Bangladeshi tech startup Arioron has unveiled OncoDetect Titan, an AI system for lung cancer detection that claims 100% sensitivity on a key international benchmark—surpassing Google Health's 94.4% on the same external test data. This positions the Dhaka-based firm as a global leader in robust medical AI designed for real hospitals, not just labs.
Early versions of OncoDetect faltered on diverse hospital scans from outdated machines, a common pitfall in medical AI. Arioron rebuilt it as a "Council of Six"—a multi-model ensemble that filters healthy scans, distinguishes cancer from pneumonia or COVID-19, and handles CT and biopsy images. Components were trained on degraded data to ensure reliability in under-resourced clinics.
The system excelled in tough tests: 100% sensitivity on lung CT benchmarks, perfect false-positive control for infections, and over 96% accuracy on low-quality biopsy images where rivals failed. While peer review and regulatory approval are pending, Arioron highlights this as rare proof of real-world consistency.
Tailored for emerging markets like Bangladesh, OncoDetect Titan could enable earlier detection amid radiologist shortages and old equipment.
The company now eyes hospital pilots and clinical trials, stating:
"We're moving from benchmarks to bedside impact."
This breakthrough underscores Bangladesh's growing tech edge in healthcare.
