Radiation plans for cervical cancer could move from weeks to just over an hour
A radiotherapy planning room where an AI-generated treatment map wraps around a pelvic tumour model while clinicians review the plan on clinical monitors.📷 AI-generated image / TECH&SPACE
- ★ARCHERY included more than 1,000 patients in four countries and tested AI planning for cervical and prostate cancer.
- ★The AI reached a high standard in more than 95% of cervical cancer cases and 85% of prostate cancer cases.
- ★Planning that can take days or weeks was cut to just over an hour, a major point for countries with limited radiotherapy access.
Radiotherapy is not a narrow technical add-on to oncology. It is one of cancer care’s core treatment pillars. That is why the result reported by MedicalXpress matters: an AI tool developed through work led by researchers at University College London and the London School of Hygiene & Tropical Medicine has shown it can effectively plan radiotherapy for cervical cancer and prostate cancer in a large international trial.
The ARCHERY trial involved more than 1,000 patients across India, South Africa, Jordan and Malaysia. That geography is not a footnote. The central question is not whether an algorithm can perform well in a controlled technical setting, but whether it can help in health systems where expert planning capacity is scarce. Radiotherapy planning decides how to deliver radiation to the tumour while limiting damage to nearby healthy tissue. In routine practice, that is complex work that can take experienced teams days or weeks.
The large ARCHERY international trial shows UCL and LSHTM software can compress complex planning work from days or weeks to just over an hour.
A closer operational view of a hospital radiotherapy workflow: treatment plan contours, dose lines and patient pathway steps moving from backlog to approved plan.📷 AI-generated image / TECH&SPACE
According to the reported results, the AI achieved high-standard planning in more than 95% of cervical cancer cases and 85% of prostate cancer cases. Planning time fell to just over an hour. That does not remove the need for clinical review or human accountability. It does suggest that one of the slowest parts of the chain could become less dependent on a small pool of highly trained specialists.
The strongest public-health signal is in cervical cancer. The supplied trial context says 94% of cervical cancer deaths occur in low- and middle-income countries, with 350,000 deaths in 2022. Access to radiotherapy is also deeply uneven: only 10% of people who need radiotherapy in low-income countries receive it, rising to about 40% in middle-income nations.
That is where AI becomes more than a software story. The WHO cervical cancer elimination initiative depends not only on vaccination and screening, but also on treatment for people who already have disease. If high-quality radiation planning can be made faster and more repeatable, hospitals with limited planning staff may be able to treat more patients without lowering the standard of care.
The limits are just as important as the promise. This is evidence about radiotherapy planning, not a claim that AI independently treats cancer. Clinical oversight, plan verification, machine availability and local implementation still decide whether patients actually receive care. But because radiotherapy helps cure about 40% of cancer cases globally, and because access is still so uneven, automating a difficult planning step could become a practical health-system intervention. The next question is how reliably this kind of system can be monitored, maintained and embedded into routine hospital workflows across very different clinical settings.
