Considerations for using potential surrogate endpoints in cancer screening trials

The requirement of large-scale expensive cancer screening trials spanning decades creates considerable barriers to the development, commercialisation, and implementation of novel screening tests. One way to address these problems is to use surrogate endpoints for the ultimate endpoint of interest, cancer mortality, at an earlier timepoint. This Review aims to highlight the issues underlying the choice and use of surrogate endpoints for cancer screening trials, to propose criteria for when and how we might use such endpoints, and to suggest possible candidates. We present the current landscape and challenges, and discuss lessons and shortcomings from the therapeutic trial setting. It is hugely challenging to validate a surrogate endpoint, even with carefully designed clinical studies. Nevertheless, we consider whether there are candidates that might satisfy the requirements defined by research and regulatory bodies.

Years of life lost due to cancer in the United Kingdom from 1988 to 2017

BACKGROUND

We investigated the application of years of life lost (YLL) in routine cancer statistics using cancer mortality data from 1988 to 2017.

METHODS

Cancer mortality data for 17 cancers and all cancers in the UK from 1988 to 2017 were provided by the UK Association of Cancer Registries by sex, 5-year age group, and year. YLL, age-standardised YLL rate (ASYR) and age-standardised mortality rate (ASMR) were estimated.

RESULTS

The annual average YLL due to cancer, in the time periods 1988-1992 and 2013-2017, were about 2.2 and 2.3 million years, corresponding to 4510 and 3823 ASYR per 100,000 years, respectively. During 2013-2017, the largest number of YLL occurred in lung, bowel and breast cancer. YLL by age groups for all cancers showed a peak between 60-64 and 75-79. The relative contributions to incidence, mortality, and YLL differ between cancers. For instance, pancreas (in women and men) made up a smaller proportion of incidence (3%) but bigger proportion of mortality (6 and 5%) and YLL (5 and 6%), whereas prostate cancer (26% of incidence) contributed 13% mortality and 9% YLL.

CONCLUSION

YLL is a useful measure of the impact different cancers have on society and puts a higher weight on cancer deaths in younger individuals.

Predictors of the experience of a Cytosponge test: analysis of patient survey data from the BEST3 trial

BACKGROUND

The Cytosponge is a cell-collection device, which, coupled with a test for trefoil factor 3 (TFF3), can be used to diagnose Barrett's oesophagus, a precursor condition to oesophageal adenocarcinoma. BEST3, a large pragmatic, randomised, controlled trial, investigated whether offering the Cytosponge-TFF3 test would increase detection of Barrett's. Overall, participants reported mostly positive experiences. This study reports the factors associated with the least positive experience.

METHODS

Patient experience was assessed using the Inventory to Assess Patient Satisfaction (IAPS), a 22-item questionnaire, completed 7-14 days after the Cytosponge test.

STUDY COHORT

All BEST3 participants who answered ≥ 15 items of the IAPS (N = 1458).

STATISTICAL ANALYSIS

A mean IAPS score between 1 and 5 (5 indicates most negative experience) was calculated for each individual. 'Least positive' experience was defined according to the 90th percentile. 167 (11.4%) individuals with a mean IAPS score of ≥ 2.32 were included in the 'least positive' category and compared with the rest of the cohort. Eleven patient characteristics and one procedure-specific factor were assessed as potential predictors of the least positive experience. Multivariable logistic regression analysis using backwards selection was conducted to identify factors independently associated with the least positive experience and with failed swallow at first attempt, one of the strongest predictors of least positive experience.

RESULTS

The majority of responders had a positive experience, with an overall median IAPS score of 1.7 (IQR 1.5-2.1). High (OR = 3.01, 95% CI 2.03-4.46, p < 0.001) or very high (OR = 4.56, 95% CI 2.71-7.66, p < 0.001) anxiety (relative to low/normal anxiety) and a failed swallow at the first attempt (OR = 3.37, 95% CI 2.14-5.30, p < 0.001) were highly significant predictors of the least positive patient experience in multivariable analyses. Additionally, sex (p = 0.036), height (p = 0.032), alcohol intake (p = 0.011) and education level (p = 0.036) were identified as statistically significant predictors.

CONCLUSION

We have identified factors which predict patient experience. Identifying anxiety ahead of the procedure and discussing particular concerns with patients or giving them tips to help with swallowing the capsule might help improve their experience. Trial registration ISRCTN68382401.

Non-speculum clinician-taken samples for human papillomavirus testing: a cross-sectional study in older women

BACKGROUND

Cervical cancer incidence and mortality are high in women aged ≥65 years, despite the disease being preventable by screening. Speculum-based screening can become more uncomfortable after the menopause.

AIM

To examine test performance and acceptability of human papillomavirus (HPV) testing on clinician-collected vaginal samples without a speculum (non-speculum).

DESIGN AND SETTING

Cross-sectional study in 11 GP practices and four colposcopy clinics in London, UK, between August 2017 and January 2019.

METHOD

Non-speculum and conventional (speculum) samples were collected from women aged ≥50 years attending for a colposcopy (following a speculum HPV-positive screening result) or women aged ≥35 years (with confirmed cervical intraepithelial neoplasia (CIN) 2+), and women aged 50-64 years attending routine screening. Sensitivity to CIN2+ was assessed among women with confirmed CIN2+ (colposcopy). Specificity to HPV relative to speculum sampling and overall concordance was assessed among women with negative cytology (routine screening).

RESULTS

The sensitivity of non-speculum sampling for detecting CIN2+ was 83.3% (95% confidence interval [CI] = 60.8 to 94.2) (n = 15/18). There was complete concordance among women with positive CIN2+ who had a speculum sample ≤91 days prior to the non-speculum sample (n = 12). Among 204 women with negative cytology, the specificity to HPV was 96.4% (95% CI = 92.7 to 98.5), with 96.6% concordant results (κ 72.4%). Seventy-one percent (n = 120/170) of women preferred a non-speculum sample for their next screen.

CONCLUSION

HPV testing on non-speculum clinician-taken samples is a viable approach that warrants further exploration in larger studies. Overall test performance was broadly comparable with that of self-sampling.

Use of a Cytosponge biomarker panel to prioritise endoscopic Barrett's oesophagus surveillance: a cross-sectional study followed by a real-world prospective pilot

BACKGROUND

Endoscopic surveillance is recommended for patients with Barrett's oesophagus because, although the progression risk is low, endoscopic intervention is highly effective for high-grade dysplasia and cancer. However, repeated endoscopy has associated harms and access has been limited during the COVID-19 pandemic. We aimed to evaluate the role of a non-endoscopic device (Cytosponge) coupled with laboratory biomarkers and clinical factors to prioritise endoscopy for Barrett's oesophagus.

METHODS

We first conducted a retrospective, multicentre, cross-sectional study in patients older than 18 years who were having endoscopic surveillance for Barrett's oesophagus (with intestinal metaplasia confirmed by TFF3 and a minimum Barrett's segment length of 1 cm [circumferential or tongues by the Prague C and M criteria]). All patients had received the Cytosponge and confirmatory endoscopy during the BEST2 (ISRCTN12730505) and BEST3 (ISRCTN68382401) clinical trials, from July 7, 2011, to April 1, 2019 (UK Clinical Research Network Study Portfolio 9461). Participants were divided into training (n=557) and validation (n=334) cohorts to identify optimal risk groups. The biomarkers evaluated were overexpression of p53, cellular atypia, and 17 clinical demographic variables. Endoscopic biopsy diagnosis of high-grade dysplasia or cancer was the primary endpoint. Clinical feasibility of a decision tree for Cytosponge triage was evaluated in a real-world prospective cohort from Aug 27, 2020 (DELTA; ISRCTN91655550; n=223), in response to COVID-19 and the need to provide an alternative to endoscopic surveillance.

FINDINGS

The prevalence of high-grade dysplasia or cancer determined by the current gold standard of endoscopic biopsy was 17% (92 of 557 patients) in the training cohort and 10% (35 of 344) in the validation cohort. From the new biomarker analysis, three risk groups were identified: high risk, defined as atypia or p53 overexpression or both on Cytosponge; moderate risk, defined by the presence of a clinical risk factor (age, sex, and segment length); and low risk, defined as Cytosponge-negative and no clinical risk factors. The risk of high-grade dysplasia or intramucosal cancer in the high-risk group was 52% (68 of 132 patients) in the training cohort and 41% (31 of 75) in the validation cohort, compared with 2% (five of 210) and 1% (two of 185) in the low-risk group, respectively. In the real-world setting, Cytosponge results prospectively identified 39 (17%) of 223 patients as high risk (atypia or p53 overexpression, or both) requiring endoscopy, among whom the positive predictive value was 31% (12 of 39 patients) for high-grade dysplasia or intramucosal cancer and 44% (17 of 39) for any grade of dysplasia.

INTERPRETATION

Cytosponge atypia, p53 overexpression, and clinical risk factors (age, sex, and segment length) could be used to prioritise patients for endoscopy. Further investigation could validate their use in clinical practice and lead to a substantial reduction in endoscopy procedures compared with current surveillance pathways.

FUNDING

Medical Research Council, Cancer Research UK, Innovate UK.

Non-speculum sampling approaches for cervical screening in older women: randomised controlled trial

BACKGROUND

Cervical cancer disproportionately affects women ≥65 years, especially those not screened regularly. Speculum use is a key barrier.

AIM

To assess if offering non-speculum clinician-taken sampling and self-sampling increases uptake for lapsed attenders aged 50-64 years.

DESIGN AND SETTING

Pragmatic randomised control trial conducted at 10 general practices in East London, UK.

METHOD

Participants were 784 women aged 50-64 years, last screened 6-15 years before randomisation. Intervention participants received a letter offering the choice of non-speculum clinician- or self-sampling. Control participants received usual care. The main outcome measure was uptake within 4 months.

RESULTS

Screening uptake 4 months after randomisation was significantly higher in the intervention arm: 20.4% (n = 80/393) versus 4.9% in the control arm (n = 19/391, absolute difference 15.5%, 95% confidence interval [CI] = 11.0% to 20.0%, P<0.001). This was maintained at 12 months: intervention 30.5% (n = 120/393) versus control 13.6% (n = 53/391) (absolute difference 17.0%, 95% CI = 11.3% to 22.7%, P<0.001). Conventional screening attendance within 12 months was very similar for both intervention 12.7% (n = 50/393) and control 13.6% (n = 53/391) arms. Ethnic differences were seen in screening modality preference. More White women opted for self-sampling (50.7%, n = 38/75), whereas most Asian and Black women and those from other ethnic backgrounds opted for conventional screening.

CONCLUSION

Offering non-speculum clinician-taken sampling and self-sampling substantially increases uptake in older lapsed attendee women. Non-speculum clinician sampling appeals to women who dislike the speculum but still prefer a clinician to take their sample. Providing a choice of screening modality may be important for optimising cervical screening uptake.

A case-control study to evaluate the impact of the breast screening programme on mortality in England

BACKGROUND

Over the past 30 years since the implementation of the National Health Service Breast Screening Programme, improvements in diagnostic techniques and treatments have led to the need for an up-to-date evaluation of its benefit on risk of death from breast cancer. An initial pilot case-control study in London indicated that attending mammography screening led to a mortality reduction of 39%.

METHODS

Based on the same study protocol, an England-wide study was set up. Women aged 47-89 years who died of primary breast cancer in 2010 or 2011 were selected as cases (8288 cases). When possible, two controls were selected per case (15,202 controls) and were matched by date of birth and screening area.

RESULTS

Conditional logistic regressions showed a 38% reduction in breast cancer mortality after correcting for self-selection bias (OR 0.62, 95% CI 0.56-0.69) for women being screened at least once. Secondary analyses by age group, and time between last screen and breast cancer diagnosis were also performed.

CONCLUSIONS

According to this England-wide case-control study, mammography screening still plays an important role in lowering the risk of dying from breast cancer. Women aged 65 or over see a stronger and longer lasting benefit of screening compared to younger women.

A state-wide population-based evaluation of cervical cancers arising during opportunistic screening in the United States

OBJECTIVE

Despite widespread cervical screening, an estimated 13,800 women will be diagnosed with cervical cancer in the United States in 2020. To inform improvements, the screening histories of women diagnosed with cervical cancer in New Mexico were assessed.

METHODS

Data were collected on all cervical screening, diagnostic tests and treatment procedures for all women diagnosed with cervical cancer aged 25-64 yrs. in New Mexico from 2006 to 2016. Women were categorized by their screening attendance in the 5-40 months (screening interval) and 1-4 months (peri-diagnostic interval) prior to cancer diagnosis.

RESULTS

Of the 504 women diagnosed between May 2009-December 2016, 64% were not screened or had only inadequate screening tests in the 5-40 months prior to diagnosis, and 90 of 182 screened women (49%) had only negative screens in this period. Only 32% (N = 162) of cervical cancers were screen-detected. Women with adenocarcinomas were more likely to have had a recent negative screen (41/57 = 722%) than women with squamous cancers (50/112 = 45%). Both older women (aged 45-64 years) and women with more advanced cancers were less likely to have been screened, and if screened, were more likely to have a false-negative outcome. Only 9% of cancers were diagnosed in women who did not attend biopsy or treatment after positive tests requiring clinical management. Screening currently prevents 35% of cancers, whereas full screening coverage could prevent 61% of cervical cancers.

CONCLUSION

Improved screening coverage has the largest potential for reducing cervical cancer incidence, though there is also a role for improved recall procedures and screening sensitivity.

Effect of mammographic screening from age 40 years on breast cancer mortality (UK Age trial): final results of a randomised, controlled trial

BACKGROUND

The appropriate age range for breast cancer screening remains a matter of debate. We aimed to estimate the effect of mammographic screening at ages 40-48 years on breast cancer mortality.

METHODS

We did a randomised, controlled trial involving 23 breast screening units across Great Britain. We randomly assigned women aged 39-41 years, using individual randomisation, stratified by general practice, in a 1:2 ratio, to yearly mammographic screening from the year of inclusion in the trial up to and including the calendar year that they reached age 48 years (intervention group), or to standard care of no screening until the invitation to their first National Health Service Breast Screening Programme (NHSBSP) screen at approximately age 50 years (control group). Women in the intervention group were recruited by postal invitation. Women in the control group were unaware of the study. The primary endpoint was mortality from breast cancers (with breast cancer coded as the underlying cause of death) diagnosed during the intervention period, before the participant's first NHSBSP screen. To study the timing of the mortality effect, we analysed the results in different follow-up periods. Women were included in the primary comparison regardless of compliance with randomisation status (intention-to-treat analysis). This Article reports on long-term follow-up analysis. The trial is registered with the ISRCTN registry, ISRCTN24647151.

FINDINGS

160 921 women were recruited between Oct 14, 1990, and Sept 24, 1997. 53 883 women (33·5%) were randomly assigned to the intervention group and 106 953 (66·5%) to the control group. Between randomisation and Feb 28, 2017, women were followed up for a median of 22·8 years (IQR 21·8-24·0). We observed a significant reduction in breast cancer mortality at 10 years of follow-up, with 83 breast cancer deaths in the intervention group versus 219 in the control group (relative rate [RR] 0·75 [95% CI 0·58-0·97]; p=0·029). No significant reduction was observed thereafter, with 126 deaths versus 255 deaths occurring after more than 10 years of follow-up (RR 0·98 [0·79-1·22]; p=0·86).

INTERPRETATION

Yearly mammography before age 50 years, commencing at age 40 or 41 years, was associated with a relative reduction in breast cancer mortality, which was attenuated after 10 years, although the absolute reduction remained constant. Reducing the lower age limit for screening from 50 to 40 years could potentially reduce breast cancer mortality.

FUNDING

National Institute for Health Research Health Technology Assessment programme.

Impact of screening on cervical cancer incidence: A population-based case-control study in the United States

Cervical cancer is widely preventable through screening, but little is known about the duration of protection offered by a negative screen in North America. A case-control study was conducted with records from population-based registries in New Mexico. Cases were women diagnosed with cervical cancer in 2006-2016, obtained from the Tumor Registry. Five controls per case from the New Mexico HPV Pap Registry were matched to cases by sex, age and place of residence. Dates and results of all cervical screening and diagnostic tests since 2006 were identified from the pap registry. We estimated the odds ratio of nonlocalized (Stage II+) and localized (Stage I) cervical cancer associated with attending screening in the 3 years prior to case-diagnosis compared to women not screened in 5 years. Of 876 cases, 527 were aged 25-64 years with ≥3 years of potential screening data. Only 38% of cases and 61% of controls attended screening in a 3-year period. Women screened in the 3 years prior to diagnosis had 83% lower risk of nonlocalized cancer (odds ratio [OR] = 0.17, 95% CI: 0.12-0.24) and 48% lower odds of localized cancer (OR = 0.52, 95% CI: 0.38-0.72), compared to women not screened in the 5 years prior to diagnosis. Women remained at low risk of nonlocalized cancer for 3.5-5 years after a negative screen compared to women with no negative screens in the 5 years prior to diagnosis. Routine cervical screening is effective at preventing localized and nonlocalized cervical cancers; 3 yearly screening prevents 83% of nonlocalized cancers, with no additional benefit of more frequent screening. Increasing screening coverage remains essential to further reduce cervical cancer incidence.

Absolute risks of cervical precancer among women who fulfill exiting guidelines based on HPV and cytology cotesting

US guidelines recommend that most women older than 65 years cease cervical screening after two consecutive negative cotests (concurrent HPV and cytology tests) in the previous 10 years, with one in the last 5 years. However, this recommendation was based on expert opinion and modeling rather than empirical data on cancer risk. We therefore estimated the 5-year risks of cervical precancer (cervical intraepithelial neoplasia grade 3 or adenocarcinoma in situ [CIN3]) after one, two and three negative cotests among 346,760 women aged 55-64 years undergoing routine cotesting at Kaiser Permanente Northern California (2003-2015). Women with a history of excisional treatment or CIN2+ were excluded. No woman with one or more negative cotests was diagnosed with cancer during follow-up. Five-year risks of CIN3 after one, two, and three consecutive negative cotests were 0.034% (95% CI: 0.023%-0.046%), 0.041% (95% CI: 0.007%-0.076%) and 0.016% (95% CI: 0.000%-0.052%), respectively (ptrend < 0.001). These risks did not appreciably differ by a positive cotest result prior to the one, two or three negative cotest(s). Since CIN3 risks after one or more negative cotests were significantly below a proposed 0.12% CIN3+ risk threshold for a 5-year screening interval, a longer screening interval in these women is justified. However, the choice of how many negative cotests provide sufficient safety against invasive cancer over a woman's remaining life represents a value judgment based on the harms versus benefits of continued screening. Ideally, this guideline should be informed by longer-term follow-up given that exiting is a long-term decision.

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

Electronic health-records (EHR) are increasingly used by epidemiologists studying disease following surveillance testing to provide evidence for screening intervals and referral guidelines. Although cost-effective, undiagnosed prevalent disease and interval censoring (in which asymptomatic disease is only observed at the time of testing) raise substantial analytic issues when estimating risk that cannot be addressed using Kaplan-Meier methods. Based on our experience analysing EHR from cervical cancer screening, we previously proposed the logistic-Weibull model to address these issues. Here we demonstrate how the choice of statistical method can impact risk estimates. We use observed data on 41,067 women in the cervical cancer screening program at Kaiser Permanente Northern California, 2003-2013, as well as simulations to evaluate the ability of different methods (Kaplan-Meier, Turnbull, Weibull and logistic-Weibull) to accurately estimate risk within a screening program. Cumulative risk estimates from the statistical methods varied considerably, with the largest differences occurring for prevalent disease risk when baseline disease ascertainment was random but incomplete. Kaplan-Meier underestimated risk at earlier times and overestimated risk at later times in the presence of interval censoring or undiagnosed prevalent disease. Turnbull performed well, though was inefficient and not smooth. The logistic-Weibull model performed well, except when event times didn't follow a Weibull distribution. We have demonstrated that methods for right-censored data, such as Kaplan-Meier, result in biased estimates of disease risks when applied to interval-censored data, such as screening programs using EHR data. The logistic-Weibull model is attractive, but the model fit must be checked against Turnbull non-parametric risk estimates.

What cervical screening is appropriate for women who have been vaccinated against high risk HPV? A simulation study

Women vaccinated against HPV16/18 are approaching the age for cervical screening; however, an updated screening algorithm has not been agreed. We use a microsimulation model calibrated to real published data to determine the appropriate screening intensity for vaccinated women. Natural histories in the absence of vaccination were simulated for 300,000 women using 10,000 sets of transition probabilities. Vaccination with (i) 100% efficacy against HPV16/18, (ii) 15% cross-protection, (iii) 22% cross-protection, (iv) waning vaccine efficacy and (v) 100% efficacy against HPV16/18/31/33/45/52/58 was added, as were a range of screening scenarios appropriate to the UK. To benchmark cost-benefits of screening for vaccinated women, we evaluated the proportion of cancers prevented per additional screen (incremental benefit) of current cytology and likely HPV screening scenarios in unvaccinated women. Slightly more cancers are prevented through vaccination with no screening (70.3%, 95% CR: 65.1-75.5) than realistic compliance to the current UK screening programme in the absence of vaccination (64.3%, 95% CR: 61.3-66.8). In unvaccinated women, when switching to HPV primary testing, there is no loss in effectiveness when doubling the screening interval. Benchmarking supports screening scenarios with incremental benefits of ≥2.0%, and rejects scenarios with incremental benefits ≤0.9%. In HPV16/18-vaccinated women, the incremental benefit of offering a third lifetime screen was at most 3.3% (95% CR: 2.2-4.5), with an incremental benefit of 1.3% (-0.3-2.8) for a fourth screen. For HPV16/18/31/33/45/52/58-vaccinated women, two lifetime screens are supported. It is important to know women's vaccination status; in these simulations, HPV16/18-vaccinated women require three lifetime screens, HPV16/18/31/33/45/52/58-vaccinated women require two lifetime screens, yet unvaccinated women require seven lifetime screens.

Risk stratification of Barrett's oesophagus using a non-endoscopic sampling method coupled with a biomarker panel: a cohort study

BACKGROUND

Barrett's oesophagus predisposes to adenocarcinoma. However, most patients with Barrett's oesophagus will not progress and endoscopic surveillance is invasive, expensive, and fraught by issues of sampling bias and the subjective assessment of dysplasia. We investigated whether a non-endoscopic device, the Cytosponge, could be coupled with clinical and molecular biomarkers to identify a group of patients with low risk of progression suitable for non-endoscopic follow-up.

METHODS

In this multicentre cohort study (BEST2), patients with Barrett's oesophagus underwent the Cytosponge test before their surveillance endoscopy. We collected clinical and demographic data and tested Cytosponge samples for a molecular biomarker panel including three protein biomarkers (P53, c-Myc, and Aurora kinase A), two methylation markers (MYOD1 and RUNX3), glandular atypia, and TP53 mutation status. We used a multivariable logistic regression model to compute the conditional probability of dysplasia status. We selected a simple model with high classification accuracy and applied it to an independent validation cohort. The BEST2 study is registered with ISRCTN, number 12730505.

FINDINGS

The discovery cohort consisted of 468 patients with Barrett's oesophagus and intestinal metaplasia. Of these, 376 had no dysplasia and 22 had high-grade dysplasia or intramucosal adenocarcinoma. In the discovery cohort, a model with high classification accuracy consisted of glandular atypia, P53 abnormality, and Aurora kinase A positivity, and the interaction of age, waist-to-hip ratio, and length of the Barrett's oesophagus segment. 162 (35%) of 468 of patients fell into the low-risk category and the probability of being a true non-dysplastic patient was 100% (99% CI 96-100) and the probability of having high-grade dysplasia or intramucosal adenocarcinoma was 0% (0-4). 238 (51%) of participants were classified as of moderate risk; the probability of having high-grade dysplasia was 14% (9-21). 58 (12%) of participants were classified as high-risk; the probability of having non-dysplastic endoscopic biopsies was 13% (5-27), whereas the probability of having high-grade dysplasia or intramucosal adenocarcinoma was 87% (73-95). In the validation cohort (65 patients), 51 were non-dysplastic and 14 had high-grade dysplasia. In this cohort, 25 (38%) of 65 patients were classified as being low-risk, and the probability of being non-dysplastic was 96·0% (99% CI 73·80-99·99). The moderate-risk group comprised 27 non-dysplastic and eight high-grade dysplasia cases, whereas the high-risk group (8% of the cohort) had no non-dysplastic cases and five patients with high-grade dysplasia.

INTERPRETATION

A combination of biomarker assays from a single Cytosponge sample can be used to determine a group of patients at low risk of progression, for whom endoscopy could be avoided. This strategy could help to avoid overdiagnosis and overtreatment in patients with Barrett's oesophagus.

FUNDING

Cancer Research UK.

Range of pathologies diagnosed using a minimally invasive capsule sponge to evaluate patients with reflux symptoms

AIMS

Reflux symptoms are highly prevalent and non-specific; hence, in the absence of alarm symptoms, endoscopy referral decisions are challenging. This study evaluated whether a non-endoscopic Cytosponge could detect benign oesophageal pathologies and thus have future potential in triaging patients with persistent symptoms.

METHODS AND RESULTS

Two complementary cohorts were recruited: (i) patients with reflux symptoms and no prior endoscopy (n = 409), and (ii) patients with reflux symptoms referred for endoscopy (n = 411). All patients were investigated using the Cytosponge and endoscopy. Significant epithelial inflammation was present in 130 (16%) Cytosponge samples, 32 of which had ulcer slough. Candida and significant inflammation was detected in a further 22 (2.3%) cases; epithelial infiltration with >15 eosinophils/high-power field reflecting possible eosinophilic oesophagitis (EOE) in five (0.6%); and viral inclusions suggestive of herpes oesophagitis in one (0.1%). No significant pathology was detected in the majority, 662 (81%), of Cytosponge samples. Cytosponge and endoscopy findings were in agreement in 574 (70%) cases, in 165 (67%) of the discordant cases one investigation showed mild inflammation while the other was negative, with an additional 22 (8.9%) differing on the extent of inflammation. Eighteen cases with severe inflammation, six with candida and two with EOE were detected only at endoscopy, while 18 with candida and significant inflammation, 13 with ulcer slough, one probable EOE and one viral oesophagitis were identified on the Cytosponge only.

CONCLUSIONS

The Cytosponge detects a range of benign oesophageal pathologies, and therefore has potential clinical utility in the triaging of patients with troublesome reflux symptoms. This warrants further investigation.

Is cervical screening preventing adenocarcinoma and adenosquamous carcinoma of the cervix?

While the incidence of squamous carcinoma of the cervix has declined in countries with organised screening, adenocarcinoma has become more common. Cervical screening by cytology often fails to prevent adenocarcinoma. Using prospectively recorded cervical screening data in England and Wales, we conducted a population-based case-control study to examine whether cervical screening leads to early diagnosis and down-staging of adenocarcinoma. Conditional logistic regression modelling was carried out to provide odds ratios (ORs) and 95% confidence intervals (CIs) on 12,418 women with cervical cancer diagnosed between ages 30 and 69 and 24,453 age-matched controls. Of women with adenocarcinoma of the cervix, 44.3% were up to date with screening and 14.6% were non-attenders. The overall OR comparing women up to date with screening with non-attenders was 0.46 (95% CI: 0.39-0.55) for adenocarcinoma. The odds were significantly decreased (OR: 0.22, 95% CI: 0.15-0.33) in up to date women with Stage 2 or worse adenocarcinoma, but not for women with Stage1A adenocarcinoma 0.71 (95% CI: 0.46-1.09). The odds of Stage 1A adenocarcinoma was double among lapsed attenders (OR: 2.35, 95% CI: 1.52-3.62) compared to non-attenders. Relative to women with no negative cytology within 7 years of diagnosis, women with Stage1A adenocarcinoma were very unlikely to be detected within 3 years of a negative cytology test (OR: 0.08, 95% CI: 0.05-0.13); however, the odds doubled 3-5 years after a negative test (OR: 2.30, 95% CI: 1.67-3.18). ORs associated with up to date screening were smaller for squamous and adenosquamous cervical carcinoma. Although cytology screening is inefficient at preventing adenocarcinomas, invasive adenocarcinomas are detected earlier than they would be in the absence of screening, substantially preventing Stage 2 and worse adenocarcinomas.

Evaluating cytology for the detection of invasive cervical cancer

OBJECTIVE

To assess the sensitivity, the number needed to screen (NNS) and the positive predictive value (PPV) of cervical cytology for the diagnosis of cancer by age in a screening population.

METHODS

A retrospective cohort of women with invasive cervical cancer nested within a census of cervical cytology. All (c. 8 million) women aged 20-64 years with cervical cytology (excluding tests after an earlier abnormality). From April 2007 to March 2010, 3372 women had cervical cancer diagnosed within 12 months of such cytology in England. The sensitivity of cervical cytology to cancer, NNS to detect one cancer and predictive values of cytology were calculated for various 'referral' thresholds. These were calculated for ages 20-24, 25-34, 35-49 and 50-64 years.

RESULTS

The sensitivity of at least moderate dyskaryosis [equivalent to a high-grade squamous intraepithelial lesion (HSIL) or worse] for cancer of 89.4% [95% confidence interval (CI) 88.3-90.4%] in women offered screening was independent of age. At all ages, women with borderline-early recall or mild dyskaryosis on cytology (equivalent to ASC-US and LSIL, respectively, in the Bethesda system) had a similar risk of cervical cancer to the risk in all women tested. The PPV of severe dyskaryosis/?invasive and ?glandular neoplasia cytology (equivalent to squamous cell carcinoma and adenocarcinoma/adenocarcinoma in situ, respectively, in the Bethesda System) were 34% and 12%, respectively; the PPV of severe dyskaryosis (HSIL: severe dysplasia) was 4%. The NNS was lowest when the incidence of cervical cancer was highest, at ages 25-39 years, but the proportion of those with abnormal cytology who have cancer was also lowest in younger women.

CONCLUSIONS

The PPV of at least severe dyskaryosis (HSIL: severe dysplasia) for cancer was 4-10% of women aged 25-64 years, justifying a 2-week referral to colposcopy and demonstrating the importance of failsafe monitoring for such patients. The sensitivity of cytology for cervical cancer was excellent across all age groups.

Impact of Screening on Breast Cancer Mortality: The UK Program 20 Years On

BACKGROUND

With changes in diagnosis, treatment, and management of breast cancer since the mammography screening trials, there is a need to evaluate contemporary breast screening programs. A case-control study was set up to assess the current impact of attendance in the English Breast Screening Program on breast cancer mortality.

METHODS

Cancer registry cases who died from primary breast cancer ages 47 to 89 years in London in 2008 to 2009 (869 women) were matched to 1 or 2 general population controls (1,642 women) with no diagnosis of breast cancer at the time of the case's diagnosis, who were alive at the case's death. Cases and controls were matched for date of birth and screening area, and had been invited to breast screening at least once prior to the case's diagnosis. ORs were estimated using conditional logistic regression. Self-selection bias was addressed using contemporaneous attendance at the cervical screening program. Sensitivity analyses were undertaken to assess the likely effect of lead time bias.

RESULTS

Attendance at breast screening resulted in a breast cancer mortality reduction of 39% [OR, 0.61; 95% confidence interval (CI), 0.44-0.85] after self-selection correction. Attendance in the last 3 years prior to diagnosis resulted in a 60% mortality reduction (OR, 0.40; 95% CI, 0.31-0.51). Lead time bias effects were negligible.

CONCLUSION

Our results suggest that community breast screening programs provide their expected benefit in terms of reducing the risk of breast cancer death among women participating.

IMPACT

Mammography is an important tool for reducing breast cancer mortality and its impact could be increased by encouraging regular attendance.

Explaining the Better Prognosis of Screening-Exposed Breast Cancers: Influence of Tumor Characteristics and Treatment

BACKGROUND

In England, population mammographic screening has been offered to women for over 20 years. Overall decrease in breast cancer mortality rates and improvements in cancer awareness and organization of medical care over this period call for a more current evaluation of the mediators behind the better prognosis of screening-exposed breast cancers.

METHODS

A case-control study was conducted within the English National Breast Screening Program. Women who died from primary breast cancer in 2008 to 2009 were matched (by year of birth, screening invitation, and area) to controls that received a diagnosis of invasive breast cancer at the time of the case diagnosis but survived the case death. Data were analyzed by unconditional logistic regression with adjustment for matching factors.

RESULTS

The unadjusted OR for dying from breast cancer associated with ever having attended breast screening was 0.44 [95% confidence interval (CI), 0.33-0.58]. After adjustment for lead time, overdiagnosis, and self-selection, the OR increased to 0.69 (95% CI, 0.50-0.94). Adjusting for tumor size, lymph node status, stage, grade, histopathology, and laterality accounted for all the screening effect (OR, 1.00; 95% CI, 0.71-1.40). Further adjustment for treatment factors only had a minimal impact on the OR (OR, 1.02; 95% CI, 0.72-1.45).

CONCLUSIONS

Our results suggest that earlier diagnosis, as reflected by tumor characteristics, remains the major mediator of the improvement in breast cancer survival due to participation in mammographic screening.

IMPACT

Mammographic screening continues to prevent breast cancer-related deaths in the epoch of adjuvant systemic therapy.

Cervical cytology and the diagnosis of cervical cancer in older women

OBJECTIVES

Most non-screen-detected cervical cancers are advanced stage. We assess the potential for cytology to expedite diagnosis when used outside of routine call and recall screening for cervical cancer.

METHODS

Two cohorts of women with cytology that did not appear to have been taken as part of routine screening, nested within a census of cervical cytology, in England between April 2007 and March 2010 were studied: 93,322 women aged 40-69 at first cytology, and 14,668 women aged ≥70. The diagnostic performance of high grade cervical squamous intraepithelial lesion (HSIL) or worse cytology was estimated. We also estimated case-fatality from stage distribution in women aged ≥66 with and without cytology in the year prior to diagnosis.

RESULTS

There were 259 cancers diagnosed in women aged 40-69 at first cytology, and 78 in women aged ≥70. The sensitivity of cytology ≥ HSIL for cancer was 89% and 83% respectively, and the number of women needed to test to identify one cancer was 404 (95% confidence interval [CI]: 355-462) and 226 (95% CI: 177-292) respectively. Women aged ≥66 with cytology within a year of diagnosis had earlier stage cancers than those without, corresponding to a 17-22% reduction in case fatality.

CONCLUSIONS

Cervical cytology is an excellent identifier of cancer among women tested outside routine screening call and recall. Its use as a triage tool, for instance in women with vague gynaecological symptoms, could facilitate earlier stage diagnosis and reduce cervical cancer mortality.

Trends in the lifetime risk of developing cancer in Great Britain: comparison of risk for those born from 1930 to 1960

BACKGROUND

Typically, lifetime risk is calculated by the period method using current risks at different ages. Here, we estimate the probability of being diagnosed with cancer for individuals born in a given year, by estimating future risks as the cohort ages.

METHODS

We estimated the lifetime risk of cancer in Britain separately for men and women born in each year from 1930 to 1960. We projected rates of all cancers (excluding non-melanoma skin cancer) and of all cancer deaths forwards using a flexible age-period-cohort model and backwards using age-specific extrapolation. The sensitivity of the estimated lifetime risk to the method of projection was explored.

RESULTS

The lifetime risk of cancer increased from 38.5% for men born in 1930 to 53.5% for men born in 1960. For women it increased from 36.7 to 47.5%. Results are robust to different models for projections of cancer rates.

CONCLUSIONS

The lifetime risk of cancer for people born since 1960 is >50%. Over half of people who are currently adults under the age of 65 years will be diagnosed with cancer at some point in their lifetime.

Evaluation of a minimally invasive cell sampling device coupled with assessment of trefoil factor 3 expression for diagnosing Barrett's esophagus: a multi-center case-control study

BACKGROUND

Barrett's esophagus (BE) is a commonly undiagnosed condition that predisposes to esophageal adenocarcinoma. Routine endoscopic screening for BE is not recommended because of the burden this would impose on the health care system. The objective of this study was to determine whether a novel approach using a minimally invasive cell sampling device, the Cytosponge, coupled with immunohistochemical staining for the biomarker Trefoil Factor 3 (TFF3), could be used to identify patients who warrant endoscopy to diagnose BE.

METHODS AND FINDINGS

A case-control study was performed across 11 UK hospitals between July 2011 and December 2013. In total, 1,110 individuals comprising 463 controls with dyspepsia and reflux symptoms and 647 BE cases swallowed a Cytosponge prior to endoscopy. The primary outcome measures were to evaluate the safety, acceptability, and accuracy of the Cytosponge-TFF3 test compared with endoscopy and biopsy. In all, 1,042 (93.9%) patients successfully swallowed the Cytosponge, and no serious adverse events were attributed to the device. The Cytosponge was rated favorably, using a visual analogue scale, compared with endoscopy (p < 0.001), and patients who were not sedated for endoscopy were more likely to rate the Cytosponge higher than endoscopy (Mann-Whitney test, p < 0.001). The overall sensitivity of the test was 79.9% (95% CI 76.4%-83.0%), increasing to 87.2% (95% CI 83.0%-90.6%) for patients with ≥3 cm of circumferential BE, known to confer a higher cancer risk. The sensitivity increased to 89.7% (95% CI 82.3%-94.8%) in 107 patients who swallowed the device twice during the study course. There was no loss of sensitivity in patients with dysplasia. The specificity for diagnosing BE was 92.4% (95% CI 89.5%-94.7%). The case-control design of the study means that the results are not generalizable to a primary care population. Another limitation is that the acceptability data were limited to a single measure.

CONCLUSIONS

The Cytosponge-TFF3 test is safe and acceptable, and has accuracy comparable to other screening tests. This test may be a simple and inexpensive approach to identify patients with reflux symptoms who warrant endoscopy to diagnose BE.

An ongoing case-control study to evaluate the NHS Bowel Cancer Screening Programme

BACKGROUND

Colorectal cancer is the third most common cause of cancer death in both males and females in England. A national bowel cancer screening programme was rolled out in England between 2006 and 2010. In the post-randomised controlled trials epoch, assessment of the impact of the programme using observational studies is needed. This study protocol was set up at the request of the UK Policy Research Unit in Cancer Awareness, Screening and Early Diagnosis to evaluate the effect of the current bowel cancer screening programme on incidence of advanced primary colorectal cancer.

METHODS/DESIGN

All incident cases of primary colorectal cancer in England will be included. Cases will be matched to controls with respect to sex, age, area of registration and year of first invitation to screening. Each evaluation round will cover a 2-year period, starting from January 2012, and ongoing thereafter. In the first instance, a pilot will be carried out in a single region. Variables related to colorectal tumour pathology will be obtained to enable selection and matching of cases and controls, and to allow analyses stratification by anatomical subsite within the bowel. Cases at Duke's stage B or worse will be considered as "advanced stage". The influence of sex will also be investigated. The incidence ratio observed in randomised controlled trials between controls (not invited) and non-attender invitees will be used to correct for self-selection bias overall. Screening participation at other national screening programmes (cervical, breast) will also be collected to derive a more contemporaneous adjustment factor for self-selection bias and assess consistency in self-selection correction in female patients.Full ethical approval was obtained from the Health Research Authority.

DISCUSSION

The case-control design is potentially prone to a number of biases. The size of the planned study, the design specifications and the development of analytical strategies to cope with bias should enable us to obtain accurate estimates of reduction in incidence of advanced stage disease. The results of analyses by sex and anatomical subsite may highlight the potential need for sex-specific recommendations in the programme.

Pregnancy outcomes after treatment for cervical intraepithelial neoplasia in a single NHS hospital

OBJECTIVE

The objective of this study was to assess the adverse pregnancy outcomes in women who had treatment for cervical intraepithelial neoplasia.

METHODS

This was a retrospective cohort using data linkage. Pathology databases from Whipps Cross University Hospital were used to identify women with a histological sample taken at colposcopy between 1995 and 2009. Births for these women were identified through the hospitals' obstetric database. A total of 876 births (from 721 women) were identified. Logistic regression was used to assess the relationship between adverse pregnancy outcomes and treatment for cervical intraepithelial neoplasia before delivery. Results were adjusted by ethnicity, deprivation, and parity.

RESULTS

After taking into account parity, socioeconomic status, and ethnicity, receiving any type of excisional treatment (single or multiple) before birth increased the risk of preterm labor compared with having a punch biopsy only (adjusted relative risk, 1.61; 95% confidence interval, 1.11-2.32). Preterm deliveries that occurred after a spontaneous onset of labor were found to be more likely after treatment for cervical disease (adjusted relative risk, 1.68; 95% confidence interval, 1.11-2.52).

CONCLUSIONS

Women receiving any type of excisional treatment before delivery are at increased risk of preterm delivery when compared with women attending colposcopy but not treated. Although we took into account the effects of parity, socioeconomic status, and ethnicity, residual confounding factors may be unidentified.

An ongoing case-control study to evaluate the NHS breast screening programme

Background

In England, a national breast screening programme (NHSBSP) has been in place since 1988, and assessment of its impact on breast cancer incidence and mortality is essential to ensure that the programme is indeed doing more good than harm. This article describes large observation studies designed to estimate the effects of the current programme in terms of the benefits on breast cancer incidence and mortality and detrimental effect in terms of overdiagnosis. The case-control design of the cervical screening programme evaluation was highly effective in informing policy on screening intervals and age ranges. We propose innovative selection of cases and controls and gathering of additional variables to address new outcomes of interest and develop new methodologies to control for potential sources of bias.

Methods/Design

Traditional case-control evaluation of breast screening uses women who have died from breast cancer as cases, and women known to be alive at the time of case death as controls. Breast screening histories prior to the cases’ date of first diagnosis are compared. If breast screening is preventing mortality from breast cancer, cases will be characterised by a lesser screening history than controls. All deaths and incident cases of primary breast cancer in England within each 2-year study period will be included in this ongoing evaluation. Cases will be age- and area-matched to controls and variables related to cancer treatment and breast tumour pathology will be obtained to investigate the interplay between screening and treatment, and the effect of screening on incidence of advanced stage disease. Screening attendance at other national screening programmes will also be collected to derive superior adjustment for self-selection bias.

The study is registered and has received full ethics approval.

What is the lifetime risk of developing cancer?: the effect of adjusting for multiple primaries

BACKGROUND

The 'lifetime risk' of cancer is generally estimated by combining current incidence rates with current all-cause mortality ('current probability' method) rather than by describing the experience of a birth cohort. As individuals may get more than one type of cancer, what is generally estimated is the average (mean) number of cancers over a lifetime. This is not the same as the probability of getting cancer.

METHODS

We describe a method for estimating lifetime risk that corrects for the inclusion of multiple primary cancers in the incidence rates routinely published by cancer registries. The new method applies cancer incidence rates to the estimated probability of being alive without a previous cancer. The new method is illustrated using data from the Scottish Cancer Registry and is compared with 'gold-standard' estimates that use (unpublished) data on first primaries.

RESULTS

The effect of this correction is to make the estimated 'lifetime risk' smaller. The new estimates are extremely similar to those obtained using incidence based on first primaries. The usual 'current probability' method considerably overestimates the lifetime risk of all cancers combined, although the correction for any single cancer site is minimal.

CONCLUSION

Estimation of the lifetime risk of cancer should either be based on first primaries or should use the new method.

A deterministic model for estimating the reduction in colorectal cancer incidence due to endoscopic surveillance

There is evidence that the removal of adenomas, by endoscopy, from the large bowel can prevent the occurrence of colorectal cancer (CRC). However, the reduction in cancer incidence due to endoscopic surveillance is difficult to estimate. Studies of cohorts of adenoma patients typically rely on comparisons with groups of historical controls. We present a model for disease progression which enables estimation of this quantity without direct comparison to a reference group. Models are applied to data from the National Polyp Study. Rates of adenoma recurrence and progression to carcinoma are estimated based on study data and relevant literature. This allows calculation of the number of cancers expected in the absence of surveillance and, thus, the number of cancers prevented. Results are compared with the original analysis. Models estimate that surveillance reduced CRC incidence by at least 97% in this cohort. The majority of the effect was due to the initial removal of adenomas rather than the follow-up surveillance. These results are similar to those produced in the original analysis when using the most appropriate reference groups. They indicate that polypectomy and follow-up surveillance can lead to large reductions in cancer incidence which may have been under-estimated in previous studies.

Avoidance of premature death: a new definition for the proportion cured

For many cancers, five-year survival is used as a synonym for cure. For some sites, such as breast, this is inappropriate since there is still considerable excess mortality 5-15 years following diagnosis. We propose using the probability of not dying prematurely as a surrogate for the probability of being cured. This is estimated without the need for information on the cause of death by comparing all-cause-mortality rates in patients with cancer to those in the general population. We also consider the probability of 'cure' in those who have survived a certain number of years since diagnosis. These quantities are estimated for various cancer sites using SEER data. The proportion not dying prematurely varies considerably for common cancers from 12% for lung cancer to 84% for prostate cancer. In women, the percentage 'cured' for colon and breast cancer are very similar at 56% and 58% respectively, but in those who survive 5 years, they are quite different--91% and 78% respectively. The proposed statistic is useful. For diseases in which excess mortality is primarily within five years of diagnosis, it agrees well with 5-year relative survival. Interpretation of differences between countries or over time is complicated and requires consideration of incidence and mortality data.

Human papillomavirus screening and cervical cancer prevention

Infection with one of several types of human papillomavirus (HPV) appears to be a necessary first step in the development of invasive cervical cancer. We cannot currently treat HPV infections; thus, the role of HPV testing is to identify women with precancerous lesions that can be removed and, in so doing, prevent progression to invasive carcinoma. Although HPV testing may help to identify women at risk of cervical cancer who might be missed by other screening tests, it is inherently nonspecific at identifying those who would otherwise develop cervical cancer. In order to avoid overtreatment of women with minor lesions with little potential for progression, HPV testing needs to be repeated or combined with Pap smears. Protocols for HPV screening have yet to be properly evaluated. Here we consider several possible applications of HPV testing in the prevention of cervical cancer. The most immediate role is as a secondary test in women with minor cytological abnormalities. Appropriate use of HPV testing as a primary screening tool depends on the setting. In a developed country without an organized screening program, HPV testing might be used in addition to Pap smears in women age 35 and over to increase sensitivity. Within an organized screening program, HPV testing might be used in combination with Pap testing, but with extended screening intervals so as to obtain the maximum advantage to women without unduly increasing costs. Where resources are strictly limited, an attractive option would be to perform visual inspection of the cervix after application of dilute acetic acid using a low threshold for referral, and to test for HPV only on those with abnormal looking lesions.

Standardized lifetime risk

The authors propose the use of two new standardized measures of risk, the standardized lifetime risk and the standardized number of years of life lost. These measures maintain the advantages of standardized rates but are more readily understood without special training. In this paper, standardizing weights based on 1992 data from England and Wales are provided, and the new measures are illustrated with a variety of examples. The new standardized rates are useful for examining trends over time; for comparing the impact of various diseases on public health; and for comparing rates of a given disease in several different countries. The authors think it is far more informative to say that 41 out of every 1,000 women die of breast cancer than to say that the standardized mortality rate is 51 per 100,000 women per year.

Supersites within superfolds. Binding site similarity in the absence of homology

A method is presented to assess the significance of binding site similarities within superimposed protein three-dimensional (3D) structures and applied to all similar structures in the Protein Data Bank. For similarities between 3D structures lacking significant sequence similarity, the important distinction was made between remote homology (an ancient common ancestor) and analogy (likely convergence to a folding motif) according to the structural classification of proteins (SCOP) database. Supersites were defined as structural locations on groups of analogous proteins (i.e. superfolds) showing a statistically significant tendency to bind substrates despite little evidence of a common ancestor for the proteins considered. We identify three potentially new superfolds containing supersites: ferredoxin-like folds, four-helical bundles and double-stranded beta helices. In addition, the method quantifies binding site similarities within homologous proteins and previously identified supersites such as that found in the beta/alpha (TIM) barrels. For the nine superfolds, the accuracy of predictions of binding site locations is assessed. Implications for protein evolution, and the prediction of protein function either through fold recognition or tertiary structure comparison, are discussed.

Antibodies to heat-shock protein 27 are associated with improved survival in patients with breast cancer

The overexpression of the heat-shock proteins hsp90, hsp70 and hsp27 in human mammary carcinomas has previously been shown to correlate with reduced overall survival. Moreover, antibodies to hsp90 were detectable in the serum of a large proportion of breast cancer patients but they were not found in normal controls. High antibody levels also correlated with reduced survival. Here, we show that antibodies to hsp27 were also detectable in the sera from breast cancer patients but not from normal controls, whereas antibodies to hsp70 were detectable in approximately one-third of both groups. The presence of antibodies to hsp27 was correlated with an improved rather than a reduced survival, particularly beyond the first 5 years. Hence, the overexpression of hsps in breast cancer cells does not provoke a generalized immune response to all the hsps. Moreover, the presence of antibodies to different hsps has distinct associations with survival. These effects are discussed in terms of the mechanisms that provoke an immune response to the hsps and the protective/non-protective effects of such a response.

From genotypes to genes: doubling the sample size

This paper considers the analysis of genetic case-control data. One approach considers the allele frequency in cases and controls. Because each individual has two alleles at any autosomal locus, there will be twice as many alleles as people. Another approach considers the risk of the disease in those who do not have the allele of interest (A), those who have a single copy (heterozygous), and those who are homozygous for A. A third approach does not differentiate between individuals with one or two copies of A. This was common when alleles were determined serologically and one could not distinguish between homozygotes and those with one copy of A and one of an unknown allele. All three approaches have been used in the literature, but this is the first systematic comparison of them. The different interpretations of the odds ratios from such analyses are explored and conditions are given under which the first two approaches are asymptotically equivalent. The chi-squared statistics from the three approaches are discussed. Both the odds ratio and the chi-squared statistic from the analysis that treats alleles rather than genotypes as individual entities are appropriate only when the Hardy-Weinberg equilibrium holds. When the equilibrium holds, the allele-based test statistic is asymptotically equivalent to the test for trend using the genotype data. Thus, analyses that treat alleles rather than people as observations should not be used. Instead, we recommend that such data should be analyzed by genotype.

Estimating the efficacy of screening by auditing smear histories of women with and without cervical cancer. The National Co-ordinating Network for Cervical Screening Working Group

The screening histories of all 348 women with invasive cervical cancer diagnosed in 1992 in 24 self-selected district health authorities and health boards in England, Wales and Scotland were compared with those of 677 age- and residency-matched controls. The controls were randomly selected from the family health services authority (FHSA) register. Screening histories, comprising the dates and results of all smears taken before the date of diagnosis of the patient's cancer, were determined from the FHSA computer and laboratory records. We estimate that the number of cases of cervical cancer in participating districts in 1992 would have been 57% (95% confidence interval 28-86%) greater if there had been no previous screening. In women under the age of 70 it would have been approximately 75% (31-115%) greater. Extrapolation of the results from this pilot suggests that screening prevented between 1100 and 3900 cases of invasive cervical cancer in the UK in 1992. Women with stage 1B cancer or worse were more likely to have no record of previous screening than controls: 47% of these women under the age of 70 had been adequately screened according to current (5 yearly screening) guidelines, compared with 75% of matched controls. Thirteen per cent of all patients under age 70 had screening histories indicative of inadequate follow-up of smears requiring colposcopy. The proportion of microinvasive cases with screening predating diagnosis was similar to the proportion of controls. There was a strong correlation between stage and age: 56% of cancers in women under 35 were microinvasive compared with just 9% in women 65 years or over. The 'relative protection' following a negative smear was greatest in the first 12 months and fell off towards the end of the fifth year. These data suggest that full adherence to current guidelines could perhaps have prevented another 1250 cases, but additional steps would have been required to prevent some of the 2300 remaining cases in women under the age of 70.