To date, no studies have evaluated adenocarcinoma of the rete testis statistically, because reports have been limited to single cases or series of 2 cases only. Univariate and multivariate analyses on disease-free survival have been performed after combining all data available in the literature with our own. Information about disease-free survival has been collected in 38 patients. As many as 40% of them died within the first year of diagnosis. Three and 5-year disease-free survival was 49% and 13%, respectively. We have not detected any difference in survival between age groups or side of the lesion. Similarly, statistical difference cannot be proved between survival of tumors with nodular infiltrating or cystic growth pattern, although it has been suggested that these two varieties represent different tumor types from a gross and microscopic morphologic point of view. Univariate analysis reveals that tumor stage, tumor size, and therapy may have an influence on survival. Tumors that are organ-confined and small lesions (testicular mass < 5 cm in maximum diameter) behave definitely better than those disseminated at diagnosis or of a bigger size. Surprisingly, tumor size is not associated with tumor stage or histologic growth pattern (nodular infiltrating versus predominantly cystic). With regard to therapy, cases in which RPLND has been performed as part of the therapy behave better in univariate analysis, while patients who receive radiation do worse. Most probably these facts reflect that patients with clearly advanced disease where local control cannot be achieved by surgery tend to undergo palliative treatment by radiation. On the other hand, RPLND tends to be performed in patients in whom there is no evidence of distant spread. Therefore, it would really be the primary stage that would set the prognosis rather than the consequent treatment, and neither radiation therapy nor RPLND would be true independent variables. Similarly, no significant difference is observed when patients receive chemotherapy. Cox's regression analysis reveals size of the testicular tumor as the only independent predictor of survival. Stage at diagnosis does not have an influence by itself and neither does any form of therapy. In this sense, the negative effect of radiotherapy is eliminated. We are aware that the results drawn from a literature review are far from ideal, but there is not enough evidence to suggest an optimal sequence of treatment for this rare malignancy. To date, no effective chemotherapy has been found. Whenever the tumor is resectable, there appears to be merit in an RPLND.(ABSTRACT TRUNCATED AT 400 WORDS)