AI: Artificial Intelligence or Augmented Interpretation for medical imaging?
The short answer, AI can be both, either or sometimes neither.
Now for the long winded, boring, I need to read something that will put me to sleep at 2 am answer*:
The BOTH: In prescreen human head CT exams for acute neurological events, an AI algorithm has been trained to BOTH analyze head CT images, prioritize cases in a radiologists worklist and alert the clinician to potentially critical findings. Just imagine how useful it would be having a system put things that are more critical in front of someone’s eyes and alerting the attending clinician.
The EITHER: The brain, a top area of current AI focus, is leading to publications on brain research to help identify and treat brain tumors. These varied studies demonstrate AI’s duality of being used for EITHER augmented interpretation or artificial intelligence. BioMind AI, an image evaluation software, was compared to a team of senior doctors in China in diagnosing brain tumors. BioMind AI came out on top correctly identifying 87% of brain tumors in a grouping of 225 cases. The doctors? Well they came in second correctly identifying 66%. A different study optimized a patient’s glioblastoma treatment plan using an AI algorithm. The algorithm was able to reduce chemotherapy and radiotherapy dosing levels by 25 to 50% without affecting treatment effectiveness. These are two studies of many that show AI’s potential usefulness as an augmented interpretation resource as well as an artificial intelligence resource to impact disease diagnosis and treatment.
The NEITHER: AI has now ventured outside of medical imaging where it is NEITHER used for augmented interpretation or artificial intelligence of medical images, but instead is being used for scheduling workflow and pre-ordering of tests. Hopefully it’s a not too distant future where your recheck exams are automatically scheduled at the perfect follow-up time, exact time of day needed for blood tests you’ll be ordering, the owner is automatically reminded of how to prepare their pet for follow-up (to feed or not to feed, time to medicate or not to medicate) and the most up to date, appropriate follow-up tests are pre-ordered for you.
It’s getting easier to see why there’s an exploding interest in AI uses in human medicine. Early research has shown incredible potential however human medicine currently faces a stumbling block that veterinary medicine doesn’t, insurance reimbursement. This impediment has led to AI being used for augmenting image interpretation and increasing efficiencies for insurance approved items. Make no mistake about it, veterinary medicine has a BIG advantage here with the pet owner still being the main payer of veterinary services. This advantage cannot be overlooked and why veterinary medicine should/will be in the mix of AI applications that enhance veterinary care.
Right now more veterinary companies are getting involved in veterinary AI, utilizing their own prior knowledge and experience to develop AI applications. From the laboratory experience, Idexx® has released an AI enhanced urine crystal analyzer software called SediVue. At our sister organization, Vetology® Innovations, we are boarded veterinary radiologists so our AI focus is on veterinary medical imaging applications such as our AI image match program, 90-second AI heart evaluation or keyword search AI program. Labs and radiologists aren’t the only ones getting into veterinary AI applications. Another company, Westside Innovations, being lead by a veterinary internal medicine specialist is believed to be putting IM expertise to work in evaluating AI enhanced blood test results.
AI, whether you’re referring to artificial intelligence or augmented interpretation, is here to enhance veterinary care. But above all else, nobody in veterinary medicine should fear being replaced by AI in their lifetime. What makes me so sure? What veterinarians and veterinary staff do on a day to day basis is so varied, technically challenging, work intensive and skilled that it’s going to take a very long time, if ever, for AI to replace rather than enhance the veterinary professional.
*It almost goes without stating that reading AAVR articles can induce sleep and must not be read while operating heavy machinery or driving.
-Seth Wallack, DVM, DACVR
Dr. Wallack, DACVR is the founder of the AAVR. He is also the founder and CEO of Vetology® Innovations and the Veterinary Imaging Center of San Diego, Inc. He has been working in veterinary radiology, teleradiology and AI for over 15 years.
Check back often for more updates and articles from AAVR.org or subscribe to be emailed the AAVR newsletter.
To learn more about Vetology® AI check out the ai.vetology.net website or check out Ask A.I.M.E. on the Vetology® platform.
AAVR has no affiliation with Idexx® or Westside Innovations. The information in this article is accurate to the best of our knowledge. Vetology® is a sister organization of the AAVR. The AAVR did not receive any payment for mentioning any company or study in this article.
Vetology® is a registered trademark of the Veterinary Imaging Center of San Diego, Inc., San Diego, hippyland California.
SediVue is an Idexx product. Idexx® is a registered trademark of Idexx Reference Laboratories, Westbrook, in the Moose state.