Robust cross-vendor mammographic texture models using augmentation-based domain adaptation for long-term breast cancer risk.

Purpose: Risk-stratified breast cancer screening might improve early detection and efficiency without comprising quality. However, modern mammography-based risk models do not ensure adaptation across vendor-domains and rely on cancer precursors, associated with short-term risk, which might limit long-term risk assessment. We report a cross-vendor mammographic texture model for long-term risk. Approach: The texture model was robustly trained using two systematically designed case-control datasets. Textural features, indicative of future breast cancer, were learned by excluding samples with diagnosed/potential malignancies from training. An augmentation-based domain adaption technique, based on flavorization of mammographic views, ensured generalization across vendor-domains. The model was validated in 66,607 consecutively screened Danish women with flavorized Siemens views and 25,706 Dutch women with Hologic-processed views. Performances were evaluated for interval cancers (IC) within 2 years from screening and long-term cancers (LTC) from 2 years after screening. The texture model was combined with established risk factors to flag 10% of women with the highest risk. Results: In Danish women, the texture model achieved an area under the receiver operating characteristic curve (AUC) of 0.71 and 0.65 for ICs and LTCs, respectively. In Dutch women with Hologic-processed views, the AUCs were not different from AUCs in Danish women with flavorized views. The AUC for texture combined with established risk factors increased to 0.68 for LTCs. The 10% of women flagged as high-risk accounted for 25.5% of ICs and 24.8% of LTCs. Conclusions: The texture model robustly estimated long-term breast cancer risk while adapting to an unseen processed vendor-domain and identified a clinically relevant high-risk subgroup.

An Artificial Intelligence-based Mammography Screening Protocol for Breast Cancer: Outcome and Radiologist Workload.

Background Developments in artificial intelligence (AI) systems to assist radiologists in reading mammograms could improve breast cancer screening efficiency. Purpose To investigate whether an AI system could detect normal, moderate-risk, and suspicious mammograms in a screening sample to safely reduce radiologist workload and evaluate across Breast Imaging Reporting and Data System (BI-RADS) densities. Materials and Methods This retrospective simulation study analyzed mammographic examination data consecutively collected from January 2014 to December 2015 in the Danish Capital Region breast cancer screening program. All mammograms were scored from 0 to 10, representing the risk of malignancy, using an AI tool. During simulation, normal mammograms (score < 5) would be excluded from radiologist reading and suspicious mammograms (score > recall threshold [RT]) would be recalled. Two radiologists read the remaining mammograms. The RT was fitted using another independent cohort (same institution) by matching to the radiologist sensitivity. This protocol was further applied to each BI-RADS density. Screening outcomes were measured using the sensitivity, specificity, workload, and false-positive rate. The AI-based screening was tested for noninferiority sensitivity compared with radiologist screening using the Farrington-Manning test. Specificities were compared using the McNemar test. Results The study sample comprised 114 421 screenings for breast cancer in 114 421 women, resulting in 791 screen-detected, 327 interval, and 1473 long-term cancers and 2107 false-positive screenings. The mean age of the women was 59 years ± 6 (SD). The AI-based screening sensitivity was 69.7% (779 of 1118; 95% CI: 66.9, 72.4) and was noninferior (P = .02) to the radiologist screening sensitivity of 70.8% (791 of 1118; 95% CI: 68.0, 73.5). The AI-based screening specificity was 98.6% (111 725 of 113 303; 95% CI: 98.5, 98.7), which was higher (P < .001) than the radiologist specificity of 98.1% (111 196 of 113 303; 95% CI: 98.1, 98.2). The radiologist workload was reduced by 62.6% (71 585 of 114 421), and 25.1% (529 of 2107) of false-positive screenings were avoided. Screening results were consistent across BI-RADS densities, although not significantly so for sensitivity. Conclusion Artificial intelligence (AI)-based screening could detect normal, moderate-risk, and suspicious mammograms in a breast cancer screening program, which may reduce the radiologist workload. AI-based screening performed consistently across breast densities. © RSNA, 2022 Online supplemental material is available for this article.

Incidence and Mortality in Upper Gastrointestinal Cancer After Negative Endoscopy for Gastroesophageal Reflux Disease.

BACKGROUND AND AIMS: Gastroesophageal reflux disease (GERD) is associated with an increased risk of cancer of the upper gastrointestinal tract. This study aimed to assess whether and to what extent a negative upper endoscopy in patients with GERD is associated with decreased incidence and mortality in upper gastrointestinal cancer (ie, esophageal, gastric, or duodenal cancer). METHODS: We conducted a population-based cohort study of all patients with newly diagnosed GERD between July 1, 1979 and December 31, 2018 in Denmark, Finland, Norway, and Sweden. The exposure, negative upper endoscopy, was examined as a time-varying exposure, where participants contributed unexposed person-time from GERD diagnosis until screened and exposed person-time from the negative upper endoscopy. The incidence and mortality in upper gastrointestinal cancer were assessed using parametric flexible models, providing adjusted hazard ratios (HRs) with 95% confidence intervals (CIs). RESULTS: Among 1,062,740 patients with GERD (median age 58 years; 52% were women) followed for a mean of 7.0 person-years, 5324 (0.5%) developed upper gastrointestinal cancer and 4465 (0.4%) died from such cancer. Patients who had a negative upper endoscopy had a 55% decreased risk of upper gastrointestinal cancer compared with those who did not undergo endoscopy (HR, 0.45; 95% CI, 0.43-0.48), a decrease that was more pronounced during more recent years (HR, 0.34; 95% CI, 0.30-0.38 from 2008 onward), and was otherwise stable across sex and age groups. The corresponding reduction in upper gastrointestinal mortality among patients with upper endoscopy was 61% (adjusted HR, 0.39; 95% CI, 0.37-0.42). The risk reduction after a negative upper endoscopy in incidence and mortality lasted for 5 and at least 10 years, respectively. CONCLUSIONS: Negative upper endoscopy is associated with strong and long-lasting decreases in incidence and mortality in upper gastrointestinal cancer in patients with GERD.

Sex differences in independence in activities of daily living early in stroke rehabilitation.

BACKGROUND: Current stroke research suggests that there are differences between females and males regarding incidence, stroke risk factors, stroke severity, outcome, and mortality. The few studies that have investigated sex differences in rehabilitation 8-12 months poststroke found that males are more independent, compared to females. OBJECTIVES: To investigate if there is a difference in the improvement of independence in activities of daily living (ADL) between females and males in the acute phase (first 2 weeks) of stroke rehabilitation in a Danish population. METHODS: A prospective cohort study enrolling patients admitted to the hospital's rehabilitation ward with a stroke diagnosis from January 1, 2016, to March 17, 2017. Baseline and follow-up data regarding the primary outcome, Barthel-100 index, were analyzed using an adjusted linear mixed model. RESULTS: The study included 206 patients (83 females). Females were older at admission and more males lived with a partner. No differences in stroke severity or any of the risk factors were found. There were no differences between female and male scores at baseline. In the adjusted linear mixed model, quantifying the difference between follow-up and baseline Barthel-100 score, females increased their Barthel-100 score by 20.8 points (95% confidence interval (CI) 15.4-26.3) and males with 29.0 points (95% CI 24.6-33.4). CONCLUSION: In a homogeneous sample of stroke survivors undergoing specialized 24-h stroke rehabilitation for 11-14 days, females were more dependent in ADL than males.

Prognostic impact of physical activity prior to myocardial infarction: Case fatality and subsequent risk of heart failure and death.

Background Animal studies indicate that exercise reduces myocardial damage during myocardial infarction by ischaemic preconditioning. Aim To determine from a prospective cohort study whether the level of leisure time physical activity (LTPA) in humans prior to myocardial infarction could modify the course of myocardial infarction by reducing case fatality and the subsequent risk of heart failure and mortality. Methods A total of 14,223 participants in the Copenhagen City Heart Study were assessed at baseline in 1976-1978; 1,664 later developed myocardial infarction (mean age at myocardial infarction 70.9 years) and were followed through registries until 2013. We explored the association of LTPA assessed before myocardial infarction with the risk of fatal myocardial infarction, heart failure and all-cause mortality after myocardial infarction. Odds ratios (ORs) and hazard ratios (HRs) were estimated by logistic and Cox proportional hazards regression models, adjusted for age at myocardial infarction and other potential confounders. Results A total of 425 (25.5%) myocardial infarctions were fatal. Higher levels of LTPA prior to myocardial infarction were associated with lower case fatality: adjusted ORs (95% confidence interval), with reference to sedentary LTPA were 0.68 (0.51-0.89) for light LTPA and 0.53 (0.38-0.74) for moderate/high LTPA. A total of 360 (29.1%) of the 1,239 myocardial infarction survivors developed heart failure and 1,033 (83.4%) died during follow-up. There was no association between LTPA levels prior to myocardial infarction and the risk of heart failure or all-cause mortality after non-fatal myocardial infarction: adjusted HRs for moderate/high versus sedentary LTPA were 1.06 (0.78-1.45) and 0.90 (0.74-1.08), respectively. Conclusion Individuals who were physically active had lower case fatality of myocardial infarction, but survivors were not protected against subsequent heart failure or mortality.

Prevalence of depression among older adults with dementia living in low- and middle-income countries: a cross-sectional study.

BACKGROUND: The prevalence of depression has been evaluated in populations of low- and middle-income (LMI) countries but the risk of depression has not been specified among persons with dementia. This cross-sectional analysis aimed to assess the prevalence and risk of depression among older people with dementia living in LMI countries. METHODS: The study analysed data from a population-based survey conducted by 10/66 Dementia Research Group in 2004. Altogether, 17 031 participants from eight different countries aged 65 years and above were assessed. Logistic regression was used to calculate prevalence and odds ratio (OR) of depression on persons with dementia. Adjustments by age and education were included in the analysis. ORs of depression on different types of dementia were determined. RESULTS: Depression was identified in 5.8% (4.4% of men, 6.6% of women) of all the 17 031 participants and in 12.4% (18.9% of men, 10.1% of women) of the 1612 persons with dementia. Persons with dementia had an increased risk of depression compared with persons without dementia, the age- and education-adjusted OR was 2.38 [95% confidence interval (CI0 1.99-2.84]); 3.86 (95% CI 2.83-5.26) for men and 1.88 (95% CI 1.51-2.35) for women. Compared with Alzheimer's disease, Lewy body [OR 2.75 (95% CI 1.40-3.72)] and vascular dementia [OR 2.35, (95% CI 1.49-3.72)] were associated with a higher risk of depression. CONCLUSIONS: Persons with dementia were twice as likely to have depression as persons without dementia. Among persons with dementia, the prevalence of depression was higher for men than women, and the risk of depression varied by the type of dementia.