Advantages and disadvantages of artificial intelligence in BreastScreen Norway

Through retrospective studies we will investigate advantages and disadvantages of using artificial intelligence in the assessment of screening mammograms in BreastScreen Norway.

Last updated: 11/20/2023

Background

BreastScreen Norway is a public health service offering screening for breast cancer every other year for women aged 50 to 69. The aim of the program is to detect breast cancer at an early stage, so that fewer women will die from the disease.

The aim of the screening program is to detect breast cancer in an early stage so that more women can receive gentler treatment and fewer women will die from the disease.

Most women, when they attend BreastScreen Norway, do not have breast cancer. Radiologists working in the screening program thus spend much of their time assessing normal mammograms with no signs of breast cancer.

In later years several studies have shown that artificial intelligence (AI) can potentially improve todays screening program, for example as a support for the radiologists in image assessment. By AI we mean digital systems trained by analyzing large amounts of data over time, through which they have learned to recognize patterns in mammograms that could be signs of breast cancer.

The Cancer Registry of Norway have initiated several studies to test different AI systems in the assessment of screening mammograms in BreastScreen Norway. The studies will use mammograms already assessed by radiologists, so-called retrospective testing.

Purpose

The aim is to explore advantages and disadvantages of different AI systems in the assessment of screening mammograms in BreastScreen Norway. The long-term goal is to offer an improved screening service to women participating in the program.

The studies will be performed as retrospective cohort studies, where results from several AI systems will be compared with the radiologists’ assessments. Seven sub-studies, using data from screening examinations performed in Norway, are planned.

The results will provide knowledge about whether the AI systems are of such quality that they can be used in BreastScreen Norway. The results will also be important when designing and performing studies where AI is used in a real screening situation (prospective studies), and to plan a possible future implementation in the program.

Data

The studies will include image data and screening information from examinations performed at different breast centers in BreastScreen Norway. Screening information includes information about attendance and result of the screening examination, as well as the radiologists’ assessment.

The following study population will be included in the sub-studies:

Sub-study I: All screening examinations performed at Haukeland University Hospital in the periods 2016-2019 (Ia) and between 2008-2015 and after 2019 (Ib).
Sub-study II: All screening examinations performed at Stavanger University Hospital from 2008 and onwards.
Sub-study III: All screening examinations performed at Førde Hospital Trust from 2017 and onwards.
Sub-study IV: A subset of screening examinations performed at Oslo University Hospital, Vestre Viken Hospital Trust, and Vestfold Hospital Trust in the period 2014-2015.
Sub-study V: All screening examinations performed at Sørlandet Hospital Trust, University Hospital of North Norway, Østfold Hospital Trust, St. Olavs Hospital, Innlandet Hospital Trust, Møre og Romsdal Hospital Trust and Vestre Viken Hospital Trust from the implementation of digital mammography at each breast center (between 2004 and 2011) and onwards.
Sub-study VI: All screening examinations included in a review study performed in BreastScreen Norway in 2016-2019. Sixteen breast centers reviewed a selection of 150 breast cancer cases each. The sub-study will include image data from all reviewed cases, in addition to all prior screening examinations of the reviewed cases.
Sub-study VII: All screening examinations performed at Akershus University Hospital, Oslo University Hospital and Vestfold Hospital Trust from the implementation of digital mammography at each breast center (between 2004 and 207) and onwards.

The studies will only use data from women who have not refused storage of their personal information related to negative screening examinations in the Cancer Registry, in accordance with the Cancer Registry Regulation. The women will not be contacted regarding the studies, and it will not be possible to identify individuals in the published results.

Organisation

The Cancer Registry of Norway is project leader and the responsible research institution. We are responsible for obtaining all necessary approvals, enter into agreements with AI system vendors, collect image data and screening information, installations of and images analyses in the AI systems, analyses of the data material, and publishing results.

The health trusts represented by the breast centers are cooperating partners in various sub-studies. They are the regional specialists in breast cancer screening and diagnostics. They will contribute with image data for the studies, as well as radiologic expertise and knowledge of practical screening.

At Stavanger University Hospital and Østfold Hospital Trust there are currently two Ph.D. students, Henrik Wethe Koch (Stavanger University Hospital) and Marit Almenning Martiniussen (Østfold Hospital Trust), working on related studies.

The vendors of the different AI systems will assist with installation of the systems and training. The vendors do not have access to personal information from the cases included in the studies, and they cannot influence the published results.

Status per August 2023

The project has obtained all necessary approvals and currently have agreements with three vendors to test AI systems from ScreenPoint Medical, Lunit Inc. and Vara.

Several of the sub-studies have started while others are in planning. There is an ongoing and continuous process of collecting image data to run through the AI systems to get results that can be used in analyses.

So far, the work has resulted in three scientific publications, and more are in progress.

Based on the results, the Cancer Registry is planning a study where AI will be included as a part of the image assessment in an otherwise normal screening situation. You can read more about this study – AIMS Norway – here.

Publications

Larsen M, Aglen CF, Lee CI, Hoff SR, Lund-Hanssen H, Lång K, Nygård JF, Ursin G, Hofvind S. Artificial Intelligence Evaluation of 122 969 Mammography Examinations from a Population-based Screening Program. Radiology. 2022 Jun;303(3):502-511. doi: 10.1148/radiol.212381. Epub 2022 Mar 29.

Larsen M, Aglen CF, Hoff SR, Lund-Hanssen H, Hofvind S. Possible strategies for use of artificial intelligence in screen-reading of mammograms, based on retrospective data from 122,969 screening examinations. Eur Radiol. 2022 Dec;32(12):8238-8246. doi: 10.107/s00330-022-08909-x.. Epub 2022 Jun 15.

Koch HW, Larsen M, Bartsch H, Kurz KD, Hofvind S. Artificial intelligence in BreastScreen Norway: a retrospective analysis of a cancer-enriched sample including 1254 breast cancer cases. Eur Radiol. 2023 May;33(5):3735-3743. doi: 10.1007/s00330-023-09461-y. Epub 2023 Mar 14.