Cancer Patients Missing Personalized Therapy Due to Quality Issues for Genetic Tests

NEW YORK (GenomeWeb) – Incorrect, inconclusive, and delayed test results for 13 cancer-linked biomarkers are hindering around 78,000 patients per year from receiving personalized, targeted treatments, a recent analysis found.

These missed opportunities mean that thousands of patients, who could have gotten drugs, such as Herceptin, Tarceva, and Xalkori, that target molecular drivers of their tumors — and which have been shown to slow cancer progression, improve quality of life, and even extend survival ― likely ended up getting more toxic chemotherapies. Additionally, the data from consulting-turned-medical technologies firm Diaceutics suggest that suboptimal testing is costing the sponsors of these drugs.

Because of inadequate investments to ensure faster turnaround times, lower false-positive rates, and sufficient sample quality around a dozen biomarker tests, pharma loses $8.3 billion in annual drug revenue in the US, and double that amount when five leading European markets (UK, France, Germany, Italy, and Spain) are added to the mix. With these data, Diaceutics is trying to drive home to its pharma clients that they can no longer treat the diagnostics that help identify which patients will benefit from their drugs as an afterthought.

"So, does testing matter? Does this prequel to the treatment matter?" posited Diaceutics CEO Peter Keeling at a Financial Times life sciences summit last week. "Our suggestion is, yes it does. It has a real economic and real patient impact, and it deserves an awful lot more focus."

He hopes this latest data will underscore the point to top level executives at pharma companies. "We're trying to convey to the C-suite that there is an economic benefit to investing in diagnostics, but there is a human imperative to go back and look at these programs and not just assume they are optimized at launch," Keeling said in an interview.

Diaceutics arrived at its estimates based on an analysis of claims data and real-time information from some 250 high-volume US and European labs that test for markers such as HER2 and PD-L1 overexpression, EGFR mutations, and ALK rearrangements. The company identified trends for the 13 biomarkers with analytics and further corroborated the findings with specific partner labs, and against what's been published in the literature about the major quality and process issues labs are facing.

Diaceutics has played a role in some 200 Rx/Dx commercialization projects, helping drug companies integrate diagnostics into their personalized medicine development pipelines. Through these projects, Keeling said that his team has tried to impress on pharma clients that personalizing a drug with the help of a diagnostic doesn't have to mean smaller market share, and that these tests, when integrated appropriately into a drug program, can increase market share.

To make this case, Keeling noted that in 2015, 36 of the 50 most profitable drugs had their labels updated with pharmacogenomics information, or are enabled by diagnostics, and together have amassed $140 billion in revenues. In the coming years, the trajectory will shift toward even greater integration of diagnostics in drug development, with an estimated 75 percent of future therapy launches having a diagnostic component from 2020 onwards.

In anticipation of that, pharma must pay more attention to quality issues related to diagnostics, advance different types of tests that get their drugs to the right patients, and educate physicians and patient advocacy organizations about testing. This responsibility must fall on pharma's shoulders, Keeling insisted in an interview, because diagnostic developers are cash strapped and, to date, have been unable to adequately drive adoption of a companion test to support uptake of a drug.

A pharmaceutical company can spend between $150 million to $200 million to launch a targeted drug, requiring investments for peer-reviewed publications, direct-to-consumer advertising, as well as physician and payor education. If the therapy relies on a test to identify the best-responder population, then the drug company can expect to spend an additional $3 million to $12 million per biomarker to engage a diagnostic developer, on getting labs to perform the test, and on raising physicians' awareness of the test, Diaceutics estimated.

However, this diagnostic investment is often an afterthought and unbudgeted. "Having worked with multiple therapy teams, multiple assets, we understand this is very complex for [pharma]. They are very focused on launching the drug, and to try and understand the diagnostic ecosystem is tough," Keeling said at the FT summit.

Within drug/diagnostic codevelopment programs, pharma companies are usually focused on aligning the development timelines for the test and therapy, so they are approved and launched together. But there is little attention paid to anticipating and accounting for the issues that are hindering adoption of the test over the long term, such as poor sample quality or long turnaround times. "So, no one is really accountable for making sure that all the patients at the right time are getting tested," Keeling said.

Despite the simultaneous launch of a targeted drug and diagnostic, the adoption of the test often lags behind the therapy. Eight years after the launch of a biomarker test, only 50 percent of the intended patient population receives it, Diaceutics estimated. In the present analysis using utilization data, Diaceutics didn't evaluate the economic impact of inadequate adoption of these 13 biomarker tests. Using other reference points, the company estimated that including lackluster test adoption would amount to $32 billion in annual lost revenues for pharma in the US and in Europe.

Other studies have also shown that patients often don't get a targeted drug because test results didn't come back on time, or the biomarker information wasn't incorporated into electronic medical records. These and other issues may be leading physicians to avoid prescribing targeted drugs altogether and opt for chemotherapy or newer immunotherapy options, recent survey data suggests.

These diagnostic challenges aren't unfamiliar to the drug industry. The uptake of the first personalized drug that required the aid of a diagnostic to identify the best responder population, Herceptin (trastuzumab), was hampered by a suboptimal testing strategy. The drug was approved in 1998 alongside Dako's HercepTest, but then Genentech invested in launching a number of companion tests that employed different technology platforms. The availability and use of multiple platforms revealed discordance, where one lab would find a sample was HER2-negative, but another would report it was HER2-positive.

At one point, studies pointed to discordance rates of around 25 percent between local and central labs in determining HER2 status. Since then, through Genentech's efforts and with the help of testing guidelines, the discordance rate has been reduced to around 4 percent, according to one study. Still, in the early days of Herceptin's availability, Diaceutics estimated that Genentech and its parent company Roche had missed out on $3 billion to $5 billion in Herceptin revenues due to test access and quality issues. 

"We went to publish that data, and prior to doing that, we checked with our colleagues at Roche and we were told that our numbers were wrong," Keeling recalled at the FT meeting. "They had done a similar analysis and their number was closer to $7 billion."

Fast forward a decade and some of the same issues around diagnostic access and quality are plaguing the latest oncology drugs on the market. In less than three years, several new immunotherapy drugs ― Keytruda (pembrolizumab), Opdivo (nivolumab), and Tecentriq (atezolizumab) ― have come to market for various indications in 17 approvals. In a handful of indications, FDA has approved these drugs with a companion or a complementary diagnostic to identify patients' PD-L1 expression.

These tests have varying cutoffs and use different antibodies, making comparisons challenging. Moreover, the marker PD-L1 is best used as an indicator of the degree to which a patient might benefit from immunotherapy. The inability of the biomarker to clearly distinguish who will respond; nuanced FDA labeling about the need for PD-L1 testing ahead of drug administration in some cases but not others; and different marketing messages from drug sponsors about the role of testing, have confused doctors about whether to test patients at all and which test to use.

Diaceutics' data suggests that 120 labs in the US are offering PD-L1 testing. But Keeling pointed out that because these labs started to offer testing once the drugs entered the market, the oncology community is only now debating the value of the PD-L1 testing. The International Association for the Study of Lung Cancer recently released a report outlining areas of clarity and debate in PD-L1 testing, but the group also recognized that healthcare providers don't have a clear understanding of which PD-L1 assay to use and if the tests are interchangeable.

This uncertainty is also resulting in missed opportunities to provide optimal patient care and in lost revenues. Just in second-line non-small cell lung cancer, Diaceutics estimated last year that the mistakes in managing PD-L1 testing integration have resulted in $40 million in lost treatment opportunities, and could balloon to $744 million over five years. 

"So, here we are traveling the same journey that we've traveled with HER2 and Herceptin," Keeling said. "Repeating some of the same mistakes."

This is in large part due to the fact that drug companies don't feel it's their responsibility to manage diagnostic-related issues, because that's what they're paying the diagnostic developer to do. "One of the big things we hear from our clients is, 'We're not a diagnostic company,' and our response is, 'No you're not, but sadly, you just put your arms around a diagnostic ecosystem you really need to manage,'" Keeling said.

Moving forward, Diaceutics imagines pharma will continue to work with diagnostic companies to develop tests, but increasingly seek partnerships with labs and organizations that can better support other aspects the diagnostic ecosystem, from test quality to education. "I see a switch. I see pharma has really woken up to the limitations of the diagnostic industry as a commercial partner," Keeling said.

Although pharma will still work with diagnostics firms for test development, he sees drug developers working with other groups to create "a community around test adoption." Diaceutics' analytics have helped pharma clients see the costs of not becoming more engaged in the diagnostics ecosystem, but Keeling sees many other uses for this type of data and is planning to delve into specific areas of the diagnostics industry, such as circulating tumor DNA testing; genetic testing in non-oncology indications, including cardiovascular and liver disease; and next-generation sequencing.

This type of data can also be useful in a healthcare system moving away from volume-based care and toward value. For example, the Medicare Access and CHIP Reauthorization Act will replace a lot of fee-for-service-based payments for doctors and start to pay them based on quality of care and efficiency. The drug industry, too, is embracing value-based pricing by entering into risk-sharing agreements with insurers for their products.

"I think we all need to get more granular about what's happening and where value is being added and not," Keeling said.

NPR reports that Turkish high school students will no longer study evolution.

Researchers report they sequenced and identified plant species in an "al fresco" laboratory.

An Australian team searches for genetic alterations linked to depression in hopes of developing personalized treatments, the Sydney Morning Herald reports.

In PNAS this week: host contributors to typhoid fever risk, effects of obesity-related variants near TMEM18, and more.