aiREAD - Accurate and Intelligent Reading for EArlier breast cancer Detection

The way that mammograms are read by breast radiologists has not really changed since the introduction of breast cancer screening programs. In this project, we will discover new knowledge on how the visual system of radiologists reacts to reading mammograms, and develop new computer methods to analyze mammograms, to improve the early detection of breast cancer with screening. In the Netherlands, all women 50 to 75 years old are invited for breast cancer screening every two years, resulting in over one million screening mammograms being acquired annually.

The Dutch screening program results in about 24,000 women per year being referred to hospitals for additional testing, of which 7,000 have cancer. In addition, another 2,000 women are diagnosed with breast cancer within the two years after a normal screening exam. Although this performance is representative of screening programs from around the world, there is room for improvement. In recent years, there have been important developments in artificial intelligence (AI) for use in medical imaging, including analyzing mammograms. We also now have new insights on how the human brain interprets what it is seeing, and how it adapts to what it is seeing. In this project, we aim to develop new AI methods of image analysis and interpretation of screening mammograms, and combine them with new, optimized methods for the breast radiologists to interpret these images.

In this way, by finding the best combination of computer analysis and human reading of the mammograms we acquire during breast cancer screening, we expect that we will find more cancers. In addition, and just as importantly, this will result in fewer false alarms, in which healthy women are told that they need to go to the hospital for additional testing. All this will be achieved while resulting in a more efficient, more economical, screening program.

Researchers:

Ioannis Sechopoulos

Sarah Verboom

Jessie Gomers

Mireille Broeders

Key Publications:

  • Gommers JJJ, Abbey CK, Strand F, Taylor-Phillips S, Jenkinson DJ, Larsen M, Hofvind S, Sechopoulos I, Broeders MJM. "Optimizing the Pairs of Radiologists That Double Read Screening Mammograms", 2023. Abstract. DOI.
  • Sarah D. Verboom, Marco Caballo, Mireille J. M. Broeders, Jonas Teuwen, Ioannis Sechopoulos. "Deep learning-based breast tissue segmentation in digital mammography: generalization across views and vendors", 2022. Abstract. DOI.
  •  Jessie J. J. Gommers, Craig K. Abbey, Fredrik Strand, Sian Taylor-Phillips, David J. Jenkinson, Marthe Larsen, Solveig Hofvind, Mireille J. M. Broeders, and Ioannis Sechopoulos. "Modeling Radiologists' Assessments to Explore Pairing Strategies for Optimized Double Reading of Screening Mammograms ", 2024. Abstract. DOI.
Tomographic Imaging

Overige afdelingen Imaging