Researchers at University College London (UCL) reanalysed data from a prior clinical trial with a bespoke artificial intelligence system and identified a subgroup of rectal cancer patients who appear to gain substantial benefit when irinotecan is added to standard chemoradiotherapy.
Study data and method
The team applied an AI specially developed for the project to diagnostic histology samples from 414 patients with rectal cancer. The system examined images containing millions of cells and classified patients into two groups based on tumour cell density.
Main findings
In the group whose tumours had a high proportion of cancer cells, adding irinotecan to standard chemoradiotherapy produced a significant advantage. Over five years of follow‑up, recurrence risk was reduced by about 43% and mortality risk by roughly 50% in that subgroup.
These results are notable because previous clinical trials that analysed the overall patient population did not demonstrate a clear benefit from adding irinotecan. The AI was able to separate those patients who truly benefited from those who showed no measurable improvement.
Clinical significance
Irinotecan can be effective but is associated with serious side effects, such as severe diarrhoea and reductions in white blood cell counts (neutropenia). Being able to predict in advance which patients are likely to benefit could spare many individuals unnecessary toxicity.
Tool release and next steps
The researchers have made an online tool called Octopath freely available; clinicians can upload digitised biopsy images for automatic AI analysis. The authors stress that these findings need confirmation in independent, prospective clinical trials before the AI system or this therapeutic approach is adopted routinely in treatment decision‑making.
Conclusion
The UCL team's work demonstrates that artificial intelligence can reveal clinically relevant patterns hidden in existing datasets. If validated prospectively, AI‑guided selection could help target irinotecan treatment to patients most likely to benefit and reduce exposure to harmful side effects for others.
