The development of more targeted therapies is moving apace, but are the payers prepared for it? Dr Sarah Houlton looks at the progress of personalised medicines
Genomic biomarkers help to predict the efficacy and safety of treatments in individuals. Qiagen markets therascreen assays in Europe for various biomarkers including KRAS, EGFR, NRAS, BRAF, PI3K, JAK2, MGMT and UGT1A.
Picture courtesy of Qiagen
A growing number of drugs, particularly in the cancer field, are designed to treat specific genetic mutations. This means that they won’t work for everyone – but it does also mean that for those patients with the correct mutation, the chances of successful therapy are much higher.
To identify these potential responders, a screening test is required. Unsurprisingly, with the rise in personalised medicines has followed an increase in the number of companion diagnostic tests designed to pinpoint those patients who have the correct genotype or phenotype. Laboratory developed tests do not require FDA approval, as they are run in the US under the auspices of the clinical laboratory improvement amendments, or CLIA, but tests sold as kits for use in labs that do not fall under these requirements have to follow a formal licensing procedure.
Although the companion diagnostics, or CDx, market remains niche, it is growing. A recent report from Research and Markets estimates the current size of the global CDx market at about US$200m, which is dwarfed by the overall size of the diagnostics market at about $50bn. The report predicts that the market size will spiral upwards in the coming years as more CDx products and associated personalised medicines are approved, to an estimated $900m in 2018. That represents a compound annual growth rate of 29% between 2012 and 2018.
In 2011 more than 40% of all new drugs approved by the FDA had an associated pharmacogenomics biomarker
The number of approved drugs with companion tests is growing at a pace, too. For example, in 2011 more than 40% of all new drugs approved by the FDA had an associated pharmacogenomics biomarker. And in 2012, drugs with biomarkers achieved total sales of more than $20bn, with six of them being big-selling blockbusters. It is estimated that about a third of all new drugs in the clinic have some form of genomic or proteomic marker associated with them, and therefore might benefit from a licensed companion diagnostic, even if one is not already available.
However, if a drug is to be promoted alongside a companion diagnostic, current FDA rules insist that the CDx must also be approved by the agency. So the tests associated with the drugs gefitinib (Iressa, AstraZeneca/Teva) and erlotinib (Tarceva/Roche, OSI Pharmaceuticals) have had lower market acceptance than the companies might have liked. This is in no small part because they could not promote the diagnostic alongside the drug as it had not been FDA approved. A CDx for erlotinib has now been approved.
Clearly, starting work on a companion diagnostic early on in a drug’s development process, and applying for approval in good time, will greatly increase its chances of commercial success. The other decision that needs to be made is whether an approved CDx product is the appropriate route to go down; it may be that providing testing on a service basis, or via a platform-based test used in a handful of labs, is more appropriate.
Probably the best known of all the drugs that target specific mutations is Roche’s trastuzumab (Herceptin). The monoclonal antibody drug targets the HER2 receptors, which are overexpressed in certain forms of breast cancer and set off uncontrollable cell proliferation. This genetic form of the disease accounts for 20–30% of all early stage breast cancers.
With the widespread uptake of trastuzumab in breast cancer treatment, it is perhaps unsurprising that numerous different commercial test kits have been approved by the FDA for the identification of HER2+ cancers – more than half of all companion diagnostics currently on the FDA approved list are for trastuzumab. Some of these are also approved for ado-trastuzumab emtansine (Kadcyla), the antibody–drug conjugate based on trastuzumab that was approved last year, and pertuzumab (Perjeta) from Genentech, licensed for use in HER2+ breast cancer in combination with trastuzumab and docetaxel, which acts by preventing the dimerisation of HER receptors.
Performing a DNA sequencing experiment (Genome Sequencer FLX): an automated, ultra-fast tool for DNA sequencing based on innovative microtechnology methods.
Picture courtesy of Roche
Immunohistochemistry (IHC) is the most traditional technique used in these kits, which include Danish company Dako’s HercepTest, Insite from Biogenix Laboratories, and Pathway from Ventana. Both identify through staining whether the cells overexpress HER2 receptors. A strong staining response indicates that at least 10% of the cells overexpress the receptors, and each of these has somewhere in the order of two million receptors.
IHC tests like these have limitations, however, not least because the results require interpretation by the scientist running the test, rather than giving a precise, quantitative number as a result. More recently, tests that use fluorescent in situ hybridisation, or FISH, have been approved, but they are much more expensive to run, not least because an image capture system and fluorescence microscopy capabilities are required. These include Inform from Ventana, and Pathvysion from Abbott Molecular.
IHC tests have limitations, not least because the results require interpretation by the scientist running the test
A third test protocol involves chromogenic in situ hybridisation, or CISH. While this protocol does not need expensive equipment, the kits themselves are much more costly. They allow the HER2 gene status of the cells to be determined quantitatively after chromogenic staining using standard light microscopy. These include Spot-Light from Life Technologies, and Dako’s HER2 CISH PharmDx kit.
Another well-known drug with a companion diagnostic is imatinib mesylate (Glivec/Gleevec) from Novartis. This small molecule drug was specifically designed to block activity of the bcr-abl fusion protein in cancers with the Philadelphia chromosome mutation, notably chronic myeloid leukaemia. While about 95% of all cases of CML are Philadelphia positive and thus it is much more likely than not to have activity in these patients, the drug is also active in gastro-intestinal stromal tumours with a different mutation, to the KIT gene. The FDA has licensed a test kit from Dako to differentiate KIT-positive GIST tumours. This IHC based kit identifies c-kit protein/CD117 antigen positive cells that might be susceptible to treatment with imatinib.
Two more cancer drugs with a companion diagnostic are cetuximab (Erbitux) and panitumumab (Vectibix). These antibodies inhibit the epidermal growth factor receptor, or EGFR, which is expressed in about three-quarters of all patients with metastatic colon cancer. Dako’s IHC kit can pinpoint cancer cells that express EGFR. However, there is some evidence that this alone is not actually a good predictor of clinical response.
The therascreen kit from Qiagen has also been approved for use with cetuximab. This PCR-based testing protocol identifies seven mutations to the KRAS gene, which encodes a protein on the EGFR pathway. Cetuximab will only work on cancer cells that do not have a mutation to KRAS, so if no mutations to KRAS are identified, cetuximab treatment can be considered.
A companion diagnostic was pivotal in Pfizer’s crizotinib (Xalkori) reaching the market just four years after the discovery process was initiated
A companion diagnostic was pivotal in Pfizer’s crizotinib (Xalkori) reaching the market just four years after the discovery process was initiated. Designed to treat non-small cell lung cancers, this small molecule kinase inhibitor works in just 4% of NSCLC patients, those who have a chromosomal rearrangement that leads to the formation of a fusion gene between ALK and EML4. Patients with this mutation are most likely to be younger than the average lung cancer patients, and to be non-smokers. With such a small proportion of patients likely to respond, being able to identify that subset which is potentially susceptible to treatment was essential for successful clinical trials. Abbott Molecular’s Vysis ALK Break Apart FISH probe kit provided this capability, providing a way to stratify the patients for eligibility for the trial. Now that the drug is approved, it is used to determine whether a NSCLC patient has the mutation that means they may respond to the treatment.
A diagnostic also proved pivotal in the development of vemurafenib (Zelboraf), Genentech and Daiichi Sankyo’s treatment for melanoma. The drug inhibits the B-Raf enzyme if it has the V600E mutation, where the valine at position 600 in the enzyme’s sequence has been replaced with glutamic acid. This mutation exists in about 60% of melanomas, so just over half of patients might be expected to respond.
The ability to identify patients with the V600E mutation is not just a case of spotting responders, however. If they don’t have V600E (or the rarer V600K mutation, where the valine is replaced by lysine), not only will it not work, it appears that it has a positive effect on normal BRAF and causes the tumour to grow more rapidly. Avoiding such patients is, clearly, essential.
Roche Molecular Systems provides the necessary test, in the form of the COBAS 4800 BRAF V600 mutation test, a qualitative real-time PFR test using Roche’s Cobas PCR system. However, McKinsey estimates about 45% of BRAF tests are still carried out via the lab-developed test route using sequencing technology, despite the availability of an approved diagnostic.
Cobas 6000 analysis module (for clinical chemistry parameters)
Picture courtesy of Roche
While cancer still has the lion’s share of the market – according to Research and Markets, in 2012 this was 96% – there are personalised medicines for non-cancer indications, and more are in development. According to McKinsey’s 2013 report, Personalised medicine, the path forward, in the next three to five years immunology and transplantation, paediatric and pre-natal, and infectious diseases are likely to represent the next frontier of personalised medicine. At a much earlier stage of development – though with much potential – are personalised medicines for CNS and cardiovascular indications.
Of course, the diagnostics for these disease areas will probably look very different. Immunology, for example, involves a complex interaction between a variety of factors such as interleukins, and McKinsey suggests that panel-based testing with a complex scoring system is most likely to prove successful. Far more straightforward would be the ability to identify early via genetic means those conditions such as Alzheimer’s disease that manifest themselves later in life.
Only one of the FDA-approved companion diagnostics is licensed for a non-cancer indication
Although various personalised medicines are available with laboratory developed tests, only one of the FDA-approved companion diagnostics is licensed for a non-cancer indication. Ferriscan from Resonance Health Analysis Services was developed for use alongside Novartis’s deferasirox (Exjade) in the identification and monitoring of thalassaemia patients being treated with the drug. Thalassaemia is an inherited autosomal recessive blood disorder caused by mutated or missing genes that reduce the amount of haemoglobin that is made, thus resulting in fewer red blood cells and, therefore, anaemia.
When given repeated blood transfusions, patients can develop chronic iron overload, and the drug chelates free iron, allowing it to be excreted. The companion diagnostic enables liver iron concentration to be measured without invasive techniques such as liver biopsy being required. Instead, it relies on MRI-based technology to produce a map of liver iron, and calculate a mean liver iron concentration measurement.
Now that PCR-based tests are starting to be approved, one might imagine that there is a good deal of potential in applying the new breed of next generation sequencing (NGS) techniques to diagnostics. Since the first genomes were sequenced in the early 2000s, companies such as Life Technologies and Illumina have greatly reduced the cost and speed at which genomes can be sequenced, and in the not-too-distant future the ability to sequence whole genomes will become cheap and routine.
However, the usefulness of bare genetic sequence data may be limited, not least because of the ability to ‘hit’ genetic targets with drug therapies, or whether mutations have any clinical relevance in the first place. Just because there is an identifiable mutation, does not mean anything can be done about it.
As with the current breed of companion diagnostics, its application in cancer is most obvious. McKinsey estimates that by 2018, NGS will be used for 50–70% of cases of lung cancer, 40–60% of colorectal cancers, 40–50% of breast cancers, and maybe 25–35% of prostate cancers. The markers likely to be tested include driver mutations and variations, drug metabolism markers, resistance markers and drug safety markers. This would put the potential market just in sequencing newly diagnosed patients north of $2bn by 2018. It could also cut clinic costs by removing the need to do individual tests for each potential mutation.
Predicting side-effects may prove extremely important in future
Indeed, predicting side-effects may prove extremely important in future. The V600E mutation test for regorafinib highlights the potential; while this is an extreme case where the wrong genetic type can accelerate cancer growth, there are many drugs that cause side-effects in subsets of patients. If these negative reactions have a genetic basis, then sequencing could identify, in advance, those who would be best to avoid the treatment, or to forewarn others about potential side-effects.
While reimbursement for diagnostic testing is common in the US, this is not the case in many other countries. In the UK, for example, the pharma company is often expected to subsidise the testing process in return for reimbursement of the associated drug; meanwhile, in China, patients usually have to pay for the test themselves. Evidently, this is another factor that a company must take into account when developing a test, and deciding whether to go down the approved CDx route. But with the potential savings to be made by not treating non-responders, the potential market for CDx and personalised medicine is clear.