[Image courtesy of Flywheel]
Minneapolis-based Flywheel has received $54 million in a Series D round backed by NVIDIA, Microsoft, Hewlett Packard Enterprise and other strategic investors.
Flywheel will invest the cash to fuel growth in core verticals, the pharmaceutical and public sector healthcare sectors. A notable Big Pharma customer is Genentech. The company also aims to ramp up expansion into emerging areas such as healthcare providers, payers, IT service providers and software vendors. Finally, Flywheel intends to extend its global footprint, especially in key markets in Europe.
Current challenges in medical AI
Medical imaging has long faced challenges with data complexity and a mosaic of metadata and formats, including DICOM (Digital Imaging and Communications in Medicine) to JPG or TIFF and modalities spanning MRI, CT, PET and ultrasound. Despite some standardization, fragmented formats prevail from different imaging modalities and vendors. The resulting data variety can make data processing a chore. “Images come with embedded metadata and inconsistent file formats across different scanning devices and vendors,” Olson noted. “It’s not the size of medical images that pose a challenge. It’s the complexity.”
Challenge number one is making sense of the data and managing it, Olson said, adding “AI and ML quality depend on well-formed, consistent, diverse data, not just quantity but diversity.”
Future prospects
Flywheel has developed a number tools that facilitate expediting drug development processes through automating data preparation and enabling collaboration. “The automation can eliminate manually preparing data for machine learning and AI analysis. All of that happens within Flywheel,” Olson said. “We then support data science teams.”
Looking ahead, Olson anticipates significant changes in the industry. “I foresee major progress and shifts in data collection, access, and developing AI, leading to breakthroughs… We remain in the early stages as an industry,” he states. His anticipation stems from Flywheel’s drive to tackle the challenges of accessing and using medical imaging data for AI.
As the field of medical AI continues to grapple with these challenges given burgeoning public interest in AI. “Looking ahead to the next six months and further into the next couple of years, I believe there will be a surge of interest in AI due to the focus on large language models such as ChatGPT.”