Was Richard Smith wrong to call precision medicine a fantasy?
Earlier this year I spoke in a debate at the Cambridge Union against the chief executive of AstraZeneca and the motion that “This house needs new drugs,” and I threw out the comment that “personalised or precision medicine is a fantasy.” Several doctors had said that to me. After the debate Ruth March, head of precision medicine and genomics at AstraZeneca, told me that I was wrong; she invited me to visit her at the company in Cambridge so she and her team could show me that I’m wrong. I went last month. Was I wrong?
As I travelled to Cambridge, two other experiences in that booming city went through my mind. Firstly, at the Cambridge Union debate one of the speakers, a pension fund manager, had described how producing a new drug costs roughly a hundred times what it did in the heyday of the 1960s. How could that be sustainable? Secondly, on another visit I’d spoken at the launch of a book Medical Nihilism, by a philosopher, in which he argues that “the medical model or theory of targeting diseases with magic bullets is unhelpful.” I was on my way to visit merchants of magic bullets.
After almost circumnavigating the large science park, which is owned by Trinity College, I arrived at one of AstraZeneca’s seven buildings in and around Cambridge. Soon all the employees will move to a new building, a Biovarium, alongside Addenbrooke’s Hospital, Papworth Hospital, Cancer Research UK, and the Medical Research Council Laboratory of Molecular Biology, where Francis Crick and James Watson first described the double helix of DNA. The move is an expression of AstraZeneca’s new culture and philosophy, bringing the company closer to science, academia, and discovery.
The reinvention and five Rs of Astra Zeneca
Catherine Priestley, director of executive communications and science engagement, explained to me how AstraZeneca, along with some other pharmaceutical companies, has been through a transformation. In 2010 the company had few new drugs in its pipeline and concentrated on drugs that would sell in high volume, many of which were “me-toos.” I remember visiting Glaxo’s research laboratories years ago and learning how they produced thousands of molecules that they then tested in animal and other models in the hope of getting a “hit,” a new drug.
Priestley explained how Astra Zeneca’s new philosophy is built around “the five Rs”: the right target, the right tissue, the right safety, the right patient, and the right commercial potential. This, as I see it, is the opposite of churning out molecules in the hope that one will be useful.
The right target means having a biological rationale behind a new drug. Often these drugs will not be “small molecules,” as are most drugs familiar to today’s doctors, but use “new modalities” (for example, like modified mRNA, antisense oligonucleotides, and bicyclic peptides) to reach targets that were considered “undruggable.” I was hearing new jargon in Cambridge.
The right tissue has led to a concentration on cancer, respiratory disease, and infections where the basic biology is better understood than, for example, neurological and mental health diseases (although there is a small “virtual group” working in neuroscience). The better the basic biology is understood, the more likely it will be possible to produce a useful drug. There is also work on cardiovascular, renal, and metabolic diseases.
The emphasis on right safety follows from the right target and tissue with better understanding of how the drug is working, but it has also led to the development of better methods of determining safety preclinically.
The old model of drug development and treatment led to huge numbers of patients being treated with drugs like antihypertensives, statins, and antidepressants, with many patients not benefiting directly, but the right patient means identifying those patients who have a much higher probability of benefiting directly. This has led to the company spending a lot of resource examining advanced diagnostic methods and partnering with companies that have methods that can identify patients most likely to benefit—methods that are better, cheaper, or quicker than current methods. AstraZeneca is not itself a diagnostic company and has no plans to become one.
The right commercial potential means thinking of the commercial possibilities from early in drug development, and by the time a drug is considered for a phase III trial there must be “clarity around the patient population, the unmet medical need, differentiation versus standard of care, payer criteria for global reimbursement, competitive environment and sales projections.”
Cultural change and collaboration
A crucial part of transforming AstraZeneca has been changing its culture to one that is much more science based, more “truth seeking,” and collaborative. Building its new headquarters in Cambridge close to Addenbrooke’s and the other science based institutions is part of this—as, indeed, is partnering with the university and sponsoring the debate in which I took part.
The traditional barrier between academia and industry is rapidly disappearing (to the concern of some and the delight of others), with the company sponsoring scientists doing PhDs and postdoc programmes and scientists moving backwards and forwards between academia and industry. I met James Hadfield, director of Precision Medicine Laboratories, who is a specialist in next generation gene sequencing and worked for more than a decade in academia. Mark Fidock, head of Precision Medicine Laboratories, who accompanied me on my tour, is an honorary senior lecturer at Cambridge University.
The company also works closely with Glasgow University and Imperial College and other universities, particularly the Karolinska Institute in Sweden. (AstraZeneca is an Anglo-Swedish company and has a plane that flies every week day from Cambridge to Gothenburg, illustrating the importance of scientific collaboration.) Altogether the company has over a thousand collaborations across the globe.
The company also has a scheme that allows external researchers to plug into the company’s systems. By the end of 2017 the company’s Open Innovation Initiative had reviewed more than 500 proposals for new drugs, and of those 26 had progressed to clinical trials and 150 are in the preclinical stage.
Precision medicine depends very much on advanced diagnostic methods, and I was shown a series of methods being tested. Helen Brown, associate principal scientist and a specialist in molecular diagnostics, showed me a machine that could identify particular genes from tiny slices of tissue. The machine can be used by a technician and avoids the need for a molecular biologist, meaning that the test can be done quickly and more cheaply. Many of AstraZeneca’s new drugs depend on identifying particular genes.
Bert Rutten, a senior scientist, showed me a point of care machine that could be used in primary care to measure eosinophils in seconds rather than days to diagnose eosinophilic asthma (about 5% of cases of asthma), for which the company has a drug. Another machine could measure uric acid from a spot of blood at home or in a GP’s surgery. Monitoring disease and drugs is an important part of precision medicine, and the company wants to help common conditions like asthma and gout be managed in primary care.
The company is moving into artificial intelligence and machine learning, working with Amazon and others, and Craig Barker, head of tissue diagnostics, and Michel Vandenberghe, an AstraZeneca postdoc, showed me how they are using artificial intelligence to help with histological diagnosis. Such methods are unlikely to replace pathologists, but might be used to diagnose the cases where the diagnosis is obvious and supplement the analysis of the pathologist with difficult cases.
Hadfield showed me a gene sequencer about the size of a mobile phone that he has helped develop. It’s currently the world’s smallest sequencer, costs about £1000, and could potentially sequence the whole human genome. It is more likely, however, to be used to sequence particular gene sequences.
AstraZeneca, although an Anglo-Swedish company, is really a global company, and—recognising that Rutten is Dutch, Vandenberghe like the chief executive Pascal Soriot French, and Mene Pangalos, the head of innovative medicines and early development and global business development who has led the turnaround of drug discovery, is a Brit of Greek origin—I wondered if Brexit might present a problem for the company, which depends heavily on being able to attract global talent.
Is all this transformation and collaboration working?
One measure of whether this transformation is working is the proportion of molecules moving from preclinical investigation to completion of phase III trials, and for AstraZeneca this has increased from 4% in 2005-10 to 19% in 2012-16. The company was below the industry average of 6% in 2005-2010 but is above the industry average of only 4% from 2003-15. (That is not, I recognise, comparing like with like, but it does show a substantial increase within AstraZeneca).
My response to March at the Cambridge debate was that I feared that even though there might be precise drugs they benefited only small numbers of patients at high cost. March emphasised to me that the cost of testing precise drugs is less than that of traditional drugs in that they didn’t need such big trials. Soriot also emphasised in his speech at the Cambridge debate that drugs are relatively quickly off patent, which brings the cost down dramatically.
It is, I recognise, still early days, but the company has used its transformed methods to produce osimertinib, olaparib, acalabrutinib, ticagrelor, lokelma, and avibactam. Osimertinib, a drug used in advanced EGFR-T790M non-small cell lung cancer, has a cost per QALY over usual treatment of about £40 000—above the usual threshold of the National Institute of Health and Care Excellence (NICE), although within the £50 000 threshold linked to NICE’s “end of life” criteria.
Olaparib is a PARP inhibitor used in ovarian cancer patients with BRCA1 or BRCA2 mutations (it is also being developed in breast cancer and prostate cancer) The treatment is approved for breast cancer in the US, is under regulatory review for breast cancer in Europe, and is in trials for prostate cancer. Used as maintenance treatment in advanced ovarian cancer, the drug’s incremental cost-effectiveness ratio (ICER) is around $100 000. This calculation came from an American study and uses different methods from those used by NICE to make judgments.
I couldn’t find data on the cost effectiveness of all the drugs, and without access to a medical library I couldn’t access some of those I did find, but unsurprisingly these drugs are mostly very expensive. They tend to be used in the later or more severe stages of disease because that’s what regulators require with new drugs. As the drugs show efficacy they can be used earlier in the disease, and osimertinib is now approved for first line treatment in EGFR lung cancer. Perhaps as they are used earlier they might be more cost effective, and potentially they will become much more cost effective when off patent. It might also be that as methods of developing precision drugs advance they could become cheaper, although producing them is always likely to be expensive with the dependence on highly qualified (and highly paid) staff.
I was wrong to argue that precision medicine is a fantasy in that increasing numbers of drugs are being produced using the methods of precision medicine. At the moment, however, the drugs tend to be very expensive and used only in patients with severe disease. Perhaps that will change as more drugs come through and some come off patent.
Richard Smith was the editor of The BMJ until 2004.
Competing interest: I’m grateful to all the people at AstraZeneca, who were charming, gave me time, and answered all my questions. I paid my own travel expenses and was treated by the company to one cup of coffee, which I’m confident (perhaps wrongly) has not corrupted me.