The drug industry’s personalized medicine problem
Patients aren’t getting the targeted therapy they need because they’re not being diagnosed
For decades, drug-industry executives have recited the personalized medicine mantra of: “right drug, right patient, right time.” It’s begun to ring hollow. With dozens of targeted drugs on the market or in late-stage trials, too many patients aren’t getting the therapy they need because they’re not being diagnosed.
And if the proximate cause is that doctors aren’t identifying all the right patients because they’re underutilizing the right tests, drug and diagnostic companies also haven’t done enough to remedy the problem.
Commercially, the challenge is becoming obvious. Targeted therapies approved for specific patient subsets based on biomarker testing appear to be underperforming. Sales of the extraordinary scientific and medical innovation from Amgen Inc. (NASDAQ:AMGN) Lumakras sotorasib, the first oral KRAS inhibitor, have been declining — less than two years after it was launched. Ipsen Group (Euronext:IPN; Pink:IPSEY) picked up Epizyme Inc. (NASDAQ:EPZM) and its targeted cancer drug Tazverik tazemetostat for $247 million, a huge comedown for a company valued barely two years before at nearly 20 times that much. Likewise disappointing — at least relative to its partner’s expectations — were sales for Blueprint Medicines Corp. (NASDAQ:BPMC) of its RET fusion drug Gavreto pralsetinib: Roche (SIX:ROG; OTCQX:RHHBY) backed out of their collaboration following hundreds of millions of dollars in collaboration funding.
Just how big a problem is this? Plenty. In non-small-cell lung cancer (NSCLC), where testing for biomarkers is most routine among cancers, 25-35% of patients with actionable mutations — that is, patients who would benefit from available, targeted therapies — aren’t getting them, according to an analysis by Diaceutics plc in JCO Precision Oncology.
No surprise: the rate of biomarker testing in NSCLC is 49% below guideline recommendations. In other cancers, the figures are much worse – 60% below guidelines recommendations in ovarian cancer, for example, based on retrospective reviews of electronic medical records in oncology practices in New Jersey and Maryland.
There are solutions. Some are governmental, though those will be slow in developing. More importantly are collaboration opportunities — unusual, testing the boundaries of most biopharmas’ comfort zones — but we believe the fastest and most direct route to patient benefits.
What’s behind the diagnostic underutilization
There are several reasons for this diagnostic underutilization: the rapid introductions of new biomarkers and therapies; the complexity of testing options and breadth of results; and reimbursement barriers.
The first two are largely challenges of education. Up to 80% of oncology patients in the U.S. are treated in community practices, where any one physician must treat multiple kinds of cancers, both solid and liquid tumors. It’s a constant challenge to keep up-to-date on the latest approvals and guidelines for all cancer types, let alone keep up with clinical trial results which haven’t yet made it into the compendia. The National Comprehensive Cancer Network (NCCN) alone updates its 84 guidelines at least once a year and, in some cases more than four times per year.
And then there’s the bewildering variety of testing possibilities. The biomarker testing recommendations have quickly evolved from just a few biomarkers for NSCLC, such as EGFR and ALK, to more than 10 genes with variants associated with drugs recommended in NCCN guidelines. Moreover, FDA has approved seven drugs for tissue-agnostic indications based on five different biomarkers. Testing for these tissue-agnostic drugs either requires or may benefit from more advanced testing chemistry — for example, comprehensive genomic profiling (CGP) and RNA plus DNA-based assays — than the targeted panels that many oncologists are used to ordering, and payers to reimbursing.
Even if the oncologist uses the right tests, the results can be confusing. For instance, she might find a known and targetable mutation — but in a tumor type not included on the targeted therapy label. Or the biomarker might require investigating clinical trial eligibility. The oncologist, and her patient, would certainly benefit from some sage advice — say from a cross-functional genomic tumor board or an expert interpretation service. But many oncologists don’t use them because they’re not reimbursed.
Payers have raised obstacles too. Medicare pays for comprehensive next-generation sequencing (NGS) for patients with advanced cancer at diagnosis. But commercial payers have been slower to adopt coverage, despite the fact that most patients do not get tested for all recommended biomarkers and treatment with a targeted therapy will provide better outcomes. The downstream implication: less wasted drug.
Only 22% of lives covered by commercial payers benefit from access to CGP at the time of advanced cancer diagnosis, where clinicians can best use the CGP information to guide therapy planning. The majority of commercial payers have restrictions in their medical policies that limit CGP access only to patients who have relapsed (53% of commercially covered lives) or with a diagnosis limited to NSCLC (5%). In fact, 20% of commercially insured lives don’t have covered access to CGP at all, only to small targeted panels or single gene tests.
Then there are some interesting, and unfortunate, diagnostic-therapeutic catch-22s: some insurers will cover a mutation-specific drug but only when confirmed by a biomarker test — which it may not pay for. Take, for example, approval for Keytruda pembrolizumab from Merck & Co. Inc. (NYSE:MRK) in unresectable or metastatic tumors with high tumor mutation burden (TMB). To figure out whether the drug might help a patient, an oncologist needs to test for TMB, which requires a CGP test. While commercial payers generally cover Keytruda for patients with high TMB, some don’t cover CGP, which means it usually doesn’t get done, and thus patients lose the chance to receive a highly effective regimen.
We’ve focused on cancer testing because that’s where most of the action is today. But not tomorrow.
There are more than 200 active pharma-sponsored trials in ClinicalTrials.gov in non-oncology indications in which genetic testing is part of the inclusion criteria — from the PCSK9 gene therapy from Verve Therapeutics Inc. in hypercholesterolemia to RLYB211 from Rallybio Corp. in fetal and neonatal alloimmune thrombocytopenia (FNAIT) to therapies from Pfizer Inc. (NYSE:PFE), Novartis AG (SIX:NOVN; NYSE:NVS), AstraZeneca plc (LSE:AZN; NASDAQ:AZN), Johnson & Johnson (NYSE:JNJ) and Ionis Pharmaceuticals Inc. (NASDAQ:IONS) in a variety of other categories. Should such drugs come to market, biomarker testing will either be required per the label or payers may require testing before reimbursing the cost of the therapies.
The solutions are (mostly) with biopharma
If the dream of personalized medicine makes both clinical and economic sense, it is clear that we will need to find solutions to the diagnostic dilemma.
Let’s start with policy. A handful of U.S. state governments actually mandate coverage of molecular diagnostics required for determining pharmacotherapy. But they don’t cover the necessary support mechanisms for clinical implementations, particularly genomic tumor board and interpretation services, or NGS-based testing in categories outside of cancer.
And many payers are still not abiding by these state mandates. Federal and state policy can and should mandate that when the FDA approves a therapy with a companion diagnostic biomarker, payers must also cover the biomarker test. Likewise, Medicaid benefits such as Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) must allow for guideline-recommended, medically necessary molecular diagnostic testing.
Other policy approaches can involve private-public programs. In Australia, for example, a public-private partnership led by the Australian Genomic Cancer Medicine Centre will provide tumor genomic profiling and access to approved and investigational therapies for more than 20,000 cancer patients, many with rare and challenging tumors. But the U.S. federal government, with its checkerboard reimbursement system, has been painfully slow to move.
In the interim, if change is going to happen, drug and diagnostic companies will have to drive it. And that will require both creativity and taking some risks. For example, Eli Lilly and Co. (NYSE:LLY), which markets Retevmo selpercatinib for RET fusion positive advanced cancers, took a small risk by agreeing to make available at no cost Tempus xT, the CGP broad-panel genomic sequencing assay from Tempus Labs Inc., for virtually any advanced NSCLC patients who haven’t had previous comprehensive biomarker testing.
Higher-risk, higher-reward: a coalition approach in which a number of drug firms, along with relevant diagnostics players, create a broader access program for advanced cancer patients that guarantees diagnostic funding if unavailable from a payer, as well as the interpretation services that might be required. Costs are shared, oncologist education supported, and more patients get the right drug at the right time.
Moreover, the outcomes data from the program provides additional evidence supporting the argument that routine comprehensive genomic profiling makes business and clinical sense.
There could even be a risk-sharing component with health plans: the consortium can agree to reimburse all or a portion of the testing if care hasn’t improved on any one of several easy-to-measure metrics.
We’ve already seen a few such models involving payers and diagnostic companies — for example, the arrangement Illumina Inc. (NASDAQ:ILMN) has with Point32Health on non-invasive prenatal testing (NIPT) and whole genome sequencing (WGS) in pediatrics. Illumina agreed to cover any increase in Point32’s non-invasive testing costs over what they were before they covered NIPT for average-risk pregnancies. The results: with a negligible increase in overall spend, Point32 saw dramatically expanded access, while the diagnostics industry could benefit from expanded access to a guideline-recommended screening approach. Together they created, and published, the results from a real-world dataset that was instantly translatable to other health plans. NIPT for average risk pregnancies is now almost always covered by U.S. payers.
With the risk-sharing structure for CGP testing, there could be direct cost-savings for payers. Identifying patients that are eligible for targeted therapy may reduce the financial burden of less effective and more expensive therapeutic options routinely offered in first-line advanced cancers. And if a test indicates utility of a drug for an indication that has yet to be approved, the patient could be put into a clinical trial — with the drug company carrying the therapeutic cost, not the payer.
But what we know now is that drug companies keep coming up with valuable targeted therapies that don’t get used when they should. The way to fix the situation may not be the way well-trodden, but the current path is simply an exercise in futility. Risk-sharing can work, we’ve seen it. And if it’s naïve to believe it will solve the entire problem, it is nonetheless the fastest road companies can take on their own towards the goal of ensuring that all patients, not just the lucky few, can benefit from our industry’s enormous innovations in personalized medicine.
Brock Schroeder is VP, market access at Illumina Inc., and Roger Longman is founder and chairman of Real Endpoints, a boutique market access consultancy.
Signed commentaries do not necessarily reflect the views of BioCentury.