Has the Promise of Precision Medicine Been Oversold?

A roundtable discussion with Edward S. Kim, MD; Vinay Prasad, MD, MPH; and Richard L. Schilsky, MD, FACP, FSCT, FASCO

Get Permission


Recently, the term “personalized medicine” in oncology care has been overtaken by the more contemporary concept of “precision medicine.” According to the National Research Council of the National Academies of Science, Engineering, and Medicine, the newer terminology shifts the focus to improving patient outcomes by identifying targeted approaches in treatment based on variations in genes, environment, and lifestyle and away from the implication (or misinterpretation) that “personalized” treatments are being developed uniquely for each patient.1 The distinction may be slight, but it is sparking debate among oncologists about whether the promise of genome-driven oncology is outpacing clinical reality and leading to disappointment among patients with advanced cancer, who may expect durable remissions and even cures from therapy based on the genomic sequencing of their tumor.

Targeted drugs for cancer date back 2 decades, with the U.S. Food and Drug Administration (FDA) approval of rituximab (Rituxan) to treat non-Hodgkin lymphoma in 1997 and trastuzumab (Herceptin) to treat breast cancer in 1998, followed by imatinib (Gleevec) to treat chronic myeloid leukemia in 2001—all of which have saved many lives. More than 150 indications for targeted therapies in 28 types of cancer have been approved since then,2 and in 2017, the FDA approved the first “tissue-agnostic” cancer treatment, pembrolizumab (Keytruda), for any solid tumor with the microsatellite instability–high or mismatch repair–deficient biomarker.3 However, the pace of increasing the number of patients who can be matched with an FDA-approved drug based on their tumor’s genomic makeup has been frustratingly slow.

Vinay Prasad, MD, MPH

Vinay Prasad, MD, MPH

Despite the abundance of approved targeted therapies, a recent study by Vinay Prasad, MD, MPH, Associate Professor of Medicine at the Oregon Health & Science University in Portland, and his colleagues found that only about 9% of patients with metastatic cancer will be eligible for a genome-targeted drug, and fewer than that—just 5%—will actually benefit from the therapy. And even when patients benefit from a therapy, many relapse after 2 years.4 The enthusiasm surrounding the promise of precision medicine needs to be tempered with the so far limited success of these therapies in most patients, especially in the setting of off-label use, according to Dr. Prasad.

Edward S. Kim, MD, FACP

Edward S. Kim, MD, FACP

Richard L. Schilsky, MD, FACP, FSCT, FASCO

Richard L. Schilsky, MD, FACP, FSCT, FASCO

In late summer, The ASCO Post held a wide-ranging and lively roundtable discussion exploring the promise vs the reality of precision medicine in current clinical care with Dr. Prasad and two other accomplished oncologists: Edward S. Kim, MD, FACP, Chair of Solid Tumor Oncology and Investigational Therapeutics and the Donald S. Kim Distinguished Chair for Cancer Research at the Levine Cancer Institute, Atrium Health in Charlotte; and Richard L. Schilsky, MD, FACP, FSCT, FASCO, Chief Medical Officer of ASCO.

Defining Precision Medicine

The National Institutes of Health (NIH) Precision Medicine Initiative defines precision medicine as “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” What is your definition of precision medicine?

Dr. Kim: I agree with the NIH definition because it gives us a broad stroke on processes that have been evolving in cancer care and that have become systematized. We have now passed the era of personalized medicine to where we have defined FDA-approved biomarkers that can be matched with therapies shown to be better than traditional non–biomarker-matched agnostic chemotherapy.

Dr. Prasad: The definition of precision medicine is a moving target, as Robert Peter Gale, MD, PhD, DSc(hc), FACP, FRSM, and I described in a previous story in The ASCO Post (January 25, 2017). It is important to link the term to the use of genomics or other “omic” technologies in oncology care.

If I have a patient with melanoma with the BRAF V600E mutation, I’m going to consider a BRAF/MEK inhibitor, so I fully embrace FDA-approved therapies that are linked to genetic markers to predict which patients are likely to benefit from a specific drug. I am more reluctant to leap forward to the broad use of next-generation sequencing for every patient’s tumor, because that is where we do not have enough proven data of patient benefit.

Dr. Schilsky: My definition of precision medicine also aligns with the NIH definition. But practically speaking, what most people think of as precision medicine is identifying a treatment based on a precise molecular abnormality thought to be a driver of a person’s cancer and using a treatment against that target, which is likely to have a more precise killing effect on the cancer and fewer side effects.

When I became ASCO President in 2008, my Presidential theme was “Personalizing Cancer Care.” I prefer the term “personalized medicine” over “precision medicine” because it incorporates patients’ goals and preferences, which is not part of the NIH definition of precision medicine.

Measuring the Effectiveness of Precision Care

How effective is precision medicine in treating patients with cancer? Is it more effective than the one-size-fits-all chemotherapy approach, and is it becoming standard of care rather than the exception in oncology care?

Dr. Kim: We want precision medicine to become standard of care for every patient, but we are not ready to apply the technique and match a therapy to every patient’s tumor. However, the field is evolving.

In 2010, we started testing for the EGFR mutation in my field of lung cancer. It was not standard practice, but because of genomic sequencing, we were able to substitute a pill for systemic chemotherapy in 15% to 25% of patients diagnosed with non–small cell lung cancer (NSCLC). Today, we test for several gene mutations, including EGFR and BRAF, programmed cell death ligand 1 (PD-L1) expression, and the ALK and ROS1 translocations, and it has had a profound effect on patient care. Patients with metastatic lung cancer are living longer and with a better quality of life. Implementing these principles of precision medicine has led to better outcomes, but the key point is we have to implement these principles effectively and safely.

Dr. Prasad: There are some well-defined situations where precision medicine is very important and more effective than cytotoxic alternatives. It’s difficult to compare genome-driven therapy with chemotherapy because, even though many of us have mixed feelings about chemotherapy, we recognize the potential of chemotherapy to treat large groups of patients and cure some cancers, especially Hodgkin lymphoma and testicular cancer.

The challenge for precision medicine is that currently only a small percentage of patients benefit from genome-driven treatments. Even though those benefits are often valuable, we can’t forget about the more than 90% of patients who aren’t benefiting from these genomic medicines at this time.

I prefer the term ‘personalized medicine’ over ‘precision medicine’ because it incorporates patients’ goals and preferences, which is not part of the NIH definition of precision medicine.
— Richard L. Schilsky, MD, FACP, FSCT, FASCO

Tweet this quote

Dr. Schilsky: Targeted therapy is emerging as being superior to conventional chemotherapy if it is possible to identify the patients whose tumors have a target. If you don’t have a biomarker selection strategy, then the targeted therapy is not necessarily a superior treatment to chemotherapy.

It is very much context dependent: what is the disease, what is the molecular target, and how good is the targeted therapy? So, it is difficult to say that precision medicine is more effective than a one-size-fits-all chemotherapy approach because the truth is it is more effective in some settings, less effective in others, and not even a possibility yet in the majority of clinical settings.

Accelerating Genome-Driven Oncology

As Dr. Prasad found in his study, only a small percentage of patients benefit from genome-driven treatment. What has to happen to accelerate progress in genome-driven oncology and improve prognostication of clinical outcomes, especially in the community oncology setting?

Dr. Kim: The results from Dr. Prasad’s study are not surprising at all. We must conduct systematic research to test our hypotheses on implementing biomarkers and genome-driven therapies. That is how we will discover whether precision medicine can become the standard of care for the majority of patients, not by doing it off-study with genomic reports. Accelerating the field will require more research focused on discovering biomarkers and therapies.


  • Angiogenesis inhibitors
  • Apoptosis inducers
  • Gene-expression modulators
  • Immunotherapies (including immune checkpoint inhibitors, immune cell therapy, monoclonal antibodies, and vaccines)
  • Monoclonal antibodies
  • Signal transduction inhibitors

When you treat patients with drugs outside of clinical trials and try to extrapolate the results using a commercial test from a diagnostic company or from an in-house facility, you are not helping to propel the field forward because the results are not measured or quantified, and you are stuck with a pile of anecdotal evidence. Precision medicine clinical trials such as the National Cancer Institute’s MATCH study ( and ASCO’s Targeted Agent and Profiling Utilization Registry -(TAPUR) study ( are meant to investigate therapies based on a patient’s tumor biomarkers to validate or refute hypotheses.

I think it is much too early to adopt a skeptical voice on precision medicine. Many areas of patient treatment have been transformed in less than a decade, and finally patients have hope that cancer treatment is improving. Technology and innovation take time to move forward.

Dr. Prasad: Among the most provocative findings from our study is that when you graph the percentage of patients who benefit from genome-targeted therapy over time, it is about half-a-percent more patients per year. So precision medicine doesn’t appear to be taking off. Before we published our paper, the rhetoric around the use of precision medicine was that it was reaching an inflection point in oncology care, but if you look at the data empirically, that doesn’t appear to be true.

We have to be committed to continuing to research precision therapies, but we have to ask the question, “What kind of research portfolio do you want to advance?” It’s easy to live in a world where we move from one innovative hype cycle to the next. So, one decade, the in-vogue treatment is high-dose chemotherapy and stem cell salvage; the next decade it is immunotherapy, and the next decade it is precision therapy. But the better way to produce more effective therapies for patients is to fund a broad portfolio of treatments that is sustained in good times and bad. I don’t think there will be one magic bullet in cancer treatment, but rather, a combination of therapies.

Dr. Schilsky: To improve outcomes in cancer we have to continue to have a better understanding of the cancer genome. Still, as Dr. Kim and Dr. Prasad said, the majority of patients don’t have a targeted therapy option available for them right now. Most of the alterations identified in genomic-profiling tests are abnormalities of unknown significance. What we need are more and better targets against which we can develop drugs.

There is no question that precision medicine is an effective approach in certain circumstances. It is highly effective in treating HER2-positive breast cancer and BRAF-mutant melanoma, for example. As Dr. Kim mentioned, it is highly effective in treating patients with NSCLC with a variety of molecular alterations for which there are now FDA-approved drugs.

It’s true we don’t yet know what the impact on survival is going to be for some targeted therapies in some contexts. But if we are going to hold that up as the gold standard, we are never going to get new drugs into clinical use.

Dr. Prasad is right in his characterization of the field in the case of a patient with far-advanced cancer and no available standard treatment options for whom a genomic-profiling test is done, essentially as a fishing expedition to see if there might be some alteration in the tumor that leads to a therapy option that hadn’t been considered. Even if that patient can be treated with a drug that is thought to be an effective option, it doesn’t work most of the time, and that’s where we’re talking about 5% of patients benefiting from treatment. But that is 5% of people with far-advanced cancer; it is not 5% of all patients with cancer, including those for whom targeted therapy is standard of care and the most effective known treatment.

Managing Patients’ Expectations

How do you manage patients’ expectations for clinical benefit and explain that a genomic-driven therapy may not be effective in halting their cancer progression and prolonging their survival?

We want precision medicine to become standard of care for every patient, but we are not ready to apply the technique and match a therapy to every patient’s tumor.
— Edward S. Kim, MD

Tweet this quote

Dr. Kim: We want to do everything we can to help our patients, and sometimes we wish we had stopped treatment earlier and discussed end-of-life preferences and hospice care. But when we are in the middle of the fight, we are trained to keep fighting and be an advocate for our patients.

With chemotherapy, there is a lot of toxicity, so it is easy to argue that patients with late-stage disease shouldn’t be overtreated. But if there is an opportunity to put a patient on a clinical trial to sequence his or her tumor and see if there are druggable mutations—and possibly get a substantial response from a therapy—it is worth trying.

When I talk with a patient with advanced cancer who is out of treatment options, I tell that patient, “We can try genomic sequencing of the tumor in the context of a clinical trial, but many times in patients with a similar circumstance, I haven’t found any targetable mutations.” The truth is that if patients are too sick to be enrolled in a study, they are often too sick to be treated.


  • Development of resistance
  • Side effects (including skin problems, hypertension, and blood-clotting disorders)
  • Cost of genomic sequencing
  • Targets affecting small populations

This is, however, a much better situation than several decades ago, when oncologists would not even get patient referrals from primary care providers because there was an overlying pessimism toward cancer therapy that it was too toxic and doesn’t work. I would be disappointed if some of Dr. Prasad’s writings began to reinforce this attitude again among primary care physicians. This may be an area he can write about: how many patients diagnosed with cancer never even see an oncologist?

Dr. Prasad: My belief that we shouldn’t overpromise a potential outcome is what has motivated me to do this work. We talk about hope and hype regarding cancer therapies, but the real balance we should strike with our patients is between hope and honesty. We have a huge problem in the clinic because we have hyped these treatments so much, it is difficult to inject honesty into conversations with patients and explain that a particular strategy may not be of any benefit. We all struggle with this conflict in our practice.

My career goal is to reduce hype and replace it with honesty. Every time we talk with patients, we have to ask ourselves, “Am I being as absolutely honest, rigorous, and evidenced based in my treatment decisions as I possibly can be, or am I being motivated by something else?”

Dr. Schilsky: I agree. As medical professionals, we need to be more circumspect in the way we describe our research findings. Sometimes what is reported in the medical literature and in the lay press are treatment results experienced by exceptional responders. Of course, there are examples of patients who have exhausted all treatment options and were given a death sentence, and then they had a genomic test that found an alteration. They were prescribed a drug for that alteration, and their tumor miraculously melted away. But these examples are a needle in a haystack, and we seldom make that clear in public communication about these issues.

We have an obligation as physicians to communicate to patients the potential benefits and risks of any therapy and give them a realistic appraisal of the possibility that a specific treatment will be effective, what the side effects might be, whether the goal of treatment is to cure or prolong survival, and how much survival prolongation they can expect. If we are doing a good job, this is the discussion we have with patients every time we introduce a new therapy, including genome-driven therapy. But what is often left unsaid is that the chances that patients with far-advanced cancer will benefit are very low and that there could be side effects.

This is an issue of full disclosure. All of us in the medical community, journalists, and patient advocates need to tone down the hype and describe genome-driven approaches in a more realistic and objective way.

Determining the Best Candidates for Genomic Sequencing

Who are the best candidates for genomic sequencing, and when and how often should a patient’s tumor be sequenced?

Accelerating the field will require more research focused on discovering biomarkers and therapies.
— Edward S. Kim, MD

Tweet this quote

Dr. Kim: When I see a patient with newly diagnosed NSCLC for the first time, I only measure the markers that are relevant to starting the patient on standard therapy: EGFR, ALK, ROS1, BRAF, and PD-L1. Once standard therapy stops working, I biopsy the growing tumor and perform a large genome-sequencing test to look for biomarkers to match with a drug in a clinical trial.

Ideally, it may be appropriate to do a genomic test of the tumor every time the cancer progresses, because it is an indication that there has been a shift in the tumor biology and how it relates to the body and the treatment it has been given. But that is not how it is routinely done now.

Dr. Prasad: It is still an open question about whether every patient’s tumor should be sequenced. I believe it is reasonable for patients who have exhausted proven options to undergo testing for specific genomic alterations that are linked to actively recruiting clinical trials for which the patient may be eligible.

Dr. Schilsky: The best candidates for genomic sequencing are patients with cancers that are known to harbor genomic alterations for which there is a known effective therapy available. For patients with [tumor types associated with] known actionable genomic targets, testing should be performed at the time of initial diagnosis of metastatic disease. For others, it is reasonable to perform genomic profiling once standard treatment options have been exhausted if the patient might be a candidate for a clinical trial that has genomic inclusion criteria.

The question of how often patients should be tested is an interesting one because we know that cancers change over time. I have reached the conclusion that patients’ tumors should be tested each time a change in therapy is contemplated, if doing the test might reveal potentially useful information.

Using Genomic Profiling in the Management of Cancer

Genomic profiling is now required in the clinical management of several cancers, including melanoma; gliomas; some sarcomas; and lung, breast, thyroid, ovarian, and colon cancers. Are patients with these cancers experiencing survival benefits as a result of their genome-guided oncology care? What strategies will be needed to engender prolonged responses to targeted therapy and mitigate the consequences of tumor heterogeneity and acquired resistance?

Dr. Kim: We still don’t know what the prognostic value is of some of the biomarkers found in tumors. Do tumors do better with treatment because the biomarker suggests a favorable overall prognosis as opposed to the treatment actually interfering with the marker? If patients with lung cancer have an EGFR mutation, it probably means they are nonsmokers and they tend to have better outcomes than smokers, so EGFR is a positive prognostic marker. Those patients will do better with a drug that targets EGFR, but they also do better on chemotherapy than patients who do not have the EGFR mutation. We still don’t understand the biology of these biomarkers and how they drive the tumor.

The diagnostic companies have to expand the types of genes they are testing and deliver information faster. Pharmaceutical companies need to develop more targeted drugs. Precision medicine gives us hope, but we shouldn’t be overselling it because, as we have said earlier, the majority of patients are not yet benefiting from this approach. Patients with cancer are desperate, and while we have to be good stewards of how we treat patients and present treatment options in a very fair and balanced way, it is also our job to give patients realistic hope. That is why I went into medicine.

Precision medicine is the standard of care now in multiple tumor types for specific patient populations. Our hope is that it will soon become standard of care for many more cancer types and improve outcomes for many more patients.

We talk about hope and hype regarding cancer therapies, but the real balance we should strike with our patients is between hope and honesty.
— Vinay Prasad, MD, MPH

Tweet this quote

Dr. Prasad: There is zero doubt that patients with melanoma who have the BRAF V600E mutation treated with a MEK and BRAF inhibitor or patients with breast cancer who have HER2 amplification treated with trastuzumab are experiencing a survival benefit by getting genome-driven care. However, just because some patients with certain targetable genes in some cancers are benefiting from precision medicine, it doesn’t mean that all patients with these genes in all cancers will also benefit. That is the crux of the dilemma.

How do we scale up to validate every gene in every cancer through testing? Right now, we have a Wild West situation in which physicians are prescribing expensive off-label drugs to patients outside of clinical trials, and no one is keeping track of their outcomes.

I commend ASCO for launching the TAPUR trial because it is capturing safety and efficacy data for FDA-approved targeted drugs in patients with advanced cancers that have a potentially actionable genomic alteration. We need to run more of these types of trials.

Patients’ tumors should be tested each time a change in therapy is contemplated, if doing the test might reveal potentially useful information.
— Richard L. Schilsky, MD, FACP, FSCT, FASCO

Tweet this quote

Dr. Schilsky: In many cases, we don’t know whether patients are experiencing a survival benefit because the necessary randomized trials with survival endpoints either haven’t been done or haven’t been completed. As Dr. Prasad said, there is evidence that patients with melanoma have experienced a survival benefit compared with patients treated with conventional chemotherapy, and the same is true for many patients with lung cancer.

Treatment resistance is a huge problem because tumors will eventually evolve a resistance mechanism to essentially all of the targeted therapies, so it is highly unlikely that any of these treatments for most solid tumors will be curative. There is hope that using combinations of targeted therapies might produce long-term control of a tumor as opposed to using a single agent because it might prevent the emergence of resistance. There is reason to believe that combination chemotherapy or combination targeted therapy could potentially cure at least some cancers, but the challenge is that it will likely require multiple different combinations uniquely tailored to each patient’s cancer.

DISCLOSURE: Dr. Kim is a consultant to AstraZeneca, Genentech/Roche, Merck, Takeda, Boehringer Ingelheim, and Pfizer. Dr. Schilsky has received institutional research funding from AstraZeneca, Bayer, Bristol-Myers Squibb, Genentech/Roche, Lilly, Merck, and Pfizer. Dr. Prasad has received royalties from his book Ending Medical Reversal; his work is funded by the Laura and John Arnold Foundation; has received honoraria from grands rounds and lectures from several universities, medical centers, and professional societies; and payments for contributions to Medscape.


1. National Institutes of Health: What is the difference between precision medicine and personalized medicine? What about pharmacogenomics? Available at Accessed September 19, 2018.

2. National Cancer Institute: Targeted cancer therapies. Available at Accessed September 19, 2018.

3. U.S. Food & Drug Administration: FDA approves first cancer treatment for any solid tumor with a specific genetic feature. May 23, 2017. Available at Accessed September 19, 2018.

4. Marquart J, Chen EY, Prasad V: Estimation of the percentage of US patients with cancer who benefit from genome-driven oncology. JAMA Oncol 4:1093-1098, 2018.