Personalized Medicine Affordability Lies in Enhanced Predictors of Response to Therapy

Wayne Kuznar

August 2015, Vol 6, No 7 - AVBCC 2015 5th Annual Conference


Washington, DC—Personalized medicine is the best way to take advantage of innovation in therapy, but the method in which it is paid for must be addressed to fully realize its potential, said Michael Kolodziej, MD, National Medical Director, Oncology Solutions, Aetna, at the Fifth Annual Conference of the Association for Value-Based Cancer Care.

He shared payers’ perspectives on payer coverage and reimbursement decisions for molecular diagnostic tests, including the clinical relevance of next-generation sequencing testing in patient care.

Reimbursement for molecular and next-generation sequencing tests depends on the evaluation of their analytical validity, clinical validity, and clinical utility, said Dr Kolodziej. Stakeholder views of these concepts often differ, posing challenges to the adoption of personalized medicine.

Payment for complex tests evolved from stacking Current Procedural Terminology (CPT) codes to a single specific code. The conversion of stacked CPT codes to specific codes was designed to improve the ability to track utilization, enable decision support tools, and to enforce coverage policy, Dr Kolodziej noted.

For example, CPT code 16460 is for EGFR mutation analysis, and the coverage policy at Aetna is for the indication of non–small-cell lung cancer (NSCLC). “We pay that code,” he said. “We don’t even question it. We just pay that code if you have lung cancer. But we don’t really know how you did the test.”

Two companion diagnostic tests exist for EGFR. Those tests have undergone scrutiny by the FDA and are tied to a specific clinical outcome in patients receiving therapy targeted to an EGFR mutation.

“We know how the test is done, because it’s a kit,” he said. “Sometimes it’s done well, and sometimes it’s probably not done well. I really can’t tell you when it’s done well, and when it isn’t done well. I know it’s being done by a CLIA-­certified lab. Beyond that, I don’t know.”

The majority of tests for this analyte are not the companion diagnostic, and are probably better than the companion diagnostic test, noted Dr Kolodziej. “What’s more important is, I don’t know whether the information is really helping the doctor make a good clinical decision. This is a problem.”

The Potential of Next-Generation Sequencing

The clinical and economic value of next-generation sequencing is not fully understood at present. Next-generation sequencing technology is likely to alter the economics of patient care, but assessing its value is complicated.

The clinical relevance of such testing lies in whether the information gleaned changes to patient management, according to Dr Kolodziej.

Next-generation sequencing provides a catalog of information and potential treatments. “You do get more mutations if you sequence the whole gene,” he said, but not all mutations are equal in an actionable gene.

Predicting Response

A patient who will benefit from an innovative treatment, based on mutation information, should receive this treatment. “If you have a HER2 mutation, you should get trastuzumab,” he said. “HER2 is validated if you have amplification.”

Most of the abnormalities in the HER2 gene are not amplification in tumors, other than breast cancer. “They’re point mutations,” he said. “Some of them are activating.”

In NSCLC, mutations in 2 regions of the EGFR gene (exons 18-24)—small deletions in exon 19 and a point mutation in exon 21 (L858R)—have been shown to predict tumor response to tyrosine kinase inhibitors.

In 2 sequential randomized trials comparing the EGFR inhibitor afatinib (Gilotrif) with standard chemotherapy for the treatment of metastatic NSCLC, progression-free survival and overall survival were superior in patients receiving afatinib who harbored either the deletion or the point mutation. “It’s something that a doctor and a patient should want to know,” he emphasized.

Identifying Phenotype Also Important

Response to therapy is not exclusively driven by genotype. Phenotypic characteristics can determine the likelihood of response to therapy. “That’s my prediction,” Dr Kolodziej said. “There are 2 tumors for which HER2 mutation makes the most difference—breast cancer and esophagogastric cancer. It is confined to certain anatomic areas, whether we like it or not.”

Although, anecdotally, acting on abnormalities in HER2 makes a difference in other tumor types, identifying appropriate subpopulations that will benefit from this knowledge is paramount.

“That identification is going to involve all the advances that we have made, and they have been tremendous advances…in genotyping tumors, our understanding of phenotype, and current clinical trial information,” said Dr Kolodziej.

The Cost of Failure

Failed therapies have a cost for all parties, Dr Kolodziej argues. Novel therapies that do not yet have specific biomarkers, such as PD-1 inhibitors in melanoma and other tumor types, become worrisome in light of their costs and length of therapy.

Nivolumab (Opdivo) is associated with a survival advantage over doce­taxel (Taxotere) in the treatment of lung cancer, and is now indicated as second-line therapy for this indication. If nivolumab receives a first-line indication for NSCLC, most patients will be candidates for it, because it is nontoxic. But the absence of a predictive biomarker means that $18 billion would be spent on futile therapy, if 33% of patients are cured; if only 15% are cured, this expense for futile therapy would jump to $24 billion, Dr Kolodziej emphasized.

Economics matter, because the total cost of care for a patient with lung cancer approaches $150,000 annually, a cost that is not sustainable. In addition, there is an opportunity cost in that making the wrong treatment decision because of a lack of a biomarker hinders response while the cost of the treatment is still realized. “We have to have a biomarker now,” he said. “It’s the only way we can survive giving this therapy to patients and make sure the right ones get it, because I want the right patient to get this drug.”

Without a biomarker, a potential solution to reimbursement is to institute value-based pricing agreements, in which the reward is tied to actual performance of the drug (Figure).

Figure

In the end, “we have to get comfortable with the fact that value and clinical utility need to catch up with technology, because that’s the only way it benefits our patients,” he said. “We need a lot less hype and a lot more modesty in the claims that we’re making.”