We demonstrate the development and application of a prototype hybrid microscopy system integrating autofluorescence (AF) and photoacoustic (PA) label-free contrast modes, for the differentiation of ocular tumors in human surgical biopsies. Hybrid imaging was performed in conjunctival nevi and uveal melanomas tissue sections to acquire quantified data for each molecular background. The AF and PA signals were spatially correlated to establish a novel malignancy indicator that could detect melanomas with high accuracy (t-test; p<0.01). The proposed methodology has the potential to simplify relevant diagnostic procedures and paves the way for the development of novel ophthalmoscopes aiming to the early diagnosis of ocular malignancies in a clinical setting.