The list of drugs that are routinely monitored is full of heavy hitters: antihypertensives, anticonvulsives, antibiotics. Digoxin. Warfarin. The list wouldn’t be out of place in a Chicago mayor’s office—it has clout.
Should anticancer drugs be popping up on the list?
The premise is promising: Oncology drugs are typically dosed according to body surface area or body weight, which gives rise to a great deal of variability. As a result, some patients might be overdosed, a troubling prospect for obvious reasons. Other patients might be underdosed, just as troubling for equally obvious reasons.
Such was the basic argument made by speakers in a session on the topic at the AACC annual meeting in July in Chicago. The session could be seen as only the latest in a long-running discussion about therapeutic drug monitoring in general. Salvatore J. Salamone, PhD, reached back through the centuries to Paracelsus, dubbed by some the father of pharmacology, who, some 500 years ago, called for doses of medicines to be properly dispensed. As is so often the case in medicine, not everyone agreed initially with the suggested protocol. “The alchemists at the university didn’t believe, and they ran him out of town,” said Dr. Salamone, CEO of Saladax Biomedical.
Likewise, TDM in oncology has yet to earn everyone’s approbation. If not viewed with suspicion on the grounds of alchemy, it still faces an uphill battle for wider acceptance, for reasons large and small, personal and institutional. The AACC audience was far more receptive, it should be noted, and Dr. Salamone was allowed to resume his seat on the podium after speaking. “If I speak to the lab community about this, they get it. Right away,” he said later, in an interview with CAP TODAY. “I’m halfway out with the explanation, and they get it.”
His explanation relies on some compelling data, including an article published in the Journal of Clinical Oncology (Gamelin E, et al. 2008;26:2099–2105). This was a phase III randomized prospective study evaluating 186 patients with metastatic colorectal cancer. Some received conventional dosing of fluorouracil (also known as 5-FU); the others had individualized dosing, based on pharmacokinetically guided adjustments. Dr. Salamone considers this study the one must-read for anyone interested in looking at TDM in oncology.
The initial dosing for both arms of the study, based on body surface area, was 1,500 mg/m2 FU during a continuous eight-hour period, once weekly. The dose remained constant in arm A unless toxicity required a reduction. In arm B, doses were adjusted weekly, based on single-point measurement of 5-FU plasma concentrations at steady state, until they reached therapeutic plasma range of 2.5 to 3 mg/L; targeted areas under the curve, or AUC, 20–25 mg•h•L-1, as established in previous studies. (Both groups also received the same leucovorin doses as part of their treatments.)
The Gamelin study is an eye-opener, Dr. Salamone contends, and not just because it strongly supports the notion of using TDM to reduce toxicity.
“The more important thing is, you do therapeutic drug management in oncology to increase dose intensity, because quite a number of patients are being undertreated,” he says. That startling statement is borne out in the Gamelin study. In the dose-adjusted group, he says, 68 percent of the individuals needed a higher dose to achieve the target threshold value; on the other hand, 17 percent of individuals in that group needed a reduction in their dose.
“An oncologist might be lulled into feeling happy that their patient is not exhibiting toxicity; well, that’s the good news,” Dr. Salamone says. The bad news: The patient could be falling far short of the optimal dose, leading to treatment failure.
Looking at the clinical results of the Gamelin study, the researchers report significant reductions in grade three and grade four toxicities, including diarrhea. More impressive, says Dr. Salamone, were the clinical data. Overall response rate was 18.3 percent for the BSA group and 33.6 percent for the pharmacokinetically guided group. Median overall survival for the BSA group was 16 months, and 22 months in the PK-guided group.
Dr. Gamelin and colleagues have also presented data on FOLFOX, a popular regimen for treating colon cancer, including an abstract (No. 356) presented at ASCO Gastrointestinal Cancers Symposium this year. This study looked at 118 patients in first-line therapy for metastatic colorectal cancer who were PK-monitored; another 39 patients, treated in parallel, were dosed based on BSA, without subsequent adjustments (barring toxic side effects).
To achieve the target level, says Dr. Salamone, 50 percent of the patients in the PK-guided arm needed an increase in dosing, while 30 percent needed a decrease. The physicians reduced toxicity, and the response rate improved from 45 to 70 percent. “Seventy percent response to FOLFOX-4 is the highest that I’ve ever seen,” says Dr. Salamone. Progression-free survival was 10 months for the BSA group and 16 months for the PK-guided group; overall survival was 22 months for the former group, 28 months for the latter.
Dr. Salamone calls the data spectacular. In his own conversations with Dr. Gamelin, Dr. Salamone has learned that the Centre Paul Papin, Angers, France, routinely does drug monitoring. Of a group of 806 patients Dr. Gamelin monitored, Dr. Salamone reports, 43 percent met the therapeutic target, 46 percent were below target, and 11 percent were above target.
Brick by brick, study by study, Dr. Salamone lays a foundation for considering TDM in oncology:
- A study from Italy (Di Paolo A, et al. Clin Cancer Res. 2008;14: 2749–2755), looked at monitoring 5-FU in an adjuvant setting. The study, involving 115 patients with colorectal cancer, found that patients at or above the cutoff for the drug had a better chance of survival after 10 years. Those with lower concentrations of 5-FU had about a 40 percent chance of disease-free survival, Dr. Salamone notes, while those with levels at or above the threshold had about a 65 percent chance of disease-free survival.
- A 1992 study (Milano G, et al. J Clin Oncol. 1992;10:1171–1175) explored variability in drug clearance among 380 patients with squamous cell carcinoma of the head and neck, who were dosed with 5-FU according to BSA. The study found a more than 100-fold difference between the patient with the lowest drug clearance and the patient with the highest clearance; within the 95 percent limit, says Dr. Salamone, there was a 35-fold difference in 5-FU clearance.
- Another study by Milano involving 186 head and neck cancer patients treated with 5-FU concluded that the greater the 5-FU systemic exposure, the longer the survival. Patients who had an AUC of above 29 mg•h•L-1 had about a 60 percent better chance of overall survival, Dr. Salamone says; those falling below that threshold had about a 25 percent overall survival rate (J Clin Oncol. 1994;12:1291–1295).
- A study in Blood (Picard S, et al. 2007; 109:3496–3499) showed that in a group of 68 patients with chronic myeloid leukemia, mean trough levels of imatinib (Gleevec) plasma levels were significantly higher in those with complete cytogenic response, and higher in the group with major molecular response.
“There’s plenty of studies,” Dr. Salamone says. “I could probably talk for several hours on this.”
To be sure, Dr. Salamone has a horse in this race—his company develops assays to measure plasma levels of common anticancer therapies. But his diligence is deliberate: When he hears physicians dismiss a role for TDM in oncology, they frequently say the evidence to support its use is thin.
Dr. Salamone disagrees. The variability in 5 FU has been shown for several decades, he says, citing a paper in Cancer Research (Kirkwood JM, et al. 1980;40:107–113) that recommended monitoring 5-FU exposure.
Another pioneer in the field is William E. Evans, PharmD, director and CEO, St. Jude Children’s Research Hospital, Memphis, Tenn. Starting in the late 1970s, he began looking at how children differed from adults in terms of dosing of anticancer agents. From there, he and his colleagues began to look at differences within children. “Typically, it could be somewhere between five- and 10-fold difference among children,” he says.
The causes were many: drug interactions; patient age; gender; organ function differences. Sometimes, the researchers couldn’t pinpoint a reason. Over time, they moved from looking at environmental and treatment-related demographic characteristics that influence drug pharmacokinetics to looking at the role of genetics.
Along the way, they also became interested in drug exposures based on blood levels, and the correlation to response to treatment. In a series of studies in children with leukemia, they found that those who eliminated the drug five or 10 times faster than average had less exposure—based on blood level—and had worse response. When that link became evident, they were spurred to try to control for it in individualized treatment, which in turn led to routine TDM of antileukemic agents (Evans WE, et al. N Engl J Med. 1986;314:471–477; Evans WE, et al. N Engl J Med. 1998;338:499–505; Pui C-H, et al. N Engl J Med. 2006;354:166–178).
He’d earlier done work showing the importance of TDM in other drugs, including antibiotics and anticonvulsives. Talking to his colleagues about considering it for anticancer drugs was a fairly easy sell. Their studies have involved a fistful of anticancer agents, including methotrexate, cytarabine, teniposide, etoposide, mercaptopurine, cisplatin, topotecan, and numerous other agents.
The rewards have been profound. The group most recently published a study in the New England Journal of Medicine this June (Pui C-H, et al. 2009;360:2730–2741), pulling together all the strands to optimize therapy: TDM, individualizing therapy, molecular diagnostics, genotyping the tumor. Drugs that had been effective in curing 75 to 80 percent of patients led to even better results—92 percent survival. “We get more mileage out of them,” says Dr. Evans. Just as notable, the risk-adjusted chemotherapy enabled physicians to eliminate prophylactic cranial irradiation, the cause of serious complications (including second cancers, cognitive deficits, and endocrinopathy) in more than two-thirds of long-term survivors.
Then there’s Gleevec, or imatinib, a prominent biologic used in treating chronic myeloid leukemia as well as gastrointestinal stromal tumors. It has a terminal half-life of 20 hours and is given orally, daily.
Among the data Dr. Salamone cites is a study by Richard Larson, MD, professor of medicine and director, Hematologic Malignancies Clinical Research Program, University of Chicago (Larson RA, et al. Blood. 2008;111:4022–4028). Dr. Larson and his colleagues measured the steady states of 351 patients with CML on day 29 of their therapy. Not too surprisingly, says Dr. Larson, patients with higher plasma drug levels were more likely to achieve complete cytogenetic response and major molecular remission. In the lowest quartile, 75.9 percent of patients had a complete cytogenetic response; in the highest quartile, 91.9 percent did.
They’ve since looked at the association of toxicity with drug levels. Dr. Larson and his colleagues looked at the incidence of grades three and four adverse events, as well as the temporary discontinuation of therapy—under the treatment protocol, patients with such toxicities had to discontinue the drug until the toxicities were resolved. As expected, high drug levels increased the likelihood of having a serious toxicity and a transient discontinuation of therapy.
Patients who discontinued the drug had a lower response rate, Dr. Larson reports. “So, there may be a ‘sweet spot’ in the curve where you have some optimal dose range for an individual patient,” a therapeutic equivalent of going yard, in which the drug has the greatest chance of being effective and the least chance of being toxic.
In the Blood study, Dr. Larson and his colleagues suggest that maintaining a trough level of =1,000 ng/mL appears to be important for achieving clinical benefit, a level that is consistent with findings in an earlier study (Picard S, et al. Blood. 2007;109:3496–3499).
In addition to citing the above studies, Dr. Salamone points to a study published this summer (Demetri GD, et al. J Clin Oncol. 2009;27:3141–3147). For anyone interested in PK-guided imatinib, “This is the one to read,” says Dr. Salamone. Like Dr. Larson’s study, this one divided patients with advanced GISTs into quartiles based on imatinib trough concentrations. The difference between patients in the lowest and highest quartiles was considerable. In the lowest quartile (<1,110 ng/mL), median time to progression was 11.3 months, compared with more than 30 months in the other three quartiles.
It’s possible to build a fairly sturdy three-legged stool from all these data: studies from Europe; studies in pediatrics; studies of Gleevec.
It’s also possible to glean from these efforts some of the many reasons why TDM has struggled to gain ground in oncology. To hear Dr. Salamone talk, oncology has remained the Turkmenistan of TDM, ignored at best, ridiculed at worst.
5-FU has an inconvenient history of being given as a bolus; only in recent years has it been given via continuous monitoring. Indeed, that’s one reason many of the TDM studies have come from France (as so many good things do). There, continuous infusion of 5-FU, rather than bolus dosing, has been standard since the 1990s, says Dr. Salamone, making drug monitoring less onerous.
Even if oncologists here had been inclined to measure 5-FU, they would have been stumped by the question, Measure it how?
Myriad Genetics hopes to answer that question, having recently launched a test, OnDose, to measure a patient’s exposure to 5-FU. Myriad developed its test based on licensed rights from Saladax, which markets its own 5-FU Personalized Chemotherapy Management assay outside of North America.
That would be one or two tests (depending on how one parses development, patents, marketing, licensing, and so on) for a drug that’s been part of the TDM literature for almost three decades. Change in medicine can sometimes seem to occur with all the velocity of caravans on the overland spice route, for reasons valid and maddening, reasonable and mysterious.
Looking to practices in France, Dr. Salamone sees a difference in habits. “In the U.S., dose monitoring just hasn’t been taught to oncologists,” he says. With 5-FU in particular, he rails against relying on body surface area to determine dose. “It’s an absurdity,” he says. “It’s a habit oncologists have,” and one he insists needs to be broken.
5-FU has been around for at least 40 years, notes Dr. Larson. “I think we still don’t fully understand how best to use it.” (It’s a little like the role of vice president in that respect.) The drug’s age illuminates another issue in TDM oncology, advocates say. Oncology, and medicine in general, are no different from other human endeavors in that the most excitement, the most propulsive force, comes from the new. iPhones are flying off the shelves, while landlines may be viewed with archaeological interest, and little more, in another generation.
That’s not to say cancer researchers are in shallow pursuit of the next shiny object. Just as it hardly makes sense to exchange mobile devices for a party line, no one wants to resurrect sulfa drugs as the best hope against cancers. The search reflects a need for something that works better.
With cancer drugs, however, the search for something better typically doesn’t include a chance to do TDM. When anticancer agents are developed, they’re first tested in small numbers of patients, at varying doses, to provide the initial safety data, says Howard McLeod, PharmD, director of the University of North Carolina’s Institute for Pharmacogenomics and Individualized Therapy. “A drug that looks boring at that stage doesn’t really get a good shake in phase II,” he says. Conversely, other classes of drugs are initially developed in hundreds of volunteers, with close attention paid to dosing and monitoring, including blood levels. That’s not likely to change in oncology. “There’s too many reasons why we do it the way we do,” he says. But as a result, “We miss out on having that kind of data,” says Dr. McLeod, who is also the Fred Eshelman Distinguished Professor, UNC Eshelman School of Pharmacy at UNC, Chapel Hill.
There are exceptions, including in pediatrics. Says Dr. Evans: “The reality is that 95 percent of adult cancers are not treated on protocols; they’re treated in private practice by sort of best clinical management. Whereas in pediatrics it’s quite the opposite. It’s like 75 percent of kids with cancer are treated on protocol.” This provides an opportunity to define monitoring criteria prospectively and to help guide caregivers. Such data are lacking in most adult cases.
If individualized cancer therapy gains more traction, and oncologists are forced to pay more attention to choosing doses, rather than, as Dr. McLeod puts it, “asking patients how they feel,” there might be a bigger push for dosing data from phase I or II trials, when the drug is used as a single agent. “Rather than waiting until phase III or, heaven forbid, phase IV, when it’s in combination, it’s different doses, it’s all over the place,” Dr. McLeod says.
Oncologists won’t feel comfortable with TDM without data, specifically the kind provided in the Gamelin study, says William Clarke, PhD, another speaker from the AACC session. But here another difficulty becomes evident. The Gamelin study, as well as other similar TDM studies, began accruing patients and clinical experience in the mid-1990s—a reflection, as noted, of the small numbers of patients in cancer studies. TDM oncology data have not been delivered by winged messengers.
Dr. Salamone sounds like he’s making up for lost time when he talks about his favorite subject, and there’s a reason for his fusillade of data. He says when he’s rebuffed by physicians, they often tell him that concentrations for maximum tolerated dose have not been established. Dr. Salamone counters this isn’t true for many drugs, including 5-FU, imatinib, and the taxanes and carboplatin. “It’s in the pharmacokinetic literature; it’s not necessarily in the oncology literature in the clinical studies,” he says.
What other resistance does he hear? Dr. Salamone has his own David Letterman-like list. File under: Physicians ’ Top Reasons for Not Using TDM.
- “I don’t need TDM because I use maximum tolerated dosing according to PI.”
Responds Dr. Salamone: Maximum tolerated dosing in package inserts is derived according to body surface area, generally with about six different individuals at each level. “There’s no way that following the package insert on body surface area is going to give you the maximum tolerated dose for everybody you treat.”
- “I don’t need TDM to control toxicity—I know how to deal with it.”
Says Dr. Salamone, “That’s half right.” Acute toxicity occurs quickly and is easy to recognize. But accumulated toxicity doesn’t and isn’t. “This is what Erick Gamelin tells me—people who are above the threshold but aren’t experiencing toxicity at first get accumulated toxicity.”
- Oncologists have “too many other protocols for new drugs that are attractive.”
This last one leaves Dr. Salamone sounding a bit defeated. Research dollars from major pharmaceutical companies speak loudly, for better and for worse. He notes that 5-FU and another drug that has been the focus of promising TDM studies, Taxol, are generic drugs; another plausible candidate, Taxotere, will come off patent fairly soon. “So there’s no push by the pharmaceutical company to do a 400-patient randomized study,” which would cost millions, he says.
That shouldn’t halt matters. “Of course, the ideal studies that we’d all like to see are never ready,” Dr. McLeod says. “It’s going to come down to a case of, Does the existing data convince you enough to make you try it in your patients?” He ticks off the drugs that routinely rely on TDM or where the momentum may be building: methotrexate, busulfan, 5-FU, Gleevec, and carboplatin. “At some point, there will be a critical mass of examples in a physician’s daily practice, which will cause them to say, ‘Why am I not doing this for every drug?’”
Besides, he says, pharmaceutical companies aren’t the only ones eyeing the bottom line. “Many of the times the companies that are producing these tests are in the mindset of high volume, low margin,” says Dr. McLeod. The profit isn’t big enough for them to launch prospective clinical trials, either.
The in vitro diagnostics business is not going to cannonball into the pool, hoping ordering physicians will follow, adds Dr. Clarke, assistant professor, pathology; director, toxicology; and director, point-of-care testing, Department of Pathology, Johns Hopkins Medical Institutions. “It’s a chicken-and-egg thing, right?” he says.
The recent Gamelin paper, in Dr. Salamone’s mind, should be a turning point for TDM, though it has not created as much stir as he’d hoped. “There are not many randomized studies comparing patients with and without drug management, and the pharmacokinetic people who look at it find it attractive.”
He says many of the oncologists he speaks to dismiss the paper, however, because it addresses an older regimen that’s not used in the United States. Dr. Salamone empathizes with them. “They’re being hit with so many new technologies, and they have to be very cautious,” he says. “But drug management has been used routinely, probably since the 1970s, in other medical disciplines. This isn’t a new technology. It’s been shown to work in medicine, and it’s being applied to the field of oncology.
“Why should oncology not follow pharmacokinetic principles?” he asks.
Like Dr. Salamone, Dr. Clarke understands the thrill of the new. Drug developers are still looking for the magic bullet, he says, and previous agents—many of the small molecules, for example—haven’t been the cure-all they hoped for. So, with limited patient populations available for clinical trials (a sad irony of cancer research, given the overall number of cancer patients), developers are going to place their bets on the new rather than try to rehabilitate the old. “They’re going to choose to pursue the more ground-breaking thing—and I can’t say that I blame them,” says Dr. Clarke. “That’s not unique to cancer, by the way,” he adds.
Dr. McLeod made a similar observation in an editorial accompanying the Gamelin study (Walko CM, McLeod HL. J Clin Oncol. 2008;26: 2078–2079). Referencing the study, he and his co-author wrote that if the researchers had been monitoring a new drug, rather than 5-FU, “it is likely we that we would be excited about this improvement in therapy. Indeed, a major obstacle to TDM-directed therapy is our own biases, well founded or not.”
Among the shadows, however, Dr. McLeod also sees sunlight. Newer cancer agents are so expensive that payers are scrutinizing quality, he says. There’s no question about the quality of the drugs per se, but rather the quality of exposure for individual patients. In Dr. McLeod’s experience, payers aren’t reluctant to pay for expensive medicines. “They would just like a better assurance that they’re going to get their money’s worth,” he says.
As a result, he’s seeing some payers starting to think about blood level monitoring in the area of quality control, making sure that “if you’re going to prescribe an expensive drug—or even a nonexpensive drug,” he says, laughing, “that your patient is going to get at least above a certain threshold of exposure.” That may not guarantee efficacy, “but at least it ups the odds to the point where they can justify paying for it.”
Gleevec is a well-turned example of all the players coming together. The drug’s maker, Novartis, has made plasma level testing available free, for the simplest of reasons, says Dr. McLeod: The company has good data saying that people with low blood levels are more likely to get resistance to their drug and move on to another agent. Patients who receive optimal doses are more likely to control their disease for longer; the drug company, not to put too fine a point on it, has patients stay on their drug longer. “This is a true win-win,” says Dr. McLeod. “The usual win-win is that the drug company and the prescriber win, but the patient gets hosed.”
As more anticancer agents are given orally, oncologists might have another incentive to turn to TDM: to monitor adherence. Dr. Larson notes that many (though not all) CML patients who stop taking imatinib will have a recurrence of disease. “If tests become available, we’ll use them,” he says. “One of the biggest concerns of oncologists is that the drugs won’t work if people don’t take them.”
Community-based oncologists could be among those chanting loudest for change. Those are the ones who are more likely to need help managing all sorts of patients on all sorts of drugs, Dr. McLeod says. Test makers, take note. “This is a very practical group of people.”
Dr. McLeod says it’s been fascinating to watch where pockets of individualized medicine have been popping up. In oncology, the community has outpaced academic centers in terms of early adoption. It stands to reason, he says—academics tend to be more focused on innovation rather than routine practice, and TDM in oncology “in some ways is less sexy, because it’s a little bit too practical.”
The UNC institute where he practices was born in part to bridge that gap. He and his colleagues were also frustrated by medicine’s caste system. Here’s how Dr. McLeod describes it: “Clinical pharmacologists convince each other that something should be done”—but no one else. “Practicing physicians, who maybe don’t care about clinical pharmacology, are in their own little sphere. Pathology and the laboratorians are in their own circle, on the other side. And then the policy people and the ethicists and the economists are across the street, in some tower that we never get to see.”
The institute pulled all those groups together (pharmacy, medicine, public health, and nursing), then got to work on a broad spectrum of TDM studies. Among their accomplishments, they’ve looked at:
- genotype-guided tamoxifen in breast cancer patients, with dose based on CYP2D6 genotype status.
- 5-FU in colon cancer patients, with dose based on blood levels.
- using genotype and blood levels to dose clopidogrel, or Plavix, in heart disease patients.
- genotype-guided warfarin dosing for various clinical indications.
“And I have a paper in front of me right now,” says Dr. McLeod. “We’re starting to work up, for routine clinical practice, using genotype to choose whether a patient gets interferon for hepatitis C.”
Dr. McLeod likes to mix consulting terms in with his medical vocabulary. For instance, he likes to use the phrase “action items” when he talks about why the institute has been able to get the TDM ball rolling. While the lab is charged with providing high-quality clinical tests, the pharmacists and physicians have “clear action items” for the results of those tests.
Nor does he shy away from putting the word “market” in front of “research.” Without it, good ideas emanating from the institute would simply vanish into the academic ether. “We do cold market research,” going into the community and asking, Is this a real problem? Often it’s not, he says. “It’s just an academic problem.” Even if a problem does turn out to be genuine, the institute folks might still have to make serious adjustments, as their own ideas collide with the realities of those in community practice. “These are not academic centers. These are folks who sometimes don’t even have a centrifuge.” Otherwise, Dr. McLeod might find himself creating another business phrase—“completely useless research,” as he puts it. This way, however, “If the research is successful, there’s an automatic handoff into the community to make it routine.”
In his work, Dr. McLeod has bumped up against some of the same resistance Dr. Salamone encounters, including clinicians who initially claim they’ve no use for TDM. “We’ve had that scenario multiple times,” Dr. McLeod says. “You talk to them, and there’s no problem! There’s no issue! And it always ends up being quite laughable in the end, because the truth eventually does come out.”
Warfarin was one such example. Dr. McLeod and his colleagues would visit anticoagulation clinics and be greeted with, We don’t need to do testing because we know how to manage our patients.
“And so the first five minutes is spent talking about how well they can manage their patients,” Dr. McLeod recalls, “and then the next 55 minutes is spent talking about all the difficult cases, and how challenging it is, and how they wish they had better tools. Basically it’s the different stages of death and dying going on,” he jokes. “Denial, then anger, then gradual acceptance.”
It’s the same with oncologists, Dr. McLeod says. He’s talked with clinical colleagues who don’t consider neutropenia to be a problem because “no one is getting killed by this.” But that ignores quality of life data, hospitalization data, and need for supportive care. “You can make the case that it is harmful.”
Other problems have eluded a TDM solution so far. For some cancer drugs, the rate-limiting step is fatigue, which is nearly impossible to measure. “Right now, we don’t have a clue what to do,” Dr. McLeod says.
The role of the laboratory has been oddly absent from the bulk of this discussion. Is oncology TDM mostly the realm of oncologists and pharmacologists?
Dr. McLeod doesn’t see it that way. He’s uncovered two lab-related issues in his work at the institute, mostly stemming from its 5-FU studies.
The first issue is obviously lab-related, since it has to do with sample timing. Often labs don’t want to “get too deep into pharmacokinetics,” he says. But they’ll need to if TDM is going to work—consider, for example, the difference between a five-minute bolus infusion versus a 72-hour continuous infusion. “Either the laboratory needs to learn more about when to sample, or they may need to reach out to others” who are more comfortable with the topic. He’s seen the latter happen at some of their community sites. “The laboratorians suddenly become fast friends with the clinical pharmacists.”
The other issue is more fraught. Laboratorians don’t want to tell their colleagues how to practice medicine. But for personalized medicine to take hold, they’re going to have to push the boundaries. “We’ve seen that happen with genetic tests and with blood level tests,” Dr. McLeod says. Labs may report a CYP2C9*3 genetic polymorphism. And then comes the inevitable phone call: Is that good or bad? He observes the same dialogue with blood levels:
“The AUC was 37.”
“I guess it’s a little bit on the high side. What do I do?”
If labs can give more guidance than usual, Dr. McLeod suggests, they’ll be turning test results into better patient management. Otherwise, “It’s at a point where oncologists would rather not have a test result because they just don’t have a clue what to do.”
All of this sounds thoroughly exhausting. But it doesn’t have to be, Dr. McLeod says, returning to the Gamelin paper. The study has caused a stir, but not a revolution, he says, which is actually a pretty handy metaphor for thinking about TDM-guided oncology.
On the one hand, “It’s been a clear reminder that the status quo is pretty silly,” Dr. McLeod says.
On the other hand, it’s a helpful reminder that TDM-guided oncology doesn’t necessarily require a seismic shift in medical practice. In many ways, it only seems like a new idea.
Dr. McLeod uses carboplatin to illustrate his point. The drug commonly is used for many solid tumors, including lung and ovarian cancers. No oncologist would dream of dosing it based on BSA, he says; instead, serum creatinine is used as a measure of kidney function to adjust dose, since kidney function correlates with blood levels, which in turn correlate with risk of platelet toxicity. In this example, creatinine is an easy surrogate for measuring blood levels, but the concept is the same. “And as we remind people of that, suddenly it’s not so foreign. It’s a case of doing something people have already done, as opposed to inventing a whole new branch of medicine.”
That drug benefited from having an assay already available, making it easy for physicians to do it routinely. When the drug was under patent, its maker, Bristol-Myers Squibb, “made sure everyone had a little slide rule, where you could take a person’s characteristics and turn that into a dose,” says Dr. McLeod.
All the stars and planets were aligned, in other words: good correlation between blood level and either toxicity or efficacy; a clinical test; and the tools to translate that test into a dose. The result was a widely accepted medical practice.
There’s another opportunity for labs to enter TDM oncology discussions, should they so choose. Dr. Clarke argues for what some have called smart TDM, which would use a pretherapeutic genetic test to select therapy, then fine-tune dose based on blood levels and clinical response.
That’s a lot of information. Unless labs can explain to clinicians what it all means, says Dr. Clarke, smart TDM could be a nonstarter. “They need to know what it means when labs say, ‘This person is a poor metabolizer’—should they reduce the dose 25 percent, or. ...? Should they not give the drug at all?”
The typical challenge for labs, that of developing tests, “is the least of our concerns,” says Dr. Clarke, especially as more academic centers gain access to mass spectrometry. It’s fairly straightforward to develop a chromatographic assay. The true challenge, he says, is to find clinical collaborators, and then generate evidence that there’s variability, and then that pharmacokinetic variability depends on more than dosing and BSA. And that there’s a desirable range for blood concentration, and that patients in that range do better than those who aren’t PK-guided. The future of TDM oncology, in short, appears to lie at the end of a long series of conjunctions.
And those are only the more immediate problems. “You’re competing with markers of chemosensitivity, you’re competing with dollars to pursue pharmacogenetics, and you’re also competing with patients signed up for new drugs,” says Dr. Clarke.
A lot of dominoes, clearly, need to fall. But labs can perhaps supply a needed nudge. At the very least, Dr. Clarke would like to see labs start talking about these issues with their colleagues. “You have to understand their needs, and where they might be going, so you can go there with them. Without some pushing from the laboratory, people are going to think, Oh, this is just another test we’re going to offer, as opposed to real opportunities to improve patient care.”
Dr. Evans, of St. Jude, agrees. “It doesn’t matter if you have a great test, if you don’t have a physician who a) embraces the concept and b) understands how to use the information.”
But it could happen. Dr. Evans says, “If you went back to the mid-’70s, you would see exactly the same picture for anticonvulsives, antibiotics, a lot of drugs where routine monitoring is now commonplace in most institutions.”
Even Dr. Salamone, whose advocacy of TDM in oncology has alternately left him nervy and nonplussed, excited and exhausted, sounds convinced.
Before Saladax, he worked in R&D at Roche, specifically helping to develop monitoring tests for drugs of abuse and therapeutic agents such as gentamicin and tobramycin. When he left Roche, he wanted to apply that expertise to cancer. “As far as I was concerned, it was the Wild West.”
It’s taking far longer than he thought. But “while it may be discouraging, I know I’m on the right side,” he says. “I lose a little sleep at night over it, but in the end, I think that we are going to be monitoring levels of most chemotherapy agents.” ?
Karen Titus is CAP TODAY contributing editor and co-managing editor.