By Chaunie Brusie

Albert Hsiao, MD, PhD, is not afraid of change. At the age of 40, Hsiao, an associate professor in-residence, radiology at UC San Diego is the co-founder of Arterys, Inc., the first cloud-native artificial intelligence (AI) medical imaging platform. From early on in his career, he embraced the role that fast-paced technology has had in imaging—and has big plans to see it through to the future.

Involved in the fields of cardiovascular imaging, oncologic imaging, lung imaging, and genomics, Hsiao’s current areas of research focus on AI with deep learning, 4D flow MRI, and spectral CT. His recognitions include the “Whitehill Prize for Excellence” from the UC San Diego Academy of Clinician Scholars (2017), “Early Career Investigators in Imaging” from the Academy for Radiology and Biomedical Imaging Research (2017), “Triton 40 Under 40” from UC San Diego (2018), and ARRS Scholar from the American Roentgen Ray Society (2019).

‘Special Forces’ in the Field

Hsiao explains that he first became fascinated with imaging during graduate school, after taking an MRI signal processing course and helping teach an anatomy course to medical students. “Much of my education in biology up to that point was cellular and molecular, which felt disconnected from organ systems and exactly where disease processes were taking place,” he says. “Imaging felt like the natural intersection of these microscopic and macroscopic processes.”

Upon returning to medical school, Hsiao says he began to see how much technology drove the field of radiology, and how advanced interventional procedures could directly target infections, cancer, and cardiovascular disease. “I had a patient tell me he thought we were like the ‘special forces’ of medicine,” Hsiao remembers. “It was exciting, fun, and amazing at the same time. There was so much intellectual detective work involved in everything in radiology, which appealed to my desire to practice medicine in a deeply analytical way, rather than simply absorbing and reciting things that people had taught me.”

Through his training, he learned of the transformation that radiology had undergone in the prior decades, thanks to the proliferation of technologies such as CT and MRI. And while at first, Hsiao lamented missing out on being an early adopter of such innovations, he has come to realize that AI is seeing the same boom that happened at the advent of CT and MRI.

“It looks like I’m smack dab in the middle of it, with the right set of skills to help carry the field forward,” he says.

Eyes on the Future

While cardiovascular imaging is arguably one of the most exciting areas of radiology, it is also one of the most challenging to perform—due to the motion of heart, lungs, and flowing blood.  This can create great disparities between what is available at high-end academic institutions and the community—a problem that Hsiao believes is worth solving.

“With deep learning, we’re beginning to see a mechanism that can encapsulate subspecialty-level physician knowledge into software algorithms,” he says. “It would have been almost impossible to do this 10 years ago, but it’s now possible for subspecialty physicians like me to train neural networks to behave as physician extenders. It would be amazing to be able to encapsulate exactly how I would use the CT or MRI into a software algorithm that could be deployed to any hospital in the world.”

Moreover, Hsiao believes that technology will be a boon to both patients and physicians in the future. For instance, in his opinion, AI applications may be used to ensure higher accuracy to diagnose and thus, treat patients, but address the increasing demand for “physician productivity” in the field.

“The number of patients certainly will increase in the coming years as the baby boomer generation ages, and the number of physicians certainly can’t be trained fast enough to accommodate,” he says. “If we can leverage AI and other new technologies to simplify and assist the practice of medicine, we might have some hope of keeping the whole system afloat.”

In fact, one of his favorite parts of his career has derived out of seeing his lab’s innovations directly impact the care of patients and watching each of the “little dreams” his team promised to deliver come to life. “We have already seen this happen with 4D flow MRI, but we are beginning to also see this with our work in spectral CT and artificial intelligence,” Hsiao says.

Learning from Failure


Hsiao (right) partakes in his favorite past-time.

Hsiao knows that innovation often comes from being able to fail—and learn from those failures.For every success I get to talk about, I feel like I’ve failed five to 10 times before to reach it,” he admits. But he says the supportive teams at Arterys, his division at UC San Diego, GE Healthcare, and, most recently, the ARRS, have each helped him continue to push through to find solutions.

“My career, I think, is just getting started,” he adds. “The big dream of an MD/PhD is to see a problem in the clinic, solve it, and come back and deliver it to patients. I’ve had the privilege of doing this a few times now and would love to keep doing it. It took me a while to realize that partnering with the industry to build these into products is the only way to do this in a tangible, long-term way.”

When the big-dreaming doctor isn’t hard at work advancing the field of cardiovascular imaging, he enjoys spending time with his family and training in martial arts. In fact, Hsiao earned his black belt in Kenpo karate last July.

The same dedication, patience, and willingness to roll with the punches that brought him to black belt status continues to carry him through his work as a radiologist. “I have high aspirations for what Arterys can do to bring AI and advanced imaging technologies to the world, and the level of care that UC San Diego could deliver to our community,” he says. “It has been a rollercoaster so far, and I’m certain we’re still at the beginning of what is to come.”

Chaunie Brusie is associate editor of AXIS Imaging News. Questions and comments can be directed to [email protected].