From differences in means between cases and controls to risk stratification: a business plan for biomarker development

Artificial intelligence-based image analysis in clinical testing: lessons from cervical cancer screening

Novel screening and diagnostic tests based on artificial intelligence (AI) image recognition algorithms are proliferating. Some initial reports claim outstanding accuracy followed by disappointing lack of confirmation, including our own early work on cervical screening. This is a presentation of lessons learned, organized as a conceptual step-by-step approach to bridge the gap between the creation of an AI algorithm and clinical efficacy. The first fundamental principle is specifying rigorously what the algorithm is designed to identify and what the test is intended to measure (eg, screening, diagnostic, or prognostic). Second, designing the AI algorithm to minimize the most clinically important errors. For example, many equivocal cervical images cannot yet be labeled because the borderline between cases and controls is blurred. To avoid a misclassified case-control dichotomy, we have isolated the equivocal cases and formally included an intermediate, indeterminate class (severity order of classes: case>indeterminate>control). The third principle is evaluating AI algorithms like any other test, using clinical epidemiologic criteria. Repeatability of the algorithm at the borderline, for indeterminate images, has proven extremely informative. Distinguishing between internal and external validation is also essential. Linking the AI algorithm results to clinical risk estimation is the fourth principle. Absolute risk (not relative) is the critical metric for translating a test result into clinical use. Finally, generating risk-based guidelines for clinical use that match local resources and priorities is the last principle in our approach. We are particularly interested in applications to lower-resource settings to address health disparities. We note that similar principles apply to other domains of AI-based image analysis for medical diagnostic testing.

Variance of age-specific log incidence decomposition (VALID): a unifying model of measured and unmeasured genetic and non-genetic risks

BACKGROUND

The extent to which known and unknown factors explain how much people of the same age differ in disease risk is fundamental to epidemiology. Risk factors can be correlated in relatives, so familial aspects of risk (genetic and non-genetic) must be considered.

DEVELOPMENT

We present a unifying model (VALID) for variance in risk, with risk defined as log(incidence) or logit(cumulative incidence). Consider a normally distributed risk score with incidence increasing exponentially as the risk increases. VALID's building block is variance in risk, Δ2, where Δ = log(OPERA) is the difference in mean between cases and controls and OPERA is the odds ratio per standard deviation. A risk score correlated r between a pair of relatives generates a familial odds ratio of exp(rΔ2). Familial risk ratios, therefore, can be converted into variance components of risk, extending Fisher's classic decomposition of familial variation to binary traits. Under VALID, there is a natural upper limit to variance in risk caused by genetic factors, determined by the familial odds ratio for genetically identical twin pairs, but not to variation caused by non-genetic factors.

APPLICATION

For female breast cancer, VALID quantified how much variance in risk is explained-at different ages-by known and unknown major genes and polygenes, non-genomic risk factors correlated in relatives, and known individual-specific factors.

CONCLUSION

VALID has shown that, while substantial genetic risk factors have been discovered, much is unknown about genetic and familial aspects of breast cancer risk especially for young women, and little is known about individual-specific variance in risk.

Clinical Value of Circulating miRNA in Diagnosis, Prognosis, Screening and Monitoring Therapy of Pancreatic Ductal Adenocarcinoma-A Review of the Literature

Pancreatic cancer (PC) is considered to be the seventh most common cause of cancer-related deaths. The number of deaths caused by PC is estimated to increase in the future. An early diagnosis of PC is crucial for improving treatment outcomes. The most common histopathological subtype of PC is pancreatic ductal adenocarcinoma (PDAC). MicroRNAs (miRNAs)-which are endogenous non-coding RNAs involved in the posttranscriptional regulation of multiple gene expression-constitute useful diagnostic and prognostic biomarkers in various neoplasms, including PDAC. Circulating miRNAs detected in a patient's serum or plasma are drawing more and more attention. Hence, this review aims at evaluating the clinical value of circulating miRNA in the screening, diagnosis, prognosis and monitoring of pancreatic ductal adenocarcinoma therapy.

Circulating HPV DNA in HPV-associated cancers

Human papillomavirus (HPV) infections are the primary cause of almost all cervical cancers, anal cancers, and a variable proportion of other anogenital tumors, as well as head and neck cancers. Circulating HPV DNA (cHPV-DNA) is emerging as a biomarker with extensive potential in the management of HPV-driven malignancies. There has been a rapid advancement in the development of techniques for analyzing cHPV-DNA for the detection, characterization, and monitoring of HPV-associated cancers. As clinical evidence accumulates, it is becoming evident that cHPV-DNA can be used as a diagnostic tool. By conducting clinical trials assessing the clinical utility of cHPV-DNA, the full potential of cHPV-DNA for the screening, diagnosis, and treatment of HPV-related malignancies can be corroborated. In this review, we examine the current landscape of applications for cHPV-DNA liquid biopsies throughout the cancer care continuum, highlighting future opportunities for research and integration into clinical practice.

Epigenetic Regulation in Exposome-Induced Tumorigenesis: Emerging Roles of ncRNAs

Environmental factors, including pollutants and lifestyle, constitute a significant role in severe, chronic pathologies with an essential societal, economic burden. The measurement of all environmental exposures and assessing their correlation with effects on individual health is defined as the exposome, which interacts with our unique characteristics such as genetics, physiology, and epigenetics. Epigenetics investigates modifications in the expression of genes that do not depend on the underlying DNA sequence. Some studies have confirmed that environmental factors may promote disease in individuals or subsequent progeny through epigenetic alterations. Variations in the epigenetic machinery cause a spectrum of different disorders since these mechanisms are more sensitive to the environment than the genome, due to the inherent reversible nature of the epigenetic landscape. Several epigenetic mechanisms, including modifications in DNA (e.g., methylation), histones, and noncoding RNAs can change genome expression under the exogenous influence. Notably, the role of long noncoding RNAs in epigenetic processes has not been well explored in the context of exposome-induced tumorigenesis. In the present review, our scope is to provide relevant evidence indicating that epigenetic alterations mediate those detrimental effects caused by exposure to environmental toxicants, focusing mainly on a multi-step regulation by diverse noncoding RNAs subtypes.

Prediction of Pancreatic Cancer in Diabetes Patients with Worsening Glycemic Control

BACKGROUND

Worsening glycemic control indicates elevated risk of pancreatic ductal adenocarcinoma (PDAC). We developed prediction models for PDAC among those with worsening glycemic control after diabetes diagnosis.

METHODS

In 2000-2016 records within the Veterans Affairs Health System (VA), we identified three cohorts with progression of diabetes: (i) insulin initiation (n = 449,685), (ii) initiation of combination oral hypoglycemic medication (n = 414,460), and (iii) hemoglobin A1c (HbA1c) ≥8% with ≥Δ1% within 15 months (n = 593,401). We computed 12-, 36-, and 60-month incidence of PDAC and developed prediction models separately for males and females, with consideration of >30 demographic, behavioral, clinical, and laboratory variables. Models were selected to optimize Akaike's Information Criterion, and performance for predicting 12-, 36-, and 60-month incident PDAC was evaluated by bootstrap.

RESULTS

Incidence of PDAC was highest for insulin initiators and greater in males than in females. Optimism-corrected c-indices of the models for predicting 36-month incidence of PDAC in the male population were: (i) 0.72, (ii) 0.70, and (iii) 0.71, respectively. Models performed better for predicting 12-month incident PDAC [c-index (i) 0.78, (ii) 0.73, (iii) 0.76 for males], and worse for predicting 60-month incident PDAC [c-index (i) 0.69, (ii) 0.67, (iii) 0.68 for males]. Model performance was lower among females. For subjects whose model-predicted 36-month PDAC risks were ≥1%, the observed incidences were (i) 1.9%, (ii) 2.2%, and (iii) 1.8%.

CONCLUSIONS

Sex-specific models for PDAC can estimate risk of PDAC at the time of progression of diabetes.

IMPACT

Our models can identify diabetes patients who would benefit from PDAC screening.

Validation of the Southend giant cell arteritis probability score in a Scottish single-centre fast-track pathway

OBJECTIVE

The aim was to provide external validation of the Southend GCA probability score (GCAPS) in patients attending a GCA fast-track pathway (GCA FTP) in NHS Lanarkshire.

METHODS

Consecutive GCA FTP patients between November 2018 and December 2020 underwent GCAPS assessment as part of routine care. GCA diagnoses were supported by US of the cranial and axillary arteries (USS), with or without temporal artery biopsy (TAB), and confirmed at 6 months. Percentages of patients with GCA according to GCAPS risk group, performance of total GCAPS in distinguishing GCA/non-GCA final diagnoses, and test characteristics using different GCAPS binary cut-offs were assessed. Associations between individual GCAPS components and GCA and the value of USS and TAB in the diagnostic process were also explored.

RESULTS

Forty-four of 129 patients were diagnosed with GCA, including 0 of 41 GCAPS low-risk patients (GCAPS <9), 3 of 40 medium-risk patients (GCAPS 9-12) and 41 of 48 high-risk patients (GCAPS >12). Overall performance of GCAPS in distinguishing GCA/non-GCA was excellent [area under the receiver operating characteristic curve, 0.976 (95% CI 0.954, 0.999)]. GCAPS cut-off ≥10 had 100.0% sensitivity and 67.1% specificity for GCA. GCAPS cut-off ≥13 had the highest accuracy (91.5%), with 93.2% sensitivity and 90.6% specificity. Several individual GCAPS components were associated with GCA. Sensitivity of USS increased by ascending GCAPS risk group (nil, 33.3% and 90.2%, respectively). TAB was diagnostically useful in cases where USS was inconclusive.

CONCLUSION

This is the first published study to describe application of GCAPS outside the specialist centre where it was developed. Performance of GCAPS as a risk stratification tool was excellent. GCAPS might have additional value for screening GCA FTP referrals and guiding empirical glucocorticoid treatment.

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.

Liquid Biopsy for Cancer Detection: Clinical and Epidemiologic Considerations

Detection of circulating nucleic acids, also referred to as liquid biopsy, has been evaluated for detection of cancer in a variety of settings. We describe important clinical and epidemiologic considerations for liquid biopsy applications in cancer early detection and for monitoring of cancer recurrence.See related article by Jeannot et al., p. 5869.

Blood DNA methylation score predicts breast cancer risk: applying OPERA in molecular, environmental, genetic and analytic epidemiology

In this issue, Kresovich and colleagues have published a hallmark paper in Molecular, Environmental, Genetic and Analytic Epidemiology. By applying artificial intelligence to the Sister Study they created a new methylation-based breast cancer risk score (mBCRS) based on blood DNA methylation. Using a prospective design and after accounting for age and questionnaire-based breast cancer risk factors, the Odds PER Adjusted standard deviation (OPERA) for mBCRS and polygenic risk score (PRS) was 1.58 (95% CI: 1.38, 1.81) and 1.58 (95% CI: 1.36, 1.83), respectively, and the corresponding area under the receiver operating curve was 0.63 for both. Therefore, mBCRS could be as powerful as the current best PRS in differentiating women of the same age in terms of their breast cancer risk. These risk scores are among the strongest known breast cancer risk-stratifiers, shaded only by new mammogram risk scores based on measures other than conventional mammographic density, such as Cirrocumulus and Cirrus, which when combined have an OPERA as high as 2.3. The combination of PRS and mBCRS with the other measured risk factors gave an OPERA of 2.2. OPERA has many advantages over changes in areas under the receiver operator curve because the latter depend on the order in which risk factors are considered. Although more replication is needed using prospective data to protect against reverse causation, there are many novel molecular and analytic aspects to this paper which uncovers a potential mechanism for how genetic and environmental factors combine to cause breast cancer.

Talc, body powder, and ovarian cancer: A summary of the epidemiologic evidence

Many women apply powder to the genital area as a drying agent. Talc, an inert mineral with a high capacity to absorb water, has historically been a major component of body powders. Due to its similarity and co-occurrence with asbestos, the association of body powder/talc use and gynecological cancer risk, specifically ovarian cancer risk, has been a long-standing research question. Retrospective case-control studies have shown associations between genital powder use and ovarian cancer risk, with summary relative risk estimates from meta-analyses and pooled analyses ranging from 1.24 to 1.35 for ever versus never use. In contrast, prospective cohort studies have not shown a statistically significant association until recently, when a pooled analysis of four large cohorts demonstrated a weak, but statistically significant association among women with patent reproductive tracts (hazard ratio 1.13). Taken together, the epidemiological data from case-control studies and cohort studies suggest that there may be a small, positive association between genital powder use and ovarian cancer. The causal factors underlying this association are not clear. Proposed factors include talc, other minerals, such as asbestos or quartz, that are known carcinogens and may contaminate talc products, or other powder ingredients that could cause inflammation of the reproductive tracts. Given the rarity of ovarian cancer in the general population, the small increase in relative risk translates to a very low increase in absolute risk. Further research is needed to understand the underpinnings of the observed association between genital powder use and ovarian cancer risk.

Expanding the search for germline pathogenic variants for breast cancer. How far should we go and how high should we jump? The missed opportunity!

Since the identification of BRCA1 and BRCA2 genes 3 decades ago, genetic testing and genetic counseling have become an integral part of routine clinical practice. The risk of breast cancer among carriers of germline pathogenic variants, like BRCA1 and BRCA2, is well established. Risk-reducing interventions, including bilateral mastectomies and salpingo-oophorectomies are both effective and have become more acceptable. Many researchers and professional societies view current guidelines as restrictive and may miss many at-risk women, and are calling to expand testing to include all patients with breast cancer, regardless of their personal or family history of cancer, while others are calling for wider adoption to even include all healthy women at age 30 or older. This review will address expanding testing in two directions; horizontally to include more patients, and even healthy women, and vertically to include more genes using next-generation sequencing-based multi-gene panel testing.

A Systematic Review of Tests for Postcolposcopy and Posttreatment Surveillance

OBJECTIVE

For the 2019 ASCCP Risk-Based Management Consensus Guidelines, we conducted a systematic review of diagnostic assays for postcolposcopy and posttreatment management.

MATERIALS AND METHODS

A literature search was conducted to identify articles reporting on tests/assays for cervical cancer screening, triage, postcolposcopy surveillance, and posttreatment surveillance published between 2012 and 2019 in PubMed and Embase. Titles and abstracts were evaluated by co-authors for inclusion. Included articles underwent full-text review, data abstraction, and quality assessment. Pooled absolute pretest and posttest risk estimates were calculated for studies evaluating management of patients after treatment.

RESULTS

A total of 2,862 articles were identified through the search. Of 50 articles on postcolposcopy, 5 were included for data abstraction. Of 66 articles on posttreatment, 23 were included for data abstraction and were summarized in the meta-analysis. The pooled posttreatment risk of cervical intraepithelial neoplasia (CIN) 2+ in all studies was 4.8% (95% CI = 3.4%-6.8%), ranging from 0.4%-19.5% (τ = 0.57) in individual studies. Among individuals testing negative for human papillomavirus (HPV) posttreatment, the risk of CIN 2+ was 0.69% (95% CI = 0.3%-1.5%); among individuals testing positive for HPV posttreatment, the risk of CIN 2+ was 18.3% (95% CI = 12.1%-26.6%) in all studies. All risk estimates were substantially higher for liquid-based cytology. The HPV-cytology co-testing provided slightly better reassurance compared with HPV alone at the cost of much higher positivity.

CONCLUSIONS

Despite a large number of published studies on postcolposcopy and posttreatment surveillance, only few met criteria for abstraction and were included in the meta-analysis. More high-quality studies are needed to evaluate assays and approaches that can improve management of patients with abnormal screening.

Risk assessment of endometrial cancer and endometrial intraepithelial neoplasia in women with abnormal bleeding and implications for clinical management algorithms

BACKGROUND

Most endometrial cancer cases are preceded by abnormal uterine bleeding, offering a potential opportunity for early detection and cure of endometrial cancer. Although clinical guidelines exist for diagnostic workup of abnormal uterine bleeding, consensus is lacking regarding optimal management for women with abnormal bleeding to diagnose endometrial cancer.

OBJECTIVE

We report the baseline data from a prospective clinical cohort study of women referred for endometrial evaluation at the Mayo Clinic, designed to evaluate risk stratification in women at increased risk for endometrial cancer. Here, we introduce a risk-based approach to evaluate diagnostic tests and clinical management algorithms in a population of women with abnormal bleeding undergoing endometrial evaluation at the Mayo Clinic.

STUDY DESIGN

A total of 1163 women aged ≥45 years were enrolled from February 2013 to May 2019. We evaluated baseline absolute risks and 95% confidence intervals of endometrial cancer and endometrial intraepithelial neoplasia according to clinical algorithms for diagnostic workup of women with postmenopausal bleeding (assessment of initial vs recurrent bleeding episode and endometrial thickness measured through transvaginal ultrasound). We also evaluated risks among women with postmenopausal bleeding according to baseline age (<60 vs 60+ years) as an alternative example. For this approach, biopsy would be conducted for all women aged 60+ years and those aged <60 years with an endometrial thickness of >4 mm. We assessed the clinical efficiency of each strategy by estimating the percentage of women who would be referred for endometrial biopsy, the percentage of cases detected and missed, and the ratio of biopsies per case detected.

RESULTS

Among the 593 women with postmenopausal bleeding, 18 (3.0%) had endometrial intraepithelial neoplasia, and 47 (7.9%) had endometrial cancer, and among the 570 premenopausal women with abnormal bleeding, 8 (1.4%) had endometrial intraepithelial neoplasia, and 7 (1.2%) had endometrial cancer. Maximum risk was noted in women aged 60+ years (17.7%; 13.0%-22.3%), followed by those with recurrent bleeding (14.7%; 11.0%-18.3%). Among women with an initial bleeding episode for whom transvaginal ultrasound was recommended, endometrial thickness did not provide meaningful risk stratification: risks of endometrial cancer and endometrial intraepithelial neoplasia were nearly identical in women with an endometrial thickness of >4 mm (5.8%; 1.3%-10.3%) and ≤4 mm (3.6%; 0.9%-8.6%). In contrast, among those aged <60 years with an endometrial thickness of >4 mm, the risk of endometrial cancer and endometrial intraepithelial neoplasia was 8.4% (4.3%-12.5%), and in those with an endometrial thickness of ≤4 mm, the risk was 0% (0.0%-3.0%; P=.01). The most efficient strategy was to perform biopsy in all women aged 60+ years and among those aged <60 years with an endometrial thickness of >4 mm, with the lowest percentage referred to biopsy while still detecting all cases.

CONCLUSION

Existing clinical recommendations for endometrial cancer detection in women with abnormal bleeding are not consistent with the underlying risk. Endometrial cancer risk factors such as age can provide important risk stratification compared with the assessment of recurrent bleeding. Future research will include a formal assessment of clinical and epidemiologic risk prediction models in our study population as well as validation of our findings in other populations.

Circulating Tumor DNA for Early Cancer Detection

BACKGROUND

Cancer cells release circulating tumor DNA (ctDNA) into the bloodstream, which can now be quantified and examined using novel high-throughput sequencing technologies. This has led to the emergence of the "liquid biopsy," which proposes to analyze this genetic material and extract information on a patient's cancer using a simple blood draw.

CONTENT

ctDNA has been detected in many advanced cancers. It has also been proven to be a highly sensitive indicator of relapse and prognosis. Sequencing the genetic material has also led to the discovery of mutations targetable by existing therapies. Although ctDNA screening is more expensive, it is showing promise against circulating tumor cells and traditional cancer biomarkers. ctDNA has also been detected in other bodily fluids, including cerebrospinal fluid, urine, saliva, and stool. The utility of ctDNA for early cancer detection is being studied. However, a blood test for cancer faces heavy obstacles, such as extremely low ctDNA concentrations in early-stage disease and benign mutations caused by clonal hematopoiesis, causing both sensitivity and specificity concerns. Nonetheless, companies and academic laboratories are highly active in developing such a test.

CONCLUSION

Currently, ctDNA is unlikely to perform at the high level of sensitivity and specificity required for early diagnosis and population screening. However, ctDNA in blood and other fluids has important clinical applications for cancer monitoring, prognosis, and selection of therapy that require further investigation.

Ability of known susceptibility SNPs to predict colorectal cancer risk for persons with and without a family history

Before SNP-based risk can be incorporated in colorectal cancer (CRC) screening, the ability of these SNPs to estimate CRC risk for persons with and without a family history of CRC, and the screening implications need to be determined. We estimated the association with CRC of a 45 SNP-based risk using 1181 cases and 999 controls, and its correlation with CRC risk predicted from detailed family history. We estimated the predicted change in the distribution across predefined risk categories, and implications for recommended screening commencement age, from adding SNP-based risk to family history. The inter-quintile risk ratio for colorectal cancer risk of the SNP-based risk was 3.28 (95% CI 2.54-4.22). SNP-based and family history-based risks were not correlated (r = 0.02). For persons with no first-degree relatives with CRC, screening could commence 4 years earlier for women (5 years for men) in the highest quintile of SNP-based risk. For persons with two first-degree relatives with CRC, screening could commence 16 years earlier for men and women in the highest quintile, and 7 years earlier for the lowest quintile. This 45 SNP panel in conjunction with family history, can identify people who could benefit from earlier screening. Risk reclassification by 45 SNPs could inform targeted screening for CRC prevention, particularly in clinical genetics settings when mutations in high-risk genes cannot be identified. Yet to be determined is cost-effectiveness, resources requirements, community, patient and clinician acceptance, and feasibility with potentially ethical, legal and insurance implications.

Quantifying risk stratification provided by diagnostic tests and risk predictions: Comparison to AUC and decision curve analysis

A property of diagnostic tests and risk models deserving more attention is risk stratification, defined as the ability of a test or model to separate those at high absolute risk of disease from those at low absolute risk. Risk stratification fills a gap between measures of classification (ie, area under the curve (AUC)) that do not require absolute risks and decision analysis that requires not only absolute risks but also subjective specification of costs and utilities. We introduce mean risk stratification (MRS) as the average change in risk of disease (posttest-pretest) revealed by a diagnostic test or risk model dichotomized at a risk threshold. Mean risk stratification is particularly valuable for rare conditions, where AUC can be high but MRS can be low, identifying situations that temper overenthusiasm for screening with the new test/model. We apply MRS to the controversy over who should get testing for mutations in BRCA1/2 that cause high risks of breast and ovarian cancers. To reveal different properties of risk thresholds to refer women for BRCA1/2 testing, we propose an eclectic approach considering MRS and other metrics. The value of MRS is to interpret AUC in the context of BRCA1/2 mutation prevalence, providing a range of risk thresholds at which a risk model is "optimally informative," and to provide insight into why net benefit arrives to its conclusion.

Current Approaches to Pancreatic Cancer Screening

Pancreatic ductal adenocarcinoma (PDAC) has a 5-year survival rate of only 8% and is estimated to be the second leading cause of cancer-related deaths by 2021. Prior convention held that screening for PDAC would not be beneficial; however, a deeper understanding of the carcinogenesis pathway supports a potential window of opportunity among the target population. Screening for PDAC is not a standard practice among the general population because of its low incidence. However, screening may be beneficial for individuals with familial history, chronic diseases with genetic predispositions, or inherited cancer syndromes, such as hereditary breast ovarian cancer syndrome, hereditary pancreatitis, Peutz-Jeghers syndrome, familial atypical multiple mole melanoma, Lynch syndrome (hereditary nonpolyposis colorectal cancer), ataxia telangiectasia, and Li-Fraumeni syndrome, all of which have been associated with an increased risk of developing PDAC. The screening strategies among these high-risk individuals are targeted to identify precursor lesions and PDAC at an early resectable stage. This review describes the risk factors for pancreatic cancer, especially the genetic risk factors in high-risk individuals and current screening strategies available for PDAC.

Association of Endometrial Cancer Risk With Postmenopausal Bleeding in Women: A Systematic Review and Meta-analysis

IMPORTANCE

As the worldwide burden of endometrial cancer continues to rise, interest is growing in the evaluation of early detection and prevention strategies among women at increased risk. Focusing efforts on women with postmenopausal bleeding (PMB), a common symptom of endometrial cancer, may be a useful strategy; however, PMB is not specific for endometrial cancer and is often caused by benign conditions.

OBJECTIVE

To provide a reference of the prevalence of PMB in endometrial cancers and the risk of endometrial cancer in women with PMB.

DATA SOURCES

For this systematic review and meta-analysis, PubMed and Embase were searched for English-language studies published January 1, 1977, through January 31, 2017.

STUDY SELECTION

Observational studies reporting the prevalence of PMB in women with endometrial cancer and the risk of endometrial cancer in women with PMB in unselected populations were selected.

DATA EXTRACTION AND SYNTHESIS

Two independent reviewers evaluated study quality and risk of bias using items from the Newcastle-Ottawa Quality Assessment Scale and the Quality Assessment of Diagnostic Accuracy Studies tool. Studies that included highly selected populations, lacked detailed inclusion criteria, and/or included 25 or fewer women were excluded.

MAIN OUTCOMES AND MEASURES

The pooled prevalence of PMB in women with endometrial cancer and the risk of endometrial cancer in women with PMB.

RESULTS

A total of 129 unique studies, including 34 432 unique patients with PMB and 6358 with endometrial cancer (40 790 women), were analyzed. The pooled prevalence of PMB among women with endometrial cancer was 91% (95% CI, 87%-93%), irrespective of tumor stage. The pooled risk of endometrial cancer among women with PMB was 9% (95% CI, 8%-11%), with estimates varying by use of hormone therapy (range, 7% [95% CI, 6%-9%] to 12% [95% CI, 9%-15%]; P < .001 for heterogeneity) and geographic region (range, 5% [95% CI, 3%-11%] in North America to 13% [95% CI, 9%-19%] in Western Europe; P = .09 for heterogeneity).

CONCLUSIONS AND RELEVANCE

Early detection strategies focused on women with PMB have the potential to capture as many as 90% of endometrial cancers; however, most women with PMB will not be diagnosed with endometrial cancer. These results can aid in the assessment of the potential clinical value of new early detection markers and clinical management strategies for endometrial cancer and will help to inform clinical and epidemiologic risk prediction models to support decision making.

A prospective study of risk-based colposcopy demonstrates improved detection of cervical precancers

BACKGROUND

Sensitivity for detection of precancers at colposcopy and reassurance provided by a negative colposcopy are in need of systematic study and improvement.

OBJECTIVE

We sought to evaluate whether selecting the appropriate women for multiple targeted cervical biopsies based on screening cytology, human papillomavirus testing, and colposcopic impression could improve accuracy and efficiency of cervical precancer detection.

STUDY DESIGN

In all, 690 women aged 18-67 years referred to colposcopy subsequent to abnormal cervical cancer screening results were included in the study (ClinicalTrials.gov: NCT00339989). Up to 4 cervical biopsies were taken during colposcopy to evaluate the incremental benefit of multiple biopsies. Cervical cytology, human papillomavirus genotyping, and colposcopy impression were used to establish up to 24 different risk strata. Outcomes for the primary analysis were cervical precancers, which included p16⁺ cervical intraepithelial neoplasia 2 and all cervical intraepithelial neoplasia 3 that were detected by colposcopy-guided biopsy during the colposcopy visit. Later outcomes in women without cervical intraepithelial neoplasia 2⁺ at baseline were abstracted from electronic medical records.

RESULTS

The risk of detecting precancer ranged from 2-82% across 24 strata based on colposcopy impression, cytology, and human papillomavirus genotyping. The risk of precancer in the lowest stratum increased only marginally with multiple biopsies. Women in the highest-risk strata had risks of precancer consistent with immediate treatment. In other risk strata, multiple biopsies substantially improved detection of cervical precancer. Among 361 women with cervical intraepithelial neoplasia <2 at baseline, 195 (54%) had follow-up cytology or histology data with a median follow-up time of 508 days. Lack of detection of precancer at initial colposcopy that included multiple biopsies predicted low risk of precancer during follow-up.

CONCLUSION

Risk assessment at the colposcopy visit makes identification of cervical precancers more effective and efficient. Not finding precancer after a multiple-biopsy protocol provides high reassurance and allows releasing women back to regular screening.

Strategies for screening and early detection of anal cancers: A narrative and systematic review and meta-analysis of cytology, HPV testing, and other biomarkers

Anal cancer incidence and mortality have been increasing over the past decade. Although the incidence in the general population remains low, it is much higher in certain subgroups, including those living with human immunodeficiency virus and men who have sex with men. Approximately 90% of anal squamous cell cancers are caused by infection with carcinogenic human papillomavirus (HPV). Given the common etiology between anal and cervical carcinogenesis, screening for anal cancer has been proposed in certain high-risk populations using strategies adapted from cervical cancer prevention. In this review, the authors discuss important differences in anal and cervical cancer regarding the populations at risk, disease natural history, and clinical procedures and outcomes that need to be considered when evaluating strategies for anal cancer screening. They also performed a systematic review and meta-analysis of the performance of anal cytology, anal HPV testing, and various biomarkers for the detection of anal precancers and cancers. The implications of these performance estimates are summarized in the context of risk-based screening and management of anal precancers, and important research gaps are highlighted that need to be addressed to fully understand the benefits and harms of anal cancer screening. Cancer Cytopathol 2018. Published 2018. This article is a U.S. Government work and is in the public domain in the USA.

Screening for human papillomavirus-driven oropharyngeal cancer: Considerations for feasibility and strategies for research

The incidence and burden of human papillomavirus (HPV)‐driven oropharyngeal cancer is expected to increase for decades, thus motivating discussions regarding possibilities for screening. This article addresses issues related to the validity and timeliness of screening for HPV‐driven oropharyngeal cancer, and raises important questions, highlights deficits and confusion in the existing literature, and proposes needed steps in the research agenda.

A novel metric that quantifies risk stratification for evaluating diagnostic tests: The example of evaluating cervical-cancer screening tests across populations

Our work involves assessing whether new biomarkers might be useful for cervical-cancer screening across populations with different disease prevalences and biomarker distributions. When comparing across populations, we show that standard diagnostic accuracy statistics (predictive values, risk-differences, Youden's index and Area Under the Curve (AUC)) can easily be misinterpreted. We introduce an intuitively simple statistic for a 2 × 2 table, Mean Risk Stratification (MRS): the average change in risk (pre-test vs. post-test) revealed for tested individuals. High MRS implies better risk separation achieved by testing. MRS has 3 key advantages for comparing test performance across populations with different disease prevalences and biomarker distributions. First, MRS demonstrates that conventional predictive values and the risk-difference do not measure risk-stratification because they do not account for test-positivity rates. Second, Youden's index and AUC measure only multiplicative relative gains in risk-stratification: AUC = 0.6 achieves only 20% of maximum risk-stratification (AUC = 0.9 achieves 80%). Third, large relative gains in risk-stratification might not imply large absolute gains if disease is rare, demonstrating a "high-bar" to justify population-based screening for rare diseases such as cancer. We illustrate MRS by our experience comparing the performance of cervical-cancer screening tests in China vs. the USA. The test with the worst AUC = 0.72 in China (visual inspection with acetic acid) provides twice the risk-stratification (i.e. MRS) of the test with best AUC = 0.83 in the USA (human papillomavirus and Pap cotesting) because China has three times more cervical precancer/cancer. MRS could be routinely calculated to better understand the clinical/public-health implications of standard diagnostic accuracy statistics.

Circulating inflammatory proteins and gallbladder cancer: Potential for risk stratification to improve prioritization for cholecystectomy in high-risk regions

BACKGROUND

Inflammatory proteins could help identify individuals most likely to have gallbladder cancer (GBC) among those waiting for cholecystectomy.

METHODS

We analyzed 49 circulating inflammation-related proteins in 144 patients with GBC and 150 patients with gallstones. We calculated age- and sex-adjusted odds ratios (ORs) and 95% CIs for protein quantiles and GBC versus gallstones. Using proteins associated with early GBC (stage 1-2) that were selected in stepwise logistic regression, we created an inflammation score and explored the potential utility for risk stratification.

RESULTS

26 proteins (53%) had P values for the trend across categories ≤0.001, with associations for a one category increase ranging from 1.52 (95% CI: 1.20-1.94) for CC motif ligand 4 to 4.00 (95% CI: 2.76-5.79) for interleukin (IL)-8. Soluble tumor necrosis factor receptor 2 (sTNFR2), IL-6, sTNFR1, CC motif ligand 20 (CCL20), vascular cell adhesion molecule 1, IL-16, and granulocyte colony-stimulating factor had P values ≤0.001 for early GBC. Of those, IL-6, IL-16, CCL20, and STNFR1 were included in the inflammation score. In a high-risk setting with a pre-test disease risk of 10% (e.g., elderly patients) and using an inflammation score cutoff that provides 90% sensitivity, 39% of patients on the waiting list would be predicted to be positive, and 23% of those would be predicted to have GBC.

CONCLUSION

These results highlight the strong associations of inflammatory proteins with GBC risk and their potential clinical utility. Larger studies are needed to identify the most effective combinations of inflammatory proteins for detecting early GBC and precursor lesions.

Sensitivity and specificity of oral HPV detection for HPV-positive head and neck cancer

BACKGROUND

The incidence of HPV-related head and neck squamous cell carcinoma (HPV-HNSCC) is increasing. Oral samples are easy and non-invasive to collect, but the diagnostic accuracy of oral HPV detection methods for classifying HPV-positive HNSCC tumors has not been well explored.

METHODS

In a systematic review, we identified eight studies of HNSCC patients meeting our eligibility criteria of having: (1) HPV detection in oral rinse or oral swab samples, (2) tumor HPV or p16 testing, (3) a publication date within the last 10 years (January 2007-May 2017, as laboratory methods change), and (4) at least 15 HNSCC cases. Data were abstracted from each study and a meta-analysis performed to calculate sensitivity and specificity.

RESULTS

Eight articles meeting inclusion criteria were identified. Among people diagnosed with HNSCC, oral HPV detection has good specificity (92%, 95% CI = 82-97%) and moderate sensitivity (72%, 95% CI = 45-89%) for HPV-positive HNSCC tumor. Results were similar when restricted to studies with only oropharyngeal cancer cases, with oral rinse samples, or testing for HPV16 DNA (instead of any oncogenic HPV) in the oral samples.

DISCUSSION

Among those who already have HNSCC, oral HPV detection has few false-positives but may miss one-half to one-quarter of HPV-related cases (false-negatives). Given these findings in cancer patients, the utility of oral rinses and swabs as screening tests for HPV-HNSCC among healthy populations is probably limited.

Identification of a Novel, EBV-Based Antibody Risk Stratification Signature for Early Detection of Nasopharyngeal Carcinoma in Taiwan

Background Epstein-Barr virus (EBV) is necessary for the development of nasopharyngeal carcinoma (NPC). By adulthood, approximately 90% of individuals test EBV-positive, but only a fraction develop cancer. Factors that identify which individuals are most likely to develop disease, including differential antibody response to the virus, could facilitate detection at early stages when treatment is most effective.Methods We measured anti-EBV IgG and IgA antibody responses in 607 Taiwanese individuals. Antibodies were measured using a custom protein microarray targeting 199 sequences from 86 EBV proteins. Variation in response patterns between NPC cases and controls was used to develop an antibody-based risk score for predicting NPC. The overall accuracy [area under the curve (AUC)] of this risk score, and its performance relative to currently used biomarkers, was evaluated in two independent Taiwanese cohorts.Findings Levels of 60 IgA and 73 IgG anti-EBV antibodies differed between stage I/IIa NPC cases and controls (P < 0.0002). Risk prediction analyses identified antibody targets that best discriminated NPC status-BXLF1, LF2,BZLF1, BRLF1, EAd, BGLF2, BPLF1, BFRF1, and BORF1. When combined with currently used VCA/EBNA1 IgA biomarkers, the resulting risk score predicted NPC with 93% accuracy (95% CI, 87%-98%) in the general Taiwanese population, a significant improvement beyond current biomarkers alone (82%; 95% CI, 75%-90%, P ≤ 0.01). This EBV-based risk score also improved NPC prediction in genetically high-risk families (89%; 95% CI, 82%-96%) compared with current biomarkers (78%; 95% CI, 66%-90%, P ≤ 0.03).Interpretation We identified NPC-related differences in 133 anti-EBV antibodies and developed a risk score using this microarray dataset that targeted immune responses against EBV proteins from all stages of the viral life cycle, significantly improving the ability to predict NPC. Clin Cancer Res; 24(6); 1305-14. ©2017 AACR.

From clinical epidemiology to practice recommendations: Knowledge gaps and uncertainty in the management of anal precancers

We discuss the existing data on anal precancers and demonstrate the impact of evidence gaps and uncertainty on a clinical decision model developed to provide clinical guidance for management of anal precancers.

Current Gaps in Ovarian Cancer Epidemiology: The Need for New Population-Based Research

With recent ovarian cancer screening studies showing no clinically significant mortality benefit, preventing this disease, identifying high-risk populations, and extending survival remain priorities. However, several challenges are impeding progress in ovarian cancer research. With most studies capturing exposure information from 10 or more years ago, evaluation of how changing patterns of exposures, such as new oral contraceptive formulations and increased intrauterine device use, might influence ovarian cancer risk and survival is difficult. Risk factors for ovarian cancer should be evaluated in the context of tumor histotypes, which have unique molecular features and cells of origin; this is a task that requires large collaborative studies to achieve meaningful sample sizes. Importantly, identification of novel modifiable risk factors, in addition to those currently known to reduce risk (eg, childbearing, tubal ligation, oral contraceptive use), is needed; this is not feasibly implemented at a population level. In this Commentary, we describe important gaps in knowledge and propose new approaches to advance epidemiologic research to improve ovarian cancer prevention and survival, including updated classification of tumors, collection of data on changing and novel exposures, longer follow-up on existing studies, evaluation of diverse populations, development of better risk prediction models, and collaborating prospectively with consortia to develop protocols for new studies that will allow seamless integration for future pooled analyses.

Plasma microRNAs as biomarkers of pancreatic cancer risk in a prospective cohort study

Noninvasive biomarkers for early pancreatic ductal adenocarcinoma (PDAC) diagnosis and disease risk stratification are greatly needed. We conducted a nested case-control study within the Prospective Investigation into Cancer and Nutrition (EPIC) cohort to evaluate prediagnostic microRNAs (miRs) as biomarkers of subsequent PDAC risk. A panel of eight miRs (miR-10a, -10b, -21-3p, -21-5p, -30c, -106b, -155 and -212) based on previous evidence from our group was evaluated in 225 microscopically confirmed PDAC cases and 225 controls matched on center, sex, fasting status and age/date/time of blood collection. MiR levels in prediagnostic plasma samples were determined by quantitative RT-PCR. Logistic regression was used to model levels and PDAC risk, adjusting for covariates and to estimate area under the receiver operating characteristic curves (AUC). Plasma miR-10b, -21-5p, -30c and -106b levels were significantly higher in cases diagnosed within 2 years of blood collection compared to matched controls (all p-values <0.04). Based on adjusted logistic regression models, levels for six miRs (miR-10a, -10b, -21-5p, -30c, -155 and -212) overall, and for four miRs (-10a, -10b, -21-5p and -30c) at shorter follow-up time between blood collection and diagnosis (≤5 yr, ≤2 yr), were statistically significantly associated with risk. A score based on the panel showed a linear dose-response trend with risk (p-value = 0.0006). For shorter follow-up (≤5 yr), AUC for the score was 0.73, and for individual miRs ranged from 0.73 (miR-212) to 0.79 (miR-21-5p).

Eurogin 2016 Roadmap: how HPV knowledge is changing screening practice

Human papillomaviruses (HPVs) are the necessary cause of most cervical cancers, a large proportion of other anogenital cancers, and a subset of oropharyngeal cancers. The knowledge about HPV has led to development of novel HPV-based prevention strategies with important impact on clinical and public health practice. Two complementary reviews have been prepared following the 2015 Eurogin Conference to evaluate how knowledge about HPV is changing practice in HPV infection and disease control through vaccination and screening. This review focuses on screening for cervical and anal cancers in increasingly vaccinated populations. The introduction of HPV vaccines a decade ago has led to reductions in HPV infections and early cancer precursors in countries with wide vaccination coverage. Despite the high efficacy of HPV vaccines, cervical cancer screening will remain important for many decades. Many healthcare systems are considering switching to primary HPV screening, which has higher sensitivity for cervical precancers and allows extending screening intervals. We describe different approaches to implementing HPV-based screening efforts in different healthcare systems with a focus in high-income countries. While the population prevalence for other anogenital cancers is too low for population-based screening, anal cancer incidence is very high in HIV-infected men who have sex with men, warranting consideration of early detection approaches. We summarize the current evidence on HPV-based prevention of anal cancers and highlight important evidence gaps.

Next Generation Sequencing of Circulating Cell-Free DNA for Evaluating Mutations and Gene Amplification in Metastatic Breast Cancer

BACKGROUND

Breast cancer tissues are heterogeneous and show diverse somatic mutations and somatic copy number alterations (CNAs). We used a novel targeted next generation sequencing (NGS) panel to examine cell-free DNA (cfDNA) to detect somatic mutations and gene amplification in women with metastatic breast cancer (MBC).

METHODS

cfDNA from pretreated patients (n = 42) and 9 healthy controls were compared with matched lymphocyte DNA by NGS, using a custom 158 amplicon panel covering hot-spot mutations and CNAs in 16 genes, with further validation of results by droplet digital PCR.

RESULTS

No mutations were identified in cfDNA of healthy controls, whereas exactly half the patients with metastatic breast cancer had at least one mutation or amplification in cfDNA (mean 2, range 1-6) across a total of 13 genes. Longitudinal follow up showed dynamic changes to mutations and gene amplification in cfDNA indicating clonal and subclonal response to treatment that was more dynamic than cancer antigen 15-3 (CA15-3). Interestingly, at the time of blood sampling disease progression was occurring in 7 patients with erb-b2 receptor tyrosine kinase 2 (ERBB2) gene amplification in their cfDNA and 3 of these patients were human epidermal growth factor receptor 2 (HER2) negative at diagnosis, suggesting clonal evolution to a more aggressive phenotype. Lastly, 6 patients harbored estrogen receptor 1 (ESR1) mutations in cfDNA, suggesting resistance to endocrine therapy. Overall 9 of 42 patients (21%) had alterations in cfDNA that could herald a change in treatment.

CONCLUSIONS

Targeted NGS of cfDNA has potential for monitoring response to targeted therapies through both mutations and gene amplification, for analysis of dynamic tumor heterogeneity and stratification to targeted therapy.

Identification of Circulating Tumor DNA for the Early Detection of Small-cell Lung Cancer

Circulating tumor DNA (ctDNA) is emerging as a key potential biomarker for post-diagnosis surveillance but it may also play a crucial role in the detection of pre-clinical cancer. Small-cell lung cancer (SCLC) is an excellent candidate for early detection given there are no successful therapeutic options for late-stage disease, and it displays almost universal inactivation of TP53. We assessed the presence of TP53 mutations in the cell-free DNA (cfDNA) extracted from the plasma of 51 SCLC cases and 123 non-cancer controls. We identified mutations using a pipeline specifically designed to accurately detect variants at very low fractions. We detected TP53 mutations in the cfDNA of 49% SCLC patients and 11.4% of non-cancer controls. When stratifying the 51 initial SCLC cases by stage, TP53 mutations were detected in the cfDNA of 35.7% early-stage and 54.1% late-stage SCLC patients. The results in the controls were further replicated in 10.8% of an independent series of 102 non-cancer controls. The detection of TP53 mutations in 11% of the 225 non-cancer controls suggests that somatic mutations in cfDNA among individuals without any cancer diagnosis is a common occurrence, and poses serious challenges for the development of ctDNA screening tests.

Risk assessment to guide cervical screening strategies in a large Chinese population

Three different cervical screening methods [cytology, human papillomavirus(HPV) testing and visual inspection with acetic acid(VIA)] are being considered in China for the national cervical screening program. Comparing risks of CIN3 and cervical cancer (CIN3+) for different results can inform test choice and management guidelines. We evaluated the immediate risk of CIN3+ for different screening results generated from individual and combined tests. We compared tests using a novel statistic designed for this purpose called Mean Risk Stratification (MRS), in a pooled analysis of 17 cross sectional population-based studies of 30,371 Chinese women screened with all 3 methods and diagnosed by colposcopically-directed biopsies. The 3 tests combined powerfully distinguished CIN3+ risk; triple-negative screening conferred a risk of 0.01%, while HPV-positive HSIL+ that was VIA-positive yielded a risk of 57.8%. Among the three screening tests, HPV status most strongly stratified CIN3+ risk. Among HPV-positive women, cytology was the more useful second test. In HPV-negative women, the immediate risks of CIN3+ ranged from 0.01% (negative cytology), 0.00% (ASC-US), 1.1% (LSIL), to 6.6 (HSIL+). In HPV-positive women, the CIN3+ risks were 0.9% (negative cytology), 3.6% (ASC-US), 6.3% (LSIL) and 38.5% (HSIL+). VIA results did not meaningful stratify CIN3+ risk among HPV-negative women with negative or ASC-US cytology; however, positive VIA substantially elevated CIN3+ risk for all other, more positive combinations of HPV and cytology compared with a negative VIA. Because all 3 screening tests had independent value in defining risk of CIN3+, different combinations can be optimized as pragmatic strategies in different resource settings.

A statistical framework to model the meeting-in-the-middle principle using metabolomic data: application to hepatocellular carcinoma in the EPIC study

Metabolomics is a potentially powerful tool for identification of biomarkers associated with lifestyle exposures and risk of various diseases. This is the rationale of the 'meeting-in-the-middle' concept, for which an analytical framework was developed in this study. In a nested case-control study on hepatocellular carcinoma (HCC) within the European Prospective Investigation into Cancer and nutrition (EPIC), serum (1)H nuclear magnetic resonance (NMR) spectra (800 MHz) were acquired for 114 cases and 222 matched controls. Through partial least square (PLS) analysis, 21 lifestyle variables (the 'predictors', including information on diet, anthropometry and clinical characteristics) were linked to a set of 285 metabolic variables (the 'responses'). The three resulting scores were related to HCC risk by means of conditional logistic regressions. The first PLS factor was not associated with HCC risk. The second PLS metabolomic factor was positively associated with tyrosine and glucose, and was related to a significantly increased HCC risk with OR = 1.11 (95% CI: 1.02, 1.22, P = 0.02) for a 1SD change in the responses score, and a similar association was found for the corresponding lifestyle component of the factor. The third PLS lifestyle factor was associated with lifetime alcohol consumption, hepatitis and smoking, and had negative loadings on vegetables intake. Its metabolomic counterpart displayed positive loadings on ethanol, glutamate and phenylalanine. These factors were positively and statistically significantly associated with HCC risk, with 1.37 (1.05, 1.79, P = 0.02) and 1.22 (1.04, 1.44, P = 0.01), respectively. Evidence of mediation was found in both the second and third PLS factors, where the metabolomic signals mediated the relation between the lifestyle component and HCC outcome. This study devised a way to bridge lifestyle variables to HCC risk through NMR metabolomics data. This implementation of the 'meeting-in-the-middle' approach finds natural applications in settings characterised by high-dimensional data, increasingly frequent in the omics generation.

Detectable Symptomatology Preceding the Diagnosis of Pancreatic Cancer and Absolute Risk of Pancreatic Cancer Diagnosis

The survival duration for pancreatic cancer is short. Given its low lifetime risk (1.5%), established factors for the disease have insufficient specificity to identify individuals at high risk of nonfamilial cancer, and prediagnostic signs and symptoms are vague and not limited to pancreatic causes. We considered whether statistical models that incorporated both risk factors and prediagnosis symptomatology could improve prediction enough to provide practical risk estimates. We combined US Surveillance Epidemiology and End Results (SEER) incidence data from 2008 to 2010 with regression models from representative case-control data from Connecticut (2005-2009) to estimate age- and sex-specific 5-year absolute risks of pancreatic cancer diagnosis. Our risk model included current cigarette smoking (adjusted odds ratio (OR) = 3.3, 95% confidence interval (CI): 2.1, 5.0), current use of proton pump-inhibitor antiheartburn medications (OR = 6.2, 95% CI: 1.7, 23), recent diagnosis of diabetes mellitus (OR = 4.8, 95% CI: 2.2, 11), recent diagnosis of pancreatitis (OR = 19, 95% CI: 3.1, 120), Jewish ancestry (OR = 1.8, 95% CI: 1.1, 3.1), and ABO blood group other than O (OR = 1.3, 95% CI: 1.0, 1.8). In total, 0.87% of controls with combinations of these factors had estimated 5-year absolute risks greater than 5%, and for some, the risks reached more than 10%. Combining risk factors for pancreatic cancer with detectable prediagnostic symptomatology can allow investigators to begin to identify small segments of the population with risks sufficiently high enough to make screening efforts among them potentially useful.

Invited Commentary: Clinical Utility of Prediction Models for Rare Outcomes--The Example of Pancreatic Cancer

Translating relative risk estimates into absolute risks is important in evaluating the potential clinical and public health relevance of etiologic discoveries. Predicting high absolute risk is challenging, particularly for rare endpoints such as pancreatic cancer. Recent efforts to develop risk prediction models for pancreatic cancer have found moderate risk levels for very small parts of the population. A new approach in which clinical symptoms and medication use are evaluated in addition to information on risk factors is presented by Risch et al. in this issue of the Journal (Am J Epidemiol. 2015;182(1):26-34). The authors estimated absolute risks based on the relative risks obtained from their case-control study. Their absolute risk estimates were higher than those from previous approaches but remained restricted to a very small proportion of the general population. In the present commentary, we address issues of absolute risk stratification (particularly for rare diseases), specific analytic methods, and how actionable information will differ based on the disease and possible intervention. We suggest that moving from cancer-specific models to broader models used to predict risk for multiple outcomes can make risk prediction for rare diseases more effective. When considering translational goals, it is important to estimate absolute risk at the early stages of etiologic research. The results can be sobering but allow focusing on the most promising goals.

HPV16 methyl-haplotypes determined by a novel next-generation sequencing method are associated with cervical precancer

We have developed and evaluated a next-generation bisulfite sequencing (NGS) assay to distinguish HPV16 cervical precancer (CIN2-3; N=59) from HPV16-positive transient infections (N=40). Cervical DNA was isolated and treated with bisulfite and HPV16 methylation was quantified by (i) amplification with barcoded primers and massively parallel single molecule sequencing and (ii) site-specific pyrosequencing. Assays were evaluated for agreement using intraclass correlation coefficients (ICC). Odds ratios (OR) for high methylation vs. low methylation were calculated. Single site pyrosequencing and NGS data were correlated (ICC=0.61) and both indicated hypermethylation was associated with precancer (ORs of 2-37). Concordant NGS and pyrosequencing results yieled ORs that were stronger when compared with using either assay separately. Within the L1 region, the ORs for CIN2-3 were 14.3 and 22.4 using pyrosequencing and NGS assays, respectively; when both methods agreed the OR was 153. NGS assays provide methylation haplotypes, termed methyl-haplotypes from single molecule reads: cases had increased methyl-haplotypes with ≥1 methylated CpG site(s) per fragment compared with controls, particularly in L1 (p=3.0×10(-8)). The maximum discrimination of cases from controls for a L1 methyl-haplotype had an AUC of 0.89 corresponding to a sensitivity of 92.5% and a specificity of 73.1%. The strengthening of the OR when the two assays were concordant suggests the true association of CpG methylation with precancer is stronger than with either assay. As cervical cancer prevention moves to DNA testing methods, DNA based biomarkers, such as HPV methylation could serve as a reflex strategy to identify women at high risk for cervix cancer.