Women's attitudes and perspectives on the use of artificial intelligence in the assessment of screening mammograms.

PURPOSE: To investigate attitudes and perspectives on the use of artificial intelligence (AI) in the assessment of screening mammograms among women invited to BreastScreen Norway. METHOD: An anonymous survey was sent to all women invited to BreastScreen Norway during the study period, October 10, 2022, to December 25, 2022 (n = 84,543). Questions were answered on a 10-point Likert scale and as multiple-choice, addressing knowledge of AI, willingness to participate in AI studies, information needs, confidence in AI results and AI assisted reading strategies, and thoughts on concerns and benefits of AI in mammography screening. Analyses were performed using χ2 and logistic regression tests. RESULTS: General knowledge of AI was reported as extensive by 11.0% of the 8,355 respondents. Respondents were willing to participate in studies using AI either for decision support (64.0%) or triaging (54.9%). Being informed about use of AI-assisted image assessment was considered important, and a reading strategy of AI in combination with one radiologist preferred. Having extensive knowledge of AI was associated with willingness to participate in AI studies (decision support; odds ratio [OR]: 5.1, 95% confidence interval [CI]: 4.1-6.4, and triaging; OR: 3.4, 95% CI: 2.8-4.0) and trust in AI's independent assessment (OR: 6.8, 95% CI: 5.7, 8.3). CONCLUSIONS: Women invited to BreastScreen Norway had a positive attitude towards the use of AI in image assessment, given that human readers are still involved. Targeted information and increased public knowledge of AI could help achieve high participation in AI studies and successful implementation of AI in mammography screening.

Comparisons of assessment pathways after abnormal mammography screening in Denmark, Norway, and Spain.

PURPOSE: To ensure high-quality screening programmes and effective utilization of resources, it is important to monitor how cancer detection is affected by different strategies performed at recall assessment. This study aimed to describe procedures performed at recall assessment and compare and evaluate the performance of the assessment in Denmark, Norway, and Spain in terms of screen-detected cancer (SDC) and interval cancer (IC) rates. METHODS: We included women aged 50-69 years from Denmark, Norway, and Spain, who were recalled for assessment after screening mammography, and recorded all procedures performed during six months after diagnosis, and the timing of the procedures. Women were followed for two years and screen-detected and interval cancer, and sensitivity of recall was calculated and compared. RESULTS: In total, data from 24,645 Danish, 30,050 Norwegian, and 41,809 Spanish women were included in the study. Most of the women had some assessment within 2 months in all three countries. SDC rates were higher in Denmark (0.57) and Norway (0.60) compared to Spain (0.38), as were the IC rates, i.e. 0.25 and 0.18 vs. 0.12, respectively. The sensitivity of the diagnostic follow-up was somewhat higher in Denmark (98.3%) and Norway (98.2%), compared to Spain (95.4%), but when excluding non-invasive assessment pathways, the sensitivities were comparable. CONCLUSION: This comparison study showed variation in the assessment procedures used in the three countries as well as the SDC and IC rates and the sensitivity of recall.

Costs and Effects of Implementing Digital Tomosynthesis in a Population-Based Breast Cancer Screening Program: Predictions Using Results from the To-Be Trial in Norway.

BACKGROUND: Although several studies from Europe and the US have shown promising screening results favoring digital breast tomosynthesis compared with standard digital mammography (DM), both costs and effects of implementing tomosynthesis in routine screening programs remain uncertain. The cost effectiveness of using tomosynthesis in routine screening is debated in the literature, and model inputs from randomized trials are lacking. Using parameters mainly from a randomized controlled trial (the To-Be trial), we simulated costs and effects of implementing tomosynthesis in the national screening program BreastScreen Norway. METHODS: The To-Be trial was performed in Bergen from 2016 to 2017 within BreastScreen Norway, where females were randomized to either digital breast tomosynthesis including synthetic mammograms (DBT) or DM. The trial was followed by a cohort study offering all females DBT in 2018-2019. The trial included over 37,000 females, and allowed for estimation of short-term costs and effects related to screening, recall examinations and cancer detection. Using these and recent Norwegian estimates for 10-year stage-specific survival and treatment costs, the cost effectiveness of replacing DM with DBT in BreastScreen Norway was simulated in a decision tree model with probabilistic sensitivity analyses. Outcomes included false-positive screening results, screen-detected and interval cancers, stage at diagnosis, all-cause deaths, life-years gained, costs at recall and treatment and incremental cost-effectiveness ratio. RESULTS: The estimated additional cost of DBT was €8.10. Simulating ten rounds of screening from 2018 and 10-year survival and costs, 500 deaths were averted and 2300 life-years gained at an additional screening cost of €29 million for females screened with DBT versus DM. Taking over-diagnosis, recall and treatment costs into account, DBT was dominant in the deterministic analysis. The incremental cost-effectiveness ratio indicated cost savings of €1400 per life-year gained. Probabilistic sensitivity analyses showed that DBT was cost effective in over 50% of the simulations at all willingness-to-pay levels per life-year gained, and in 80% of the simulations at levels above €22,000. If willingness-to-pay levels up to €35,000 were assumed, DBT would be cost effective in over 50% of the simulations for additional costs of DBT of up to €32, almost four times the estimated additional cost of €8.10. CONCLUSION: DBT may be cost effective if implemented in BreastScreen Norway. However, generalizability of results could depend on factors varying between countries, such as recall rates, program sensitivity and specificity, treatment cost and willingness-to-pay levels.

Assessment of breast positioning criteria in mammographic screening: Agreement between artificial intelligence software and radiographers.

OBJECTIVES: To determine the agreement between artificial intelligence software (AI) and radiographers in assessing breast positioning criteria for mammograms from standard digital mammography and digital breast tomosynthesis. METHODS: Assessment of breast positioning was performed by AI and by four radiographers in pairs of two on 156 examinations of women screened in Bergen, April to September 2019, as part of BreastScreen Norway. Ten criteria were used; three for craniocaudal and seven for mediolateral-oblique view. The criteria evaluated the appearance of the nipple, breast rotation, pectoral muscle, inframammary fold and pectoral nipple line. Intraclass correlation and Cohen's kappa coefficient (κ) were used to investigate the correlation and agreement between the radiographer's assessments and AI. RESULTS: The intraclass correlation for the pectoral nipple line between the radiographers and AI was >0.92. A substantial to almost perfect agreement (κ > 0.69) was observed between the radiographers and AI on the nipple in profile criterion. We observed a slight to moderate agreement for the other criteria (κ = 0.06-0.52) and generally a higher agreement between the two pairs of radiographers (mean κ = 0.70) than between the radiographers and AI (mean κ = 0.41). CONCLUSIONS: AI has great potential in evaluating breast position criteria in mammography by reducing subjectivity. However, varying agreement between radiographers and AI was observed. Standardized and evidence-based criteria for definitions, understandings and assessment methods are needed to reach optimal image quality in mammography.

Self-reported symptoms among participants in a population-based screening program.

BACKGROUND: A limited number of studies have explored the association between self-reported symptoms and the risk of breast cancer among participants of population based screening programs. METHODS: We performed descriptive statistics on recall, screen-detected and interval cancer, positive predictive value and histopathological tumour characteristics by symptom group (asymptomatic, lump, and skin or nipple changes) as reported from 785,642 women aged 50-69 when they attended BreastScreen Norway 1996-2016. Uni- and multivariable mixed effects logistic regression models were used to analyze the association between symptom group and screen-detected or interval cancer. Results were presented as odds ratios and 95% confidence intervals (CI). RESULTS: A lump or skin/nipple change was reported in 6.2% of the 3,307,697 examinations. The rate of screen-detected cancers per 1000 examinations was 45.2 among women with a self-reported lump and 5.1 among asymptomatic women. Adjusted odds ratio of screen-detected cancer was 10.1 (95% CI: 9.3-11.1) and 2.0 (95% CI: 1.6-2.5) for interval cancer among women with a self-reported lump versus asymptomatic women. Tumour diameter, histologic grade and lymph node involvement of screen-detected and interval cancer were less prognostically favourable for women with a self-reported lump versus asymptomatic women. CONCLUSION: Despite targeting asymptomatic women, 6.2% of the screening examinations in BreastScreen Norway was performed among women who reported a lump or skin/nipple change when they attended screening. The odds ratio of screen-detected cancer was higher for women with versus without symptoms. Standardized follow-up guidelines might be beneficial for screening programs in order to take care of women reporting signs or symptoms of breast cancer when they attend screening.

Cost differences between digital tomosynthesis and standard digital mammography in a breast cancer screening programme: results from the To-Be trial in Norway.

BACKGROUND: Several studies in Europe and the US have shown promising results favouring digital breast tomosynthesis compared to standard digital mammography (DM). However, the costs of implementing the technology in screening programmes are not yet known. METHODS: A randomised controlled trial comparing the results from digital breast tomosynthesis including synthetic mammograms (DBT) vs. DM was performed in Bergen during 2016 and 2017 as a part of BreastScreen Norway. The trial included 29,453 women and allowed for a detailed comparison of procedure use and screening, recall and treatment costs estimated at the individual level. RESULTS: The increased cost of equipment, examination and reading time with DBT vs. DM was €8.5 per screened woman (95% CI 8.4-8.6). Costs of DBT remained significantly higher after adding recall assessment costs, €6.2 (95% CI 4.6-7.9). Substantial reductions in either examination and reading times, price of DBT equipment or price of IT storage and connectivity did not change the conclusion. Adding treatment costs resulted in too wide confidence intervals to draw definitive conclusions (additional costs of tomosynthesis €9.8, 95% CI -56 to 74). Performing biopsy at recall, radiation therapy and chemotherapy was significantly more frequent among women screened with DBT. CONCLUSION: The results showed lower incremental costs of DBT vs. DM, compared to what is found in previous cost analyses of DBT and DM. However, the incremental costs were still higher for DBT compared with DM after including recall costs. Further studies with long-term treatment data are needed to understand the complete costs of implementing DBT in screening.

Validity and reliability of self-reported health indicators among women attending organized mammographic screening.

AIMS: Risk factors for breast cancer are often used for adjustment in epidemiological studies, including in the evaluation of early performance measures in mammographic screening. Information about risk factors among participants in the Norwegian Breast Cancer Screening Program has been collected since 2006. We aimed to examine the validity of self-reported history of breast cancer and mammographic screening, as well as the reliability of weight and height amongt women attending the program. METHODS: Information from a questionnaire handed in by participants in the program, 2006-2015, was linked to outcomes from the Cancer Registry of Norway. Sensitivity, specificity, and positive predictive values (PPV) were calculated for self-reported histories of breast cancer and screening. Results were stratified by five-year age groups and evaluated using the χ2 statistic. The reliability of self-reported weight and height were assessed using descriptive statistics, histograms, and mean differences. RESULTS: A total of 538,907 of 611,711 (88%) women attending the program during the study period returned at least one part of the questionnaire. The overall sensitivity, specificity, and PPV for breast cancer and mammography were 96.5%, 99.8%, and 81.3%; and 99.9%, 84.4%, and 97.6% respectively. The mean difference in self-reported weight was 0.35 kg and for height was -0.14 cm, over a period of up to 10 years. CONCLUSIONS: Norwegian women attending the screening program are reasonably accurate in self-reporting their breast cancer and mammography histories. On average, women consistently reported weight and height within one kg/cm.