AI-Supported Breast Cancer Screening - New Results Suggest Even Higher Accuracy

The new findings are published in The Lancet Digital Health. The initial results of the Mammography Screening with Artificial Intelligence (MASAI) study* - a randomised trial to evaluate whether AI can improve mammography screening - were published in August 2023. The study started in spring 2021, and the final report will be written next year. A second report has now been published, and Kristina Lång, who is responsible for the study, is pleased to be able to show strong figures.

"Since the first report last year, the number of cancers detected by AI-supported screening has gone from being 20 per cent more to 29 per cent more than those found by traditional screening," says Kristina Lång, researcher and associate professor of diagnostic radiology at Lund University, Sweden and consultant at the Unilabs Mammography Unit in Malmö.

Every year, around one million women are called for mammography screening in Sweden. The mammogram are reviewed by two breast radiologists - a medical skill that is currently in short supply. In the MASAI trial, AI had to identify mammograms with an increased risk of breast cancer. These cases were then reviewed by two breast radiologists. Other mammograms were only reviewed by one breast radiologist. In all these cases, the radiologist(s) was/were assisted by AI, which highlighted suspicious findings on the image.

The new research report is based on results from almost 106,000 women screened for breast cancer. Half of them were randomly assigned to undergo traditional mammography screening, while the other half received AI-supported screening. AI-supported screening was found to pick up not only more cancers overall (in 338 people compared with 262) but also 24 per cent more early-stage invasive cancers (in 270 people compared with 217).

"They also included relatively more aggressive cancers that are particularly important to detect early. At a later stage, the prognosis may have deteriorated and more intensive treatment is often required," says Kristina Lång, pointing to potential benefits in terms of reduced suffering, higher survival rates and lower economic costs if these cancers can be detected more frequently at an early stage.

Pre-cancerous lesions, known as in situ cancers, were also more likely to be detected with AI - 51 per cent more such cases were found (68 people compared with 45). A large proportion of the additional in situ cases found were of the more severe type which also benefits from early detection.

Importantly in this context, the number of false positives did not rise despite an increase in cancer detection.

"A false positive is when a woman is recalled for work up but is then found not to have cancer. This can be a stressful experience for women participating in screening. But only seven more people, corresponding to a one per cent increase, received these false alarms in the AI-supported group compared with the control group," says Kristina Lång.

As in the previous report from last year, AI-supported screening was again found to significantly reduce breast radiologists’ workload by 44 per cent.

Sweden has a generous screening programme by international standards. All women aged 40-74 years are invited for a mammogram every 1.5-2 years. However, the interval between two screening visits can be long enough for cancer to be diagnosed - even if the last screening was considered normal. These so called interval cancers is the next group of cancers that Kristina Lång and her colleagues will analyse.

"This December, the 106,000 women have been followed up for two years, allowing researchers to see how common it is to receive a breast cancer diagnosis between two screening visits. Our hope is that AI will prove helpful here too," says Kristina Lång.

Results from the MASAI trial have already contributed to the implementation of AI support in several regional screening programs in Sweden.

Hernström V, Josefsson V, Sartor H, Schmidt D, Larsson AM, Hofvind S, Andersson I, Rosso A, Hagberg O, Lång K.
Screening performance and characteristics of breast cancer detected in the Mammography Screening with Artificial Intelligence trial (MASAI): a randomised, controlled, parallel-group, non-inferiority, single-blinded, screening accuracy study.
Lancet Digit Health. 2025 Feb 3:S2589-7500(24)00267-X. doi: 10.1016/S2589-7500(24)00267-X

* The MASAI trial has been funded with support from the Swedish Cancer Society, Lund University/ALF, the Confederation of Regional Cancer Centres and MAS Cancer.

Most Popular Now

AI-Powered CRISPR could Lead to Faster G…

Stanford Medicine researchers have developed an artificial intelligence (AI) tool to help scientists better plan gene-editing experiments. The technology, CRISPR-GPT, acts as a gene-editing “copilot” supported by AI to help...

Groundbreaking AI Aims to Speed Lifesavi…

To solve a problem, we have to see it clearly. Whether it’s an infection by a novel virus or memory-stealing plaques forming in the brains of Alzheimer’s patients, visualizing disease processes...

AI Spots Hidden Signs of Depression in S…

Depression is one of the most common mental health challenges, but its early signs are often overlooked. It is often linked to reduced facial expressivity. However, whether mild depression or...

AI Model Forecasts Disease Risk Decades …

Imagine a future where your medical history could help predict what health conditions you might face in the next two decades. Researchers have developed a generative AI model that uses...

AI Tools Help Predict Severe Asthma Risk…

Mayo Clinic researchers have developed artificial intelligence (AI) tools that help identify which children with asthma face the highest risk of serious asthma exacerbation and acute respiratory infections. The study...

AI Model Indicates Four out of Ten Breas…

A project at Lund University in Sweden has trained an AI model to identify breast cancer patients who could be spared from axillary surgery. The model analyses previously unutilised information...

Smart Device Uses AI and Bioelectronics …

As a wound heals, it goes through several stages: clotting to stop bleeding, immune system response, scabbing, and scarring. A wearable device called "a-Heal," designed by engineers at the University...

AI Distinguishes Glioblastoma from Look-…

A Harvard Medical School–led research team has developed an AI tool that can reliably tell apart two look-alike cancers found in the brain but with different origins, behaviors, and treatments. The...

ChatGPT 4o Therapeutic Chatbot 'Ama…

One of the first randomized controlled trials assessing the effectiveness of a large language model (LLM) chatbot 'Amanda' for relationship support shows that a single session of chatbot therapy...

Overcoming the AI Applicability Crisis a…

Opinion Article by Harry Lykostratis, Chief Executive, Open Medical. The government’s 10 Year Health Plan makes a lot of the potential of AI-software to support clinical decision making, improve productivity, and...

Dartford and Gravesham Implements Clinis…

Dartford and Gravesham NHS Trust has taken a significant step towards a more digital future by rolling out electronic test ordering using Clinisys ICE. The trust deployed the order communications...