About

Margaret M. Madeleine, MPH, PhD, is a Research Associate Professor in the Department of Epidemiology at the University of Washington's School of Public Health and an Associate Member of the Program in Epidemiology at the Fred Hutchinson Cancer Research Center. She received her MPH in Epidemiology from Yale University in 1991 and her PhD from the University of Washington in 1996. Her research focuses on the molecular epidemiology of pathogens and the immune response to them that may be associated with cancer development and prognosis. She is particularly interested in the common drivers behind chronic subclinical inflammation, such as age, viral infection, and UV exposure, and their roles in cancer. Her current projects include post-licensure trials of the HPV vaccine and studying the uptake of HPV vaccination in Washington State. Madeleine has been a key member of the Seattle HPV research group and has contributed to understanding the epidemiology of HPV-related cancers, inflammation, and immune genetics in cancer prevention.

Research topics

• Medicine
• Internal medicine
• Bioinformatics
• Oncology
• Biology
• Pathology
• Dermatology
• Biochemistry
• Microbiology
• Genetics

Selected publications

• Nonavalent <scp>HPV</scp> vaccine to prevent recurrent anal or vulvar high‐grade squamous intraepithelial lesions ( <scp>VIVA</scp> trial): A randomized, double‐blind, placebo‐controlled trial

  International Journal of Cancer · 2026-02-13

  articleOpen accessSenior author

  The nonavalent human papillomavirus (9vHPV) vaccine protects against HPV infection and high-grade squamous intraepithelial lesions (HSIL) when administered prior to exposure, but evidence supporting its potential therapeutic benefit has been inconsistent. This randomized, double-blind, placebo-controlled trial evaluated whether the 9vHPV vaccine reduces recurrence of HSIL or HPV persistence in 27-69-year-old persons previously treated for anal or vulvar HSIL. Participants were HSIL-free at enrollment and received 9vHPV or placebo at months 0, 2, and 6. High-resolution anoscopy or vulvoscopy was performed at months 18 and 36, and anal or vulvar swabs were collected at months 0, 18, 24, and 36 for HPV DNA detection. The primary endpoints were HSIL recurrence and HPV persistence (≥2 consecutive positive swabs for the same 9vHPV-type). Of 185 participants included in the intent-to-treat analysis, 91 received vaccine and 94 received placebo. The DSMB recommended early termination for futility. The 9vHPV vaccine was not more effective than placebo in preventing recurrent HSIL, with 16 HSIL among 9vHPV recipients versus 21 HSIL in placebo recipients (incidence 8.1 vs. 10.1/100 person-years; p = .54). HPV persistence was 21% in vaccine versus 31% in placebo recipients (p = .20). The 9vHPV vaccine delivered after treatment of anal or vulvar HSIL did not reduce HSIL recurrence or HPV detection. Our study underlines the importance of HPV vaccine administration prior to HPV exposure and the need for novel treatments for HSIL with high recurrence potential.

  PublisherOA PDFDOI

• Population-level cancer trends among solid organ transplant recipients in the United States during 1995-2021

  American Journal of Transplantation · 2026-03-01

  articleOpen access

  Solid organ transplant recipients (SOTRs) experience elevated cancer risk from immunosuppression and underlying medical conditions. Medical management has improved over time, and SOTRs are living longer. We used registry data covering 693,718 SOTRs in the United States (US) to evaluate population-level cancer trends during 1995-2021. Compared with SOTRs in 1995-2003, those in 2013-2021 were living at older ages and were followed at a longer time since their transplant. Based on 65,081 cancers, cancer incidence in SOTRs was higher during 2013-2021 than 1995-2003 (unadjusted incidence rate ratio [IRR], 1.29; 95% confidence interval [95% CI], 1.26-1.32). However, cancer incidence was lower in 2013-2021 after adjustment for age (IRR, 0.93; 95% CI, 0.91-0.95) and multivariable adjustment (0.93, 0.90-0.96). Results for the 6 most common cancer types showed varying trends during 1995-2021. Overall cancer incidence was higher in SOTRs than in the US general population during 1995-2021 and, most recently, in 2013-2021 (standardized incidence ratio, 1.66; 95% CI, 1.64-1.67). In conclusion, after accounting for age, there was an encouraging decline in cancer incidence among US SOTRs during 1995-2021. However, incidence remained elevated compared with the general population in 2013-2021. Measures are needed to reduce the cancer burden as SOTRs live longer after transplantation and the population ages.

  PublisherDOI

• Generalizable deep neural networks for image quality classification of cervical images

  Scientific Reports · 2025-02-21 · 5 citations

  articleOpen access

  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.

  PublisherOA PDFDOI

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

  JNCI Journal of the National Cancer Institute · 2025-03-03 · 8 citations

  articleOpen access

  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.

  PublisherOA PDFDOI

• ULACNet-301, OPTIMO protocol: optimizing HPV vaccination regimen for cancer prevention in children and adolescents living with HIV

  BMC Cancer · 2025-01-27

  articleOpen access

  BACKGROUND: Persistent infection with human papillomavirus (HPV) is associated with most cervical and anal cancer cases and a large fraction of other anogenital and oropharyngeal cancers. The prophylactic HPV vaccines are known to prevent HPV infections and HPV-associated disease, although there is evidence of reduced response to the HPV vaccination among individuals living with HIV. Prior studies among individuals without HIV suggest that a single HPV vaccine dose induces humoral immune responses that, while lower than those induced by two or three doses, still confer protection against HPV infection. Current recommendations for HPV vaccine include a single-dose schedule for children 9-14-years-olds without HIV. Although two to three doses are recommended for children living with HIV (CLWH), there is very limited data comparing responses to one vs. 2-3 doses in CLWH. METHODS: The OPTIMO study will compare immune responses to HPV vaccination in CLWH by measuring antibody and memory B cell (Bmem) responses after 1, 2, or 3 doses of the 9-valent HPV (9vHPV) vaccine, Gardasil-9. A comparison group of children without HIV will receive one dose of the vaccine. The durability of the response will be assessed at 24 months after the last dose of a given regimen. The OPTIMO trial will take place among CLWH from low and middle-income country (LMIC) settings in Peru, Brazil, and Haiti. DISCUSSION: Previous studies of single-dose regimens in individuals without HIV raise questions about whether one dose would suffice for CLWH and, if not, whether two or three doses are needed to provide protection against HPV-related cancers. These questions have operational consequences in LMICs given the barriers to delivering multiple doses, uneven availability, and intermittent shortages of HPV vaccines. In addition, information on HIV status for children and adolescents is rarely available during vaccination campaigns based in schools or public health clinics, so CLWH may receive a single dose despite policy recommendations that they receive two or three. This study will provide evidence on the optimal number of doses needed for CLWH that can inform HPV vaccination campaigns in LMICs, especially those with a higher burden of HIV infection and higher incidence of HPV-related cancers. TRIAL REGISTRATION: ClinicalTrials.gov NCT04265950.

  PublisherOA PDFDOI

• Generalizable deep neural networks for image quality classification of cervical images

  Research Square · 2024-01-18 · 1 citations

  preprintOpen access
  PublisherOA PDFDOI

• Tables S1-3 from Risk of Second Malignancies in Solid Organ Transplant Recipients Who Develop Keratinocyte Cancers

  2023-03-31

  preprintOpen access

  &lt;p&gt;This file includes Supplementary tables describing 1) Concordance of Medicare claims and SRTR malignancy reports, 2011-2013; 2) SEER diagnosis codes and groupings; and 3) Hazard ratios for selected SEER diagnoses by diagnosis of KC among 118,440 solid organ transplant recipients.&lt;/p&gt;

  PublisherDOI

• Data from Investigation of Epstein–Barr Virus as a Potential Cause of B-Cell Non-Hodgkin Lymphoma in a Prospective Cohort

  2023-03-31

  preprintOpen access

  &lt;div&gt;Abstract&lt;p&gt;&lt;b&gt;Background:&lt;/b&gt; We hypothesized that poor control of Epstein–Barr virus (EBV) infection, leading to reactivation of the virus, increases the risk of non-Hodgkin lymphoma (NHL) in the general population of primarily immunocompetent persons.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Methods:&lt;/b&gt; We conducted a case–control study nested within the Women's Health Initiative Observational Study cohort in which we measured antibodies to EBV antigens [immunoglobulin G (IgG) to viral capsid antigen (VCA), nuclear antigen (EBNA1), and early antigen-diffuse (EA-D)] and EBV DNA load in prediagnostic samples of 491 B-cell NHL cases and 491 controls.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Results:&lt;/b&gt; We found no association with established EBV infection, based on seropositivity for VCA. Seropositivity for EBNA1 was associated with decreased risk of B-cell NHL, overall [OR = 0.5; 95% confidence interval (CI), 0.3–0.8] and for each of the histologic subtypes examined. Increased risk of chronic lymphocytic leukemia (CLL) and related subtypes was observed with higher levels of EBV DNA and antibody to EA-D, both markers reflective of reactivation. These associations were strongest for cases with the shortest time interval between blood draw and diagnosis.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Conclusions:&lt;/b&gt; In balance, these results do not provide strong evidence of EBV playing a causal role in B-cell NHL in general population women. The associations we observed may reflect increased risk of NHL with underlying immune impairment or could be due to reverse causation.&lt;/p&gt;&lt;p&gt;&lt;b&gt;Impact:&lt;/b&gt; Further characterization of the subtype-specific association with CLL is warranted. Exclusion of cases with preclinical disease markers (such as monoclonal B-lymphocytosis for CLL) may help rule out reverse causation in future studies. &lt;i&gt;Cancer Epidemiol Biomarkers Prev; 22(10); 1747–55. ©2013 AACR&lt;/i&gt;.&lt;/p&gt;&lt;/div&gt;

  PublisherDOI

• Data Supplement from Downregulation of MHC-I Expression Is Prevalent but Reversible in Merkel Cell Carcinoma

  2023-04-03

  preprintOpen access

  &lt;p&gt;Allred proportion (ranging from 0-5) and intensity (ranging from 0-3) scores for each tumor. The median score among triplicate TMA cores and reviewers was utilized. Most MCC tumors with an Allred score of 7 continued to express MHC class I in all tumor cells, however the intensity of expression was significantly reduced as compared to those with an Allred score of 8.&lt;/p&gt;

  PublisherDOI

• Using Haplotype-Based Artificial Intelligence to Evaluate SARS-CoV-2 Novel Variants and Mutations

  JAMA Network Open · 2023-02-21 · 3 citations

  articleOpen access

  Importance: Earlier detection of emerging novel SARS-COV-2 variants is important for public health surveillance of potential viral threats and for earlier prevention research. Artificial intelligence may facilitate early detection of SARS-CoV2 emerging novel variants based on variant-specific mutation haplotypes and, in turn, be associated with enhanced implementation of risk-stratified public health prevention strategies. Objective: To develop a haplotype-based artificial intelligence (HAI) model for identifying novel variants, including mixture variants (MVs) of known variants and new variants with novel mutations. Design, Setting, and Participants: This cross-sectional study used serially observed viral genomic sequences globally (prior to March 14, 2022) to train and validate the HAI model and used it to identify variants arising from a prospective set of viruses from March 15 to May 18, 2022. Main Outcomes and Measures: Viral sequences, collection dates, and locations were subjected to statistical learning analysis to estimate variant-specific core mutations and haplotype frequencies, which were then used to construct an HAI model to identify novel variants. Results: Through training on more than 5 million viral sequences, an HAI model was built, and its identification performance was validated on an independent validation set of more than 5 million viruses. Its identification performance was assessed on a prospective set of 344 901 viruses. In addition to achieving an accuracy of 92.8% (95% CI within 0.1%), the HAI model identified 4 Omicron MVs (Omicron-Alpha, Omicron-Delta, Omicron-Epsilon, and Omicron-Zeta), 2 Delta MVs (Delta-Kappa and Delta-Zeta), and 1 Alpha-Epsilon MV, among which Omicron-Epsilon MVs were most frequent (609/657 MVs [92.7%]). Furthermore, the HAI model found that 1699 Omicron viruses had unidentifiable variants given that these variants acquired novel mutations. Lastly, 524 variant-unassigned and variant-unidentifiable viruses carried 16 novel mutations, 8 of which were increasing in prevalence percentages as of May 2022. Conclusions and Relevance: In this cross-sectional study, an HAI model found SARS-COV-2 viruses with MV or novel mutations in the global population, which may require closer examination and monitoring. These results suggest that HAI may complement phylogenic variant assignment, providing additional insights into emerging novel variants in the population.

  PublisherOA PDFDOI

Recent grants

• Administrative and Coordinating Core

  NIH · $20.3M · 2019–2026

• NIH Grant R01CA116786

  NIH · $1.9M · 2012

• Post GWA Analysis of Inflammation Pathways in Barrett's and Esophageal Cancer

  NIH · $484k · 2014–2017

• Novel Pathogen Associated Cancers (PQ12)

  NIH · $2.6M · 2012–2017

• NIH Grant P01CA042792

  NIH · $42.7M · 2013

Frequent coauthors

• Eric A. Engels

  National Cancer Institute

  207 shared

• Charles F. Lynch

  University of Iowa

  180 shared

• Brenda Y. Hernandez

  University of Hawaii System

  172 shared

• Jack L. Finch

  172 shared

• Karen Pawlish

  New Jersey Department of Health

  171 shared

• Joseph F. Fraumeni

  Division of Cancer Epidemiology and Genetics

  171 shared

• Christina A. Clarke

  Grail (United States)

  171 shared

• Zaria Tatalovich

  National Cancer Institute

  169 shared

Labs

• Cancer EpidemiologyPI

Education

• PhD, Epidemiology

  University of Washington

  1996

• MPH, Epidemiology

  Yale University

  1991

Awards & honors

• Research Team Honored for Innovative Science to Advance Canc…