Breast cancer remains the leading cause of cancer-related death among women, with more than 2.3 million new cases annually worldwide (WHO, 2024).
In this context, discovering methods capable of predicting the risk of breast cancer before it appears represents one of the most powerful promises of modern medicine.
In 2019, researchers from Massachusetts General Hospital (MGH) and MIT CSAIL created a deep learning model called Mirai, which analyzes mammograms and estimates the likelihood of developing breast cancer within the next five years.
The results are impressive: in published studies, the model was trained on over 62,000 patients and achieved an accuracy between 75–84%, surpassing traditional risk assessment tools such as Tyrer–Cuzick.
But beyond the technological excitement lies a crucial question:
How do we know these results will translate into real benefits for patients?
The answer lies in one essential concept: clinical trials.
Every medical innovation — whether a drug, a surgical technique, or an AI algorithm — must undergo scientific validation.
In breast cancer, clinical trials are the bridge between laboratory discovery and real-world impact.
1. Proving Effectiveness and Safety
An AI model may look revolutionary in a lab, but only testing it on real patients can show whether:
its predictions reduce mortality,
improve diagnostic timelines,
and avoid unnecessary interventions.
For example, MGH and MIT researchers are currently conducting prospective clinical trials, where women identified as “high risk” by the Mirai model are monitored for years to confirm whether the predictions hold true.
2. Enabling Truly Personalised Medicine
Modern clinical trials no longer compare only treatments — they investigate individual risk profiles.
In breast oncology, this means tailoring screening and care based on:
age,
breast density,
family history,
and, increasingly, AI-generated risk assessments.
Women with higher predicted risk may receive more frequent mammograms or additional imaging (MRI, 3D ultrasound), while those at lower risk can avoid unnecessary radiation exposure and anxiety.
A vital role of clinical trials is ensuring fairness and accuracy.
AI models must work equally well across diverse groups — European, Asian, African, Latin American, and more.
Only by including diverse populations in clinical trials can we prevent biased predictions and inequities in care.
Artificial intelligence does not replace clinical trials — it enhances them.
Instead of waiting years for observational patterns to emerge, algorithms can rapidly analyze millions of images and medical records, identifying trends worth investigating in a controlled study.
AI can also:
accelerate patient recruitment for trials,
process complex imaging and genetic data in minutes,
identify subgroups of responders that human analysis might overlook.
This leads to clinical trials that are faster, more precise, and more effective.
According to a study published in Science Translational Medicine (2022), the Mirai model was validated on data from the U.S., Sweden, Israel, Taiwan, and Brazil.
The model demonstrated consistent performance across populations, with an AUC between 0.75 and 0.84 — showing that AI can reliably predict future breast cancer risk even in ethnically diverse groups.
However, researchers stress that only long-term clinical trials can confirm whether these predictions ultimately lead to saved lives.
Predicting breast cancer five years before onset is no longer science fiction. It is an emerging scientific reality.
But for this reality to become everyday clinical practice, we need strong collaboration among researchers, clinicians, patients, and regulators.
Artificial intelligence brings speed and precision.
Clinical trials bring trust and proof.
Together, they can transform medicine from reactive to predictive and preventive, offering women more time, more hope, and more control over their health.