Editor's note: In the original version of this article there were some errors in the reporting of information made during a presentation at the 2008 RSNA meeting by Dr. David Levin, professor and chair emeritus of the radiology department at Thomas Jefferson University Hospital in Philadelphia. A corrected version follows.

As percutaneous transluminal angioplasty and stent placement replace bypass surgery in the treatment of peripheral arterial disease, cardiologists and surgeons now are performing far more of the procedures than radiologists.

A study using Medicare databases for 1996 through 2006 showed that the number of percutaneous transluminal angioplasty and stent procedures increased by 213% (from 240 to 751 per 100,000 patients) in that time. During the same decade, bypass surgery saw a 47% decline.

The utilization rate of peripheral angioplasty and stenting among surgeons who do the procedures mushroomed by 1095% (from 21 to 251 per 100,000 Medicare beneficiaries). Among cardiologists who perform the procedures, the rate grew by 395% (from 55 to 272 per 100,000). Meanwhile, the utilization rate among radiologists increased by only 24% (from 151 to 187).

In 1996, radiologists performed far more of the procedures than either cardiologists or surgeons.

But the research showed that the market share for radiologists declined in the study period from 63% to 25%, while for cardiologists it climbed from 23% to 36%, and for surgeons the share rose dramatically, from 9% to 33%.

Dr. David Levin, who oversaw the study and is professor and chair emeritus of the radiology department at Thomas Jefferson University Hospital in Philadelphia, said of the shift, "The only logical explanation would appear to be self-referral."

Levin added that fewer radiology residents are going into the intervention subspecialty.

He also said radiologists should set up admitting services and vascular clinics of their own and then market them to primary care doctors in their regions.

But Dr. R. Torrance Andrews, an associate professor and chief of vascular and interventional radiology at the University of Washington Medical Center in Seattle, disagreed with Levin's conclusion, asserting that surgeons and cardiologists are simply becoming more aggressive in going after such patients.

"We're not reaching out and mining these patients," Andrews said. "They're just better at getting patients into the system than we are."

Levin responded that, unlike interventional radiologists, cardiologists and surgeons are primary care physicians.

But Andrews said interventional radiologists must adopt the tactics of cardiologists and surgeons if they are to reverse the downward trend.

"The reason it has become critical now and no longer optional is the people who traditionally have been gatekeepers and triage specialists are now doing the procedures themselves," Andrews said. "They're becoming interventional radiologists, so we need to become primary care doctors. That's the reality."

Internists and primary care doctors have traditionally sent their patients to vascular surgeons "because that's the way they've always done it," he said.

"But what they don't realize is the high quality that their patients had been getting through interventional radiology is now being compromised because it's being done by people who basically have limited training and experience in endovascular techniques, and I don't think they realize it," Andrews said. "We need to make sure people understand the old practice model doesn't lead to the same end point."

He also took issue with Levin's assertion that the drop-off can simply be attributed to increased self-referral.

"It's a buzzword that implies violation of the Stark law," he said.

What worries the radiology community, Andrews said, is orthopedic surgeons and neurologists who buy their own MRIs and cardiologists who buy their own CT scanners.

"But I think using image guidance to perform surgery, which is basically what we and vascular surgeons do, that's not self-referral," he said. "That's just using imaging as a tool for doing a procedure, and I don't think there's any way that radiologists can make the argument that it's not acceptable and constitutes actual self-referral."

Radiologists can't have it both ways, Andrews said.

"I think if you say that nobody who orders an angioplasty procedure can actually perform it because that constitutes self-referral, then that puts interventional radiologists in the same boat -- we can't do it either," he said. "If interventional radiologists in clinics want to order an MRI of the pelvis for a fibroid patient and that constitutes self-referral, then interventional radiologists can't be part of a radiologic group."

Part of the problem, Andrews said, is that radiologists are often stymied by their own practice groups, which usually think radiologists' time in clinics should be spent reading CT scans.

From his own experience, Andrews said he has worked out that imaging or procedures that he has ordered have generated over $1 million in billing over a two-year period, which he estimates to equal about $2500 per hour in imaging billing.

"Although IRs are not reading studies in the clinic, they're actually generating imaging revenue by ordering studies for patients they're caring for and by ordering intervention, which they perform," he said.