HETEROGENEITY IN MARKOV MODELS OF THE NATURAL HISTORY OF DISEASE: INVESTIGATING ONCOGENICITY

   Purpose: Uptake of the human papillomavirus (HPV) vaccine in the United States will decrease cervical cancer rates by lowering infection rates of highly oncogenic genotypes. The purpose of this research is to investigate the heterogeneity of risk in cervical cancer based on underlying HPV infection genotype.

   Methods: We constructed a microsimulation model to investigate the natural history of cervical cancer from first infection with a particular HPV type to cancer death. We simulated cohorts of women defined by age at infection and type of infection (HPV-16, HPV-18, HPV-31, HPV-52, and other high-risk types). For the purposes of this model, it was assumed that a woman would not develop a secondary infection during her lifetime. Parameters were based on a review of the literature and included type-specific time-varying transition probabilities for progression and regression of the initial HPV infection and low-grade cervical intraepithelial neoplasia. High-grade neoplasia and cancer were modeled according to established best practices.

   Results: For women first infected with an HPV strain at age 18, risk of cancer death was 0.4% in those infected with HPV-16, twice as high as those infected with other high-risk strains (0.2%). At age 15, the risk of cancer death was 0.3%, 0.2%, 0.2%, 0.2%, and 0.1% for HPV-16, -18, -31, -52, and other high-risk genotypes, respectively. Life expectancy estimates varied by as much as 0.77 years between different genotypes for women infected at age 21.

   Conclusions: Heterogeneity of risk due to underlying HPV infection genotype can have significant ramifications for risk and life expectancy in cervical cancer modeling. As vaccination changes the underlying distribution of HPV genotypes in the US population, models should separately consider each strain of HPV in order to accurately reflect the associated risks.