Can Society Afford Not to Personalize Medicine in Oncology?

October 2015, Vol 5, No 7

Personalized medicine gained a national platform with the January 2015 proposal by President Barack Obama to analyze genetic information from more than 1 million American volunteers, as part of a new initiative to understand human disease and develop medicines targeted to an individual’s genetic makeup. The president included a $215-million line item in his proposed 2016 budget to fund this initiative. The initial focus would be on cancer and how to individualize treatment for a range of cancer types. By April 2015, Roche had invested $250 million to become the majority shareholder in Foundation Medicine, with the goal to optimize the development of and access to novel treatment options for patients with cancer and to advance personalized healthcare in oncology.

At the 2015 Association for Value-Based Cancer Care conference, Michael A. Kolodziej, MD, National Medical Director, Oncology Solutions, Aetna, discussed how big investments in personalized medicine seem to make a compelling argument that this is the way of the future.

Dr Kolodziej’s biggest concern is that there are several challenges regarding personalized medicine, and that the payer and the provider communities need to recognize those challenges. Then we need to fix them, and very quickly. The challenges holding us back occur at every level, including the testing, data, doctors, patients, and cost.

Questions abound regarding the actual accuracy and relevance of genomic tests, with no real standards for comparison and benchmarking. Clinical validity and utility must be measured, assessed, and proved in the context of reliable analytic performance and related economic value.

Coding and Payers

He said that payers do not receive enough information via billing codes to understand what tests are being done, and how they are being done. Kits and companion diagnostic tests are performed, but there is no way to know if they are being conducted accurately, or even in what setting. Often, coding required the listing of all the steps taken in the process of a given test, which is known as stacking codes.

Payers had no way of knowing what tests were being done or why, but the stacked codes were often paid anyway. Recent coding changes have allowed for more precise coding when there is a code that is specific for an analyte.

However, even this process leaves payers with a vacuum of information about the value and depth of oncology management and any contribution of the test and the resulting information. The current coding leaves payers paying for tests without having confidence that the test actually changed the oncology management of the patient.

We can currently test approximately 50 genes; however, as the number of tests increases, particularly for next-generation sequencing, we could soon be testing 300 or more genes. The results from these tests yield tables and catalogs of information on genetic mutations. Our challenge is to look only at the mutations that can be considered actionable, with actions that we know work at least sometimes in patients with a particular mutation. Other possible actions should, appropriately, fall below our attention level, because we have no data that they will work. All mutations are not equal in an actionable gene, and we cannot afford to study every possible treatment. Some of the current tests focus on hot spots and are designed to include relevant information (eg, companion diagnostic mutations), but they do not look at the whole gene, and thus are missing a large part of the picture.

Providers and Personalized Medicine

When we look at the whole gene, some mutations will matter, and some will not. As physicians, we need to understand the whole picture. We need answers to identify the few patients for whom a drug will be lifesaving, so that we can help get the drug to them. But who are those patients?

We want to personalize medicine for them, but we cannot afford to do it blindly, and universally, in the hope that we will find them inside a group for which there will be no benefit. We have an incomplete set of information that is hindering optimal patient care.

Population Health

As researchers, payers, and providers, we need to collect more information and figure out the best way to get the right treatments to patients.

We now know that further information will help us understand what is working, and more important, why it works. Maybe it is a gene, a protein, an expression analysis, or even characteristics of a phenotype. We currently live in a world of anecdotes, and we need to move into a world where we can take a population, identify an appropriate subpopulation, and then integrate all of the advances we have made in genotyping tumors, in understanding phenotypes, and from current clinical trial information.

Low Confidence in Genetic Testing

We need confidence in these tests. Stacy W. Gray, MD, AM, of the Dana-Farber Cancer Institute, Boston, recently asked physicians at Dana-Farber their thoughts on using somatic mutation testing in their patients (Gray SW, et al. J Clin Oncol. 2014;32:1317-1323). Their responses mirror what is often seen in private practices. Even at 1 of the top 5 cancer centers in the world, there were not many physicians who were confident about their ability to use this somatic mutation testing: 22% of the physicians reported low confidence in their genomic knowledge, 25% anticipated testing most patients, and 18% anticipated testing patients infrequently.

One significantly large oncology group looked at the uptake of Oncotype DX, which is frequently accepted and covered by payers. The group used an objective measure of the percent of patients who would be eligible for the test based on a particular indication. The results showed that the prescribing behavior of physicians varied widely. Unfortunately, if the physicians did not order a test very often, they might not have used it wisely. If a physician orders a test often, they are more likely to understand how to use it in their medical decision-making.

Action Points

Dr Kolodziej summarized what needs to be fixed as we move toward personalized medicine, noting that each one is quite fixable.

Dr Kolodziej noted that we need better management and a better understanding of the cost of therapy (more than only the test), and that as targeted therapies become more cost-effective, the more we can define the appropriate subpopulation for them, by reviewing the mutations and the adaptive response to the mutation. We cannot afford to treat large populations of patients with targeted therapies in hopes that perhaps 15% of patients are cured.

If we add a new targeted therapy to the first-line treatment of lung cancer, although there may be a distinct population for whom this could be curative, without biomarkers to help identify that subpopulation of patients with lung cancer, we could add as much as $28 billion to our country’s spending for patients with this disease. Currently, the total cost of care for a patient with lung cancer is between $100,000 and $120,000. Adding a broad-spectrum application of a new treatment that may cost $150,000 per patient is clearly unaffordable.

He ended by noting the need to ensure that the right patients get these wonder drugs. We cannot afford not to. However, we also know that not all sequences work. If a therapy fails, we have spent money and received no value. This comes at a cost to patients, physicians, payers, and society in general. Opportunity costs will result if the wrong sequence of drugs is chosen first; we might have made a wrong decision because of a lack of personalization information and may have hurt the patient in the process with unwarranted side effects and costs.

We cannot afford not to move forward with personalized medicine. As a society, as physicians, and as individuals, this is the only way we can continue to embrace and adopt true innovation. Personalized medicine requires that we know the genome, as well as the treatment’s value and the clinical utility. We need to collect all of the information we can about our patients to learn what is actually working and why it is effective.

Related Articles

© Amplity Health. All Rights Reserved. Privacy Policy.