Frontotemporal dementia (FTD) has a debilitating effect on patients and their caregivers and leads to substantial economic costs. 15-30% of patients have familial FTD caused by known pathogenetic mutations. For the other 70-85% of patients, termed sporadic FTD, diagnosis is slow (~3.6 years) with frequent misdiagnosis due to clinical, genetic and molecular heterogeneity. Thus, there is great need for biomarkers for early diagnosis of sporadic FTD and its pathological subtypes. In PREDICTFTD, we will validate a set of biomarkers and create a diagnostic tool for early diagnosis of familial and sporadic FTD, which will facilitate tailored support and symptomatic treatments and care. We will apply several new approaches to achieve this: 1) we combine 11 geographically diverse cohorts of sporadic and familial FTD with retrospective and prospective longitudinal liquid biopsy samples and extensive clinical and behavioural data; 2) we are the first to use multimodal clinical and liquid biomarker data to train an AI-algorithm as a diagnostic tool for quick and early clinical FTD diagnosis; and 3) we implement a novel robust two-stage strategy for biomarker and AI algorithm validation, where phase I validates biomarkers and algorithms on a cohort of genetic and autopsied cases and phase II assesses biomarker value for diagnosis of sporadic FTD and at-risk pre-symptomatic mutation carriers. We will apply this two-stage validation strategy to address three critical clinical challenges: i) To distinguish sporadic FTD from (non-) neurodegenerative disorders that show significant clinical/symptomatic overlap, ii) To robustly detect FTD pathological subtypes in sporadic FTD and iii) pre-symptomatic identification of FTD onset. Thus, PREDICTFTD will transform FTD diagnosis, offering potential for early disease confirmation, guiding treatment decisions, facilitating patient recruitment for clinical trials, guidance of patients and caregivers, and enabling preventive measures.

A key problem in Mental Health is that up to one third of patients suffering from major mental disorders develop resistance against drug therapy. However, patients showing early signs of treatment resistance (TR) do not receive adequate early intensive pharmacological treatment but instead they undergo a stepwise trial-and-error treatment approach. This situation originates from three major knowledge and translation gaps: i.) we lack effective methods to identify individuals at risk for TR early in the disease process, ii.) we lack effective, personalized treatment strategies grounded in insights into the biological basis of TR, and iii.) we lack efficient processes to translate scientific insights about TR into clinical practice, primary care and treatment guidelines. It is the central goal of PSYCH-STRATA to bridge these gaps and pave the way for a shift towards a treatment decision-making process tailored for the individual at risk for TR. To that end, we aim to establish evidence-based criteria to make decisions of early intense treatment in individuals at risk for TR across the major psychiatric disorders of schizophrenia, bipolar disorder and major depression. PSYCH-STRATA will i.) dissect the biological basis of TR and establish criteria to enable early detection of individuals at risk for TR based on the integrated analysis of an unprecedented collection of genetic, biological, digital mental health, and clinical data. ii.) Moreover, we will determine effective treatment strategies of individuals at risk for TR early in the treatment process, based on pan-European clinical trials in SCZ, BD and MDD. These efforts will enable the establishment of novel multimodal machine learning models to predict TR risk and treatment response. Lastly, iii.) we will enable the translation of these findings into clinical practice by prototyping the integration of personalized treatment decision support and patient-oriented decision-making mental health boards.