You can call it personalised medicine, genomic medicine, stratified medicine or precision medicine, or other terms – but whatever you call it, the name is personal. The final aims are a more precise approach to a patient, better results, and long-lasting treatment.
‘My eHealth’ is a patient’s digital health information portal which gets almost a million requests every month (let’s remind ourselves that the Estonian population is just 1.3 million!). The patient’s portal was launched in 2009, and since then 17 per cent of the population has visited ‘My eHealth’ – which is a remarkable rate.
‘My eHealth’ is part of a health information system that is an inseparable part and backbone of a personalised or precise medicine. Estonians simply term it eHealth; everybody understands that this means the system incorporating Electronic Health Records, Digital Registration, Digital Image, and Digital Prescriptions.
But let’s return to the two significant facts noted above: the million requests on eHealth and 17 per cent of the population. Why do people have such a keen interest on their health records? ‘The idea behind eHealth is that no matter where you are located your health records and visits to different doctors are safely accessible,’ says Peeter Ross, professor Tallinn University of Technology, radiologist and an expert at the Estonian eHealth Foundation.
eHealth is widely used by healthcare workers to access patients’ health information, eg. the attending doctor or nurse gets quick access to patient’s complete health information all over Estonia. This in turn decreases the number of unnecessary appointments, eg. prescription refills, duplicate lab tests and screenings, effectively using the patient as a courier to make inquiries about getting an appointment with a specialist etc.
Granted, this increases the efficiency of healthcare system and raises the quality of medical service, but what is personalised for the patient about that?
‘It means that the doctor no longer makes the decision only based on what the patient is saying or showing at the appointment. eHealth gives us a more complete picture of the patient’s precise health based on a variety of health records,’ explains Peeter Ross. And that is more accurate and personalised than just a single and fleeting visit to the doctor.
Medical workers are just one group of users, however. eHealth is ‘tailored’ to function quite personally for the patient. Peeter Ross reveals one possible reason for this: curiosity. ‘People come to see what the price of their personal health care is.’ These sorts of individual calculation were made accessible for the patient in the first half of 2016, which aroused the ‘curiosity’ levels significantly – the percentage of users rose from 10 per cent to 17 per cent.
Personal eHealth can be compared with personal banking. ‘People don’t use electronic banking just for seeing their account balance. They want to have more than that – make money transfers, sign contracts, get loans, and use other banking services,’ Peeter Ross says by way of example. It is a similar case with eHealth. For example you can fill in an application for a drivers’ licence health certificate and then pop by your family doctor to give blood test. It saves both the patient’s and doctor’s time.
Doctors: medicine has always tried to be personal!
Doctors tend to get irritated when the topic of personalised medicine comes on the agenda. They say that they have always looked at and treated every patient as an individual person. Therefore it is only fair to ask what makes personalised medicine personal, isn’t it?
‘Personalised medicine doesn’t take away anything that we have practiced so far,’ explains Peeter Ross, practicing radiologist. Last year the Estonian Ministry of Social Affairs in cooperation with the University of Tartu, Tallinn University of Technology and the main hospitals conducted an analysis on eHealth services, which also included the defining personalised medicine to a wider scale and scope. Compared with traditional medical practice, personalised or precision medicine has two additional components. One of these is perhaps already widely-known – The Estonian Genome Centre of the University of Tartu and personal genome-wide data. We will come back to that later on. But the second inseparable component of personalised medicine is information and communications technology (ICT). Currently there are two hot topics on the ICT-related side of personalised medicine: ‘digital decision support system’ and ‘crowd diagnosing’.
The digital decision support system (DDSS) compares genome-wide data with the phenotypic, or simply patients’, medical data. The objective is to find patterns of connections and parallel the findings with lifestyle and health habits. Based on this data analysis DDSS is able to make recommendations to the attending doctor and denote to possible obstacles in treatment – breast cancer, diabetes, and cardiovascular diseases, for example. DDSS should be able to model recommendations based on genotypic and phenotypic data of similar previously-treated patients.
Crowd diagnosing, an application of DDSS, enables us to create advice or alerts based on prior evidence-based knowledge in order to analyse large volumes of data gathered from other patients in existing medical databases. For instance in the case of breast cancer patient´s genome, data is compared with genome data of patients with similar diseases and applied treatment. It can be characterised as gathering patterns and creating models from the whole population in order to personally approach one patient.
Peeter Ross is a radiologist and in his everyday work personalised medicine means more standardised and precise descriptions of diseases. Take a tumour as an example. An Estonian radiologist may describe the size of the tumour comparing it with ‘a chestnut’. A Finnish doctor, at the same time, would say that the tumour is the ‘size of an egg’. The implementation of ICT in order to standardise and analyse the descriptions of X-ray images would help to make more precise health care decisions, not to mention the possibility of sharing digitally X-ray, magnetic resonance imaging (MRI) or other images with other doctors involved in the treatment of the patient.
Personalised medicine in our genes
Did you know that for more than 90 per cent of commonly-prescribed drugs, the desired effect is only achieved in 30-50 per cent of the population? 20 to 50 per cent of those individuals diagnosed with depression respond to the standard dose of antidepressant, for example. ‘But there is about one third of patients who do not respond to the treatment at all,’ says Lili Milani, Senior Researcher at the Estonian Genome Center of the University of Tartu (EGCUT).
Depression is a very heterogeneous diagnosis – the causes of the disease are manifold and vary from person to person. As for the drug response, one of the factors is undoubtedly genetics, states Milani. She explains that the genes that encode the enzymes which are supposed to metabolise, ie. activate or eliminate the drugs from the body, are particularly polymorphic. Therefore some people metabolise the drug quickly or even ultra-rapidly. At the other end of the scale there are people who are slow metabolisers and the required drug-dosing between these two groups may be up to a 100-fold difference. For ultra-rapid metabolisers, the standard drug dose would simply be eliminated from the body long before it has had a chance to have an effect. Slow metabolisers, on the other hand, bear the risk of drug induced side effects.
Since to a great extent this variation is based on genetic differences in drug metabolism enzymes, drug transporters, or drug targets, it can be approached personally based on polymorphisms in related genes. For this reason the researchers at the EGCUT are studying the genetics of drug response to see which genes are related to which response. This, in turn, can be implemented in the medical system in the near future. Lili Milani gives an example from the US, where five hospitals are already implementing pharmacogenomics for at least five gene types which can metabolise up to 90 per cent of the standard drugs prescribed today.
The EGCUT suggests that in Estonia, everyone aged 35 to 65 should be offered the opportunity to get genotyped – ie. have their personal gene chip. So far, 52 000 Estonians have donated their blood to create the Estonian Biobank. The data generated in the biobank enables us to use individual genomic variation obtained from genetic analysis and computational methods to predict and prevent diseases, and to optimize drug treatment.
The most common limitation – money
The data collected at the biobank has been further improved by incorporating data from the nation-wide health database of the Estonian National Health Information System and other more specific registries. These extensive health records and molecular profiling data of the biobank participants are used to calculate disease risk and the associated likely drug response. Since both databases are continually improving, estimates of the disease risks and probable drug response must be re-calculated regularly.
Asking Lili Milani whether every new-born should have their DNA sequenced, her answer is a firm ‘yes’! Every child should get its genotypes analysed. Of course it raises the ethical question of sequencing the whole genome of the child. What if the child later in life does not want to be aware of its genetic risks? Lili Milani proposes a solution – it would be wise to look for mutations in certain known disease-causing genes in order to avoid situations, where the child ends up in the intensive care and only then the doctors turn to EGCUT for genetic testing.
This is actually a real-life situation Lili Milani describes: 18-month-old child is in intensive care in critical condition. The doctor has ordered for genetic testing from EGCUT and is impatiently calling for results. But it takes at least two weeks from sample delivery to final report. The whole genome sequencing costs about 1 000 Euros. At the same time testing for precise mutation for example certain cancer type may cost up to 5 000 Euros.
‘But this is relatively cheap if we compare it to the cost of treatment or the risk of maltreatment in case,’ states Hele Everaus, hematology-oncology professor at the University of Tartu. She has studied cancer treatment for 30 years. ‘Back then they were already very hopeful of finding cancer cure in the near future,’ says Everaus. In reality there are treatments, but they are effective on only 40 per cent of the cancer types.
Consequently the current cancer treatment is partly ineffective and also expensive. Oncologists tend to say that there is no lung or colon cancer, there are tumours of different genetic profile in specific micro ecosystem affected by cigarette smoke and food consumption, for example. Molecular profiling of cancer means that the gene test is made on tumour cells. ‘The developing mechanisms of the cancer are somehow universal, but we don’t know microenvironment the cancer develops in,’ explains Hele Everaus.
For example the lung cancer can be caused by 200 different genetic mutations and their combinations, adds Lili Milani. The challenge is to find out the gene leading to the development of the cancer. If we look at the cancer treatment from the point of view of pharmacogenomics and precision medicine, in oncology it means precisely saving lives. It means the optimum treatment according to a patient’s medical history, physiological status and on the tumour’s genetic peculiarities.
Prof. Hele Everaus describes how, depending on the country, the genetic testing in question may cost around 4 000 to 5 000 Euros. But the test gives us the certainty that the selected treatment may well work. ‘But compared with cancer therapy that may cost up to 100 000 Euros, it is a significantly lower amount of expenses, especially if the drug is inefficient or there are adverse drug responses for that precise cancer,’ says Everaus.
She hopes that the research and development will turn more and more into concrete diagnostics in order to discover tumours in very early stages of development. The scientists are moving towards finding novel potential biomarkers for early stage detection of cancer.
Using ‘junk-DNA’ in cancer diagnostics
Based on the nature of tumour where cells are growing rapidly, causing increased cell death rate compare to normal tissues, the genomic DNA from dead tumour cells is released to patient blood system. This ‘junk-DNA’, also known as cell-free DNA (cfDNA), carries specific mutations which already have diagnostic value in the early phase of recurred tumour after treatment. The cfDNA based screening of relapsed tumour has great potential in healthcare.
Scientist in Competence Centre on Health Technologies (CCHT) in Tartu are developing a cfDNA based analytical tool to evaluate the precise percentage of tumour mutations among patient cfDNA. The precise laboratory method is a missing link between the usage of cfDNA in molecular screening and cost-effective healthcare. Current research-intensive methods for cfDNA analysis are not quantitative if certain mutations need to be tracked after surgery.
Though the fact that cfDNA can be used as a potential biomarker for cancer screening and diagnostics by the researchers at CCHT, the same research group’s main field of research is reproductive medicine. There are a lot of possibilities for personalised medicine practices in the fertility studies.
Triin Laisk-Podar is a researcher at the CCHT and Women’s Clinic of the University of Tartu. She tells us that assessing female fertility and predicting reproductive aging is a new challenge in the current situation, where more and more women are postponing motherhood, evidenced by the fact that age at the first birth is increasing both in Estonia (currently 26.5 years) and in Europe more generally (28.7).
‘Although menopause is usually considered the time point that marks the loss of female fertility, in reality, fertility starts to decrease decades before and subfertility can manifest, in fact 10 years earlier,’ she adds. Women who experience early (before the age of 45) or premature (before the age of 40) menopause are therefore at risk of earlier reproductive senescence in their thirties, which combined with the fact that age at first childbirth is increasing can result in age-related infertility.
For this reason, markers that could predict age at menopause have been extensively sought for, but all currently used hormonal or ultrasound markers reflecting ovarian function have several shortcomings. Age at menopause also has a strong genetic component, which is illustrated by the fact that mother’s age at menopause is one of the best predictors for daughter’s menopausal age.
The knowledge on the genetic variants affecting menopausal age was the basis for developing the Fertify test, which uses polygenic risk scores for predicting the risk of early menopause and concomitant earlier decrease in fertility. Genetic risk profiling has several advantages over conventional markers, such as robustness, stability and usability in young women who do not show remarkable changes in hormone dynamics or have no information on their mother’s menopausal age. ‘Genetic risk profiling for reproductive aging can be used to offer women personalised evidence-based advice for family planning, and therefore is most beneficial for women in their early twenties,’ Triin Laisk-Podar brings as an example of personalised medicine in reproductive medicine.
Estonia is currently in a position to be one of the first countries in the world to start the implementation of personalised medicine on a national scale. Implementing personal genetic information together with automated decision support systems will help physicians greatly and have a huge impact on disease prediction, prevention and treatment. Personalised medicine should be viewed just as a new, additional instrument available for the physicians.