Recurrence in patients with penile carcinoma occurs in about one third of cases, usually due to insufficient surgery or positive resection margins. An evaluation of surgical resection margins in penectomy specimens was performed to determine precise anatomic sites of tumor involvement, hoping to advance knowledge concerning the local routes of spread of penile carcinomas. A pathologic study of 80 partial penectomies revealed 14 positive margins. Margins were examined after their separation from the main specimen as follows: 1) proximal urethra and surrounding tissues consisting of urethral epithelium with Litree glands, lamina propria, corpus spongiosum, and penile fascia (periurethral cylinder); 2) proximal shaft with corresponding corpora cavernosa separated and surrounded by the tunica albuginea and penile fascia; and 3) skin of shaft with underlying corporal dartos. In 9 patients, only one site was involved by carcinoma, and in 5 there were multiple contiguous sites (for a total of 20 anatomic sites). The distribution of the various sites involved by carcinoma was as follows: urethral epithelium, 4 cases (2 in situ and 2 invasive carcinomas including intraluminal spread); lamina propria, 5 cases; corpus spongiosum, 3 cases; penile fascia, 6 cases; and corpora cavernosa and skin, 1 case each. One of the in situ lesions was discontinuous with the main glans tumor, and the other one was continuous with it. The penile fascia was the most commonly involved site followed by the urethral lamina propria and epithelium. Dissemination to outer skin, corpora cavernosa, and corpus spongiosum was less frequent. The highly vascularized and innervated loose connective tissue of the penile fascia appears to facilitate tumor spread. The urethra is either a pathway for in situ tumor progression from glans to urethra or part of a field prone to malignant transformation. The infrequent involvement of corpora cavernosa is probably due to the tunica albuginea acting as a barrier preventing tumor spread. Based on these observations and the examination of hundreds of penectomy specimens, we are proposing five probable routes of local spread for penile cancer: 1) horizontal and superficially spreading from one epithelial mucosal compartment (glans, coronal sulcus, and foreskin) to the other; 2) following the penile fascia; 3) through spaces created by feeding vessels in the tunica albuginea; 4) vertical spreading involving step-by-step different penile anatomic compartments; and 5) along the urethral epithelium.

Difficulty in foreskin retraction and phimosis are risk factors for penile carcinoma that may be related to the anatomically variable length of the foreskin. This observation has stimulated us to postulate the hypothesis that foreskin length is related to penile cancer. To compare the foreskin in the general population and patients with penile cancer, an anatomic classification of foreskin was designed. We examined the foreskin of 215 uncircumcised males without cancer (age range 15-93 years) and the foreskin of 23 patients with cancer (age range 31-90 years). Foreskin types were classified as long (with the preputial orifice located beyond glans meatus and entirely covering the glans), medium (with the preputial orifice located between meatus and glans corona), and short (with the preputial orifice located between corona and coronal sulcus). Phimosis was defined as a nonretractable prepuce of the long type. We found that 77% of noncancer population cases had long foreskin and that only 7% of these cases were phimotic. Cancer patients showed long foreskin in 78% of the cases, and phimosis was significantly frequent in this group (52%) as compared with the other (p <0.001). Coexistence of a long foreskin and phimosis may explain the high incidence of penile cancer in some geographic regions. To better document these findings, a comparison of foreskin types in countries with high and low incidence of penile cancer will be interesting. However, because phimosis appears to be a major factor, the presence of long foreskin may be a necessary but not a sufficient condition for cancer development. For these reasons we support preventive circumcision in patients with long and phimotic foreskins living in high-risk areas. Cancers not related to long foreskins and phimosis may be causally different.