Comparing quality of breast cancer care in the Netherlands and Norway by federated propensity score analytics.

PURPOSE: The aim of the study was to benchmark and compare breast cancer care quality indicators (QIs) between Norway and the Netherlands using federated analytics preventing transfer of patient-level data. METHODS: Breast cancer patients (2017-2018) were retrieved from the Netherlands Cancer Registry and the Cancer Registry of Norway. Five European Society of Breast Cancer Specialists (EUSOMA) QIs were assessed: two on magnetic resonance imaging (MRI), two on surgical approaches, and one on postoperative radiotherapy. The QI outcomes were calculated using 'Vantage 6' federated Propensity Score Stratification (PSS). Likelihood of receiving a treatment was expressed in odds ratios (OR). RESULTS: In total, 39,163 patients were included (32,786 from the Netherlands and 6377 from Norway). PSS scores were comparable to the crude outcomes of the QIs. The Netherlands scored higher on the QI 'proportions of patients preoperatively examined with breast MRI' [37% vs.17.5%; OR 2.8 (95% CI 2.7-2.9)], the 'proportions of patients receiving primary systemic therapy examined with breast MRI' [83.3% vs. 70.8%; OR 2.3 (95% CI 1.3-3.3)], and 'proportion of patients receiving a single breast operation' [95.2% vs. 91.5%; OR 1.8 (95% CI 1.4-2.2)]. Country scores for 'immediate breast reconstruction' and 'postoperative radiotherapy after breast-conserving surgery' were comparable. The EUSOMA standard was achieved in both countries for 4/5 indicators. CONCLUSION: Both countries achieved high scores on the QIs. Differences were observed in the use of MRI and proportion of patients receiving single surgery. The federated approach supports future possibilities on benchmark QIs without transfer of privacy-sensitive data.

Prolonged screening interval due to the COVID-19 pandemic and its association with tumor characteristics and treatment; a register-based study from BreastScreen Norway.

OBJECTIVE: During the COVID-19 pandemic Norway had to suspend its national breast cancer screening program. We aimed to investigate the effect of the pandemic-induced suspension on the screening interval, and its subsequent association with the tumor characteristics and treatment of screen-detected (SDC) and interval breast cancer (IC). METHODS: Information about women aged 50-69, participating in BreastScreen Norway, and diagnosed with a SDC (N = 3799) or IC (N = 1806) between 2018 and 2021 was extracted from the Cancer Registry of Norway. Logistic regression was used to investigate the association between COVID-19 induced prolonged screening intervals and tumor characteristics and treatment. RESULTS: Women with a SDC and their last screening exam before the pandemic had a median screening interval of 24.0 months (interquartile range: 23.8-24.5), compared to 27.0 months (interquartile range: 25.8-28.5) for those with their last screening during the pandemic. The tumor characteristics and treatment of women with a SDC, last screening during the pandemic, and a screening interval of 29-31 months, did not differ from those of women with a SDC, last screening before the pandemic, and a screening interval of 23-25 months. ICs detected 24-31 months after screening, were more likely to be histological grade 3 compared to ICs detected 0-23 months after screening (odds ratio: 1.40, 95% confidence interval: 1.06-1.84). CONCLUSIONS: Pandemic-induced prolonged screening intervals were not associated with the tumor characteristics and treatment of SDCs, but did increase the risk of a histopathological grade 3 IC. This study provides insights into the possible effects of extending the screening interval.

Matrix factorization for the reconstruction of cervical cancer screening histories and prediction of future screening results.

BACKGROUND: Mass screening programs for cervical cancer prevention in the Nordic countries have strongly reduced cancer incidence and mortality at the population level. An alternative to the current mass screening is a more personalised screening strategy adapting the recommendations to each individual. However, this necessitates reliable risk prediction models accounting for disease dynamics and individual data. Herein we propose a novel matrix factorisation framework to classify females by the time-varying risk of being diagnosed with cervical cancer. We cast the problem as a time-series prediction model where the data from females in the Norwegian screening population are represented as sparse vectors in time and then combined into a single matrix. Using novel temporal regularisation and discrepancy terms for the cervical cancer screening context, we reconstruct complete screening profiles from this scarce matrix and use these to predict the next exam results indicating the risk of cervical cancer. The algorithm is validated on both synthetic and registry screening data by measuring the probability of agreement (PoA) between Kaplan-Meier estimates. RESULTS: In numerical experiments on synthetic data, we demonstrate that the novel regularisation and discrepancy term can improve the data reconstruction ability as well as prediction performance over varying data scarcity. Using a hold-out set of screening data, we compare several numerical models and find that the proposed framework attains the strongest PoA. We observe strong correlations between the empirical survival curves from our method and the hold-out data, and evaluate the ability of our framework to predict the females' next results for up to five years ahead in time using only their current screening histories as input. CONCLUSIONS: We have proposed a matrix factorization model for predicting future screening results and evaluated its performance in a female cohort to demonstrate the potential for developing prediction models for more personalized cervical cancer screening.

Artificial Intelligence Evaluation of 122 969 Mammography Examinations from a Population-based Screening Program.

Background Artificial intelligence (AI) has shown promising results for cancer detection with mammographic screening. However, evidence related to the use of AI in real screening settings remain sparse. Purpose To compare the performance of a commercially available AI system with routine, independent double reading with consensus as performed in a population-based screening program. Furthermore, the histopathologic characteristics of tumors with different AI scores were explored. Materials and Methods In this retrospective study, 122 969 screening examinations from 47 877 women performed at four screening units in BreastScreen Norway from October 2009 to December 2018 were included. The data set included 752 screen-detected cancers (6.1 per 1000 examinations) and 205 interval cancers (1.7 per 1000 examinations). Each examination had an AI score between 1 and 10, where 1 indicated low risk of breast cancer and 10 indicated high risk. Threshold 1, threshold 2, and threshold 3 were used to assess the performance of the AI system as a binary decision tool (selected vs not selected). Threshold 1 was set at an AI score of 10, threshold 2 was set to yield a selection rate similar to the consensus rate (8.8%), and threshold 3 was set to yield a selection rate similar to an average individual radiologist (5.8%). Descriptive statistics were used to summarize screening outcomes. Results A total of 653 of 752 screen-detected cancers (86.8%) and 92 of 205 interval cancers (44.9%) were given a score of 10 by the AI system (threshold 1). Using threshold 3, 80.1% of the screen-detected cancers (602 of 752) and 30.7% of the interval cancers (63 of 205) were selected. Screen-detected cancer with AI scores not selected using the thresholds had favorable histopathologic characteristics compared to those selected; opposite results were observed for interval cancer. Conclusion The proportion of screen-detected cancers not selected by the artificial intelligence (AI) system at the three evaluated thresholds was less than 20%. The overall performance of the AI system was promising according to cancer detection. © RSNA, 2022.

A hidden Markov model for population-level cervical cancer screening data.

The Cancer Registry of Norway has been administrating a national cervical cancer screening program since 1992 by coordinating triennial cytology exam screenings for the female population between 25 and 69 years of age. Up to 80% of cancers are prevented through mass screening, but this comes at the expense of considerable screening activity and leads to overtreatment of clinically asymptomatic precancers. In this article, we present a continuous-time, time-inhomogeneous hidden Markov model which was developed to understand the screening process and cervical cancer carcinogenesis in detail. By leveraging 1.7 million individual's multivariate time-series of medical exams performed over a 25-year period, we simultaneously estimate all model parameters. We show that an age-dependent model reflects the Norwegian screening program by comparing empirical survival curves from observed registry data and data simulated from the proposed model. The model can be generalized to include more detailed individual-level covariates as well as new types of screening exams. By utilizing individual screening histories and covariate data, the proposed model shows potential for improving strategies for cancer screening programs by personalizing recommended screening intervals.

Bayesian multitask learning regression for heterogeneous patient cohorts.

Multitask learning (MTL) leverages commonalities across related tasks with the aim of improving individual task performance. A key modeling choice in designing MTL models is the structure of the tasks' relatedness, which may not be known. Here we propose a Bayesian multitask learning model that is able to infer the task relationship structure directly from the data. We present two variations of the model in terms of a priori information of task relatedness. First, a diffuse Wishart prior is placed on a task precision matrix so that all tasks are assumed to be equally related a priori. Second, a Bayesian graphical LASSO prior is used on the task precision matrix to impose sparsity in the task relatedness. Motivated by machine learning applications in the biomedical domain, we emphasize interpretability and uncertainty quantification in our models. To encourage model interpretability, linear mappings from the shared input spaces to task-dependent output spaces are used. To encourage uncertainty quantification, conjugate priors are used so that full posterior inference is possible. Using synthetic data, we show that our model is able to recover the underlying task relationships as well as features jointly relevant for all tasks. We demonstrate the utility of our model on three distinct biomedical applications: Alzheimer's disease progression, Parkinson's disease assessment, and cervical cancer screening compliance. We show that our model outperforms Single Task (STL) models in terms of predictive performance, and performs better than existing MTL methods for the majority of the scenarios.

Risk stratification in cervical cancer screening by complete screening history: Applying bioinformatics to a general screening population.

Women screened for cervical cancer in Sweden are currently treated under a one-size-fits-all programme, which has been successful in reducing the incidence of cervical cancer but does not use all of the participants' available medical information. This study aimed to use women's complete cervical screening histories to identify diagnostic patterns that may indicate an increased risk of developing cervical cancer. A nationwide case-control study was performed where cervical cancer screening data from 125,476 women with a maximum follow-up of 10 years were evaluated for patterns of SNOMED diagnoses. The cancer development risk was estimated for a number of different screening history patterns and expressed as Odds Ratios (OR), with a history of 4 benign cervical tests as reference, using logistic regression. The overall performance of the model was moderate (64% accuracy, 71% area under curve) with 61-62% of the study population showing no specific patterns associated with risk. However, predictions for high-risk groups as defined by screening history patterns were highly discriminatory with ORs ranging from 8 to 36. The model for computing risk performed consistently across different screening history lengths, and several patterns predicted cancer outcomes. The results show the presence of risk-increasing and risk-decreasing factors in the screening history. Thus it is feasible to identify subgroups based on their complete screening histories. Several high-risk subgroups identified might benefit from an increased screening density. Some low-risk subgroups identified could likely have a moderately reduced screening density without additional risk.

The epidemiology of post-traumatic stress disorder in Norway: trauma characteristics and pre-existing psychiatric disorders.

PURPOSE: The prevalence of PTSD differs by gender. Pre-existing psychiatric disorders and different traumas experienced by men and women may explain this. The aims of this study were to assess (1) incidence and prevalence of exposure to traumatic events and PTSD, (2) the effect of pre-existing psychiatric disorders prior to trauma on the risk for PTSD, and (3) the effect the characteristics of trauma have on the risk for PTSD. All stratified by gender. METHOD: CIDI was used to obtain diagnoses at the interview stage and retrospectively for the general population N = 1634. RESULTS: The incidence for trauma was 466 and 641 per 100,000 PYs for women and men, respectively. The incidence of PTSD was 88 and 31 per 100,000 PYs. Twelve month and lifetime prevalence of PTSD was 1.7 and 4.3 %, respectively, for women, and 1.0 and 1.4 %, respectively, for men. Pre-existing psychiatric disorders were risk factors for PTSD, but only in women. Premeditated traumas were more harmful. CONCLUSION: Gender differences were observed regarding traumatic exposure and in the nature of traumas experienced and incidences of PTSD. Men experienced more traumas and less PTSD. Pre-existing psychiatric disorders were found to be risk factors for subsequent PTSD in women. However, while trauma happens to most, it only rarely leads to PTSD, and the most harmful traumas were premeditated ones. Primary prevention of PTSD is thus feasible, although secondary preventive efforts should be gender-specific.

Better survival after breast-conserving therapy compared to mastectomy when axillary node status is positive in early-stage breast cancer: a registry-based follow-up study of 6387 Norwegian women participating in screening, primarily operated between 1998 and 2009.

BACKGROUND: Recent registry studies on early-stage breast cancer have shown better survival rates when women underwent breast-conserving therapy (BCT) compared with mastectomy (MTX). The aim of this study is to investigate women participating in screening, in all four stages of early breast cancer (T1N0M0, T2N0M0, T1N1M0, and T2N1M0), as to whether there is a survival benefit when women undergo BCT compared to MTX. METHOD: A cohort of 6387 women aged 50-69, with primary-operated breast cancer from January 1998 to December 2009, participating in screening and followed-up until the end of 2010. Life tables were calculated by stages (pT1N0M0, pT2N0M0, pT1N1M0, and pT2N1M0), surgery groups (BCT and MTX), and screening detection (first screening, later screening, or interval cancer). Cox regression was used to calculate hazard ratios (HR) between BCT and MTX in crude and adjusted analyses. RESULTS: In stage T1N1M0, women who underwent MTX had an HR of 2.91 (95% CI 1.30-6.48) for breast cancer death compared to women who underwent BCT, after adjusting for screening detection, years of diagnosis, age at diagnosis, histology, grade, and hormone receptor status. For all other TNM categories of early breast cancer, there was no difference in survival. 10-year breast cancer-specific survival (BCSS) in T1N0M0 was 98% for women undergoing BCT and 96% for women undergoing MTX. 10-year BCSS in T1N1M0 was 97% for women undergoing BCT and 89% for women undergoing MTX. CONCLUSIONS: For women participating in screening, there is a benefit of BCT over MTX in stage T1N1M0. No such effects were observed in the other early stages of breast cancer.

Survival is Better After Breast Conserving Therapy than Mastectomy for Early Stage Breast Cancer: A Registry-Based Follow-up Study of Norwegian Women Primary Operated Between 1998 and 2008.

BACKGROUND: Breast-conserving therapy (BCT) and mastectomy (MTX) has been considered to have a similar long-time survival. However, better survival in women undergoing BCT compared with MTX is found in two recent register studies from the United States. The purpose of this study was to compare survival after BCT and MTX for women with early-stage breast cancer in Norway. METHODS: Women with invasive, early-stage breast cancer (1998-2008) where BCT and MTX were considered as equally beneficial treatments were included for a total of 13,015 women. Surgery was divided in two main cohorts (primary BCT, primary MTX) and five subcohorts. Analyses were stratified into T1N0M0, T2N0M0, T1N1M0, T2N1M0, and age groups (<50, 50-69, ≥70). Overall survival and breast cancer-specific survival (BCSS) were calculated in life tables, hazard ratios by Cox regression, and sensitivity analyses. RESULTS: Five-year BCSS for women who underwent primary BCT or primary MTX was 97 and 88 %, respectively. Women who underwent primary MTX had a hazard ratio of 1.64 (95 % confidence interval 1.43-1.88) for breast cancer death compared with women who underwent primary BCT after adjusting for the year of diagnosis, age at diagnosis, stage, histology, and grade. CONCLUSIONS: Survival was better or equal after breast-conserving therapy than mastectomy in all early stages, surgical subcohorts, and age groups. This advantage could not only be attributed to differences in tumor biology.

Cross-population evaluation of cervical cancer risk prediction algorithms.

BACKGROUND: Cervical cancer is a preventable disease, despite being one of the most common types of female cancers worldwide. Integrating existing programs for cervical cancer screening with personalized risk prediction algorithms can improve population-level cancer prevention by enabling more targeted screening and contrive preventive healthcare innovations. While algorithms developed for cervical cancer risk prediction have shown promising performance in internal validation on more homogeneous populations, their ability to generalize to external populations remains to be assessed. METHODS: To address this gap, we perform a cross-population comparative study of personalized prediction algorithms for more personalized cervical cancer screening. Using data from the Norwegian and Estonian populations, the algorithms are validated on internal and external datasets to study their potential biases and limitations when applied to different populations. We evaluate the algorithms in predicting progression from low-grade precancerous cervical lesions, simulating a clinically relevant application of more personalized risk stratification. RESULTS: As expected, our numerical experiments show that algorithm performance varies depending on the population. However, some algorithms show strong generalization capacity across different data sources. Using Kaplan-Meier estimates, we demonstrate the strengths and limitations of the algorithms in detecting cancer progression over time by comparing to the trends observed from data. We assess their overall discrimination performance in personalized risk predictions by analyzing the accuracy and confidence in individual risk estimates. DISCUSSION AND CONCLUSION: This study examines the effectiveness of personalized prediction algorithms across different populations. Our results demonstrate the potential for generalizing risk prediction algorithms to external populations. These findings highlight the importance of considering population diversity when developing risk prediction algorithms.

Performance of an Artificial Intelligence System for Breast Cancer Detection on Screening Mammograms from BreastScreen Norway.

Purpose To explore the stand-alone breast cancer detection performance, at different risk score thresholds, of a commercially available artificial intelligence (AI) system. Materials and Methods This retrospective study included information from 661 695 digital mammographic examinations performed among 242 629 female individuals screened as a part of BreastScreen Norway, 2004-2018. The study sample included 3807 screen-detected cancers and 1110 interval breast cancers. A continuous examination-level risk score by the AI system was used to measure performance as the area under the receiver operating characteristic curve (AUC) with 95% CIs and cancer detection at different AI risk score thresholds. Results The AUC of the AI system was 0.93 (95% CI: 0.92, 0.93) for screen-detected cancers and interval breast cancers combined and 0.97 (95% CI: 0.97, 0.97) for screen-detected cancers. In a setting where 10% of the examinations with the highest AI risk scores were defined as positive and 90% with the lowest scores as negative, 92.0% (3502 of 3807) of the screen-detected cancers and 44.6% (495 of 1110) of the interval breast cancers were identified with AI. In this scenario, 68.5% (10 987 of 16 040) of false-positive screening results (negative recall assessment) were considered negative by AI. When 50% was used as the cutoff, 99.3% (3781 of 3807) of the screen-detected cancers and 85.2% (946 of 1110) of the interval breast cancers were identified as positive by AI, whereas 17.0% (2725 of 16 040) of the false-positive results were considered negative. Conclusion The AI system showed high performance in detecting breast cancers within 2 years of screening mammography and a potential for use to triage low-risk mammograms to reduce radiologist workload. Keywords: Mammography, Breast, Screening, Convolutional Neural Network (CNN), Deep Learning Algorithms Supplemental material is available for this article. © RSNA, 2024 See also commentary by Bahl and Do in this issue.

Can I trust my fake data - A comprehensive quality assessment framework for synthetic tabular data in healthcare.

BACKGROUND: Ensuring safe adoption of AI tools in healthcare hinges on access to sufficient data for training, testing and validation. Synthetic data has been suggested in response to privacy concerns and regulatory requirements and can be created by training a generator on real data to produce a dataset with similar statistical properties. Competing metrics with differing taxonomies for quality evaluation have been proposed, resulting in a complex landscape. Optimising quality entails balancing considerations that make the data fit for use, yet relevant dimensions are left out of existing frameworks. METHOD: We performed a comprehensive literature review on the use of quality evaluation metrics on synthetic data within the scope of synthetic tabular healthcare data using deep generative methods. Based on this and the collective team experiences, we developed a conceptual framework for quality assurance. The applicability was benchmarked against a practical case from the Dutch National Cancer Registry. CONCLUSION: We present a conceptual framework for quality assuranceof synthetic data for AI applications in healthcare that aligns diverging taxonomies, expands on common quality dimensions to include the dimensions of Fairness and Carbon footprint, and proposes stages necessary to support real-life applications. Building trust in synthetic data by increasing transparency and reducing the safety risk will accelerate the development and uptake of trustworthy AI tools for the benefit of patients. DISCUSSION: Despite the growing emphasis on algorithmic fairness and carbon footprint, these metrics were scarce in the literature review. The overwhelming focus was on statistical similarity using distance metrics while sequential logic detection was scarce. A consensus-backed framework that includes all relevant quality dimensions can provide assurance for safe and responsible real-life applications of synthetic data. As the choice of appropriate metrics are highly context dependent, further research is needed on validation studies to guide metric choices and support the development of technical standards.

Impact of the COVID-19 pandemic on breast cancer incidence and tumor stage in the Netherlands and Norway: A population-based study.

BACKGROUND: Comparing the impact of the COVID-19 pandemic on the incidence of newly diagnosed breast tumors and their tumor stage between the Netherlands and Norway will help us understand the effect of differences in governmental and social reactions towards the pandemic. METHODS: Women newly diagnosed with breast cancer in 2017-2021 were selected from the Netherlands Cancer Registry and the Cancer Registry of Norway. The crude breast cancer incidence rate (tumors per 100,000 women) during the first (March-September 2020), second (October 2020-April 2021), and Delta COVID-19 wave (May-December 2021) was compared with the incidence rate in the corresponding periods in 2017, 2018, and 2019. Incidence rates were stratified by age group, method of detection, and clinical tumor stage. RESULTS: During the first wave breast cancer incidence declined to a larger extent in the Netherlands than in Norway (27.7% vs. 17.2% decrease, respectively). In both countries, incidence decreased in women eligible for screening. In the Netherlands, incidence also decreased in women not eligible for screening. During the second wave an increase in the incidence of stage IV tumors in women aged 50-69 years was seen in the Netherlands. During the Delta wave an increase in overall incidence and incidence of stage I tumors was seen in Norway. CONCLUSION: Alterations in breast cancer incidence and tumor stage seem related to a combined effect of the suspension of the screening program, health care avoidance due to the severity of the pandemic, and other unknown factors.

Hierarchical continuous-time inhomogeneous hidden Markov model for cancer screening with extensive followup data.

Continuous-time hidden Markov models are an attractive approach for disease modeling because they are explainable and capable of handling both irregularly sampled, skewed and sparse data arising from real-world medical practice, in particular to screening data with extensive followup. Most applications in this context consider time-homogeneous models due to their relative computational simplicity. However, the time homogeneous assumption is too strong to accurately model the natural history of many diseases including cancer. Moreover, cancer risk across the population is not homogeneous either, since exposure to disease risk factors can vary considerably between individuals. This is important when analyzing longitudinal datasets and different birth cohorts. We model the heterogeneity of disease progression and regression using piece-wise constant intensity functions and model the heterogeneity of risks in the population using a latent mixture structure. Different submodels under the mixture structure employ the same types of Markov states reflecting disease progression and allowing both clinical interpretation and model parsimony. We also consider flexible observational models dealing with model over-dispersion in real data. An efficient, scalable Expectation-Maximization algorithm for inference is proposed with the theoretical guaranteed convergence property. We demonstrate our method's superior performance compared to other state-of-the-art methods using synthetic data and a real-world cervical cancer screening dataset from the Cancer Registry of Norway. Moreover, we present two model-based risk stratification methods that identify the risk levels of individuals.

Towards a data-driven system for personalized cervical cancer risk stratification.

Mass-screening programs for cervical cancer prevention in the Nordic countries have been effective in reducing cancer incidence and mortality at the population level. Women who have been regularly diagnosed with normal screening exams represent a sub-population with a low risk of disease and distinctive screening strategies which avoid over-screening while identifying those with high-grade lesions are needed to improve the existing one-size-fits-all approach. Machine learning methods for more personalized cervical cancer risk estimation may be of great utility to screening programs shifting to more targeted screening. However, deriving personalized risk prediction models is challenging as effective screening has made cervical cancer rare and the exam results are strongly skewed towards normal. Moreover, changes in female lifestyle and screening habits over time can cause a non-stationary data distribution. In this paper, we treat cervical cancer risk prediction as a longitudinal forecasting problem. We define risk estimators by extending existing frameworks developed on cervical cancer screening data to incremental learning for longitudinal risk predictions and compare these estimators to machine learning methods popular in biomedical applications. As input to the prediction models, we utilize all the available data from the individual screening histories.Using data from the Cancer Registry of Norway, we find in numerical experiments that the models are strongly biased towards normal results due to imbalanced data. To identify females at risk of cancer development, we adapt an imbalanced classification strategy to non-stationary data. Using this strategy, we estimate the absolute risk from longitudinal model predictions and a hold-out set of screening data. Comparing absolute risk curves demonstrate that prediction models can closely reflect the absolute risk observed in the hold-out set. Such models have great potential for improving cervical cancer risk stratification for more personalized screening recommendations.

Federated Networks for Distributed Analysis of Health Data.

Access to health data, important for population health planning, basic and clinical research and health industry utilization, remains problematic. Legislation intended to improve access to personal data across national borders has proven to be a double-edged sword, where complexity and implications from misinterpretations have paradoxically resulted in data becoming more siloed. As a result, the potential for development of health specific AI and clinical decision support tools built on real-world data have yet to be fully realized. In this perspective, we propose federated networks as a solution to enable access to diverse data sets and tackle known and emerging health problems. The perspective draws on experience from the World Economic Forum Breaking Barriers to Health Data project, the Personal Health Train and Vantage6 infrastructures, and industry insights. We first define the concept of federated networks in a healthcare context, present the value they can bring to multiple stakeholders, and discuss their establishment, operation and implementation. Challenges of federated networks in healthcare are highlighted, as well as the resulting need for and value of an independent orchestrator for their safe, sustainable and scalable implementation.

Improving five-year survival prediction via multitask learning across HPV-related cancers.

Oncology is a highly siloed field of research in which sub-disciplinary specialization has limited the amount of information shared between researchers of distinct cancer types. This can be attributed to legitimate differences in the physiology and carcinogenesis of cancers affecting distinct anatomical sites. However, underlying processes that are shared across seemingly disparate cancers probably affect prognosis. The objective of the current study is to investigate whether multitask learning improves 5-year survival cancer patient survival prediction by leveraging information across anatomically distinct HPV related cancers. Data were obtained from the Surveillance, Epidemiology, and End Results (SEER) program database. The study cohort consisted of 29,768 primary cancer cases diagnosed in the United States between 2004 and 2015. Ten different cancer diagnoses were selected, all with a known association with HPV risk. In the analysis, the cancer diagnoses were categorized into three distinct topography groups of varying specificity. The most specific topography grouping consisted of 10 original cancer diagnoses differentiated by the first two digits of the ICD-O-3 topography code. The second topography grouping consisted of cancer diagnoses categorized into six distinct organ groups. Finally, the third topography grouping consisted of just two groups, head-neck cancers and ano-genital cancers. The tasks were to predict 5-year survival for patients within the different topography groups using 14 predictive features which were selected among descriptive variables available in the SEER database. The information from the predictive features was shared between tasks in three different ways, resulting in three distinct predictive models: 1) Information was not shared between patients assigned to different tasks (single task learning); 2) Information was shared between all patients, regardless of task (pooled model); 3) Only relevant information was shared between patients grouped to different tasks (multitask learning). Prediction performance was evaluated with Brier scores. All three models were evaluated against one another on each of the three distinct topography-defined tasks. The results showed that multitask classifiers achieved relative improvement for the majority of the scenarios studied compared to single task learning and pooled baseline methods. In this study, we have demonstrated that sharing information among anatomically distinct cancer types can lead to improved predictive survival models.

Validity of self-reported Pap smear history in Norwegian women.

OBJECTIVES: To assess the validity of self-reported Papanicolau (Pap) smear history in Norwegian women and to identify characteristics that influence the validity. METHODS: Interview data from a sample of 16,574 Norwegian women, aged 18-45, in 2004-2005, was compared with information from the population-based cytology register. Crude validity in the self-reports with respect to ever/never having taken a Pap smear was summarized. The validity of the reported interval since last Pap smear was assessed by a smoothed distribution of the reported interval, stratified by the registered interval. Characteristics of influence on validity were identified by logistic regression for true positives (sensitivity and positive predictive value), true negatives (specificity and negative predictive value) and for more than one year discrepancy in time since last Pap smear, between reported and registered interval. RESULTS: Overall validity was summarized by: concordance = 0.9, sensitivity = 0.97, positive predictive value = 0.92, specificity = 0.55, negative predictive value = 0.78 and report-to-records ratio = 1.51. The variance in the reported interval increased proportionally with the registered interval, and women tended to underestimate the interval (telescoping). Age and registered number of years since last Pap smear had the strongest influence on ever/never and time interval validity, respectively. CONCLUSIONS: Estimated screening rates, based on self-reporting without organized screening, are biased. Telescoping leads to increased risk for developing invasive disease, because women will postpone their next Pap smear.

Selection effects in psychiatric epidemiology: a 14-year prospective study of the Hopkins Symptom Check List-25 as a predictor of mortality in the Norwegian general population.

BACKGROUND: Inconsistent findings in studies examining the association between mental health problems and mortality have suggested study design as a source of discrepancy. This study investigates if selection bias is introduced by an extensive personal interview, recruiting a healthier population, and furthermore examines the association between mental distress and mortality. METHOD: The OsLof study consists of a random population-based sample of 2,014 persons above 18 years that participated in an interview in 1989, including the Hopkins Symptom Check List-25 (HSCL-25). A second material was drawn as an approximate 1% sample of the Norwegian population (N = 31.907). Both samples were followed up until 2003, emigration or death. Hazard rates were estimated by Cox regression and piecewise exponential survival regression. RESULTS: Mortality rates were similar regardless of HSCL, and lower than in the 1% population sample, in the first 2 years of follow-up. A higher mortality rate in those with HSCL > or = 1.75 appeared after 8 years of follow-up, and after 12 years the mortality rate was 1.57 times higher than in the population sample. CONCLUSION: Studies of predictors ought to have a long follow-up, as several years might pass before differences in mortality are revealed. Due to selection effects, short follow-up might lead to erroneously believing predictors to be protective. This study was consistent with other studies with long follow-up, showing higher mortality associated with mental distress.