College of American Pathologists
Printable Version

  Heart-to-heart discussion of test's value, use




November 2007
Feature Story

Karen Titus

Figuring out the best use of the AlloMap test is a little like assembling the evidence in the board game Clue. Instead of Colonel Mustard in the library with the lead pipe, the winning combination right now is stable heart transplant patients with no evidence of rejection more than two months after transplant and normal heart function by echocardiography.

The permutations appear endless, however, and researchers and users of the test, as well as its maker (XDx Inc., Brisbane, Calif.), continue to look at other possible uses. Truth be told, these other prospects are creating a little bit more excitement among pathologists and cardiologists than the current application.

For now, though, evidence continues to pour in that this is a fine test to confirm absence of rejection in quiescent heart transplant patients, a use that first emerged based on the CARGO study (Deng MC, et al. Am J Transplant. 2006;6:150-160). A consensus paper (Starling RC, et al. J Heart Lung Transplant. 2006;25:1389-1395) recently underscored that recommendation. Since CARGO's publication, use of the test has spread to more than 60 of the 160 U.S. heart transplant centers.

The resulting clinical experience has put more clues on the board, which the consensus paper also explores. The field is moving in a slightly different direction, beyond identifying and following low-risk heart transplant patients a year post-transplant, says one of the paper's authors, who suggests the test might have a useful predictive signature. Others wonder if the test could serve as a surrogate biomarker to titrate immunosuppression in transplant patients, or if a similar classifier could be generated for lung transplant patients.

"This is not the ultimate guideline," says E. Rene Rodriguez, MD, director of cardiovascular pathology and director of the autopsy service at Cleveland Clinic, and another of the consensus paper's authors. AlloMap, in short, is a good test in search of even better applications.

The recent consensus statement was put forth by the Working Group on Molecular Testing in Cardiac Transplantation, which reviewed the first 200 AlloMap tests done after publication of CARGO, says Mario Deng, MD, the lead author on the CARGO study and a member of the working group. That should open the door to wider clinical use of the test.

At Columbia University Medical Center/New York-Presbyterian Hospital, New York, where Dr. Deng practices—he's director of cardiac transplantation research, Department of Medicine, Division of Cardiology—AlloMap is starting to stick.

The first step was a cautious introduction, using the test in quiescent patients alongside endomyocardial biopsy, or EMB. Along the way, Dr. Deng says, he and his colleagues observed that the new blood test reflected the transplanted heart's function beyond ruling out cellular rejection: The more nonquiescent the score, the lower the cardiac function detected by pulmonary artery saturation, the higher the filling pressure of the right atrium, and the longer the so-called corrected QT time on the routine ECG, an indicator of myocardial edema. This information was presented as an abstract at this year's American College of Cardiology meeting.

The second phase of implementation is ongoing, with the AlloMap test being used in place of EMB for patients who are stable two months after transplant.

Dr. Deng says most of his transplant cardiologist colleagues have started using the test, with the patient load increasing gradually. Most of the early patients were those who had a systemic coagulopathy and related venous clotting disorders, whose myocardium could not be accessed via EMB. Patients with other access problems followed, and now the test is being used in patients who prefer noninvasive monitoring. "They like the idea of the less frequent penetration of their neck veins with the heart muscle biopsy catheter," he says.

Dr. Deng, though co-initiator and co-principal investigator of the project, remains, in his words, "anal" about restricting its use to this clearly defined subset of stable patients as outlined by CLIA and reimbursed by Medicare, warding off its inappropriate use as he conducts grand rounds on the test throughout the country.

The test is not meant for patients who present with new onset shortness of breath or with other dysfunction that may indicate rejection. Nor is it meant for lung transplant recipients, he continues, though an ongoing study, LARGO, will attempt to generate a similar classifier in this population; completion is about three years away. And it's not meant for the kidney transplant population, though that too is ripe for investigation. "We're continuously working on this: What's the evidence basis? What's the appropriate use? What's not appropriate use?" says Dr. Deng.

Dr. Deng is busy, albeit prudently so, at one end of the spectrum, working with those who want to use the test. At the other end are those who agree the test is good, then turn their attention back to EMB. That's the case at the Cleveland Clinic, Dr. Rodriguez reports. "The test has not been used very much right now. It's my understanding that the clinicians are not ordering it as often as probably the company would like."

There's no mystery here, he says: In the right setting, EMB can provide more information, more quickly. "If you have a patient with an experienced team of transplant physicians who can do biopsies in a very, very safe way, the turnaround time for biopsies that evaluate cellular rejection and antibody-mediated rejection is about six hours. And nobody can compete with this," he says. The AlloMap test is a send-out, to XDx, and it tests for cellular, not antibody-mediated, rejection.

The test could certainly be valuable for patients who don't live near a transplant center or hospital that can easily perform followup EMB, Dr. Rodriguez says. "For them, it's probably easier to go to their local hospital, get their blood drawn, and sent to XDx." In patients who are doing well, the test could offer great reassurance.

In terms of safety, he says, a needlestick trumps heart biopsies. "But in experienced hands, there's almost no difference in terms of [patient] discomfort."

Dr. Rodriguez has discovered some patients are sticklers for quick turnaround times. At the Cleveland Clinic, he usually provides EMB results between 2 PM and 5 PM. "If I'm late, by 5:15 the nurse practitioners are calling me because the patients are calling them. They're used to the fact that, 'I got my biopsy this morning—I need to know tonight: Am I doing OK?' Friday afternoon we get a lot of calls, because you don't want to have patients waiting all weekend for their results. And the pediatric patients—my goodness, the poor moms. They need that information. I don't know that they can sleep without it."

At Cleveland Clinic, current protocol calls for one biopsy a week for the first month post-transplant, then biweekly the second month. If the patient is doing well, EMB is done monthly up to six months, followed by three-month intervals. By year two, biopsies are halted unless the patient has symptoms of dysfunction, frank heart failure, or other symptoms that prompt a heart biopsy. Other centers are apparently tireless. "At my previous institution, we did biopsies 20 years out, and we would do at least one biopsy a year to see how the patients were doing," Dr. Rodriguez says. Within the transplant community, he adds, the general consensus is that biopsies can tail off two or three years after transplant in compliant patients who have not had a bad episode of rejection.

At other institutions, AlloMap and EMB perform the equivalent of a job share, with the goal of reducing (but not eliminating) the number of post-transplant biopsies. Such is the case at the University of Maryland Medical Center, which traditionally has done 13 to 15 EMBs in the first year following transplant. With relatively conservative use of AlloMap in months one through 12, the number of biopsies may decrease by about 60 percent, to five to seven biopsies the first year, says Mandeep Mehra, MD, the Herbert Berger professor and head of cardiology, University of Maryland School of Medicine, Baltimore, and chief of cardiology, University of Maryland Medical Center.

Neither test is a closed case. EMB drew heavy scrutiny during the CARGO study. Dr. Deng says some transplant teams and patients trust it blindly, even though the CARGO study showed significant variability in readings between local pathologists and a team of central readers. (The best concordance, 77 percent, was for grade 3, while grade 2 was 44 percent.) Dr. Deng's point: "The assumption that biopsy is an absolute gold standard needs to be questioned to begin with, before even asking, 'What alternatives do we have?'"

But questioning EMB isn't the same as showing it the door. During CARGO, "We kept looking to the cardiologists, thinking, 'There must be some more functional aspect, filling pressures or cardiac output, or something,'" recalls Charles Marboe, MD, who, as one of the three study pathologists for CARGO, helped uncover limitations of EMB (Marboe CC, et al. J Heart Lung Transplant. 2005;24:[suppl] S219-S226). "And in most instances, they said, 'None of those is either specific or sensitive enough to be used alone.'"

He adds: "You need the biopsy for management, but dubbing it 'the gold standard' turned out to be a little misleading."

Part of the reason EMB gained its foothold, he suggests, was because it was crucial in transplantation's pre-cyclosporine era. In one sense, biopsies were easier to read then, since the presence of any cells in the biopsy and any interstitial edema was called rejection and had to be treated aggressively. Cyclosporine changed the picture for pathologists. "All of a sudden cells were present in the biopsy, but because cyclosporine was interfering in interleukin-2 production and IL-2 receptor expression, nothing was really happening," Dr. Marboe says. "And with the introduction of Quilty lesions, it became confusing: You'd see these dense infiltrates that, to a casual glance, looked horrible and ominous, but in truth were also benign."

To perform a head-to-head comparison between biopsy-based and gene-expression-profiling-based monitoring, the university hospitals at Stanford, Columbia, Cleveland, Pittsburgh, Utah, Kansas, Chicago, and UPenn are completing a clinical study called IMAGE (Invasive Monitoring Attenuation for Gene Expression). The hypothesis, Dr. Deng explains, is that a strategy-based molecular profiling, compared to the biopsy, is not inferior in detecting rejection in heart transplant patients. The gene expression arm does not have routine biopsies, he notes; clinical rejection is defined as a decrease of the ejection fraction of the left ventricle by 25 percent, or the onset symptoms of heart failure. IMAGE targets 500 patients at the eight U.S. centers.

But physicians continue to lean on the biopsy. "We know they rely on it, because I spend Friday night here and come in on weekends#8212;they really want to know the result of the biopsy," says Dr. Marboe, who is vice chair of pathology and director of the pathology residency training program, Department of Pathology, Columbia University.

If AlloMap's use widens, an unintended consequence might even be a rise in number of biopsies. AlloMap assesses cellular rejection, but labs are also taking note of noncellular forms of rejection, including antibody-mediated rejection, which requires EMB for identification. What may eventually happen, says Dr. Marboe, is if some patients are biopsied less frequently, "the percentage of interesting biopsies we get, with either more cells or biopsies for AMR, may go up."

The revision of the International Society for Heart & Lung Transplantation guidelines helped drive home a heretofore unappreciated point: that AMR is a legitimate clinical issue. It may not affect many patients, says Dr. Rodriguez, but when it does, it can be deadly. "That's why we evaluate every single patient on every single biopsy for both cellular rejection and antibody-mediated rejection."

EMB may have limitations, Dr. Rodriguez agrees, but they don't necessarily limit practitioners. EMB's weakest point, he says, is sampling error caused by limited access—in patients with lots of scarring, for example, or difficult access through their veins.

Occasionally physicians will encounter patients with a severe depression of cardiac function, who have no evidence of cellular or antibody-mediated rejection on biopsy, at which point the biopsy again is moot. Ditto for the echocardiogram. "You've ruled out cellular rejection, AMR, allograft vasculopathy, infection. It's real, but we haven't learned yet to identify what that particular entity truly is," Dr. Rodriguez says. He calls such patients "very, very challenging—and, luckily, extremely rare."

Pathologists and other physicians are adjusting to the new ISHLT guidelines for grading EMB, puzzling out what the various AlloMap and EMB results mean in different settings and combination.

If the AlloMap score is high but the EMB score low, AlloMap could be an important clue that rejection is starting and building, Dr. Marboe says, suggesting this would then be reflected in the peripheral blood before cells accumulate in sufficient numbers in the biopsy to be graded 3A or above.

Essentially, a high AlloMap score in the first six months, when coupled with no rejection on current biopsy, indicates a six in 10 chance of at least one treatable episode of rejection in the next three months, says Dr. Mehra, an author on both the CARGO study and the consensus paper. The significance of this finding two or more years later is not yet clear, he adds. "And we don't know if there's a signal that denotes adaptation." In other words, could a high AlloMap score denote a benefit late after transplantation with such patients more likely to survive?

At least one abstract and one small published study suggest that a high AlloMap score is associated with coronary artery disease that develops in the transplanted heart. "So we may be going beyond just the definition of traditional rejection to looking at chronic rejection," says Dr. Mehra, who is also president-elect of the International Society for Heart & Lung Transplantation.

The latest incarnation of the ISHLT guidelines attempts to eliminate the murky area of grades 1B and 2 rejections—both are now downgraded to the same level, 1R, a near-0 category that requires no heightened treatment.

How does AlloMap fit into that context? The test has always been shadowed by questions regarding the former 1B and 2 categories, since it was built to address the extreme ends of the rejection spectrum—when AlloMap was developed, its scores were correlated with grades 0/1A and grades 3A or higher histology. "For some time," says Dr. Mehra, "there had been a raging debate as to the value of the AlloMap test, because we see a lot of 1Bs and 2s in general." The downgrading of 1Bs and 2s to a nontreatable category has made matters relatively simpler, he says. "The usefulness of the AlloMap has probably been heightened by this."

Dr. Marboe adds a twist to the discussion, noting that at the time the ISHLT system was revised, no hard data existed to suggest that the former 1A and 1B categories were different. The AlloMap data might offer another take on this, he says. Looking at average AlloMap scores for 1A and 1B, he says, the 1A scores resemble grade 0 scores while the 1B scores resemble 3A scores. This was first noticed in pediatric patients, says Dr. Marboe, but can also be seen in adults. "This is the first intriguing suggestion that 1B histology is actually early rejection. Unfortunately," he says with a laugh, "I think we revised the grading system too soon."

Dr. Rodriguez, however, says the evidence is scanty that 1B may denote early rejection. Instead, it may be that grade 1B cases are simply being confused with 3A scores. Experienced pathologists should be able to identify the endocardial infiltrate extensions that can be present in grade 1B cases, which distinguish them from true higher rejection grade, he says.

Dr. Mehra and his colleagues are exploring whether AlloMap's molecular signals predate pathological manifestation of rejection by weeks or months.

He says it does. If the AlloMap score for patients who are two to six months post-transplant is, say, 18 (the scale ranges from 0 to 40), the chance of that patient developing a treatable rejection in the next three months is close to zero, he says. On the other hand, he says, if an AlloMap score in the first six months is 30 or greater, and the current biopsy is negative, then, based on the CARGO prediction study data, the chance of a positive biopsy within the next 12 weeks is nearly 60 percent. "So six out of 10 people will have significant rejection on a subsequent biopsy," he says.

That derives from pure, or quantitative, scores. But Dr. Mehra and his colleagues have started looking at the qualitative aspects of the 11 genes in the test that drive this predictive power and whether they differ from those implicated in current rejection.

The answer is "yes," he says. They've noted a group of three genes in this 11-gene algorithm that are steroid-responsive—uniquely related and associated with the undulations in the doses of corticosteroids that are given to patients. These same genes, in their altered expression, are largely responsible for driving the predictive score, Dr. Mehra says. "So we may have, for the first time, hit upon a method to assess adequacy of corticosteroid dosing in transplantation.

"The second thing we've shown," he continues, "is that as you look at the steroid-responsive genes, when rejection does become manifest, there's a switch in the qualitative nature of the genes, where we start to see a different set of genes become evident. These are more so in the alloimmune cascade pathway, as well as in the response of the bone marrow to impending or overt rejection." These are probably representative of a counter-regulatory attempt by the body to counter the rejection episode, he says.

Dr. Marboe is keen to see AlloMap take up the problem of measuring immunosuppression. "One of the things that always bothered me about heart transplants was the absence of a bioassay for calcineurin inhibitors or any immunosuppressive drugs," he says, noting the kidney damage done by the former and the post-transplant lymphoproliferative disorders due to the latter. "We see so many of the nasty effects of drugs in the heart transplant population. So wouldn't it be great if you could use this as your surrogate marker to start dropping the doses of steroids and everything else as soon as you can?"

All these ideas are the future. In the here and now, AlloMap has become more sophisticated. For the present to merge with the future, more validation and replication studies are needed. Given the complexity of genomic and gene expression data, this remains a task worthy of Daedalus. "To the best of my knowledge, there is no statistical test that can truly help us define and deal with genetic information in a reasonable way except independent demonstration of reproducibility across different patient populations," Dr. Mehra says.

Eventually, clinicians will need to perform randomized controlled trials that avoid EMB in one arm of the study. One such trial is underway, Dr. Mehra reports, looking at the AlloMap/EMB question one year post-transplant.

He and others would also like to see AlloMap become more sophisticated qualitatively. The current 0-40 scoring is "an actionable but relatively unintelligent way of looking at complicated genetic information." Imagine, he says, a test that reported corticosteroid dose appeared to be low, for example, as evidenced by a change in the gene expression profile, as well as the up-regulation of specific alloimmune genes.

Dr. Mehra compares the development of AlloMap to that of cars. That would mean the test is far better than a Crosley Hotshot, but nowhere near the sophistication of today's models. He places the test in the 1960s or '70s, which, alas, filled the highways with Gremlins, Chevettes, and the Chrysler Imperial LeBaron Two-Door Hardtop. "We need to modernize our concepts a lot more," Dr. Mehra says.

The collaborative research continues, as Dr. Deng and those in Columbia University's Genome Center analyze the CARGO DNA-array raw data to study the heart transplant recipient's cellular networks. Using the tools of computational and systems biology, Dr. Deng and other researchers can systematically dissect the white blood cell pathways involved in the rejection process and design even more sophisticated immunomonitoring strategies.

In time, the current tantalizing clues may fall into place, allowing researchers to solve not a whodunit, exactly, but rather a who-should-do-it mystery.

Karen Titus is CAP TODAY contributing editor and co-managing editor.