May 25, 2015
18 min read

Despite promise, ‘maximum potential’ of precision medicine yet to be reached

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact

Precision medicine has garnered tremendous public attention since President Barack Obama established it as a national priority during his State of the Union address.

The Precision Medicine Initiative — created with an initial $215 million investment — is intended to help researchers determine which treatments best target specific characteristics of each individual patient’s disease. Leaders expect to gain insights from tissue samples, as well as biological, environmental, lifestyle and behavioral information provided by at least 1 million volunteers.

“The initiative is an investment to support research that includes sharing ‘big data’ across U.S. health care systems and that focuses on certain diseases like cancer, a problem that costs the U.S. about $120 billion a year,” Sameek Roychowdhury, MD, PhD, medical oncologist and specialist in cancer genomics at The Ohio State Comprehensive Cancer Center — Arthur G. James Cancer Hospital and Richard J. Solove Research Institute, told HemOnc Today. “Cancer is a natural entry into precision medicine because the new technologies allow us to look at genetic changes, and cancer happens to be a disease driven by genetic changes. Precision medicine research in cancer can demonstrate how we could employ these approaches in other human diseases.”

The ability to tailor cancer treatment based on genomic knowledge offers a therapeutic advantage, and it also may spare patients from unnecessary toxicities or costs incurred from less effective treatments. This strategy appears particularly promising for patients with advanced cancers, for whom standard-of-care approaches may no longer work. The same premise also can inform clinicians about treatments for early-stage disease and may even play a role in cancer prevention.

Still, experts emphasize the field is in its infancy, and they suggest such large potential payoffs may create unjustified hype or false hope for patients. In reality, ideal genetic targets may be relatively uncommon in cancer, and the rapidly expanding range of available genomic data may further strain clinicians.

Sameek Roychowdhury, MD, PhD

Sameek Roychowdhury

“These approaches are pretty new, and some of this information is difficult to interpret,” Roychowdhury said. “We need specialization to interpret that information. Further, some of the information is not usable yet and it isn’t ready for prime time.”

HemOnc Today spoke with oncologists and genomic specialists about how precision medicine is incorporated into cancer treatment, whether excitement should be tempered due to the relatively limited amount of usable data available to help target a rapidly increasing number of identified actionable mutations, and the potential that precision medicine can inform decisions about first-line therapies or cancer prevention strategies.

One gene, big data

The principles of precision medicine are changing the way clinicians define and treat cancer.

“Precision medicine has the potential to revolutionize cancer treatment decision-making at every level,” Cory Abate-Shen, PhD, professor of urology and pathology and associate director of the Herbert Irving Comprehensive Cancer Center at Columbia University, said in an interview. “It is as if we’ve been treating patients blindly without the knowledge of the molecular properties of an individual’s tumor. The consequence is perhaps missing the opportunity to give the right drugs to the right patients.”

Patients should be treated based on the genetic makeup of their disease, not their cancer type, Roychowdhury said.

“It’s important to start thinking about an agnostic approach to cancer, meaning not just where the cancer arose — from the breast, lung or blood cells — but its molecular underpinnings,” he said. “We might find that the same gene that is important in lung cancer is also important in a certain kind of lymphoma, and the therapies are going to be similar to target that gene alteration. The point is, we need to cure all cancers, period.”

At the American Association for Cancer Research Annual Meeting in April, Nickerson and colleagues presented data on genomic alterations found in 54 patients with bladder cancer. Researchers identified somatic alterations in the BRCA1-associated protein-1 (BAP1) gene in 15% of tumors. The finding suggests these patients may benefit from treatment with a PARP inhibitor, as some patients with BRCA1- and BRCA2-mutant breast, ovarian and prostate cancers do.


Identifying targets may be particularly helpful for rare malignancies, as well as cancers in children, Arul M. Chinnaiyan, MD, PhD, director of the Michigan Center for Translational Pathology and professor of urology at the University of Michigan, told HemOnc Today.

“Pediatric malignancies are intriguing for precision medicine approaches because often times there are fewer drivers contributing to the disease than in adult tumors, where the mutational landscape is quite complex,” Chinnaiyan said. “Additionally, we might find something clinically actionable through genomic sequencing for rare cancers or cancers of unknown origin.”

Ross and colleagues conducted genomic profiling on 200 samples of cancer of unknown primary site. Results, published in February in JAMA Oncology, showed 96% of samples harbored at least one genetic alteration and 85% harbored a clinically relevant genomic alteration.

Identifying effective treatment combinations is one of the biggest challenges of precision medicine, according to Arul M. Chinnaiyan, MD, PhD.

Identifying effective treatment combinations is one of the biggest challenges of precision medicine, according to Arul M. Chinnaiyan, MD, PhD. “Comprehensive sequencing-based approaches are applied in patients with advanced cancers who have been treated with multiple regimens, so they have developed different resistance mechanisms,” he said. “You generally don’t get a durable response with a single agent.”

Source: Photo courtesy of Liz Walker, University of Michigan

Because a detected mutation may be uncommon for any individual patient’s cancer type, sharing data will be essential.

“A patient may have breast cancer but have a mutation so rare within breast cancer that it’s hard to study unless we work together,” Roychowdhury said. “The data has to be collected and shared so we can learn from that as a community.”

Many cancer centers have initiated precision medicine data collection and sharing programs. For example, Robert H. Lurie Comprehensive Cancer Center, Northwestern Medicine Developmental Therapeutics Institute and Northwestern Memorial Hospital formed Northwestern Onco-SET. The “SET” acronym represents the program’s three steps — sequence, evaluate and treat.

“We plan to develop research efforts around the Onco-SET,” Leonidas Platanias, MD, PhD, director of Lurie Cancer Center, said in an interview. “For instance, there may be specific research efforts to develop compounds for new mutations that will be identified in the process.”

Further, IBM Watson Health recently announced a collaboration with a dozen cancer centers that will use Watson — IBM’s supercomputer — to analyze patients’ genetic information and identify their appropriate treatment options by the end of this year.

Although patients and clinicians may be optimistic about the potential benefits of precision medicine, many of these projects are still in their initial development stages.

“This really is a work in progress,” Alex Adjei, MD, PhD, FACP, the Katherine Anne Gioia chair in cancer medicine and senior vice president of clinical research at Roswell Park Cancer Institute, told HemOnc Today. “Although we can decipher the molecular aberrations in a particular cancer, we have effective treatments for only a few of these.”

Actionable findings

There are several examples of precision medicine targets in oncology, such as HER-2 for breast cancer, BRAF for melanoma and ALK for lung cancer.

According to a blog post by Janet Woodcock, MD, director of the FDA’s Center for Drug Evaluation and Research, the agency has approved 30 targeted therapies for a variety of indications since the approval of trastuzumab (Herceptin, Genentech) for HER-2–positive breast cancer in 1998.

The use of imatinib (Gleevec, Novartis) for chronic myeloid leukemia often is viewed as the epitome of precision medicine. Yet, it is unclear whether the success with CML — the cause of which can be traced to the BCR-ABL fusion gene — can be replicated in other malignancies.

“With CML and BCR-ABL, often the ABL kinase fusion is the only driver for that disease, so that’s why these patients get such a remarkable response,” Chinnaiyan said. “CML has really been the poster child for precision medicine. But that’s the extreme, exceptional responder-type example of precision medicine.”

Identifying a single mutation within another cancer type may not necessarily induce such dramatic responses.

“Although a number of aberrations respond to inhibitors regardless of tumor type — such as the ALK translocation, which responds to ALK inhibitors in a variety of cancers — BRAF mutations in colon cancer are not very responsive to single-agent BRAF inhibitors,” Adjei said.


Researchers observed the BRAF V600E mutation — present in about 40% of patients with melanoma and targeted by vemurafenib (Zelboraf; Genentech, Daiichi Sankyo) — in a subset of patients with colorectal cancer. Thus, Kopetz and colleagues sought to evaluate vemurafenib in 21 patients with BRAF-mutated metastatic colorectal cancer. The results — presented at the 2010 ASCO Annual Meeting — showed only one patient achieved a partial response.

Researchers also may need to identify potential resistance mutations.

“Imatinib was a very straightforward model because BCR-ABL was clearly the culprit of CML,” Platanias said. “Solid tumors are more complicated. Due to negative feedback regulatory pathways, targeted therapy may lead to the escape of a new class of cells and activation of other cellular pathways. Therefore, you may need to combine targeting to block the escape mechanism of a cancer cell.”

Identifying appropriate combinations is one of the biggest challenges of precision medicine, Chinnaiyan said.

“These comprehensive sequencing-based approaches are applied in patients with very advanced cancers who have been treated with multiple different regimens, so they have developed different resistance mechanisms,” Chinnaiyan said. “Even when we do find potentially actionable driver genes, it becomes quite difficult to target because they have other resistance mechanisms in place. You generally don’t get a durable response with a single agent.”

Studying how HIV and tuberculosis are treated may provide insights into how best to attack these resistance mechanisms, Roychowdhury said.

“HIV and TB can be treated with a curative intent with three or four drugs, each of which targets a different part of the life cycles of that organism or virus,” Roychowdhury said. “We need to stop cancer escaping therapy before it even happens in order to reach our curative goals.”

Although Roychowdhury suggested 40% of patients have actionable mutations, Platanias said only about 15% do.

“In reality, precision medicine is still research,” Platanias said. “It’s not standard of care, but we learn during the process. We have by no means seen the maximum potential of precision medicine.”

Implications for patients, clinicians

Because few patients have actionable mutations, the expanse of unusable knowledge may overwhelm physicians and disappoint patients.

Cory Abate-Shen, PhD

Cory Abate-Shen

“I understand the concern: ‘I have all of this information, what do I do with it?’” Abate-Shen said. “We should flip that around and say, ‘Wow, it’s great to have information as opposed to little or no information.’”

It may take time for the clinical community to adopt that mindset.

Gray and colleagues surveyed 160 medical oncologists (57%), surgeons (29%) and radiation oncologists (14%) to assess their attitudes about genomic testing.

Results, published in 2014 in Journal of Clinical Oncology, showed 22% of participants lacked confidence in their genomic knowledge, 14% had little to no confidence in their ability to explain genomic testing to patients, and 26% had little to no confidence in their ability to use genomics to choose treatments.

“Standard education does not prepare us for how to sit down and talk to patients about this highly scientific concept,” J. Philip Kuebler, MD, an oncologist with Columbus Oncology and Hematology Associates and HemOnc Today’s section editor for independent/private community practice, said in an interview.

Many academic centers and hospitals conduct tumor board meetings so clinicians can decipher genetic information in a collaborative setting.

An observational analysis presented by Kehl and colleagues at the 2014 Quality Care Symposium showed patients of physicians who participated in tumor boards were more likely to enroll on clinical trials (OR = 1.6; 95% CI, 1.1-2.2). Weekly tumor board participation also was linked to prolonged survival for patients with extensive-stage small cell lung cancer or stage IV colorectal cancer compared with less frequent participation (P .05).


“At our institution, these precision medicine tumor boards have helped convey appropriate options for individual patients and also to educate our medical oncologists in the use of precision medicine approaches,” Chinnaiyan said.

Yet, tumor boards and education efforts may not be as accessible in the community practice setting.

“Some of the sequencing companies send pages of results, and they will make recommendations or at least mention possible drugs you can use,” Kuebler said. “That helps a little bit for a physician who doesn’t have the molecular biologic knowledge.”

A lack of clinical trial data may be the primary source of difficulty.

Alex Adjei, MD, PhD, FACP

Alex Adjei

“The challenges stem from the fact that often there are no clinical data to indicate whether a drug will be effective against a certain aberration,” Adjei said. “A number of these abnormalities are relatively infrequent, making clinical studies challenging to perform.”

Patients also need to be well-informed.

“We have to warn patients ahead of time that precision medicine sounds great when it’s written up all over the Internet, but it’s expensive, the testing is not covered by insurance and — if you have a ‘druggable’ mutation — it’s not easy to obtain the drug,” Kuebler said. “Often the test reveals a mutation we have never heard of for which there is certainly no drug available. Other times there is no mutation, and precision medicine isn’t an option for those patients.”

There are certain abnormalities that may predict for resistance rather than sensitivity, Adjei said. For example, RAS in colon cancer eliminates cetuximab (Erbitux; Bristol-Myers Squibb, Lilly) and panitumumab (Vectibix, Amgen) as treatment options. This knowledge prevents exposure to side effects and cost without benefit; however, patients likely are hoping genomic sequencing will lead to a treatment option, experts said.

Results of genomic sequencing also may reveal genetic predisposition for other health risks.

Yushak and colleagues presented survey data from 855 patients with breast, gastrointestinal and thoracic cancers at the AACR Annual Meeting. Results showed 57% would want to know about a gene variant that would cause a serious illness other than cancer, and 47% would want to know about a detected variant of unknown significance.

Cost also is a concern.

The cost of sequencing an individual genome decreased from more than $95.2 million in 2001 to less than $5,000 in 2014, according to data from the National Human Genome Research Institute Genome Sequencing Program. Still, insurance may not cover this cost, nor the cost of an off-label agent.

“Even if the insurance company covers the drug, there’s a copay,” Kuebler said. “I have patients who have such large copays, it’s a problem.”

Although precision medicine approaches may focus on a small percentage of patients with a certain mutation, the information gleaned from such analyses may be invaluable.

“Should we treat every patient with breast cancer with the same treatment, knowing that most of them won’t respond to it?” Roychowdhury said. “It may seem like a precision medicine approach is expensive, but the opposite is just as costly and we’re wasting time. If we can complete a clinical trial by molecularly selecting and enriching patients who have the right genetics for a specific therapy, that trial can be smaller, the differences in benefit can be greater, and there is cost savings there. What we have been doing isn’t working.”

‘Precision prevention’

Concepts of precision medicine may prevent overtreatment or the development of cancer.

“We don’t invest enough time, energy, effort or money into prevention,” Abate-Shen said. “Even if people don’t die of cancer, there are considerable survivorship issues, many of which can be addressed by prevention, better modes of detection and determining which cancers don’t need to be treated.”


Improved detection methods have led to an “embarrassment of riches,” Abate-Shen said. Estimates indicate 250,000 men are diagnosed with prostate cancer each year, yet only 10% of those men have lethal disease.

“The problem is that 90% of the population does not really need aggressive treatment,” she said. “But if you don’t treat anyone, we may end up in a situation where you have missed the chance to treat those [with potentially lethal disease] early.”

Precision medicine can help clinicians work through this scenario.

“What I call ‘precision prevention’ is not just identifying tumors at the earliest possible stages, but then deciding which ones do and don’t need to be treated,” Abate-Shen said. “Even within a person’s tumor, there may be sub-characteristics that have more or less aggressiveness. We can tease out this heterogeneity and understand how it contributes to the potential to develop metastasis. You want to treat not just the individual tumor, but the individual bad tumor.”

Abate-Shen and colleagues have developed a prostate cancer panel composed of three indolent and two aggressive markers of disease. Irshad and colleagues published the indolent signature — composed of FGFR1, PMP22 and CDKN1A — in 2013 in Science Translational Medicine. Aytes and colleagues published the aggressive FOXM1 and CENPF signature in 2014 in Cancer Cell.

“This is a two-part test, so you can look for aggressiveness and indolence at the same time,” Abate-Shen said. “Year after year, the clinician can look at these molecular markers to help with decision-making and incorporate the precision component of prevention.”

Earlier use of precision medicine approaches also may lead to improved efficacy of targeted agents.

“We have to bring these precision medicine approaches earlier in the course of cancer development before the patient has been treated with multiple different chemotherapeutics and before they have developed different resistance mechanisms,” Chinnaiyan said.

Results of the MATCH trial, scheduled to open this year, may support earlier implementation of precision medicine. The trial is designed to assign patients with advanced solid tumors and lymphomas to different treatment arms based on their genetic alterations.

“If MATCH is a positive trial and we get some data that shows precision medicine is better than standard care, then it’s going to move forward very quickly,” Kuebler said. “But, we haven’t actually proven that targeting a patient’s tumor is better in terms of response, survival or PFS than my best choice as a physician. I’m hopeful and I think this is the way to go, but we don’t have proof of concept yet.”

In his State of the Union address, Obama predicted precision medicine may one day help prevent diseases.

“Precision medicine efforts that include germline sequencing provide some insight into other cancers to which the patient might be susceptible,” Chinnaiyan said. “It also suggests their family members might be at risk and we should enhance screening. In that aspect, precision medicine efforts that incorporate normal sequencing are, in a way, studying cancer prevention because they evaluate individual predispositions to cancer.”

The association between Lynch syndrome and colorectal cancer is another area in which precision medicine may inform cancer prevention, Roychowdhury said.

“This leads back to early detection and hopefully a better opportunity for a curative approach,” he said. “It’s a good example of how basic bench research has been brought to the clinic because of investment in translational research. I’m hopeful that precision medicine research will eventually include more strategies for prevention and early detection, in addition to the treatment of patients with advanced cancer.” – by Alexandra Todak


Aytes A, et al. Cancer Cell. 2014;doi:10.1016/j.ccr.2014.03.017.

Gray SW, et al. J Clin Oncol. 2014;doi:10.1200/JCO.2013.52.4298.

Irshad S, et al. Sci Transl Med. 2013;doi:10.1126/scitranslmed.3006408.

Kehl KL. Abstract 179. Presented at: Quality Care Symposium; Oct. 17-18, 2014; Boston.

Kopetz S, et al. Abstract 3534. Presented at: ASCO Annual Meeting; June 4-8, 2010; Chicago.

NHGRI Genome Sequencing Program. DNA Sequencing Costs. Available at: Accessed May 7, 2015.

Nickerson ML, et al. Abstract 1105. Presented at: American Association for Cancer Research Annual Meeting; April 18-22, 2015; Philadelphia.

Ross JS, et al. JAMA Oncol. 2015;doi:10.1001/jamaoncol.2014.216.

Woodcock J. FDA continues to lead in precision medicine. Available at: Accessed May 7, 2015.

Yushak ML, et al. Abstract 3879. Presented at: American Association for Cancer Research Annual Meeting; April 18-22, 2015; Philadelphia.

For more information:

Cory Abate-Shen, PhD, can be reached at Columbia University, 1130 St. Nicholas St., ICRC Room 217A, New York, NY 10032; email:

Alex Adjei, MD, PhD, FACP, can be reached at Roswell Park Cancer Institute, Elm and Carlton streets, Buffalo, NY 14263; email:


Arul M. Chinnaiyan, MD, PhD, can be reached at Michigan Center for Translational Pathology, 1400 E. Medical Center Drive, 5309 CCC, Ann Arbor, MI 48109; email:

J. Philip Kuebler, MD, can be reached at Columbus Oncology and Hematology Associates, 810 Jasonway Ave., Suite A, Columbus, OH, 43219; email:

Leonidas Platanias, MD, PhD, can be reached at Robert H. Lurie Medical Research Center, Room 3-125, 303 E. Superior St., Chicago IL 60611; email:

Sameek Roychowdhury, MD, PhD, can be reached at The Ohio State University Comprehensive Cancer Center, Biomedical Research Tower, Room 508, 460 W. 12th Ave., Columbus, OH 43210;; email:

Disclosure: Chinnaiyan reports scientific advisory board roles with Life Technologies, MolecularMD and Paradigm. Abate-Shen, Adjei, Kuebler, Platanias and Roychowdhury report no relevant financial disclosures.



Should all patients with cancer undergo genomic sequencing?


Sequencing should be part of the routine management of patients with cancer.

Howard L. McLeod, PharmD

Howard L. McLeod

As a way of managing patients, sequencing needs to become less novel and something we do on a routine basis. Looking forward — as the cost goes down and value goes up — this is the type of information for which we will need baseline data, actionable or not, to be able to build the future sequencing data into our clinical decision-making.

There are already areas where genomic sequencing should be routine, such as for patients with cancers for which there is no standard of care, patients who have run out of evidence-based options, or patients who never had options due to having a rare malignancy. In the past, we would refer these patients to a clinical trial or request the off-label use of a drug. Now, the standard should be that we use sequencing data to either enroll patients in clinical trials or have a more objective indication for off-label drug use.

Many sequencing strategies provide both germline and somatic data. Some commercial approaches only provide somatic data, but a lot of the strategies used at academic centers provide germline and somatic information. With that data comes opportunity to not only look for beneficial therapeutic options, but also to look at the risk part of the risk–benefit ratio. These data may be able to identify patients who have a heightened risk for peripheral neuropathy or other attributes that would allow us to make a more precise therapeutic decision. The choice may be between an excellent therapy and a not-so-excellent therapy. In that case, it would take amazing data to shift the decision. However, often it is between a choice of equals — there are two options that seem to have a similar probability of offering benefit — and just a little data could shift that discussion. Part of the value of genomic sequencing that we are seeing has been informing our choice amongst equals.

If we consider the use of EGFR antibodies in colorectal cancer, whether the patient is positive for KRAS mutations is really a yes-or-no type of decision. If the answer is “yes,” that excludes the patient from that treatment. This is the type of model we have had, but now we are moving into an era in which the sequencing data will allow us to make a more refined decision on which of two apparently equal options should be the place to start. Further, when we evaluate a patient’s sequencing data, we not only make suggestions about where to start therapy, but also about what therapy to start next under the understanding that most therapies buy a little time but are unlikely to be a major cure. Thus, the data will hopefully provide some information to make a choice today, but it will also give us some direction for tomorrow.

Howard L. McLeod, PharmD, is medical director at the DeBartolo Family Personalized Medicine Institute at Moffitt Cancer Center. He can be reached at Disclosure: McLeod reports no relevant financial disclosures.


Genomic sequencing is a useful and valuable technique, but it is not ready for prime time.

Daniel Hertz, PharmD, PhD

Daniel Hertz

Most patients have common tumor types for which there are first-line treatments that have proven efficacy in large randomized clinical trials. There is no evidence that foregoing these first-line treatments in favor of a targeted agent would benefit these patients. A study called M-PACT (Molecular Profiling-based Assignment of Cancer Therapeutics) is designed to address this question by comparing targeted treatment vs. standard of care. In terms of applying sequencing to the broad patient population, we are going to have to wait to see the results of those “test-the-strategy” studies to determine the clinical utility of genome sequencing for patients with cancer.


Genomic sequencing can benefit some patients, such as those with rare tumor types or patients who are on their second or third line of therapy. However, initial reports of research-based sequencing in these patients suggest a lack of clinically actionable findings. Researchers from the Michigan Oncology Sequencing Center (MI-ONCOSEQ) will report data from their first 200 enrolled patients at ASCO this year (Cobain F, et al. Abstract 11057. Scheduled for presentation at: ASCO Annual Meeting; May 29-June 2, 2015; Chicago). Based on their data, most tumors harbor biologically relevant genetic aberrations but very few patients — 10% to 20% — can be matched to a targeted agent approved in any tumor type or available on an investigational clinical trial. This 10% to 20% is far lower than what is being advertised by companies that provide tumor sequencing services, and the expectations of patients. The MI-ONCOSEQ group recently reported that 72% of patients agreed that the results of their tumor sequencing would help them enroll on a clinical trial, with half of those patients expressing strong agreement (Roberts JS, et al. Giving patients incidental information from genomic sequencing. Presented at: Annual Meeting of the Society of Behavioral Medicine; April 22-25, 2015; San Antonio). Finally, to offer this technology to all patients with cancer but only inform treatment decisions for 10% to 20% of them does not appear to be an efficient use of limited health care resources. At this time, cost and bioanalytical/bioinformatic capability preclude widescale adoption of tumor sequencing. Continued research efforts will overcome these challenges and the number of patients for whom tumor sequencing should be offered undoubtedly will expand rapidly.

Germline information derived from genomic sequencing has lifelong value to the patient. Unfortunately, similar to tumor sequencing, at this time actionable variants are found in a very small subset of the population. Based on what I have seen, the proportion of patients with cancer who carry an actionable mutation in a cancer-risk gene is around 3%, and clinical use of germline genetics to inform treatment decisions is still in its infancy. I am just as excited as others in the clinical, patient and research communities to move these technologies forward to improve patient care. However, even when considering the added lifelong benefit derived from concurrent germline analysis, at this time we do not obtain sufficient clinically useful information to warrant genomic sequencing for all patients.

Daniel Hertz, PharmD, PhD, is research assistant professor of pharmacy at the University of Michigan. He can be reached at Disclosure: Hertz reports no relevant financial disclosures. He is not personally affiliated with MI-ONCOSEQ and his views do not represent the views of the project and/or its leaders.