Optimising therapeutic strategy in psoriasis

As global life expectancy and the prevalence of chronic diseases have increased, scientific research has delivered an expanding repertoire of treatments. An emerging challenge is ensuring the most effective, appropriate drug is selected early in each patient's journey.

Aim:
This study will determine whether a model that predicts response to treatments for psoriasis improves patient outcomes.

Background:
Psoriasis is a common, debilitating skin disease that affects between 2 and 4 in every 100 people worldwide. Patients develop red, scaly patches of skin that are itchy and painful. It is associated with heart and joint disease, can affect an individual's ability to work, and impacts on quality of life are comparable to heart disease or cancer. A third of patients report clinically significant anxiety and depression, and 1 in 10 have contemplated suicide.

Recent research has led to several 'biologic' treatments for psoriasis, however almost one third of patients fail to respond. Biologics are prescribed in a 'trial and error' manner until disease control is achieved. This process is expensive, since biologics cost ~£10,000/year/patient. Current practice also places patients at risk of long periods with poorly controlled disease and exposure to drug side effects. It is therefore vital to understand which patients will benefit from a particular biologic.

Since we now recognise that each patient has multiple unique characteristics (e.g. changes in their genetic code) that contribute to their psoriasis, there is an urgent need to use medications intelligently. The MRC-funded Psoriasis Stratification to Optimise Relevant Therapy (PSORT) study has uncovered markers of treatment response in psoriasis, however it is not known whether this knowledge will translate into better patient outcomes.

Methods:
Based on PSORT's findings, I will develop a model that predicts which biologic will be the most effective for a particular patient. The model will use information that is unique to the patient, including their genetic code ('DNA'; measured from a blood sample) and presence of arthritis. The model will then use blood tests (measuring drug levels) to guide dose adjustment during treatment.

I will evaluate the model's accuracy by analysing an independent, multicentre psoriasis database called BADBIR, which has periodically followed up patients receiving biologics, and recorded detailed genetic and clinical information. I will next design an innovative clinical trial to determine whether the model will lead to better outcomes for patients, including clearance of psoriasis and improved quality of life.

Patient and public involvement:
I sought patient and public involvement (PPI) in the development of this study through focus groups and surveys supported by my collaborators the Psoriasis Association (the UK's leading national charity and membership organisation for individuals affected by psoriasis). I will set up a patient and public advisory panel and meet with PPI focus groups six-monthly to provide progress updates and ascertain feedback.

Dissemination:
The results will potentially significantly impact on patients and national clinical-care pathways, so all relevant stakeholders will be updated with key findings. I will disseminate results at national and international conferences and through high-impact peer-reviewed papers. Results will be broadcast via university press offices and the information channels of the Psoriasis Association (website, social-media, magazine). I will also highlight the importance of this research to other medical specialties prescribing biologics such as rheumatology and gastroenterology. It will thus help a wide range of patients with chronic immune-mediated diseases maintain better control of their health, and save money for the NHS.

Psoriasis affects 2-4% of the population and NHS costs are >£750 million/year, largely due to biologics costing £10,000/year/patient. NICE recommend 10 biologics targeting TNF or IL-23/T17 cytokines with equal first-line weighting. Real-world data suggest 20-30% of patients fail to respond, highlighting inter-individual variation in driver immune pathways. We thus urgently need to understand which biologic best suits each patient. Whereas biomarkers guide cancer care, such strategies are lacking in inflammatory diseases. Given research by the PSORT biomarker discovery programme, psoriasis presents an ideal disease model for this work.

Hypothesis: An algorithm which stratifies patients to receive TNF or IL-23/T17 inhibitors on the basis of genetic and clinical biomarkers, followed by dose optimisation using therapeutic drug monitoring (TDM), will improve real-world outcomes.

Aims:
1. Develop a treatment algorithm to personalise biologic use
2. Evaluate the algorithm's clinical utility in silico using a multicentre real-world observational pharmacovigilance dataset
3. Design a prospective biomarker-strategy trial to assess the algorithm against standard care

Methods: I will develop an algorithm, using PSORT's existing dataset and logistic regression, that stratifies patients to receive TNF or IL-23/T17 inhibitors based on established genetic (HLA-C*06:02) and clinical (e.g. psoriatic arthritis) biomarkers of response. TDM will prompt dose modification in those with supra/subtherapeutic drug levels. The algorithm's clinical utility will be determined in silico in an existing independent longitudinal outcome bioresource containing genetic, clinical and drug level data on patients receiving biologics, using area under the ROC curve. A trial will be designed to assess the algorithm.

By personalising treatment selection, this study will improve outcomes and cost efficiency of care and provide a blueprint for the management of other inflammatory diseases.

This research will accelerate the introduction of personalised medicine into the management of chronic inflammatory diseases using psoriasis as an exemplar. It will develop and validate personalised data-driven biologic treatment selection for patients with psoriasis. It aims to find the right class of biologic for the right patient and provide an evidence-based alternative to current 'trial and error' prescribing. It will translate the results of a sustained biomarker discovery research programme (PSORT) into improved care pathways for psoriasis, thus delivering concrete health benefits for a high-need patient group.

While biologic treatments have shown exciting efficacy in trials, disease treatment remains challenging in real-world settings, owing to variable response rates and long-term toxicity. This reflects inter-individual variation in underlying driver immune pathways (TNF versus IL-23/T17) and drug pharmacokinetics (as indicated by serum drug levels). Patients thus often cycle through multiple biologics before remission is achieved, resulting in excess cost, morbidity and potential adverse events. Given the elevated incidence of psoriasis (the Psoriasis Association estimates >1.5 million individuals have psoriasis in the UK alone), the severity of associated co-morbidities (e.g. psoriatic arthritis, cardiovascular disease, depression) and increased mortality, there is an urgent need for the development of intelligent treatment strategies to enable targeting of the critical immune pathways that are deregulated in each individual. A personalised biologic algorithm that uses multi-dimensional biomarkers to predict response will optimise treatment, limit adverse effects and improve cost-effectiveness, thus impacting on many stakeholders including patients, health professionals and the public.

The research will benefit psoriasis patients and family members who are frequently involved in their care. 75% develop psoriasis before age 35 and biologic treatment can be lifelong. Finding the most effective biologic early in each patient's treatment journey will increase their quality of life, confidence in the healthcare system, and impact on co-morbidities. Inadequately treated disease leads to a greater burden on the NHS through increased visits to healthcare services. Further, the average sickness absence for a worker with moderate-severe psoriasis is 14 days/year, with total costs of >£1billion/year for absenteeism and reduced work productivity. Ensuring that each patient receives the correct biologic at the optimal dose will reduce their healthcare visits, and increase their employment prospects and ability to work, leading to wider gains to the NHS, society and the economy.

The findings will be applicable to patients receiving the same biologics for other diseases e.g. inflammatory bowel disease and rheumatoid arthritis, and enhance research efforts in these disciplines, such as that of the MATURA consortium. By developing more effective, efficient strategies to use expensive biologics in high-need patient populations, this work will inform evidence-based policy makers, professional bodies developing clinical guidelines (e.g. NICE), and lead to healthcare gains and cost savings.

This research, which will capitalise on recent genetic, clinical and pharmacological biomarker discoveries, is likely to generate considerable interest within the pharmaceutical industry, where the search for optimised biologic treatment strategies and algorithms to predict response is a very active field. Greater insights on the implementation of personalised medicine will inform the future learning of clinicians and researchers. The research will optimise trial methodology for evaluating personalised treatment algorithms in real-world settings. The proposed in silico evaluation of the algorithm's clinical utility is also highly generalisable given the pharmacovigilance datasets available in other diseases (e.g. rheumatoid arthritis).