Conclusions about human papillomavirus-related malignancies.

Although cancer death rates and overall cancer incidence in the United States have recently declined, human papillomavirus (HPV)-associated cancers of the oropharynx and anus have the ominous distinction of increasing in incidence. The article by Skinner et al therefore addresses an important topic — what is the relationship between multiple HPV-related malignancies in a single individual? This question has previously been addressed using epidemiological approaches, but not in a comprehensive chart review. Skinner and Beadle present a retrospective review of patients treated at their institution over a 60-year period starting in 1949 with the intent of understanding which patients are at risk for HPV-related malignancies, specifically, are patients with one HPV-related malignancy more likely to have additional HPV-related malignancies. It is logical to assume that HPV exposure may predispose to a phenotype of multiple cancers in susceptible organs. Using administrative data, others have shown that patients with a history of cervical cancer are at risk for a head and neck cancer primary (1.7 incidence ratio). Conversely, patients with oropharyngeal cancer are at increased risk for cervical cancer. Skinner et al have investigated additional tumor sites and determined latency and associated clinical characteristics. Among other findings, the authors identify significant differences in latency periods between first and second HPV-related cancers depending on the initial cancer site. Related to methodological problems, the authors acknowledge that the results “should be interpreted as hypothesis-generating” and recognize some other deficiencies in the discussion; however, these shortcomings and potential interpretations warrant additional comment. Although this article addresses an important topic, there are several methodological issues that should be recognized in the data collection. The head and neck cancer population included all sites rather than the oropharynx and, as a result, probably significantly overestimates the number of HPV-related cancers. Elimination of patients with disease outside the oropharynx or preferably some method of HPV testing tissue (p16 immunohistochemistry or HPV in situ hybridization) of archival tumor specimens, would significantly improve identification of HPVrelated malignancies. It is fair to say that many patients with oropharyngeal disease get primary medical therapy and, as a result, there is often limited tissue available for laboratory analysis. However, limiting the study to patients with oropharyngeal disease and eliminating patients with laryngeal or oral cavity disease, especially over the time frame of the study, would be very appropriate although it would also shrink the overall study numbers. Because HPV-related head and neck malignancies have trended upward over time, it is likely that, over the 60-year study period, the incidence of HPV-related oropharyngeal cancer has significantly increased just as the rates of tobacco-related disease have decreased. This may explain the impact of tobacco on the latency in their study population. Furthermore, as recognized by the authors, because cervical cancer is common, the study population remains largely female whereas head and neck cancer and HPV-associated head and neck squamous cell carcinoma (HNSCC) is a predominately male disease, particularly over the time period studied. Unfortunately, these deficiencies significantly impact the authors’ stated goals to “determine the patterns of development, clinical characteristics, and latency” of multiple HPV-associated cancers and to provide “clinical information . . . important to counsel newly diagnosed patients.” Because HPV-status of HNSCCs included in the study is not known (and includes non-HPV-related HNSCC sites), we feel that the data are too immature to be used for counseling of newly diagnosed patients. In addition, conclusions related to latency between tumors, as well as associations of multiple potentially HPV-related cancers with sex, marital status, age at initial diagnosis, site of initial diagnosis, and smoking status should be interpreted cautiously. Finally, comparisons between HPV-related malignancies should be taken in the context of the differences between these clinical entities. HPV-associated malignancies of the oropharynx and genital regions have several molecular similarities suggesting shared pathogenesis once cells are infected with HPV; however, there are several key molecular, clinical, response, demographic, and epidemiologic differences suggesting that HPV-mediated cancer of the cervix and oropharynx are phenotypically distinct. For example, head and neck cancer has no identifiable intermediate stage and more likely to metastasize *Corresponding author: E. L. Rosenthal, Division of Otolaryngology, BDB Suite 563, 1808 7th Avenue South, Birmingham, AL 35294-0012. E-mail: oto@uab.edu