viannia are more aggressive than ulcers caused by parasites of the subgenus Leishmania. Another problem is distinguishing between true Leishmania infection and other skin diseases in endemic areas, where cutaneous lesions and a single positive Montenegro intradermal test are enough to submit patients to specific treatment for cutaneous leishmaniasis. This study evaluated the efficacy of PCR in detecting in Leishmania in patients with cutaneous lesions. Leishmania (V.) braziliensis complex was determined by a primer pair from the multicopy spliced leader RNA. The results were compared to those of traditional methods. We analyzed biopsies of 109 patients with cutaneous lesions in the second most endemic region of Sao Paulo State, Brazil. Definitive diagnosis was established by clinical and "consensus laboratory criteria" (positive culture, stained tissue smears or PCR). Of 52 patients with cutaneous leishmaniasis, 96% had positive PCR, 69%, positive parasitological tests and 100%, positive Montenegro intradermal tests. Histopathological examination (only in 32 samples) were positive in 14 samples, suggestive in 14 and negative in 4 samples. All 57 patients with other etiologies had negative results in parasitological methods, PCR and histopathological examination (in 39 samples), but Montenegro intradermal tests were positive in 35%. PCR was highly sensitive and specific for L. (V.) braziliensis complex detection compared with other laboratory methods. Despite the specificity of the parasitological tests, the sensitivity was less than 70%. Montenegro intradermal reaction was highly sensitive, but with low specificity, only 65%. As suggestive results in histopathological examinations were shown in 14 samples, it was difficult to determine the true result. PCR applied to biopsies proved to be useful for differential diagnosis of cutaneous lesions of other etiologies in patients living in endemic areas. The advantages are most striking in clinical specimens with scarce amastigotes for which conventional methods have low sensitivity and should be considered for clinical and epidemiological patterns. On the other hand, both Montenegro intradermal test and parasitological methods are only modestly effective in cutaneous leishmaniasis diagnosis.