Histologic grade has been reported as an important pathologic parameter predictive of nodal metastases and outcome in patients with penile squamous cell carcinoma. There is no consensus about the criteria for grading and the proportion of anaplastic cells required to classify a tumor as high grade. The incidence and management of heterogeneous tumors (tumors harboring more than 1 histologic grade) are not well established. The purposes of this study were to present a grading model for penile cancer, to test the practicality of the system by comparing a visual ("naked-eye") estimation of the proportions of grades with a digitally guided measuring system and to determine the influence on nodal metastasis of the various proportions of grades. A total of 117 penectomy and circumcision specimens with bilateral inguinal lymph node dissections were studied and 62 heterogeneous tumors were identified (53%). The following steps were taken: (1) design of a grading system model; (2) determination of proportions of histologic grades by naked-eye evaluation and by digital measurement; (3) evaluation of metastasis according to proportions of grades; (4) determination of the influence of site of grade 3 in nodal metastasis; (5) design of a Gleason-like scoring system; and (6) statistical evaluation. We designed a 3-tier grading system. Grade 1: well-differentiated cells, almost undistinguishable from normal squamous cells except for the presence of minimal basal/parabasal cell atypia. Grade 3: tumors predominantly composed of anaplastic cells. Grade 2: all tumors not fitting into criteria described for grade 1 or 3. A visual and digital-based (slides scanned and the corresponding areas measured with an image-editing software) proportions of grades were estimated and the metastatic rate between them were confronted using different proportions of grade 3. To evaluate the influence of site of grade 3 on nodal metastasis, we selected 20 heterogeneous tumors. We established 3 sites: superficial, at or within the main tumor and deep at front of invasion. There was no significant difference between the visual estimation and the digital measurement systems. Heterogeneous tumors comprised about half of penile squamous cell carcinomas. The majority of the heterogeneous tumors were composed of a combination of grades 2 and 3 (68%). No statistical differences were noted in the incidence of nodal metastasis when comparing tumors with various proportions of anaplastic cells (P>0.10 in all cases). Metastatic rate was significantly more frequent in tumors harboring any proportion of grade 3 as compared with tumors without grade 3 (58% vs. 14%, P=0.04). Any proportion of grade 3 was equally associated with a significant risk of nodal metastasis. A Gleason-like system showed a correlation of higher scores and rate of nodal metastasis. No predictive advantage was found when comparing the Gleason-like model with the proposed 3-tier grading system. The proposed grading system emphasized both ends of the differentiation spectrum and was based on easily recognizable morphologic criteria. When histologically evaluating penile carcinomas, we recommend a careful search of areas of grade 3. Any focus of grade 3 should be sufficient to grade the neoplasm as a high-grade tumor.