Bringing Personalized Medicine to Cancer Patients

By Alissa Poh 

Dec 10, 2008 | WALTHAM, Mass.--At a recent roundtable hosted by AstraZeneca on the topic of personalized medicine, representatives from industry, academia, and patient advocacy groups gathered to discuss ways to translate the alluring prospects of individualized medicine into therapeutic realities.

The question was summed up by the roundtable chair, Alan Barge, AstraZeneca’s vice president for clinical oncology: “What, practically speaking, can we do to bring personalized medicine to cancer patients?”

Barge noted that personalized medicine has become a much overused term in scientific lingo, but also one that is evolving within pharma.

“We’ve undergone a major shift in our thinking [at AstraZeneca],” said Jeff Hanke, Astrazeneca’s vice president for cancer discovery. “Historically, we’d pick up journal publications, find interesting biological targets to work on, and then try to fit these targets into cancer patients down the line. Now we’re focused on identifying the relevant targets to pursue in individual cancer patients, as well as the right drugs, which is where pharma has a lot of horsepower to throw at the problem.”

Brent Vose, the company’s head of oncology and infection therapy, said that every project emerging from discovery at Astrazeneca is now systematically assessed to determine whether a personalized medicine approach is appropriate. “[Each project] is tied to an attempt to put a toolkit together that allows us to ask, throughout the whole process, ‘Which patients are actually going to benefit, and who can really be targeted for therapy?’” Vose remarked.

Different View

Daniel Haber, director of Mass General’s Cancer Center, summarized three basic hurdles to be overcome for personalized medicine to become more than a castle in the air for cancer research.

“We need a different view in academia – for example, we’ll have to change the way pathologists study tumors,” he said. “It’s no longer simply about peering through micoscopes; they’ll need genetic studies on DNA, RNA, chromosomal abnormalities…all of which are very expensive, so it’ll be a whole new way of doing business.”

Clinicians, too, have been accustomed to enrolling patients in clinical trials “based on who’s available, and those [patients] who can’t be offered anything else.” Instead, it will be necessary to select the appropriate patients for each trial, in order to obtain meaningful data. “Even finding the right [drug] dose for an individual patient may differ,” Haber noted. “What’s right for one, with a particular activating genetic mutation, may not be right for another patient.” The standard recipe applies all the latest research and innovative treatments after the standard of care has failed. Sometimes, Haber said, “that standard isn’t so great to begin with, which is something we need to consider more often, and early on.”

As for pharma, a rather large paradigm shift is required as it adjusts to the idea of a segmented cancer market comprising many “little” disease subsets, each with a reduced profit margin and blockbuster potential. “But if that’s what cancer is, then that’s what drug companies have to be willing to do,” Haber added.

Cancer research won’t be able to move forward in the personalized medicine arena if academia and pharma cannot forge early partnerships with patient advocacy groups either, said Carlos Arteaga, director of the Vanderbilt-Ingram Cancer Center. “They need to review our [clinical trial] consent forms and buy into our ideas, which may not always be good ones.”

Participants representing patient advocacy groups, not surprisingly, were more focused on issues of communication and education. Access to the latest cancer detection methods still varies by insurance status when it should be universal, according to Thomas Sellers of the American Cancer Society. “It doesn’t matter how good the diagnostic procedure is, if it isn’t universally accessible, you won’t have the necessary, or right, patients available,” Sellers said. Patients also need a clearer grasp of personalized treatment possibilities as they emerge, added Winfield Boerckel from the Cancer Care of Long Island. Patients, he said, often wonder: “What if the right drug is out there, but I have no means of identifying it for myself, and can’t pay for it?”

Good questions emerged in the closing discussion. The obvious scientific and technical issues are whether there’ll ever be sufficient discriminatory power in science to enable selection of cancer patients who will or won’t respond to treatment and whether clinical trials, as currently conducted (by industry and academia alike), aren’t only cumbersome, but obsolete. The financial implications are also huge – in other words, who’s going to pay for bringing personalized medicine to oncology, and who will then seek the return on this very large investment?

But success breeds success, Barge noted. “We need to keep up with matching the right [drug] agents to the right targets in cancer subsets, even if each one is small.” He also acknowledged the need for better dissemination of information to patients, adding that “pharma is keen to play a larger role here, excepting that anything we do is tightly regulated, and often widely viewed as self-serving, so we need help.” It will be essential, then, for industry, academia, and advocacy groups to continue collaborating – not just around the table, but in real time.

Click here to login and leave a comment.  


  • Avatar

    There are plenty of targeted treatments in the pipelines of big Pharma these days. The immediate challenge is to find the right populations of patients for which these treatments will be most effective. The magnitude of this problem can be easily illustrated on the example of EGFR inhibitors which were effective only in a minor population of patients in the initial clinical trials. Now, through the retrospective analysis of clinical data, supported by laboratory experiments, we have learned that activated b-raf and KRAS are predictors of resistance to EGFR inhibitors in patients with activated EGFR pathway, thus Iressa or Tarceva would not be a treatment of choice for such patients. However, the key to successful clinical trials of new, targeted therapies is our ability to predict such resistance. Our targets may, or may not drive/sustain proliferation of cancer cells, depending on the “background” of other genetic alterations in these cells. Recently published analysis of the genome of various clinical solid tumors (as a part of the cancer genome sequencing effort) indicated that each and every cancer analyzed harbored multiple, but not necessarily overlapping, activated oncogenes, as well as numerous inactivated tumor suppressor genes. This is why it is critical to understand the genetic “background” of individual tumors in order to select effective treatment. The fact that the tumor has activated EGFR does not necessarily mean that it will depend on the EGFR activity for survival. The best way to dissect the effects of different genetic backgrounds on the role of our target is to start from model systems where the role of a single oncogene/tumor suppressor gene alone in oncogenic transformation can be tested first, followed by revealing the importance of the target gene in the presence of another activated oncogene (or inactivated tumor suppressor gene), finally moving to more complex systems where multiple, defined oncogenes/tumor suppressor genes are altered. I

  • Avatar

    A neglected aspect of personalized medicine is the wide variations in the ways in which individual patients adhere to prescribed drug dosing regimens. This point is not new: see Science 321: 769, 2008. A now well-documented, frequent basis for drug nonresponse is clinically unrecognized dose omissions or early discontinuation of dosing.
    Thus, a basic measurement needed for effective and economical personalized medicine concerns the patient's use of prescribed drug(s). The website shows the range of drug dosing histories in >17000 patients in a wide range of therapeutic fields.
    A key lesson taught by reliable dosing history data is that there is no disease yet recognized, including AIDS, post-transplant immune reactions, and various cancers that can be counted upon to enforce universally strict adherence to prescribed, life-saving drug regimens. The template for ambulatory pharmacotherapy is provided by the dosing histories seen in >4700 patients during their 1st year of treatment for hypertension with pharmaceutically ideal drugs that work, have convenient once-daily dosing regimens, and side-effect profiles no different from placebo control (BMJ 2008;336;1114-17).
    The reality is that the ranges of variations in dosing histories are about the same from one major disease state to another, with the practical, clinical question being: how is the patient I'm presently treating actually taking the drugs I prescribe? The answer to the question is: reliable measurement, i.e., reliable dosing history data. Lacking such information is tantamount to flying blind and destined to lead attempts at personalized medicine into many costly blind alleys.

    John Urquhart, MD FRCPE FRSE FAAAS (
    Chief Scientist, AARDEX Group (
    Prof of Biopharmaceutical Sciences, Center for Drug Development Sci UCSF
    Emeritus Prof of Pharmacoepidemiology, Maastricht (NL) University
    Office: 975 Hamil

  • Avatar

    The very recent article below from PNAS on opportunities for colon cancer therapeutics gives some hope that a paradigm shift is underway in drug development, through pharma/academia and patient advocacy cooperation.

    A genomic approach to colon cancer risk stratification yields biologic insights into therapeutic opportunities
    Katherine S. Garmana et al.

Add Comment

Text Only 2000 character limit

Page 1 of 1