May 2012, Vol 1, No 1
Personalized Medicine in Oncology: The Landscape of the Next Generation of Cancer CareUncategorized
Personalized medicine. Innovation’s 21st century poster child … and bastion of medical uncertainty, wishful thinking, and unfulfilled aspirations striking terror into the ledger sheets of payers unable to cope with conventional (empirical) medicine. Mastering personalized medicine (PM) requires an awareness of its governing dynamics, understanding that it exists in many forms on multiple planes. It possesses components as diverse as individual biomarkers (KRAS mutations and HER2 expression), broader prognostic nomograms, multivariate models (OncotypeDx), systems biology approaches utilizing patterns of genes, proteins or metabolites, and individualized strategies addressing subjective conditions of patient lifestyle. Some PMs are treatments themselves: eg, tumor vaccines, image-guided therapies based on fiducial markers, nano – particles with targeted delivery, etc. But whatever its scope and form, PM has captured the mind of the medical community and laity alike, who are determined to peel back its layers and harness its immense power to defeat cancer. For each ephemeral success in the PM experience – be it clinical, research, or policy – elevates the process of care to heights previously relegated to fantastical visionary dreams. And so the multiple stakeholders within healthcare’s clinical, business, and policy sectors reorganize their systems in anticipation of the ultimate paradigm shift to PM, believing beyond doubt that somehow PM has already begun redefining their processes, capabilities, strategic usefulness, and viability.
Personalized Medicine and Value
PM is more goal than reality, yet it nonetheless represents the beginning of the fulfillment of a promise made throughout the 20th century to “lick cancer.” PM holds this elevated status for the simple reason – and very little concerning PM is simple – that it involves specific, targeted eradication of cancer cells rather than blunderbuss diagnostics and medicines too broadly invasive to heal the tumor at hand and ill-suited to tumor landscape particulars or the genetics driving the cancer. PM technology is helping fulfill that promise, making it the centerpiece of healthcare policy, clinical practice, and financial management. Indeed, its first successes have caused healthcare system professionals to question whether they have traded the problem of incurable acute diseases for the responsibility of long-term maintenance therapy at unsustainable cost levels. This would present healthcare with a dilemma instead of a solution, since healthcare operates under the fundamental “Iron Triangle” of value: the balance of cost, quality, and access. Therefore, for PM to be regarded as progress in cancer care, it must demonstrate value rather than just quality. And gathering, categorizing, understanding, and synthesizing new and old categories of data make the PM challenge no less than the greatest medical challenge ever to face the scientific, clinical, business, public, and policy healthcare community. This consideration must precede the acceptance of PM. It caught the healthcare community by surprise with what appeared to be this good news/bad news scenario. Yet upon closer examination, the bad news – higher drug cost of cancer treatment added to the transformation of intrinsically acute fatal diseases into chronic ones requiring long-term treatment at higher costs than conventional/empirical medicine – was not so bad, or at least it need not become so if all parties to the process work collaboratively to ensure that balance accompanies this paradigm shift. At the heart of the matter is the understanding of PM dynamics and conversely appreciating just how imprecise pre-PM healthcare intervention often was. The implications abound for physician treatment strategies, payer cost management strategies, patient prognostic expectations, research and development strategies, and policymaker legislation.
The central questions of PM in oncology are less about whether the new treatments will work than whether we can afford them and what utilization efficiency measures are possible. Healthcare is reeling from excessive resource utilization, much of it wasteful, and most of it for outpatient treatment of patients with chronic diseases. What makes healthcare costs unsustainable is not simply the demographic increase in patients requiring treatment, but also their readiness to be treated and the reality that neither they nor physicians nor payers exercise sufficient vigilance to avoid resource misapplication. Healthcare disparities research is a growth industry, with every stakeholder, be they from the government, medical profession, private sector, or patient population, demonstrating a capacity to misapply or deny precious healthcare resources and wasting health and wealth in the process. Properly managed, PM can fill an efficiency gap in healthcare like nothing before it by satisfying the demand for certainty in resource allocation – targeting only the most likely candidates for treatment and, of equal importance, disqualifying from treatment attempts those many patients whose PM profile shows them to be incapable of responding favorably to a specific therapy.
PM is not simply about new biological drugs, but drugs and surgery linked to systematic diagnostic evaluation that prequalifies patients for therapy. Pre-PM (empirical) research and clinical treatment involved drugs and surgical procedures capable of effectiveness in unspecified members of the population but not on the individual, amounting to blindfold therapy where physicians administered treatment with a high likelihood for failure in the patient at hand. The basis for PM affordability lies first in the biological qualification of patients for treatment. This eliminates waste utilization expected in empirical pharmacotherapy, which would make biological drug utilization an exercise in economic futility, distancing it forever from mainstream affordability.
PM provides value-based care precisely because it goes only where it is capable of success – in some cases, guaranteed success. This has been the goal of healthcare cost management: value-based benefit design, wellness-based healthcare (the cycle of prevention, intervention, and innovation), meaningful use, and healthcare disparities reform (different treatments for patients with identical conditions). PM is not just high technology, novel agents/procedures likely to “surprise” the diseases it encounters. It is a methodical scientific process that determines, before treatment is administered, whether it can or even definitely will work based on the individual patient’s biology. This makes everything new. It ushers in a new era, replacing the present one characterized by delays in finding the right therapy, which robs the cancer patient not only of money but of something far more imperative: time lost pursuing treatments incapable of success that permits cancer to run its deadly course unimpeded. By the time trial-and-error stumbles upon the right treatment choice, the disease may have established a beachhead impervious to the right treatment delivered too late.
The History of Personalized Medicine
To know just where we stand regarding PM in oncology, a chronology of its development is in order. We previously understood some things about patient populations, separating them according to phenotypic features: more fit, less fit, older, younger, able/unable to access care, associated with comorbidities. The next phase was being able to distinguish patient differences at the level of the light microscope and the particle lab. This enabled physicians to understand prognostic features and separate patients into subgroups based on prognosis.
As science progressed, researchers became adept at devising strategies to enrich subgroups within a disease population who may be susceptible to specific therapy. First they were prognostic factors; now they are both prognostic and predictive factors – prognostic in terms of behavior, predictive in terms of outcomes association with therapy.
The modus operandi of PM is the ability to use biomarkers to assess prognosis and predict both the likelihood of response and, importantly, the potential for nonresponse to therapy. Not just conventional biologics in the sense of some kind of protein, be it antibodies and cytokines, but also oral small molecules – considered as oncologic reagents as opposed to biologics – and surgery. This is important core terminology since there is a tendency to equate PM with biologics, a most incomplete appreciation of its scope of operations, with implications for the ability of generics to emerge. Both large and small molecules are part of the PM revolution. So too is surgery with all its nuances: PM techniques can identify which breast cancer patient is suited for mastectomy vs lumpectomy, taking much of the guesswork out of the likelihood of recurrence following breast-conserving surgery via patient pretreatment stratification into high, moderate, and low risk for local failure based on gene expression profiling–based subtypes (HER2 and basal subtypes vs luminal A or luminal B subtypes). “High-risk patients may benefit from a more aggressive surgery, such as unilateral mastectomy and contralateral prophylactic mastectomy rather than the standard breast-conserving surgery for localized disease. Recent data show a dramatic increase of more aggressive surgery in the United States to prevent this local failure. Instead, this generalized surgical overtreatment, a personalized aggressive surgery only to highrisk patients, may prevent local failure and improve survival in these women while sparing unnecessary complications of an aggressive surgery in low-risk patients.” 1 Thus, all treatment options – conventional chemotherapy, biologics, and surgery – fall under the umbrella of personalized medicine, which relies so heavily on pretreatment risk stratification.
The second aspect of PM strategies involves understanding who harbors a disease potentially susceptible to therapy and who harbors a disease that is not. The value of PM lies in both pictures, a form of fast-fail strategy. The history of PM predates the notion of molecular targeting: eg, oophorectomy in the premenopausal woman, knowing in advance of that procedure whether she was still having menses. Then the concept of hormone receptor–positive tumor came into use – before scientists knew anything about the entire molecular signaling cascade (upregulation, downregulation of that receptor). Using immunohistochemistry determinations, physicians could assert high likelihood that an individual whose tumor did not express a certain feature would not benefit from that intervention. The reverse was not always true – one could not be certain that the presence of that feature predicted response to treatment – but the likelihood increased. Researchers were on their way to PM as a science, even if it were not a silver bullet that never missed its target. But excluding two-thirds of premenopausal women from needless intervention represented a real stride forward toward the predictive characteristic/promise of PM. One could focus treatment based on hormone receptor responsiveness. The advent of the biomarkers had arrived, which was to become the mainstay of PM.
HER2 testing is demonstrative of the nuances of personalized medicine in clinical practice. Amplification or overexpression of HER2 predict for benefit of Herceptin. But this HER2 amplification makes response to Herceptin only possible, not inevitable, and either gene amplification or protein overexpression is equally predictive of efficacy. So the next step of enrichment is to exclude individuals in whom the ability to have the reagent and treat them with it amounts to a Pyrrhic victory. These patients are at risk for the side effects of a therapy but impervious to its benefit. This knowledge still predated an understanding of gene mechanics, being based only on the empiric response to treatment. The next step away from empirical medicine was the concept of fusion genes, the knowledge that gene signaling drives the pathophysiology of chronic myelogenous leukemia. As the medical community became more sophisticated, physicians strived for the setting where they could identify the molecular driver and the ability to target a therapy, enhancing the likelihood that this therapy worked in essentially 100% of selected patients. In the epidermal growth factor receptor (EGFR) setting in people with lung cancer, one of the emerging strategies is that if the EGFR gene is mutated, then the receptor is hypersensitive and can signal, making susceptibility to treatment high.
This highlights what researchers are now looking for: the setting in which the expression of a molecular biomarker can give physicians a yes/no answer to whether a patient should receive treatment with a specific agent or a specific type of surgical procedure.
The implications for value, patient benefit, patient selection, therapy sequencing, and next-generation molecular development all come out of that. The ability to develop reagent strategies is within the realm of our current strategies. This is breakthrough medicine, and it is happening, one biomarker, diagnostic test, and drug or surgical procedure at a time. The fact that these breakthroughs have begun must be recognized for what they are and are not. They are representative of a superior process of care for cancer, but this kind of care is still in the planning stages for most cancers. This underscores the need for all stakeholders to continue to make conditions favorable to continued breakthroughs – or more directly, to remove all obstacles within the clinical, business, and policy healthcare sectors to PM innovation and implementation.
The cost equation thus becomes viewed not in terms of the population as a whole but only in terms of the approachable population. The more oncologists understand a patient’s individual biology at the tumor level, the greater the likelihood of success with the therapy used. Value-based care thus replaces empiric trial-anderror treatment that not only burdens the healthcare system with wasted resources but also insinuates a culture of waste. This leads directly to the problem of healthcare disparities. A culture cannot rise above its own expectations for excellence. It will not take the High Road if it regards it as the stuff of foolhardy dreams. PM not only does its job effectively at its own site of activity, it also enriches the expectations of all participants to succeed and to get things right the first time! Thepath to efficient, successful resource allocation is being paved with PM building blocks.
Value: The Intersection of Science and Financial Viability
PM thereby becomes an extension of efficiency reform measures that include healthcare disparities reform, value-based medicine, patient-centered care, and wellness-based medicine (prevention, intervention, and innovation). Instead of fearing the empiric costs of care, PM elevates economics to a level of efficiencies heretofore impossible with empirical medicine. When PM testing identifies 85% of patients as ineligible for therapy and the remaining 15% as having a 90% chance of success with therapy, value sets in – and value is the ultimate goal of medicine … and the ultimate deliverable of PM. The value from identifying ineligible patients is immense, transforming actuarial models even as radically as the molecular imaging technology models are transforming medical textbooks.
In the pre-PM era, pioneering cancer meta-analyst Sir Richard Peto stated, “Subgroup analysis kills people.” And in that era of the 1970s this was true – the science was not there to avoid unwarranted subgroup carve-outs done safely. PM not only allows subgroup analysis, it thrives on it. What PM does is avoid putting patients through unnecessary and often critically time-wasting therapy with no chance of success. Nor does PM torture the data with subgroup analysis that yields inaccurate projections of which patients are likely to succeed. Pharma has expended untold billions of dollars exploring unenriched patient populations at great cost. Thus, PM will help change the unsustainable research and development costs of new treatments, as well as saving money at the treatment level.
Tactical Success: Keeping Clinicians Current Amidst the Personalized Medicine Information Explosion
Research success is one thing, clinical translation quite another. The oncology community has embraced PM and follows its dictates, but with varying levels of sophistication. At the 2011 World Health Care Congress, a survey of physicians revealed that 98% of them acknowledged that PM is essential to their practice of medicine, but less than 20% were satisfied with their level of PM acumen. This calls to mind the gap between research discoveries and their application at the clinical level, stated forcefully in the Institute of Medicine’s document, “To Err Is Human,” which reported the dread statistic of 17 years from research discovery to mainstream clinical practice. While the pace of innovation uptake in cancer is faster than overall medicine, it is imperative that PM find usage that is as accelerated as it is well informed, reflecting guidelines and adoption of best practices. PM is centered on efficiency; hence, delayingdeployment of PM findings is inimical to the premise of PM as the crown of healthcare efficiency.
Exploring the Personalized Medicine Frontier
There are areas where understanding of PM principles is universal. For example, the status of KRAS biomarker is a dichotomizing factor in colorectal cancer patients, and no oncologist would treat patients exhibiting KRAS-mutated tumors with anti-EGFR antibodies. Conversely, for patients with KRAS wild type they would proceed with this treatment – not with a guarantee of outcomes, but of patient eligibility for treatment. This is known universally. Other areas are much sketchier; the validation of PM hypotheses is uncertain, iterative. It becomes important that the clinician knows what it is that we do not know as he navigates clinical paths occasionally with no precedent, the only guide being a global knowledge of biological and genomic dynamics.
An example of what we might term “perfectly certain uncertainty” involves the genomic breast cancer assay, Oncotype DX, used to predict response to adjuvant therapy in women with breast cancer that is hormone receptor positive. Controversy ranges around this assay, with some doctors deciding a priori not to order the test, despite robust data on the strategy it offers. Oncotype DX is approved as a predictor of outcome to adjuvant tamoxifen in patients with estrogen receptor (ER)-positive tumors. It is not approved as a predictor of adjuvant therapy in general. While supportive data for the test are indeed robust, they are retrospective. A prospective trial was recently completed that will help position the value of the test a bit more. Thus the test now finds usage, but with an evidentiary basis that is less than compelling to elevate it to mandatory status for all patients with ERpositive disease. The plot has a further twist, as the very prospective trial vindicating Oncotype DX has now been used to validate another test, Mammostrat, which appears to be as effective as Oncotype DX.
This example demonstrates that the search for PM standards is distressingly short on absolutes, and that PM aptitude relies on quick thinking, an open mind, sensitivity to nuance, and strong listening skills.
Personalized Medicine and Healthcare Disparities: A Supercharged Bell-Shaped Adoption Curve
Uncertainty breeds disparity, and in medicine, PM brings in its wake some inevitable baggage: the epidemiological discipline called healthcare disparities, which has been studied extensively, and from diverse, complementary standpoints, by the Agency for Healthcare Research and Quality (AHRQ) and the Dartmouth Atlas. Healthcare disparities may be defined as “The condition of uneven levels of care for different patients with the same condition.” It can be expected that PM technology involving commonplace clinical practice conditions will be applied somewhat consistently, though still adhering to the classic bell-shaped product- adoption curve that features, left to right, Innovators, Early Adopters, Early Majority, Late Majority, Late Adopters, Therapeutic Nihilists, and those about to retire at the end of the month. PM technology for uncommon conditions must also follow a bell-shaped curve, but it will be more radically and erratically shaped. This presents a gap analysis unique to PM: to bring out the totality of an ever-changing body of knowledge by experts constantly questioning their own discoveries and revising best practices in light of fresh research findings on a weekly rather than annual basis. The heterogeneous composition of cancer disease states, coupled with the flourishing discoveries and difficulties in replicating research findings in biologics, constitute core challenges of capitalizing on PM. Uptake of PM technology is dependent on ease of access to the tests, insurance coverage, and the work environment of the physician. The difficulty in assigning bestpractices in terms of quality will be paralleled by determinations of overall PM value (cost + quality + access). Like the rest of PM, no precedent exists, and we must slog through a trying course, keeping the patients’ interests central. We can expect a litany of choruses warning of the destructive impact of PM on “the system” – to which the reply might be, “But the system was always meant to be replaced by something better, and it has arrived, and it is called PM, and it has new financial and scientific characteristics. Adapt or perish.”
Complexity Necessitates Specialization
The simplicity of the PM premise and strategy belies the immense complexity of its execution, and so now that the premise of PM is largely accepted as praiseworthy, the order of the day for clinicians is to get on with understanding PM technologies and the new systems for informatics involved in the process of PM care: ie, what level of knowledge of which categories of PM technology must be understood in order to participate in this new type of medical care?
The specifics of this informational challenge give focus to the unique PM mission: a quest to answer why some patients respond to therapies and others do not. Dr Edward Benz of Dana-Farber Cancer Institute described the biological complexity as one element confronting the oncology community in a presentation delivered at the 2011 World Health Care Congress.2
This complexity has given rise to medical specialization needed to synchronize patient, tumor, diagnostics, and treatment. New technological disciplines are emerging. One such discipline is systems biology. In the course of explaining the role of systems biology in PM, researchers from M.D. Anderson Cancer Center point out how profoundly PM has changed the rules of engagement by subordinating the preeminence of randomized controlled trials, the gold standard of evidence, which “…are designed to determine the best approach for the average populations and not for specific individuals. The development of molecular profiling technologies to assess DNA, RNA, protein, and metabolites provides the potential to tailor medical care, both at tumor and patient levels. These approaches have the potential to fulfill the promise of delivering the right dose for the right indication to the right patient at the right time.”3
Tempering Enthusiasm: Coexistence of Personalized Medicine and Empirical Medicine
The advance of PM can connote a false sense of the obsolescence of empirical science. Approximately 85% of patients do not have a personalized signature to make them eligible for PM treatment, and hyperbole regarding the current range of PM only hinders its actual implementation, which takes place alongside empirical medicine in a form of peaceful coexistence. They are not one another’s arch enemies, but complements. Physicians will not be abandoning empirical medicine anytime soon, and it will continue to play a role for that part of the cancer population that still requires it. Examples of limitation to PM include situations where access to tissue is limited, where heterogeneity within the patient’s tumor burden exists, and situations where mechanisms of resistance are not yet known (especially with novel agents). Clinicians recognize the realities of adding PM to empirical medicine one step at a time.
The tactical execution of optimizing PM application at the oncology clinical level entails organizing and showcasing the vast, highly complex body of scientific research on personalized medical treatment of hematological and oncological diseases, translating innovation into clinical progress that delivers value – the balance of cost, quality, and access. Clinicians must understand the basic tenets of PM diagnosis, treatment, and informatics technologies in order to maintain a working grasp of PM that is realistic and grounded in current scientific updates. This will permit clinicians to include their patients in a dialogue on how to capitalize on the ability of PM to tailor cancer treatment according to their personal life goals – something that this targeted technology is only now beginning to make feasible, for PM is still more a promise than a reality. Currently we are witnessing a translational exposition of important new technologies in PM treatment of cancers. What remains to be seen is how skillfully and creatively this translational process is executed, what new avenues for PM open up before the ceaseless efforts of researchers, and how clinicians and their patients adapt to a culture of extraordinary change. Cultivating practitioner and patient clinical acumen in PM diagnosis and treatment will yield the long-awaited condition of healthcare value and triumphant outcomes that PM holds for cancer patients.
Roukos DH, Lykoudis E, Liakakos T. Genomics and challenges toward
personalized breast cancer local control. J Clin Oncol. 2008;26:4360-4361.
Edward J. Benz Jr. Cancer research and care in the genomic age: toward
personalized medicine. Presented at the 8th Annual World Health Care
Congress; April 4-6, 2011; Washington, DC.
Gonzalez-Angulo AM, Hennessy BT, Mills GB. Future of personalized
medicine in oncology: a systems biology approach. J Clin Oncol.
The shift from dose-intense standard chemotherapy to therapies targeting specific signaling pathways, molecular targets, or elements of the tumor microenvironment presents a number of challenges to the oncology professional and the patient. Cancer diagnosis and treatment requires precision in the diagnostic evaluation (asking all the critical diagnostic questions at the [ Read More ]
The inaugural conference of the Global Biomarkers Consortium brought together an international panel of oncology experts to explore the rapidly evolving field of biomarker research. Cochairs of the event were Hope S. Rugo, MD, director of Breast Oncology and Clinical Trials Education at the University of California San Francisco, and [ Read More ]