Challenges in risk estimation using routinely collected clinical data: The example of estimating cervical cancer risks from electronic health-records

Untreated cervical intraepithelial neoplasia grade 2 and subsequent risk of cervical cancer: population based cohort study

OBJECTIVE

To describe the long term risk of cervical cancer in women with untreated (that is, undergoing active surveillance) or immediately treated cervical intraepithelial neoplasia grade 2 (CIN2).

DESIGN

Nationwide population based historical cohort study.

SETTING

Danish healthcare registries.

PARTICIPANTS

Women with CIN2 diagnosed in 1998-2020 and aged 18-40 years at diagnosis, who had either active surveillance or immediate treatment with large loop excision of the transformation zone (LLETZ). Women with a previous record of CIN2 or worse or LLETZ were excluded.

MAIN OUTCOME MEASURE

A Weibull survival model for interval censored time-to-event data was used to estimate the cumulative risk of cervical cancer. Inverse probability treatment weighting was used to adjust estimates for age, index cytology, calendar year, and region of residence.

RESULTS

The cohort included 27 524 women with CIN2, of whom 12 483 (45%) had active surveillance and 15 041 (55%) had immediate LLETZ. During follow-up, 104 cases of cervical cancer were identified-56 (54%) in the active surveillance group and 48 (46%) in the LLETZ group. The cumulative risk of cervical cancer was comparable across the two groups during the active surveillance period of two years. Thereafter, the risk increased in the active surveillance group, reaching 2.65% (95% confidence interval 2.07% to 3.23%) after 20 years, whereas it remained stable in the LLETZ group at 0.76% (0.58% to 0.95%).

CONCLUSIONS

Undergoing active surveillance for CIN2, thereby leaving the lesion untreated, was associated with increased long term risk of cervical cancer compared with immediate LLETZ. These findings show the importance of continued follow-up of women having active surveillance.

The combined finding of HPV 16, 18, or 45 and cytologic Atypical Glandular Cells (AGC) indicates a greatly elevated risk of in situ and invasive cervical adenocarcinoma

BACKGROUND

Cervical screening has not effectively controlled cervical adenocarcinoma (AC). Human papillomavirus (HPV) testing is recommended for cervical screening but the optimal management of HPV-positive individuals to prevent AC remains a question. Cytology and HPV typing are two triage options to predict the risk of AC. We combined two potential biomarkers (atypical glandular cell, AGC, cytology and HPV-types 16, 18, or 45) to assess their joint effect on detecting AC.

METHODS

Kaiser Permanente Northern California (KPNC) used triennial co-testing with cytology and HPV testing (positive/negative) for routine cervical screening between 2003 and 2020. HPV typing of a sample of residual HPV test specimens was performed on a separate cohort selected from KPNC (Persistence and Progression, PaP, cohort). We compared risk of prevalent and incident histologic AC/AIS (adenocarcinoma in situ) associated with preceding combinations of cytologic results and HPV typing. Risk of squamous cell cancer (SCC)/cervical intraepithelial neoplasia grade 3 (CIN3) (SCC/CIN3) was also included for comparison.

RESULTS

Among HPV-positive individuals in PaP cohort, 99% of prevalent AC and 96% of AIS were linked to HPV-types 16, 18, or 45 (denoted HPV 16/18/45). Although rare (0.09% of screening population), the concurrent detection of HPV 16/18/45 with AGC cytology predicted a highly elevated relative risk of underlying histologic AC/AIS; the absolute risk of diagnosing AC/AIS was 12% and odds ratio (OR) was 1341 (95%CI:495-3630) compared to patients with other high-risk HPV types and normal cytology. Cumulatively (allowing non-concurrent results), approximately one-third of the AC/AIS cases ever had HPV 16/18/45 and AGC cytology (OR = 1785; 95%CI:872-3656). AGC was not as strongly associated with SCC/CIN3.

CONCLUSION

Detection of HPV 16/18/45 positivity elevates risk of adenocarcinoma, particularly if AGC cytology is also found.

The Improving Risk Informed HPV Screening (IRIS) Study: Design and Baseline Characteristics

BACKGROUND

Cervical cancer screening with high-risk HPV (HrHPV) testing is being introduced. Most HrHPV infections are transient, requiring triage tests to identify individuals at highest risk for progression to cervical cancer. Head-to-head comparisons of available strategies for screening and triage are needed. Endometrial and ovarian cancers could be amenable to similar testing.

METHODS

Between 2016-2021 discard cervical cancer screening specimens from women ages 25-65 undergoing screening at Kaiser Permanente Northern California were collected. Specimens were aliquoted, stabilized, and stored frozen. HPV, cytology and histopathology results as well as demographic and co-factor information were obtained from electronic medical records. Follow-up collection of specimens was conducted for 2 years and EMR-based data collection was planned five years.

RESULTS

Collection of enrollment and follow-up specimens are complet and EMR-based follow-up data collection is ongoing. At baseline, specimens were collected from 54,971 HPV-positive, 10,219 HPV-negative/Pap-positive and 12,751 HPV-negative/Pap-negative women. Clinical history prior to baseline was available for 72.6% of individuals, of which 53.9% were undergoing routine screening, 8.6% recently had an abnormal screen, 30.3% had previous colposcopy, and 7.2% had previous treatment. As of November 2020, 55.6% had one or more colposcopies, yielding 5,515 CIN2, 2,735 CIN3 and 147 cancer histopathology diagnoses.

CONCLUSIONS

This robust population-based cohort study represents all stages of cervical cancer screening, management, and post-treatment follow-up.

IMPACT

The IRIS study is a unique and highly relevant resource allowing for natural history studies and rigorous evaluation of candidate HrHPV screening and triage markers, while permitting studies of biomarkers associated with other gynecological cancers.

Influential Usage of Big Data and Artificial Intelligence in Healthcare

Artificial intelligence (AI) is making computer systems capable of executing human brain tasks in many fields in all aspects of daily life. The enhancement in information and communications technology (ICT) has indisputably improved the quality of people's lives around the globe. Especially, ICT has led to a very needy and tremendous improvement in the health sector which is commonly known as electronic health (eHealth) and medical health (mHealth). Deep machine learning and AI approaches are commonly presented in many applications using big data, which consists of all relevant data about the medical health and diseases which a model can access at the time of execution or diagnosis of diseases. For example, cardiovascular imaging has now accurate imaging combined with big data from the eHealth record and pathology to better characterize the disease and personalized therapy. In clinical work and imaging, cancer care is getting improved by knowing the tumor biology and helping in the implementation of precision medicine. The Markov model is used to extract new approaches for leveraging cancer. In this paper, we have reviewed existing research relevant to eHealth and mHealth where various models are discussed which uses big data for the diagnosis and healthcare system. This paper summarizes the recent promising applications of AI and big data in medical health and electronic health, which have potentially added value to diagnosis and patient care.

The confounding effect of multi-type human papillomavirus infections on type-specific natural history parameter identification

The natural history of human papillomavirus (HPV) from infection to cervical cancer differs between HPV types. Accordingly, type-specific natural history parameters are crucial for the mathematical models used to optimize the nearly life-long series of disease prevention measures. These parameters are estimated from genotyped data from trials and population level screening programs, typically one type at a time, which requires projecting the multiple-type data to the single type. To analyze impacts of such projection methods on the estimates, we compared estimating one type at a time using different projection methods with estimating all types together. We simulated genotyped data with chosen parameter values for two HPV types and analyzed the identifiability of the chosen values using the different estimation methods. We found the success of estimating one type at a time to be excessively sensitive to the data projection method, with potential to falsely identify the parameters at wrong values. Estimating all types together identified the parameters well. Our results were consistent both when trial and population level data were used. In conclusion, the potential confounding by multi-type infections has to be considered when choosing an estimation method for type-specific natural history parameters.

A Study of Partial Human Papillomavirus Genotyping in Support of the 2019 ASCCP Risk-Based Management Consensus Guidelines

INTRODUCTION

The 2019 ASCCP Risk-Based Management Consensus Guidelines include recommendations for partial human papillomavirus (HPV) genotyping in management of abnormal cervical cancer screening results. The guidelines are based on matching estimates of cervical intraepithelial neoplasia (CIN) 3+ risk to consensus clinical action thresholds. In support of the guidelines, this analysis addresses the risks predicted by individual identification of HPV 16 and HPV 18.

METHODS

Risk estimates were drawn from a subset of women in the Kaiser Permanente Northern California screening program, whose residual cervical specimens were HPV typed as part of the HPV Persistence and Progression study. We calculated risk of CIN 3+ to assess how identification of HPV 16, HPV 18, or 12 other "high-risk" HPV types would influence recommended clinical management of new abnormal screening results, taking into account current cytologic results and recent screening history. Immediate and/or 5-year risks of CIN 3+ were matched to clinical actions identified in the guidelines.

RESULTS

Identification of HPV 16 at the first visit including HPV testing elevated immediate risk of diagnosing CIN 3+ sufficiently to mandate colposcopic referral even when cytology was Negative for Intraepithelial Lesions or Malignancy and to support a preference for treatment of cytologic high-grade squamous intraepithelial lesion. HPV 18 less clearly elevated CIN 3+ risk.

CONCLUSIONS

Identification of HPV 16 clearly mandated consideration in clinical management of new abnormal screening results. HPV 18 positivity must be considered as a special situation because of established disproportionate risk of invasive cancer. More detailed genotyping and use beyond initial management will be considered in guideline updates.

Risk Estimates Supporting the 2019 ASCCP Risk-Based Management Consensus Guidelines

Objective

The 2019 American Society for Colposcopy and Cervical Pathology Risk-Based Management Consensus Guidelines for the management of cervical cancer screening abnormalities recommend 1 of 6 clinical actions (treatment, optional treatment or colposcopy/biopsy, colposcopy/biopsy, 1-year surveillance, 3-year surveillance, 5-year return to regular screening) based on the risk of cervical intraepithelial neoplasia grade 3, adenocarcinoma in situ, or cancer (CIN 3+) for the many different combinations of current and recent past screening results. This article supports the main guidelines presentation¹ by presenting and explaining the risk estimates that supported the guidelines.

Methods

From 2003 to 2017 at Kaiser Permanente Northern California (KPNC), 1.5 million individuals aged 25 to 65 years were screened with human papillomavirus (HPV) and cytology cotesting scheduled every 3 years. We estimated immediate and 5-year risks of CIN 3+ for combinations of current test results paired with history of screening test and colposcopy/biopsy results.

Results

Risk tables are presented for different clinical scenarios. Examples of important results are highlighted; for example, the risk posed by most current abnormalities is greatly reduced if the prior screening round was HPV-negative. The immediate and 5-year risks of CIN 3+ used to decide clinical management are shown.

Conclusions

The new risk-based guidelines present recommendations for the management of abnormal screening test and histology results; the key risk estimates supporting guidelines are presented in this article. Comprehensive risk estimates are freely available online at https://CervixCa.nlm.nih.gov/RiskTables.

2019 ASCCP Risk-Based Management Consensus Guidelines: Methods for Risk Estimation, Recommended Management, and Validation

Supplemental digital content is available in the text.

Objective

To manage cervical screening abnormalities, the 2019 ASCCP management consensus guidelines will recommend clinical action on the basis of risk of cervical precancer and cancer. This article details the methods used to estimate risk, to determine the risk-based management, and to validate that the risk-based recommendations are of general use in different settings.

Methods

Based on 1.5 million patients undergoing triennial cervical screening by cotesting at the Kaiser Permanente Northern California from 2003 to 2017, we estimated risk profiles for different clinical scenarios and combinations of past and current human papillomavirus and cytology test results. We validated the recommended management by comparing with the estimated risks in several external data sources.

Results

Risk and management tables are presented separately by Egemen et al.¹ and Demarco et al.² Risk-based management derived from the Kaiser Permanente Northern California largely agreed with the management implied from the estimated risks of the other data sources.

Conclusions

The new risk-based guidelines present management of abnormal cervical screening results. By describing the steps used to develop these guidelines, the methods presented in this article can provide a basis for future extensions of the risk-based guidelines.

Cumulative risk of cervical intraepithelial neoplasia for women with normal cytology but positive for human papillomavirus: Systematic review and meta-analysis

Most women positive for human papillomavirus (HPV) are cytology normal. The optimal screen-management of these women is unclear given their risk of developing precancer. We performed a systematic review and meta-analysis of progression rates to precancer and cancer for HPV-positive, cytology normal women. We searched MEDLINE, EMBASE and Scopus for prospective studies measuring the cumulative incidence of precancer and cervical cancer in HPV-positive, cytology/histology normal women. Record screening was performed independently by two reviewers. We modeled the cumulative incidence over time using a multilevel random-effects meta-regression model. We used the model to predict HPV type-specific risks of precancer and cancer over follow-up. Data from 162 unique records were used in our analysis. The average incidence rate of cervical intraepithelial neoplasia grade 3 or cancer (CIN3+) in high-risk HPV positive but cytology/histology normal women was 1.0 per 100 women-years (95% CI: 1.0-1.1). This corresponds to an average cumulative risk at 1, 3 and 5 years of 2.1% (95% prediction interval 0.0-9.5), 4.3% (95% prediction interval 0.0-11.5) and 6.4% (95% prediction interval 0.0-13.5). HPV type was a strong predictor of the risk of oncogenic progression. There was substantial heterogeneity in the background precancer risk across studies (P-value < .0001). Our HPV type-specific progression risk estimates can help inform risk-based cervical cancer screening guidelines for HPV-positive women. However, precancer and cervical cancer risks are highly variable and may not be generalizable between populations.

A study of type-specific HPV natural history and implications for contemporary cervical cancer screening programs

BACKGROUND

HPV testing is replacing cytology for cervical cancer screening because of greater sensitivity and superior reassurance following negative tests for the dozen HPV genotypes that cause cervical cancer. Management of women testing positive is unresolved. The need for identification of individual HPV genotypes for clinical use is debated. Also, it is unclear how long to observe persistent infections when precancer is not initially found.

METHODS

In the longitudinal NCI-Kaiser Permanente Northern California Persistence and Progression (PaP) Study, we observed the clinical outcomes (clearance, progression to CIN3+, or persistence without progression) of 11,573 HPV-positive women aged 30-65 yielding 14,158 type-specific infections.

FINDINGS

Risks of CIN3+ progression differed substantially by type, with HPV16 conveying uniquely elevated risk (26% of infections with seven-year CIN3+ risk of 22%). The other carcinogenic HPV types fell into 3 distinct seven-year CIN3+ risk groups: HPV18, 45 (13% of infections, risks >5%, with known elevated cancer risk); HPV31, 33, 35, 52, 58 (39%, risks >5%); and HPV39, 51, 56, 59, 68 (23%, risks <5%). In the absence of progression, HPV clearance rates were similar by type, with 80% of infections no longer detected within three years; persistence to seven years without progression was uncommon. The predictive value of abnormal cytology was most evident for prevalent CIN3+, but less evident in follow-up. A woman's age did not modify risk; rather it was the duration of persistence that was important.

INTERPRETATION

HPV type and persistence are the major predictors of progression to CIN3+; at a minimum, distinguishing HPV16 is clinically important. Dividing the other HPV types into three risk-groups is worth considering.

Five-Year Risk of Cervical Precancer Following p16/Ki-67 Dual-Stain Triage of HPV-Positive Women

Importance

As cervical cancer screening transitions to primary human papillomavirus (HPV) testing, effective triage and management of HPV-positive women is critical to avoid unnecessary colposcopy referral and associated harms while maintaining high sensitivity for cervical precancer. Triage with p16/Ki-67 dual-stain (DS) testing has shown high sensitivity and specificity for detection of cervical precancers; however, longitudinal studies are needed to determine the long-term risk of precancer following a negative DS result.

Objective

To evaluate the longitudinal performance of p16/Ki-67 DS triage for detection of cervical precancer in HPV-positive women over 5 years of follow-up in the context of clinical management thresholds.

Design, Setting, and Participants

Prospective cohort study of HPV-positive women 30 years or older undergoing routine cervical cancer screening in 2012 with HPV and Papanicolaou (hereinafter "cytology") co-testing within the Kaiser Permanente Northern California health care system. Follow-up of medical records was conducted through 2017.

Exposures

All p16/Ki-67 DS testing was performed on residual SurePath material, and slides were evaluated for p16/Ki-67 positivity.

Main Outcomes and Measures

Histological end points were ascertained from the clinical database through 2017. We estimated 5-year cumulative risks of cervical intraepithelial neoplasia grades of 2 or worse (≥CIN2) or grades 3 or worse (≥CIN3) by baseline DS and cytology at yearly intervals using Logistic Weibull models. Risks were compared with clinical management thresholds for colposcopy referral and a 1-year return interval.

Results

Among the 1549 HPV-positive women in this study, the mean age at enrollment was 42.2 years, and the median follow-up time was 3.7 years (range, 0.2-5.4 years). Positive DS results were associated with significantly higher cumulative 5-year risks of ≥CIN2 compared with abnormal cytology (31.0%; 95% CI, 27.2%-35.3% vs 25.0%; 95% CI, 21.7%-28.7%; P = .03). Women with DS-negative findings had significantly lower 5-year risks of ≥CIN2 compared with women with normal cytology (8.5%; 95% CI, 6.5%-11.1% vs 12.3%; 95% CI, 9.8%-15.4%; P = .04). In DS-negative women, the risks of both ≥CIN2 and ≥CIN3 remained below the colposcopy referral threshold for all 5 years, crossing the 1-year return threshold at 3 years.

Conclusions and Relevance

Triage with p16/Ki-67 DS provides better long-term risk stratification than cytology over 5 years. The low risk of cervical precancer in p16/Ki-67 DS-negative women permits safe extension of follow-up intervals for 3 years.

Assessing the time dependence of prognostic values of cytology and human papillomavirus testing in cervical cancer screening

Accurate assessment of risks for developing cervical intraepithelial neoplasia of grade 2 or worse (CIN2+) after a given set of screening test results is instrumental to reaching valid conclusions and informing cervical cancer screening recommendations. Using data from the Canadian Cervical Cancer Screening Trial (CCCaST), we assessed prognostic values of enrollment screening test results to predict CIN2+ among women attending routine cervical screening using multivariable Cox proportional hazards (PH) regression and its flexible extension during each of two follow-up periods (protocol-defined and extended). Nonproportional (time-dependent (TD)) and/or nonlinear effects were modeled, as appropriate. Women with abnormal cytology had hazard ratios (HRs) for CIN2+ detection of 17.61 (95% CI: 11.25-27.57) and 10.46 (95% CI: 5.41-20.24) relative to women with normal cytology during the protocol-defined and extended follow-up periods, respectively. High-risk human papillomavirus (HR-HPV) positivity was an even stronger predictor of CIN2+ risk, with significant TD effects during both follow-up periods (p <0.001 for both TD effects). Risks among women co-testing HR-HPV+ with and without abnormal cytology (relative to women co-testing negative) were highest immediately after baseline, and decreased significantly thereafter (p <0.001 for both TD effects). HRs for HPV16+ and HPV18+ women (relative to those testing HR-HPV-) did not vary significantly over time (HR = 182.96; 95% CI: 95.16-351.77 and HR = 111.81; 95% CI: 44.60-280.31, respectively). Due to TD effects, conventional Cox model estimates considerably underestimated adjusted HRs associated with positive HR-HPV testing results early on in the follow-up periods.

Population-based e-records to evaluate HPV triage of screen-detected atypical squamous cervical lesions in Catalonia, Spain, 2010-15

Equivocal lesions (ASC-US) are common abnormalities in cervical cancer screening exams. HPV testing helps to stratify the risk of progression to high-grade squamous intraepithelial lesions or more (HSIL+). Population-based medical electronic data can be used to evaluate screening recommendations. The study uses routine electronic data from primary health centers to estimate the impact of HPV testing in a 3- and a 5-year risk of HSIL+ after an ASC-US. The study includes data derived from medical electronic information from 85,775 women who first attended a cervical cancer screening visit at the National Health System facilities of Catalonia, Spain, during 2010-11 and followed up to 2015. Included women were aged between 25-65 years old, having at least one follow-up visit, and a cervical cytology of ASC-US (N = 1,647). Women with a first result of low-grade squamous intraepithelial lesions (LSIL) (N = 945) or those with negative cytology (N = 83,183) were included for comparison. Those with a baseline HSIL+ were excluded. Incident HSIL+ was evaluated by means of Kaplan-Meier curves and multivariate regression models. HPV test results were available for 63.4% of women with a baseline ASC-US. Among all ASC-US, 70 incident HSIL+ were identified at 5 years. ASC-US HPV positive women had a high risk of HSIL+ compared to women with negative cytology (adjusted HR = 32.7; 95% CI: 23.6-45.2) and a similar risk to women with baseline LSIL (HR = 29.3; 95% CI: 22.4-38.2), whereas ASC-US HPV negative women had no differential risk to that observed in baseline negative cytology. Women with ASC-US and no HPV test had an average HSIL+ risk (HR = 14.8; 95% CI: 9.7-22.5). Population-based e-medical records derived from primary health care centers allowed monitoring of screening recommendations, providing robust estimates for the study outcomes. This analysis confirms that HPV testing improved risk stratification of ASC-US lesions. The information can be used to improve diagnosis and management of screen detected lesions.