The solution, used along with CT scans and integrated directly into the PACS system, helps highlight abnormal or hyperdense findings, such as hemorrhages or tumors, for radiologists, helping them to prioritize urgent cases.
Tom Valent, the vice president of business development for Aidoc, said that instead of going pathology by pathology, the solution detects a group of visual abnormalities representing the different pathologies.
“The AI is trained to detect those types of visual abnormalities,” Valent told HCB News during an interview at RNSA. “So, anything the radiologist would put in his or her report, that would be classified from a visual perspective as, in terms of the head, intracranial hyperdensity. That represents a big volume of cases.”
For cases in which there is no pathology suspected, the AI solution can help radiologists prioritize, with processing time of under two minutes.
“We already have the Aidoc findings, typically, before the radiologist has even seen the scan,” Valent said. “Oftentimes they don’t know what to expect, so they don’t know if it is urgent or not. We can already indicate to the radiologist, we believe that it’s positive and thus we recommend you look at it.”
The analyzed information is added to the PACS system, which Valent says is an advantage.
“We’re not a secondary type of reader with our own platform,” Valent said. “We inject our information into the existing PACS.”
While the company’s first product is for CT, Valent said the product is scalable and they hope to eventually capture more imaging modalities.
“AI can help with time, help prioritize, and time is of the essence for those trauma cases,” Valent said. “That’s why we’ve begun there.”
The company expects FDA clearance of its product next year and is conducting research at U.S. hospitals, including at Cedars-Sinai Medical Center in Los Angeles, where an assessment of Aidoc's solution earlier this year resulted in accuracy in scan analysis.
“In our clinical trial, Aidoc's technology has demonstrated its ability to enhance our radiologists’ workflow, as abnormal scans can be prioritized and more carefully reviewed,” Dr. Barry Pressman, chairman of imaging at Cedars-Sinai Medical Center, said in a statement. “Our firsthand experience has led me to believe in the technology’s potential to achieve a significant increase in our radiologists’ productivity and accuracy. It’s a win both for our physicians and our patients. Aidoc’s AI-powered solution will help our radiologists be their best, and streamline their workflow.”
Valent said radiologists shouldn’t fear AI solutions as a replacement for their expertise.
“If anything, AI will help radiologists replace radiologists who don’t use AI,” Valent said. “What we’re doing is looking for those areas where we can provide value. There’s a big bottleneck in radiology in the U.S. and pretty much everywhere in the world where the amount of imaging is increasing rapidly, but the number of radiologists is pretty much stagnant. And that means that their workload is ever growing. We’re trying to look for places where we can help optimize the workflow.”