PAVE Resources

Design of the HPV-automated visual evaluation (PAVE) study: Validating a novel cervical screening strategy abstract:

Background: The HPV-automated visual evaluation (PAVE) Study is an extensive, multinational initiative designed to advance cervical cancer prevention in resource-constrained regions. Cervical cancer disproportionally affects regions with limited access to preventive measures. PAVE aims to assess a novel screening-triage-treatment strategy integrating self-sampled HPV testing, deep-learning-based automated visual evaluation (AVE), and targeted therapies.

Methods: Phase 1 efficacy involves screening up to 100,000 women aged 25–49 across nine countries, using self-collected vaginal samples for hierarchical HPV evaluation: HPV16, else HPV18/45, else HPV31/33/35/52/58, else HPV39/51/56/59/68 else negative. HPV-positive individuals undergo further evaluation, including pelvic exams, cervical imaging, and biopsies. AVE algorithms analyze images, assigning risk scores for precancer, validated against histologic high-grade precancer. Phase 1, however, does not integrate AVE results into patient management, contrasting them with local standard care.

Phase 2 effectiveness focuses on deploying AVE software and HPV genotype data in real-time clinical decision-making, evaluating feasibility, acceptability, cost-effectiveness, and health communication of the PAVE strategy in practice.

Results: Currently, sites have commenced fieldwork, and conclusive results are pending. 

Conclusions: The study aspires to validate a screen-triage- treat protocol utilizing innovative biomarkers to deliver an accurate, feasible, and cost-effective strategy for cervical cancer prevention in resource-limited areas. Should the study validate PAVE, its broader implementation could be recommended, potentially expanding cervical cancer prevention worldwide.

Funding: The consortial sites are responsible for their own study costs. Research equipment and supplies, and the NCI-affiliated staff are funded by the National Cancer Institute Intramural Research Program including supplemental funding from the Cancer Cures Moonshot Initiative. No commercial support was obtained. Brian Befano was supported by NCI/ NIH under Grant T32CA09168.

Generalizable deep neural networks for image quality classification of cervical images abstract:

Successful translation of artificial intelligence (AI) models into clinical practice, across clinical domains, is frequently hindered by the lack of image quality control. Diagnostic models are often trained on images with no denotation of image quality in the training data; this, in turn, can lead to misclassifications by these models when implemented in the clinical setting. In the case of cervical images, quality classification is a crucial task to ensure accurate detection of precancerous lesions or cancer; this is true for both gynecologic-oncologists’ (manual) and diagnostic AI models’ (automated) predictions. Factors that impact the quality of a cervical image include but are not limited to blur, poor focus, poor light, noise, obscured view of the cervix due to mucus and/or blood, improper position, and over- and/or under-exposure. Utilizing a multi-level image quality ground truth denoted by providers, we generated an image quality classifier following a multi-stage model selection process that investigated several key design choices on a multi-heterogenous “SEED” dataset of 40,534 images. We subsequently validated the best model on an external dataset (“EXT”), comprising 1,340 images captured using a different device and acquired in different geographies from “SEED”. We assessed the relative impact of various axes of data heterogeneity, including device, geography, and ground-truth rater on model performance. Our best performing model achieved an area under the receiver operating characteristics curve (AUROC) of 0.92 (low quality, LQ vs. rest) and 0.93 (high quality, HQ vs. rest), and a minimal total %extreme misclassification (%EM) of 2.8% on the internal validation set. Our model also generalized well externally, achieving corresponding AUROCs of 0.83 and 0.82, and %EM of 3.9% when tested out-of-the-box on the external validation (“EXT”) set. Additionally, our model was geography agnostic with no meaningful difference in performance across geographies, did not exhibit catastrophic forgetting upon retraining with new data, and mimicked the overall/average ground truth rater behavior well. Our work represents one of the first efforts at generating and externally validating an image quality classifier across multiple axes of data heterogeneity to aid in visual diagnosis of cervical precancer and cancer. We hope that this will motivate the accompaniment of adequate guardrails for AI-based pipelines to account for image quality and generalizability concerns.

Initial evaluation of a new cervical screening strategy combining human papillomavirus genotyping and automated visual evaluation: the Human Papillomavirus-Automated Visual Evaluation Consortium abstract:

The HPV-Automated Visual Evaluation Consortium is validating a cervical screening strategy enabling accurate cervical screening in resource-limited settings. A rapid, low-cost human papillomavirus (HPV) assay permits sensitive HPV testing of self-collected vaginal specimens; HPV-negative women are reassured. Triage of positive participants combines HPV genotyping (4 groups in order of cancer risk) and visual inspection assisted by automated cervical visual evaluation that classifies cervical appearance as severe, indeterminate, or normal. Together, the combination predicts which women have precancer, permitting targeted management to those most needing treatment. We analyzed CIN3+ yield for each HPV-Automated Visual Evaluation risk level (HPV genotype crossed by automated cervical visual evaluation classification) from 9 clinical sites (Brazil, Cambodia, Dominican Republic, El Salvador, Eswatini, Honduras, Malawi, Nigeria, and Tanzania). Data from 1832 HPV-positive participants confirmed that HPV genotype and automated cervical visual evaluation classification strongly and independently predict risk of histologic CIN3+. The combination of these low-cost tests provided excellent risk stratification, warranting pre-implementation demonstration projects.