Inferring the natural history of HPV from global cancer registries: insights from a multi-country calibration

Sep 21, 2026

Uploaded by Violet

Human papillomavirus (HPV) drives nearly all cervical cancer, yet the progression from initial infection to invasive disease is largely unobserved, limiting insight into HPV’s natural history. Modeling offers additional context across regions with differing prevalence and host immune factors such as HIV and tobacco use. This study builds calibrated models for 30 populous Sub-Saharan African countries, using country-specific demographic and behavioral inputs. It shows that identical natural history parameters cannot fit data; adding one immunocompetence parameter per country improves fit, suggesting immune variation influences cross-country differences in cervical cancer burden.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Introduction

HPV burden and natural history challenges

Modeling approaches and identifiability issues

Study objective

Explaining variation across countries

Minimal parameter requirements

[www. nature.com/scientificreports](www. nature.com/scientificreports) OPEN Inferring the natural history of HPV from global cancer registries: insights from a multi‑country calibration Robyn M. Stuart1*, Jamie A. Cohen2, Romesh G. Abeysuriya 3, Paula Sanz‑Leon4, Cliff C. Kerr2, Darcy Rao1 & Daniel J. Klein2 Human papillomavirus (HPV) is the cause of almost all cases of cervical cancer, a disease that kills some 340,000 women per year. The timeline from initial infection with HPV to the onset of invasive cervical cancer spans decades, and observational studies of this process are limited to settings in which treatment of precancerous lesions was withheld or inadequate. Such studies have been critical for understanding the natural history of HPV. Modeling can shed additional insight on the natural history of HPV, especially across geographical settings with varying prevalence offactors known to affect the host‑side immune response to HPV, such as HIV and tobacco use. In this study, we create models for the 30 most populous countries in Sub‑Saharan Africa, each with country‑specific demographic, and behavioral inputs. We found that it was not possible to fit the data if we assumed that the natural history parameters were exactly identical for all countries, even after accounting for demographic and behavioral differences, but that we could achieve a good fit with the addition of a single immunocompetence parameter for each country. Our results indicate that variation in host immune responses may play a role in explaining the differences in the burden of cervical cancer between countries, which in turn implies a greater need for more geographically diverse data collection to understand the natural history of HPV. Human papillomavirus (HPV) is an extremely common sexually transmitted infection, with an estimated lifetime incidence of around 80% in the United States1. More than 90% of infections clear on their own within 2 years, but in some cases high-risk HPV infection persists, leading to pre-cancerous lesions which may progress to invasive cancer2,3. HPV has been implicated as the causative agent in more than 90% of anal and cervical cancers, about 70% of vaginal and vulvar cancers, and more than 60% of penile cancers4. Of these, cervical cancer forms the largest burden, with more than 600,000 new cases and 342,000 deaths estimated in 20205, making it the fourth most common cancer among women globally. While the association between HPV and cervical cancer is well-established, the probability of, and time taken for, progression from HPV infection through stages of cervical intraepithelial neoplasia (CIN) and to invasive cervical cancer is unobserved. Two broad categories of natural history estimates exist: panel-based estimates and model-driven estimates. Panel-based estimates typically come from settings in which treatment of precancerous lesions was withheld for some reason6. Model-driven estimates, on the other hand, are generated when models of HPV and cervical disease are calibrated to available data, such as HPV prevalence, the age distribution of cervical cancer, and sexual behavior data7–10. Model-driven estimates of the natural history of HPV overcome several of the challenges of observational studies, e.g. they are much cheaper and simpler to conduct. However, they also have several drawbacks. Chief among these is an identifiability problem. Models of HPV and cervical disease often contain hundreds of parameters, while the dimensionality of the underlying data is much smaller. As a result, there are infinitely many parameter combinations that would replicate the observed data. In order to make it feasible to fit such a model to data, a decision must be made as to which parameters to vary. Demographic, behavioral, and programmatic parameters are clearly context-dependent, but it may be reasonable to assume natural history is not, i.e. that a 1Gender Equality Division, Bill & Melinda Gates Foundation, Sea

FAQs

Why is HPV natural history difficult to observe directly?

The probability and timing of progression from HPV infection through cervical intraepithelial neoplasia to invasive cancer are not directly observed, and observational studies are restricted by settings where precancer treatment was withheld or inadequate.

What modeling approach was used to study HPV natural history across countries?

The study calibrated HPV and cervical disease models to available data such as HPV prevalence, the age distribution of cervical cancer, and sexual behavior data, using country-specific demographic and behavioral inputs for 30 Sub-Saharan African countries.

What key finding did the study report about natural history parameters across countries?

A single set of natural history parameters identical across all countries could not fit the data, even after accounting for demographic and behavioral differences; adding a single immunocompetence parameter per country enabled a good fit.

Sep 21, 2026

Uploaded by Violet