A Real World Evidence approach to develop a better understanding of clinical outcomes of patients with myeoloproliferative neoplasms (MPNs)

Lead Supervisor
Dr Mieke Van Hemelrijck
Reader
Translational Oncology & Urology Research, Cancer and Pharmaceutical Sciences, FoLSM, King’s College London
mieke.vanhemelrijck@kcl.ac.uk

Co-supervisor
Dr Shahram Kordasti
Consultant Haematologist, Guy’s and St. Thomas’ (GSTT)

Project Details

Unprecedented advances in machine learning methods have enabled the synergy of artificial intelligence and digital pathology, which offers image-based diagnosis possibilities that were once limited only to radiology. In this project, we will use digitised whole slide images (WSI) from a very aggressive breast cancer type, namely triple negative breast cancer (TNBC). Patients of TNBC have currently limited treatment options, as this type of cancers is clinically and molecularly very heterogeneous. So far, we have only identified few subgroups of patients for whom targeted therapies are available.

Myeloproliferative Neoplasms (MPN) are stem cell neoplasms displaying autonomous proliferation associated with JAK/STAT pathway activation. Following acquisition of an MPN-initiating mutation MPN-development requires clonal expansion, myeloproliferation and a self-reinforcing hematopoietic niche. The JAK2V617F mutation, leading to constitutive activation of JAK2 and upregulated JAK-STAT signalling, is found in the majority (95%) of polycythaemia vera (PV) patients and in 50-60% of essential thrombocythaemia (ET) and myelofibrosis (PMF) which comprise the MPN entities. 

Chronic inflammation is an established feature of MPN and JAK-STAT pathway activation is considered a major driver. Presence of specific mutations such as DNMT3, TET2 and ASXL1 is associated with higher risk of haematological malignancies as well as inflammatory associated diseases such as coronary heart disease and ischemic stroke. Chemotherapy and/or radiation therapy for solid tumours may create an environment that selects for pre-existing mutant clones and lead to treatment related MPN (t-MPN). Most of these patients are older and may have other health problems which contribute to adverse events and overall outcome. The current project is set out to get a better understanding of the clinical outcomes of these patients both in terms of disease progression, adverse events (e.g. thrombosis) as well as symptom and multimorbidity control. 

We will use Guy’s Cancer Cohort to identify all patients with MPN treated since 2009. The PhD project will specifically address the following clinically relevant questions using real world data for healthcare analytics:

  1. Correlation between myeloid derive inflammation/ immune signature(s) and common symptoms like fatigue and night sweat.
  2. Presence of specific myeloid related gene mutations and their effect on disease progression and thrombosis.
  3. Validate the specific immune signature(s) which could potentially predict adverse clinical events (prospective part of study)
  4. Assessment of how comorbidities may affect oncological outcomes

In addition, a comprehensive database of demographic and detailed clinical data of patients with t-MPN will be created and combined with somatic mutation findings as well as inflammation biomarkers such as cytokine levels and over/under expression of inflammasome pathway related genes. We will use this database to evaluate the predictive value of inflammatory environment on evolution to t-MPN as secondary cancer.

A combination of common multivariate regression models (e.g. Cox proportional hazards and logistic) will be used as well as more advanced analytical tools based on latent class analyses and Bayesian survival analysis (Rowley, Garmo, Van Hemelrijck et al) to answer these questions. Handling of covariates or confounding factors as well as temporal logics will also be an important component when processing the data. The project will provide the opportunity to develop analytical skills using real world data and hence allow the student to set up a career in big data analytics.

Datasets

Guy’s Cancer Cohort (REC: 18/NW/0297; PI: Van Hemelrijck M); About 150 newly diagnosed patients per year, with follow-up for about 40/50 patients on a weekly basis. 

Keywords

big data; health analytics; real world evidence; Myeloproliferative Neoplasms