Dollars and Sense: The Economics of PET

For as long as anyone can remember, positron emission tomography (PET) always has had a bright future. The problem has been reimbursement, but even that luck has changed.

PET has never had an easy ride from the folks in charge of reimbursement. It’s a bit like the youngest child syndrome in reverse. In the 1980s, CT and MRI came blazing out of the gate with numerous clinical indications, but instead of “the parents” becoming more lenient by the time PET came along in the late 1990s, they (CMS) introduce a whole new and tougher set of standards (evidence-based and outcomes-based medicine).

PET’s initial cancer indications and reimbursement were the result of “pure political force,” according to one Capital Hill insider. Apparently, Sen. Ted Stevens of Alaska, a ranking member of the health committee, had a personal family experience with PET, deemed it valuable, and through pure political will got HCFA (now CMS) to pay for PET in 1997. In January, Sen. Stevens appeared on John McLaughlin’s “One on One,” television broadcast about reimbursement for diagnosing Alzheimer’s Disease with PET. Medicare currently does not reimburse for a PET scan diagnosis of Alzheimer’s, even though several studies show that it significantly outperforms the traditional diagnostic methods.

“I really think it’s sort of a sabotage of a new system because the people who are involved on that advisory committee [MCAC] all have interests in the existing diagnostic systems,” Sen. Stevens said. “We believe that PET has a 93 percent accuracy in the early diagnosis, and that compares to about 50 percent on the traditional diagnostic devices. But this advisory committee has consistently recommended against using it. I just think it’s one of those things where the people [on the MCAC] — neurologists and others — may benefit from charging patients under the old system. They’ve got their investments in the old systems, and they just don’t want to move forward.”

Despite the challenges, PET finally may be coming into its own.

In April, a bipartisan bill, the Medicare Innovation and Responsiveness Act, was introduced in both the U.S. House and Senate with the aim of eliminating the serious delays in the amount of time it takes Medicare to make new technologies and procedures available to beneficiaries. The bill targets hospital inpatient payments, clinical trials access, Medicare coverage of new technologies and timely coding and payment for new medical technologies.

On the technology side, with the help of its older, faster and more popular brother, CT, PET is making inroads. There is industry consensus that PET-CT, which was approved in late 2002, will overtake the sales of stand-alone PET within one to two years. The PET-CT revenue trickle of 2001 turned into a surge in 2002. The real-time functional imaging of PET combined with the anatomical mapping advantages of CT make an integrated PET-CT scanner a powerful diagnostic tool for radiologists and oncologists who cite PET-CT’s superiority in detecting cancers and in providing information that often results in treatment protocol changes for patients.

The introduction and adoption of PET-CT this year seriously impeded the adoption of stand-alone PET, according to a recent Frost and Sullivan report; however, “the adoption of PET-CT has been sharp and steep and has been aided by the infrastructure set in place by PET over the past several years.” Frost and Sullivan also cite relatively stable reimbursement and the growing awareness of PET by medical specialties as growth drivers.

The Frost and Sullivan numbers tell an upbeat story. In 2002, the combined PET and PET-CT markets generated $481 million in revenue at an annual growth rate of 55 percent. The PET market accounted for approximately 55 percent of revenue, while PET-CT quickly grabbed control of a whopping 45 percent of the market. The prime competitors in PET and PET-CT are Siemens Medical/CTS Innovation, GE Medical Philips and Positron. CTI Molecular Imaging and Hitachi have entered the market via distribution agreements with CTS Innovations.

Among the several trends driving the PET and PET-CT market, according to Frost and Sullivan, is the proliferation of FDG distribution through radio-pharmacies. In 1998, for example, there were only 14 FDG-focused radio pharmacies in the United States. By 2002, there were approximately 62 sites, not counting independent radio-pharmacies and facilities with in-house cyclotrons. Naturally, the increased competition decreased the average dose price of FDG.

Another PET/PET-CT driver is the increased number of applications for clinical indications applied for and approved by CMS and other third-party payers. (See time line for details.) In October 2002, for example, FDG PET was approved for reimbursement for the staging and re-staging of recurrent breast cancer and for monitoring treatment of locally advanced tumors when a change in treatment is considered.

Also pushing the market forward are the more numerous PET scanner models, with features and prices to suit wide-ranging customer budgets and preferences. This trend coincides with another tendency. While the types of facilities that purchased PET in 2002 remained essentially the same as in previous years (academic and community hospitals, general and PET-focused imaging centers, select radiology and oncology practices and mobile service providers), the percent by type shifted notably. In 2000, an analysis of a sample of the installed base of PET by Frost and Sullivan revealed that approximately 50 percent of scanners resided in hospitals. By 2002, this percentage fell to approximately 42 percent, as the adoption became more widespread outside early adopter hospitals.

Technology Advantage of PET-CT CTI Molecular Imaging of Knoxville, Tenn., is a key manufacturer of PET technology and the only vendor that makes its own BGO and LSO crystals for its scintillators. CTI makes its own private-label PET scanners and is the PET OEM partner for Siemens Medical Solutions. CTI Vice President of Marketing Brad Herrington said that in 1992, the company starting working on a new type of crystal that it just released a couple of years ago. The new LSO (lutetium oxyorthosilicate) scintillator is denser than the old BGO (bismuth germinate oxide) crystal and affords the opportunity to 3D imaging. Herrington notes that because the older type scintillators are slow, that manufacturers put lead in front of the scintillators to block scatter. But in doing so, they forfeit sensitivity. “By going to a scintillator that is seven times faster, we were able to get the lead out, literally,” Herrington says. “We joke about it, but that’s exactly what we did. We were able to remove the lead and use the inherent speed of the scintillator to do true electronic collimation. We were able to speed up the scanners by about a factor of two in clinical usage.” Adding a CT scanner speeds up the PET exam by 30 to 50 percent, Herrington says. “That takes an average 40-minute scan using a BGO scintillator down to about 25 minutes. We see scan times of seven to 15 minutes, so you can see the progression when you’re using CT and an LSO crystal ring PET.” Right now CTI is the only company with the LSO crystals, according to Herrington. The patents run out on it later this decade, but it is not the only company with a fast scintillator. Philips/Adac also has begun work on a fast scintillator. Despite the many battles fought for PET reimbursement, Herrington is optimistic. “More important than current reimbursement numbers, the number of indications for PET continue to rise,” Herrington says. That’s the exciting part. Physicians are recognizing more and more the different cancers that are accessible by PET. So even if reimbursement does go down, which in inevitable with all modalities, the good news is that the indications are expanding at about twice the rate.”

“While the percentage of scanners located in imaging centers has remained relatively stable, the increase was primarily noted within the category labeled other, which consists of mobile PET services, government facilities and unclassifiable institutions,” according the Frost and Sullivan report. “This shift demonstrates over time the opportunities for PET adoption lie not only in the hospital base but in a number of facilities as well. This trend bodes well for the future of the market and represents ample opportunity for market penetration.”

For many years, R. Edward Coleman, M.D., chief of nuclear medicine at the Duke University Medical Center in North Carolina, has worked extensively with CMS and its technology assessment group in getting reimbursement for PET.

“We were able to do this first successfully in 1998 when Medicare paid for some limited indications in lung cancer,” Coleman said. “Then we were able to send them more data and had a town hall meeting with them. After that they expanded the indication to colorectal cancer, lymphoma and melanoma. Then we sent them more data. We applied for full coverage of all PET and oncology, and they approved several other cancers — head and neck cancers, esophageal cancer — and now we have breast cancer covered. Today, we’re supposed to hear that thyroid and sarcoma are covered. Next week, we’re meeting with them to start a new round of indications for other cancers. We’ve been working with them to get Alzheimer’s Disease covered. There’s going to be an announcement today that after their first review, they’re not going to pay for Alzheimer’s, but they’re starting immediately a second review under more limited indications.”

Although PET has come along way since 1998 and the number of indications continues to climb, Coleman agrees that PET has had a rough reimbursement road. “When CT and MR came along, they were reimbursed without the hassles that we’re going through for PET,” Coleman noted. “But times are different. One of the reasons we’re having to go through all these hassles is that a lot of people think MRI and CT should not be reimbursed for all the things they’re being used for now. The technology assessors want to make sure that there is more than adequate data to support the indications that are going to be reimbursed. But it’s not just PET. Any new technology or any new utilization is having to go through a process similar to what PET’s going through.”

Duke owns a dedicated PET scanner and saw the arrival of it first PET-CT in May. Coleman uses the new PET-CT scanner on whole-body oncology patients while reserving the older, dedicated PET for brain patients and research. The PET-CT’s very fast transmission scan, less than one minute, will save about 20 minutes per patient. With their dedicated PET, Duke patients also receive a low-resolution CT scan that helps correct for the absorption of positron-emitting radio-nucleotides as they come out of the body. That takes 15 to 25 minutes. The exact alignment of the anatomic and metabolic information afforded by PET-CT also provides better quality diagnostic information.

Oncologists, Coleman said, are not doing the number of CT and MRI scans they they once were. It used to be that patients with lymphoma or melanoma would receive frequent CT scans to see how treatment was going and to find out if the tumor had spread. Today they get PET scans instead.

Why not add a PET and fusion software to an in-place CT rather than acquire a full-blown integrated PET-CT scanner? Clinical research has shown that using PET-CT in radiation therapy planning has resulted in a significant number of treatment protocol changes due to its superior ability to localize cancer. The same cannot be said for combining PET and CT images with fusion software. Also, the fusion software method takes technologist and/or radiologist time to overlay the CT and PET images. With an integrated system, alignment is perfect and immediate. “Now the radiation therapist can deliver a higher radiation dose to the tumor and lower radiation dose to normal tissue,” Coleman said. “There are already some studies to show that combining PET and CT information improves outcomes for patients having radiation therapy.”

Highlights From — A Cost Analysis of Positron Emission Tomography AJR September 2001 CTI Molecular Imaging of Knoxville, Tenn., is a key manufacturer of PET technology and the only vendor that makes its own BGO and LSO crystals for its scintillators. CTI makes its own private-label PET scanners and is the PET OEM partner for Siemens Medical Solutions. CTI Vice President of Marketing Brad Herrington said that in 1992, the company starting working on a new type of crystal that it just released a couple of years ago. The new LSO (lutetium oxyorthosilicate) scintillator is denser than the old BGO (bismuth germinate oxide) crystal and affords the opportunity to 3D imaging. Herrington notes that because the older type scintillators are slow, that manufacturers put lead in front of the scintillators to block scatter. But in doing so, they forfeit sensitivity. “By going to a scintillator that is seven times faster, we were able to get the lead out, literally,” Herrington says. “We joke about it, but that’s exactly what we did. We were able to remove the lead and use the inherent speed of the scintillator to do true electronic collimation. We were able to speed up the scanners by about a factor of two in clinical usage.” Adding a CT scanner speeds up the PET exam by 30 to 50 percent, Herrington says. “That takes an average 40-minute scan using a BGO scintillator down to about 25 minutes. We see scan times of seven to 15 minutes, so you can see the progression when you’re using CT and an LSO crystal ring PET.” Right now CTI is the only company with the LSO crystals, according to Herrington. The patents run out on it later this decade, but it is not the only company with a fast scintillator. Philips/Adac also has begun work on a fast scintillator. Despite the many battles fought for PET reimbursement, Herrington is optimistic. “More important than current reimbursement numbers, the number of indications for PET continue to rise,” Herrington says. That’s the exciting part. Physicians are recognizing more and more the different cancers that are accessible by PET. So even if reimbursement does go down, which in inevitable with all modalities, the good news is that the indications are expanding at about twice the rate.”

Medicare’s zero-sum game
When we talk about increasing reimbursement for one medical specialty versus another, it’s important to remember that the Medicare pot contains a finite dollar amount that must be spread across the entire system. That’s why hospital administrators flinch when they hear about so-called “increased” reimbursement for a new technology or clinical indication. That only means that the same amount of money is going to be subtracted from somewhere else in the Medicare budget. And that includes APC coverage for outpatient reimbursement.

“Medicare is always trying to reduce the amount they pay for anything, and particularly for high-priced items such as PET scanners,” Coleman said. “So I’m sure there will be an attempt to decrease reimbursement amounts for 2004 just like there were for 2003.” Fortunately, Coleman and his colleagues were able to demonstrate the actual costs of PET in 2002, so that CMS kept reimbursement rates relatively stable in 2003. “I think we’re in a good position this year, if they try to lower them, to again point out what the costs are.”

Currently there is no reimbursement for interpreting the combined PET-CT image while there is an individual reimbursement code for a separate reading on each. That’s something the professional societies are working to change. “Getting reimbursed for the fused or combined image is something we’ll be working on over the next year or two,” Coleman noted.

Peter S. Conti, M.D., Ph.D., director of the University of Southern California PET Center, agrees that reimbursement for PET has not been easy, but he believes it’s more a function of the times than the technology.

“CT and MR grew up in a time when there wasn’t much demand for outcomes-based analysis and evidence-based medicine criteria,” he said. “However, criteria are evolving rapidly, and in their effort to provide the highest quality of assessment, some believe they’ve [those responsible for reimbursement] hindered development of medical imaging technology. Primarily because the established criteria are more in line with those associated with therapies as opposed to diagnostics.” Conti believe CMS and other payers should offer relief in areas such as imaging where it’s more a question of altering steps on a decision tree rather than a final therapy option.

“To rely purely on outcomes analysis as a way of approving or disapproving imaging technology is a little too harsh,” Conti said. “Whether or not a patient survives a cancer may be more dependent upon what treatment options were available at the stage of the disease at diagnosis, not whether they had a PET scan or not.”

The point is that a lot of the decisions made based on imaging information are usually made in the short term— to treat or not to treat and which treatment to use. Those are more appropriate in terms of approval or disapproval for imaging technology. Does it {the technology} perturb the algorithm or decision-making process or the immediate patient care as opposed to the final outcome? PET has been caught in this interest in evidence-based and outcome analysis as arbiters, and it’s really hurt the technology. On the other hand, CT and MR grew up in an era when there was rampant proliferation of imaging technology, before CONs.”

Alzheimer’s Disease is a prime example of what Conti is talking about. In Alzheimer’s, for example, there is one therapy available, a pill, and even that only helps delay or diminish symptoms for about six months. “A PET scan can diagnose Alzheimer’s correctly 90 percent of the time while the currently reimbursed alternative diagnostic protocol has a track record of about 50 percent,” Conti said. “What about the advantages that PET diagnosis provides in terms of patient management and quality of life?” CMS won’t pay for a PET scan for the diagnosis of Alzheimer’s because currently there is no treatment besides the inexpensive pill that is largely ineffective.

USC’s PET-CT has been installed since August, so Conti does not yet have a full year of use. However, he sees some distinct advantages with the technology. “If you’re a newly diagnosed lung cancer patient and you have to get staged, it’s really pointless to get a CT today,” he said. “You’re better off having a PET-CT because you get the benefit of having a more accurate staging of the disease, plus you retain all the original qualities of the CT scan.”

Conti would like CMS to take a second look at some of the traditional tests that are no longer performed, but continue to be reimbursed for or are on the books as reimbursable studies. “They need to take a retrospective look at imaging technologies that are currently approved and currently being paid out where there are better technologies,” Conti said. “For lymphoma patients, say, a PET-CT vs. a chest-abdomen-pelvis CT or a gallium scan. They ought to be focusing on paying more for a primary PET-CT test first as opposed to getting a CT or a gallium scan. I would stop paying for gallium scans. I would get an FDG scan, because it makes more sense given what we know about lymphoma patients and PET.

Someone needs to start doing some homework within CMS. I understand their problem. They can’t just keep adding technologies, because there is a limited amount of money in the pot. It’s time to start looking at some of these older technologies or some of the not-as-powerful technologies and begin to say, we’re not paying for these anymore.”