African-led CHIL AI Lab has unveiled a promising new AI platform that’s poised to transform breast cancer treatment monitoring, enabling much earlier detection of treatment failure and giving oncologists greater confidence in decision-making.
Developed under the leadership of Dr. Nabuuma Shamim Kaliisa, a physician and founder of the Uganda-based health tech company (also known as Chil Artificial Intelligence Lab or Chil Femtech), the platform tackles one of oncology’s most frustrating challenges: the long wait times—often up to six months—for patients and doctors to learn whether a treatment regimen is effective.
The problem it solves
In many settings, especially in resource-constrained regions like parts of Africa, breast cancer patients undergo chemotherapy or other therapies without timely feedback on response. This delay can mean months of ineffective treatment, worsening outcomes, increased side effects, and lost opportunities to switch to better options.
Traditional medical AI tools often function as “black boxes,” delivering binary yes/no predictions without insight into reliability. CHIL AI Lab’s innovation flips this script by incorporating a confidence score for each assessment, quantifying how certain the system is about its analysis of patient scans or data.
Key innovations that set it apart
- Honest Uncertainty: If the AI detects ambiguity in a scan, it openly flags low confidence and escalates the case for immediate human review—preventing risky guesses.
- Ultra-Early Warnings: The platform identifies potential treatment non-response as soon as the first 21 days, a dramatic improvement over conventional timelines.
- Doctor-Centric Design: Rather than rigid outputs, it provides oncologists with probabilistic ranges and explanations, empowering faster, more informed decisions about adjusting treatment plans.
Dr. Nabuuma Shamim Kaliisa emphasized the human impact: “In oncology, an ‘I don’t know’ from a tool can be just as life-saving as a ‘Yes.’ By building a system that can quantify its own certainty, we are giving doctors a tool they can actually trust. We are moving away from guessing and toward a future where we can pivot treatment the moment the math shows a high probability of failure.”
Rigorous testing and validation
The breakthrough stems from a 24-month study involving 306 patients across multiple cancer centers in Africa. Results demonstrated that the AI’s early confidence scores reliably predicted final treatment responses, validating its clinical utility in real-world settings.
This work highlights how African innovators are addressing local healthcare gaps with globally relevant technology—building solutions born from on-the-ground realities rather than imported models.
