Perhaps “personalised/precision medicine” will become the norm in years to come, but I think it is timely to remember the words of the great Canadian physician William Osler: “The good physician treats the disease. The great physician treats the patient.” Nevertheless there has been a lot of discussion about “precision medicine” in the printed media recently. The early success of treating Hodgkin disease and childhood leukaemia in the 1960s with cytotoxic drugs (drugs which kill dividing cells) heralded a new age in medicine, and many thought that if similar approaches were taken with other cancers they would be cured. However, combination chemotherapy is relatively toxic as the drugs damage normal as well as tumour cells. Sadly therefore, the majority of cancers remain incurable. The lack of success of this approach might in part be due to the rather unsubtle application of a “one size fits all” approach. Now we are embarking on a different approach known as “precision medicine”.
Many people believe that precision medicine will lead to a fundamental understanding of the complex interplay between genetics, epigenetics, nutrition, environment and clinical presentation, thus facilitating the development of effective prevention and treatment. I will attempt to define precision medicine and contextualise it for non-medical readers. Needless to say the term originated in the USA, where many medical issues are referred to using military terminology. (A good example is “the war on cancer”.) Precision medicine (PM) proposes the customisation of healthcare, with medical decisions, practices, and/or products tailored to an individual patient’s disease. “Collateral damage” should be minimised. Diagnostic testing can be used to select appropriate and optimal therapies based on a patient’s genetic makeup or other molecular or cellular analyses. A similar term, “personalised medicine”, is also in vogue. The National Cancer Institute (NCI), a division of the National Institutes of Health (NIH) in Bethesda, Maryland, defines personalised medicine as
A form of medicine that uses information about a person’s genes, proteins, and environment to prevent, diagnose, and treat disease. In cancer, personalized medicine uses specific information about a person’s tumor to help diagnose, plan treatment, find out how well treatment is working, or make a prognosis. Examples of personalised medicine include using targeted therapies to treat specific types of cancer cells, such as HER2-positive breast cancer cells, or using tumor marker testing to help diagnose cancer.
Although both terms may be subtly different from each other, in practice they are used interchangeably. Therefore where the individual characteristics of a patient are sufficiently distinct, interventions can be concentrated on those who will benefit, sparing expense and side effects to those who will not. Unfortunately only about ten per cent of breast cancers are HER2-positive.
How is this different from the type of medicine we have known for over a hundred years? The treatment of leukaemia and cancer with cytotoxic drugs depends on the observation that these drugs kill more cancer cells than normal cells. However the dose of these drugs is limited because of damage to normal cells known as “toxicity”. The idea behind precision medicine in contrast depends on defining the molecular characteristics of the cancer and developing a drug aimed at this molecular target thereby, it is hoped, limiting damage to normal cells.
When did it all begin? We could take 1953 as the defining moment, as it was the year in which James Watson and Francis Crick published their seminal paper “A structure for deoxyribose nucleic acid” (DNA) in Nature (it is of note that this seminal publication was not peer-reviewed!). Rosalind Franklin’s name was not included in the paper although her photographs of the crystalline structure of DNA were intrinsic to their description of what is commonly called the “double helix”. Many years after Franklin’s premature death in 1968, Watson included an apology for the omission of her name from his and Crick’s 1953 publication in his book, The Double Helix: A Personal Account of the Discovery of the Structure of DNA. Watson revealed, at a biochemistry seminar in Trinity College Dublin in 2013, that Crick and he were greatly influenced by Erwin Schrödinger’s book What is Life? Schrödinger had been invited to Dublin by former president Éamon de Valera to establish the Dublin Institute for Advanced Studies in 1940. During his time in Dublin, Schrödinger gave a series of lectures in Trinity College which morphed into the book What is Life? Watson and Crick’s discovery of the structure of DNA was to lead to the establishment of the era of “molecular medicine”.
Somewhat surprisingly the beginning of “molecular medicine” was the unravelling of a benign disease, sickle cell anaemia, rather than cancer. Vernon Ingram observed that the abnormal haemoglobin in sickle cell disease was due to a single amino acid substitution which alters the solubility of haemoglobin within red cells with catastrophic results for those who inherit the abnormal gene. Many investigators were involved, including Linus Pauling and Max Perutz, in the understanding of this devastating disorder. A number of other inherited disorders such as haemophilia have been well defined at a molecular level but as yet there is no widespread and inexpensive way of providing a cure. However, for others the story is more positive. The prenatal diagnosis of a disease called thalassaemia (an inherited disorder of the production of the haemoglobin molecule) has resulted in the elimination of that disease in some countries.
The development of precision medicine depended entirely on molecular medicine but also detailed knowledge of the structure and function of human genes provided by the “Human Genome Project” (HGP). This international scientific research project involving scientists from the USA, Australia, the UK, Japan, France and China, set out to determine the sequence of nucleotide base pairs of genes that make up human DNA. It further attempted to identify and map all of the genes of the human genome from both a physical and a functional standpoint. The HGP had a long gestation. Following a number of false starts, including a seminar in Cold Spring Harbor Laboratory initiated by James Watson in 1986, the project was supported by inclusion in President Ronald Reagan’s budget in 1988. In 1990 the Department of Energy and the NIH agreed to spend $13 billion on the project. It was completed two years ahead of schedule in April 2003.
The HGP covered ninety-nine per cent of the euchromatic (active) human genome with 99.99 per cent accuracy. This makes up more than ninety-five per cent of the genome. In terms of medicine, the HGP hoped to identify mutations linked to different forms of cancer and to design specific medications and more accurate prediction of their effects. It had some degree of success. It found genes localised to specific chromosomal regions that predisposed to diseases such as Huntington’s disease, amyotrophic lateral sclerosis, neurofibromatosis types 1 and 2, myotonic dystrophy, and fragile X syndrome. Genes that conferred a predisposition to common diseases such as breast cancer, colon cancer, hypertension, diabetes and Alzheimer’s disease were also localised to specific chromosomal regions. While these findings are encouraging, a word of caution should be added at this point. As Sir Gordon Duff has said: “We have the alphabet; we now need the language.” Although the English alphabet contains twenty-six letters, there are over one million words. Thus the complexity of disease process is important to bear in mind, especially the fact that although cancer is an acquired genetic disease, in most cases the disease is not inherited. Even if we know that the genetic abnormalities found in cancer cells are acquired, in the majority of cases we do not yet know why they occur.
Although precision medicine has not provided all the answers, there have been some spectacular successes. A form of leukaemia known as Chronic Myeloid Leukaemia (CML) is treated with tablets with great success. There is no doubt that the development of an inhibitor drug for tyrosine kinase (an enzyme in cells which acts as an on/off switch for many cellular functions) in CML by Brian Druker and Joerg Zimmerman has turned a universally fatal disease into a chronic illness. This drug (Imatinib Mesylate/Gleevec/Glivec) was featured on the cover of Time magazine on May 28th, 2000. The accompanying headlines again used military allusions: “bullets: war on cancer”. Gleevec and its second- and third-generation iterations are probably the best examples today of precision medicine. The reason for the success is that the disease, CML, lends itself to such intervention, as the genetic abnormality is the same in all patients. Unfortunately the success of tyrosine kinase inhibitors in the treatment of CML has not been translated into effective therapy for other leukaemias and common cancers (the incidence of CML is one to two per 100,000 of the population annually). Some patients with breast cancer and lung cancer will benefit from precision medicine, but the majority of cancers remain elusive and, as yet, are not curable.
Will widespread surveillance of large numbers of patients with precision medicine lead the way forward, and will this approach raise questions about patient confidentiality? The work of Kári Stefánsson, an Icelandic physician who founded DeCode Genetics in 1996, serves as an example that this question has moved beyond the theoretical. His ambition was to create a large centralised healthcare database for the commercial use of health information and genetics. In the midst of concerns about privacy he believed that Iceland, which contained a relatively homogeneous population, was the ideal country in which to collect genetic data. The company specialises in looking for genetic associations and predisposition for different diseases. By 2003, almost 100,000 of 285,000 Icelandic citizens had consented and given blood to DeCode Genetics for analysis. The idea was to link people’s genetics profiles with medical records and genealogy information. This led to the discovery of the neuregulin-1 (NRG1) gene’s association with schizophrenia. The company went bankrupt in 2009, but resumed data collection in 2014 after it was purchased by Amgen. Another 100,000 envelopes were added to the 120,000 samples already taken. In spite of the discovery of NRG1 one must be careful not to confuse association with the cause of a disease such as schizophrenia. A review in 2016 in Neuroscience and Biobehavioral Reviews claimed that a lot more research needed to be undertaken before the precise relationship between NRG1 and schizophrenia could be understood. This is because the cause of disease is often linked to more than one associated element.
So far, relatively few cancers are amenable to a personalised medicine approach. This may in part be because many cancers are quite heterogeneous in terms of their molecular abnormalities. Cells have multiple genetic abnormalities and may undergo “clonal evolution” (change their molecular makeup) over time. So in spite of the huge amount of genetic, epigenetic and environmental data we have amassed which has yielded a lot of information about the pathogenesis (mechanism) of disease, the aetiology (cause) of most cancers continues to elude us. However, the drive continues to achieve this objective. In 2015 the NCI planned a “precision medicine initiative in cancer” with an initial budget of $215 million. They planned to use genomics to identify and target the molecular vulnerabilities of individual cancers. They initiated 2,400 sites and wanted to enrol 3,000 patients. Biopsies would be taken of the tumour and detailed molecular analyses carried out. If molecular abnormalities were discovered then the clinical effect of new agents, aimed at the molecular abnormality, could be evaluated. Yet precision medicine on this scale raises serious questions about patient privacy and confidentiality which have always been fundamental concepts in medical practice.
Will what was private become public when large amounts of biological and clinical data are collected and shared between laboratories, clinics and perhaps even the pharmaceutical industry? The problems for patient confidentially are already evident in the US, where some employers are asking employees for confidential medical data before agreeing to pay for medical insurance. As Anna Laakman pointed out in The New York Times in October 2015, “we need a balance between proprietary rights and the public domain”. But as yet there is no legal framework for data-sharing globally.
Public debates are increasingly airing the possibilities and the problems provided by precision medicine. Collecting genomic data on tumours is generally accepted but the notion of collecting data on the whole genome of patients is hotly disputed. One of these debates took place in London recently. In January 2015 the third Astellas Innovation Debate (Astellas is a Japanese pharmaceutical company founded in 2005) took place at the Royal Institute in London (the debate is available on YouTube). The title was “What the DNA and data revolution means for our health”. The debate was moderated by journalist Jonathan Dimbleby and panellists were Baroness Helena Kennedy QC, vice-president of the Patients’ Association, Rolf A Stahel, president of the European Society for Medical Oncology, Lionel Tarassenko, head of engineering, University of Oxford, and Leroy Hood, director of the Institute of Systems Biology in Seattle, Washington. The debate examined the pros and cons of personalised healthcare with particular emphasis on patient confidentiality. They discussed miniaturised diagnostic testing of blood samples. During the debate George Freeman MP, life aciences minister, said that the NHS would collect molecular/genetic data on 100,000 patients and match these with the patients’ clinical notes. The information would be shared with the pharmaceutical industry in an anonymised fashion. Many people have doubts about the wisdom of agencies such as the NHS collecting large amounts of data on individuals, as they fear that the results might be used for unethical purposes or fall into the wrong hands. Dimbleby concluded the debate by asking the audience if they approved of the idea of whole genome sequencing at birth. The audience was overwhelmingly opposed to such an idea but interestingly most believed that such sequencing would eventually happen.
There are other areas where a note of caution needs to be added to the phenomenon of precision/personalised medicine. Laboratories need to be strictly regulated to protect the public from unscrupulous operations and exaggerated claims of genetic testing. What the above panellists did not know at the time was that in 2016 Theranos, a silicon valley company which claimed it could carry out multiple tests on small blood samples for a fraction of the costs of “standard” laboratories, was suspended by federal regulators and the CEO Elizabeth Holmes banned from owning or running a medical laboratory for two years. The action came about because the company had violated a number of federal regulations and their tests were deemed to be unreliable. Similarly a Google-backed DNA testing company, 23andMe Inc, was forced to stop selling its personalised health reports in 2013, after the Food and Drug Administration (FDA) said the tests fell under federal testing laws. The company had produced more than 250 test reports that purported to tell users if they were likely to develop diseases such as Alzheimer’s and Parkinson’s.
Should politicians interfere in medical research? In 2015, following the death of his son from brain cancer, US vice-president Joe Biden understandably wanted to get involved. His “Moonshot Program” proposed performing full genomic sequencing on 100,000 cancer patients to create a vast database for supercomputer analysis. He received endorsement from many senior, well-respected investigators, but some remained sceptical, noting the lack of success of President Nixon’s “War on Cancer” forty-five years earlier in spite of vast resources and research.
Whether data can be truly anonymised is another matter for debate. Some people would suggest that anonymised data is an oxymoron. It should be said that attitudes to data-sharing and privacy seem to be generational. Young people do not seem to have the same attitude to privacy as the over-60s, as witnessed by the posting of private information on social media. Therefore the collection of large amounts of data on patients and sharing it between investigators and the pharmaceutical industry might not pose an insurmountable problem for future generations.
Whatever about the pros and cons of the debate about personalised/precision medicine, the concept will have an enormous impact on the doctor-patient relationship, the education of medical students and the public. As Professor Tarassenko pointed out at the third Astellas Debate, the ability of a sophisticated and presumably highly educated audience to understand the mathematics of probability and concurrent probabilities is about five per cent. How therefore can we expect doctors and the public to understand these statistical concepts? At present genetic and molecular medicine take up a relatively small fraction of undergraduate medical education. This will need to be radically changed if the personalised medical revolution becomes a widespread reality. It is estimated that at least thirty per cent of patients check health care matters online. Therefore the molecular revolution, the era of personalised/precision medicine, may be upon us already.
Another development apart from precision medicine is the encroachment of smartphone technology into medical practice. Can smartphones provide medical diagnoses and treatment and bypass doctors? An app called Babylon (founded by Ali Parsa), according to Madhumita Murgia, writing in the Financial Times magazine (January 14th/15th, 2017) is trying to create the world’s largest repository of medical knowledge, a superhuman doctor who can triage, diagnose and even treat you via your mobile phone. The company hopes that the app’s new version, to be launched in April 2017, will be the first robot to be clinically certified by the UK’s Medicines and Healthcare Products Regulatory Agency to provide medical diagnoses. Keith McNeill, chief clinical information officer of the NHS, is quoted as saying:
In five years’ time smartphones will take the burden away from the limited number of specialists we have. People will get really intelligent triage that’s personalised to them from their phones, or be empowered to look after their own chronic conditions, like diabetes, via home monitoring.
Smartphones will be used to examine heart rhythms. The patient/individual will become the custodian of their own health. All your data will be analysed by algorithms. The smartphone will remind you to change your lifestyle, diet, physical fitness etc. Do you want to be reminded of these things constantly? Are patients qualified to manage their own illness? Is the day of a visit to your general practitioner coming to an end? If patients are to manage their own illness, what about the central element of trust in the doctor/patient relationship?
Of course modern technology has a place in the practice of medicine, and perhaps will become more prominent over time; however, it has to be questioned whether undue reliance on technology over human experience needs to be more balanced. Gillian Tett in the Financial Times magazine recently addresses the question. She was discussing the film Sully (which I have not seen) about the pilot who landed a US Airways flight in the Hudson River after a bird strike had crippled the plane’s engines. Initially hailed as a hero, Captain Chesley “Sully” Sullenberger was subsequently pilloried by the National Transport Safety Board (NTSB) because he did not behave in the way a computer-derived algorithm suggested he should. Sully fought back. He pointed out that the NTSB would have made a terrible mistake by ignoring human factors in its computer models which assumed that as soon as the bird strike damaged the plane’s engines, the pilots would calmly and immediately head for the nearest airport. However if the time needed for the pilots to assess the damage was taken into account, the plane would have crashed before they had time to land at the nearest airport. To me, at least, it seems clear that computer models are useful and algorithms have their place in medicine; however the human factor remains critical. As a practitioner for over thirty years I know that patients, for a variety of reasons, do not always tell the doctor what is really worrying them. A good doctor will intuitively spot this and delve further until he/she delineates the real reason for the consultation. A computer or an algorithm are unlikely to do this. The importance of listening, touching and empathy have been part of medical practice for millennia, based fundamentally on human trust between doctors and patients.
Is the era of medicine as we know it coming to an end? I hope not, but as Niels Bohr, Nobel Prize laureate for physics in 1922, said: “Prediction is difficult, especially if it’s about the future.”
Shaun R McCann Hon FTCD is Professor Emeritus of Haematology and Academic Medicine, St James’s Hospital and Trinity College Dublin and the author of A History of Haematology: from Herodotus to HIV.