What You Ought to Know:
- Voio, a frontier AI lab spun out of UC Berkeley and UCSF, has secured $8.6M in seed funding led by Laude Ventures and The Home Fund to launch a unified studying platform
- The corporate’s open-source mannequin, Pillar-0, demonstrates unprecedented accuracy—outperforming main fashions from Google and Microsoft by 10%–17% AUC—to handle radiologist burnout and diagnostic delays.
Fixing Fragmentation and Burnout in Radiology
Voio’s core mission is to unravel the fragmentation that plagues radiology reporting, which at present requires fixed context-switching between picture viewers, reporting software program, EHRs, and a number of AI instruments.
Voio’s first focus is constructing a unified studying platform powered by frontier vision-language fashions.
Seamless Reporting: The platform interprets full exams and drafts high-quality studies, permitting radiologists to overview and finalize sooner with out sacrificing accuracy.
Restoring Stability: By designing the reporting expertise from the bottom up, Voio goals to scale back unsatisfying grunt work and let radiologists give attention to affected person care.
This unified method goals to handle the essential workforce scarcity in radiology, which, with 375 million CT scans carried out yearly, results in longer turnaround occasions and diagnostic delays.
Pillar-0: Unprecedented Accuracy in Medical Picture Interpretation
Pillar-0, the preliminary iteration of the corporate’s AI, is benchmarked because the world’s most correct AI mannequin for medical imaging. Key efficiency highlights embody:
Accuracy Lead: Demonstrated a ten% to 17% AUC accuracy enchancment over main fashions from Google, Microsoft, and Alibaba.
Broad Protection: Achieved a 0.87 AUC throughout 350+ findings in chest CT, stomach CT, mind CT, and breast MRI scans.
Predictive Energy: Positive-tuning of P0 led to a 7% enchancment over the state-of-the-art in predicting future lung most cancers (Sybil-1) throughout an exterior validation research at Massachusetts Basic Hospital.
“Radiologists shouldn’t have to decide on between velocity and high quality,” mentioned Dr. Maggie Chung, Co-Founder and Medical Lead of Voio. “Our purpose is to make radiology reporting seamless by drafting full studies and connecting photographs, historical past, and prior exams into one clever platform that feels pure to make use of. By re-designing the reporting expertise from the bottom up, we are able to cut back unsatisfying grunt work and let radiologists give attention to affected person care.”
