Combining routine morphology, p16INK4a immunohistochemistry, and in situ hybridization for the detection of human papillomavirus infection in penile carcinomas: A tissue microarray study using classifier performance analyses

Objectives ion by high-risk human papillomavirus (HR-HPV) plays an important role in the pathogenesis of penile cancer in approximately 50% of the patients. The gold standard for human papillomavirus (HPV) detection is the polymerase chain reaction (PCR) assay. However, technical requirements and associated costs preclude the worldwide use of PCR assays on a routine basis. Herein, we evaluated the predictive abilities of tumor morphology, immunohistochemistry for p16INK4a expression, and in situ hybridization (ISH) for HR-HPV detection in defining HPV status, as established by PCR. als and methods samples from 48 patients with HPV-positive penile squamous cell carcinoma (SCC) were included in 4 tissue microarrays (TMA). s ivities and specificities were as follows: tumor morphology, 70% and 68%; p16INK4a immunohistochemistry, 65% and 90%; HR-HPV ISH, 47% and 100%. Regarding combinations of the predictors, the best performance was seen when HR-HPV ISH and p16INK4a immunohistochemistry were combined, regardless of the tumor morphology: sensitivity, 88%; specificity, 64%; area under the receiver-operating characteristic (AUC) curve, 0.83. Combinations of tumor morphology with p16INK4a immunohistochemistry or with HR-HPV ISH performed similarly well. sions ile SCC, both p16INK4a immunohistochemistry and ISH for HR-HPV increase the predictive ability of routine morphology in defining HPV status. These tests can be interpreted differentially, depending on the necessity of a higher sensitivity or a higher specificity. For research/screening studies, we recommend combining tumor morphology, p16INK4a immunohistochemistry, and HR-HPV ISH. To increase sensitivity, positivity in any of these predictors should be considered as indicative of HPV infection. For routine diagnosis of clinical cases, criteria should be more stringent, and, to achieve the highest specificity in classifying a case as HPV-related, all predictors should be consistently positive. The data generated in the present study could be used in algorithms for defining HPV status in penile carcinomas.