Artificial intelligence is beginning to reshape how cancer treatment decisions are made. In a significant development, the United States Food and Drug Administration (FDA) has cleared a new AI-powered test designed to help doctors assess whether certain breast cancer patients may need aggressive chemotherapy or could safely avoid it.
The tool, called ArteraAI Breast, is being described as the first FDA-cleared artificial intelligence-based risk stratification platform for patients with early-stage hormone receptor-positive (HR+), HER2-negative invasive breast cancer. The AI model analyses digital pathology slides and clinical data to predict the likelihood of cancer spreading or recurring, helping oncologists personalise treatment intensity.
The development comes at a time when breast cancer continues to be the most commonly diagnosed cancer among women worldwide. According to the World Health Organisation, around 2.3 million women were diagnosed with breast cancer in 2022, and approximately 670,000 deaths were linked to the disease globally.
Breast cancer treatment is rarely one-size-fits-all. Depending on the type and stage of the disease, patients may undergo surgery, chemotherapy, radiation, hormone therapy, targeted drugs, or a combination of these approaches. But one of the biggest challenges for oncologists has been determining which patients genuinely benefit from chemotherapy and which patients may be exposed to unnecessary toxicity.
ArteraAI Breast aims to bridge that gap. The platform uses scanned images of tumour tissue obtained during surgery and combines them with clinical variables to generate an AI-based risk score. The model was reportedly trained using data from more than 8,500 breast cancer patients enrolled in clinical trials.
The final assessment categorises patients into low-risk or high-risk groups for distant metastasis, the spread of cancer to other organs.
“AI is helping early diagnosis of breast cancer mainly by improving mammogram interpretation. Computex suspicious regions earlier reduced breast cancer and improved tuning accuracy, especially in dense breast tissue. In terms of accuracy, modern AI systems are showing performance close to expert breast metabolism studies, but currently AI works best as a support tool alongside doctors rather than replacing them completely," says Dr Amol Akhade, Consultant Medical Oncology at Fortis Hospital Mulund.
Regarding cost advantage, AI does not replace mammography itself, but it can reduce unnecessary recalls, improve work-life efficiency, and potentially reduce long-term treatment costs through earlier diagnosis. As far as adaptation is concerned, many centres in Europe, the UK, and the US are already using AI to treat breast screening. "In India, adaptation is still at an early stage, and we are still not yet in that zone where we can use AI to treat," says Dr Akhade.
Currently, one of the most commonly used tools to predict recurrence risk in HR+ breast cancer patients is the Oncotype DX genomic test. The test analyses tumour biology and provides a recurrence score that helps determine whether chemotherapy is likely to help.
However, these genomic assays can be expensive, may take several weeks to process, and are not always easily accessible, particularly outside major cancer centres or in low-resource settings.
The new AI model could potentially offer a faster alternative because it relies on pathology slides that are already generated during routine diagnosis. Experts say that if validated further in real-world settings, such tools could help streamline treatment decisions while reducing costs and delays.
For many women with early-stage HR+ breast cancer, especially post-menopausal women with node-negative disease, the decision to undergo chemotherapy often falls into a clinical “grey zone”.
"AI-based testing also helps increase detection rates while maintaining low false positives. AI is prompt in detecting aggressive cancers earlier, improving the quality of life of the patients. It reduces the workload of the radiologists, making screening faster and more effective. So, it's reliable, improves accuracy and consistency in cancer diagnosis," says Dr Adwaita Gore, Senior Consultant Medical Oncologist, Zen Multi-Speciality Hospital, Chembur.
Doctors may worry about undertreating aggressive cancers, while patients fear the physical and emotional toll of chemotherapy. Side effects can include nerve damage, fatigue, infections, infertility, heart complications, and long-term quality-of-life issues.
AI-based predictive tools may help bring greater clarity to these decisions. If patients identified as low-risk can safely skip chemotherapy, it could spare them from avoidable toxicity and financial stress. At the same time, patients flagged as high-risk may benefit from more intensive treatment earlier in the disease course.
Experts urge caution
Despite the enthusiasm surrounding the FDA clearance, cancer specialists say the technology will need extensive validation before it becomes routine clinical practice. Doctors are expected to closely evaluate how the AI model performs against existing gold-standard tests and whether its recommendations translate into long-term survival benefits.
Questions around cost, insurance coverage, integration into hospital workflows, and transparency of the AI algorithm also remain.
Some oncologists have pointed out that AI systems in healthcare must avoid becoming “black boxes”, where doctors receive risk predictions without understanding how conclusions were reached.
India continues to witness a rising breast cancer burden, with increasing numbers of younger women being diagnosed in urban centres. Access to advanced genomic testing remains uneven across the country, particularly outside metropolitan hospitals.
If AI-driven pathology tools prove reliable and cost-effective, experts believe they could eventually improve cancer care access in resource-constrained settings by enabling faster risk assessment using existing pathology infrastructure.
However, Indian oncologists note that data generated primarily from Western populations may need validation among Indian patients before widespread adoption.
While AI is unlikely to replace oncologists anytime soon, developments such as ArteraAI Breast signal a broader shift towards precision medicine, where treatment decisions are increasingly guided by data-driven prediction models tailored to individual patients.
