Abstract
The identification of cross-diagnostic and diagnosis-specific biotype-informed subgroups of patients and the translational dissection of those subgroups may help to pave the way toward precision medicine with artificial intelligence-supported tailored interventions and treatment. This aim is particularly important in psychiatry, a field where innovation is urgently needed because specific symptom domains, such as negative symptoms and cognitive dysfunction, and treatment-resistant symptoms in general are still difficult to treat.
Introduction: The Challenge of Heterogeneity in Psychiatry
Current psychiatric diagnostics for severe mental illnesses (SMI), such as schizophrenia (SZ), bipolar disorder (BD), and major depressive disorder (MDD), rely on symptom clusters defined by the DSM-5 or ICD-10. However, these categories are biologically heterogeneous; two patients sharing a diagnosis may possess vastly different underlying neurobiology and treatment responses. This "translational gap" prevents laboratory findings from improving clinical outcomes. Approximately 25% of schizophrenia patients exhibit treatment resistance early in their illness, and domains like cognitive dysfunction remain difficult to address.
The Munich Clinical Deep Phenotyping (CDP) study is a large-scale initiative designed to move psychiatry beyond symptom-based diagnosis toward "biotype-informed" precision medicine. By collecting a massive, multimodal dataset, researchers aim to identify biological signatures that predict treatment response and guide personalized interventions.
Methodology: The Deep Phenotyping Approach
The CDP study adopts the Research Domain Criteria (RDoC) framework, shifting focus from traditional diagnoses to functional domains across a spectrum from normal to pathological. The "deep phenotyping" approach involves collecting an unusually broad array of data from every participant:
- Clinical & Psychometric Assessment: Utilizing standardized scales (PANSS, CDSS, YMRS) and neurocognitive batteries (BACS).
- Multimodal Brain Imaging (mMRI): Employs 3T scanners with protocols aligned with the Human Connectome Project (HCP), including structural MRI, DTI, resting-state/task-based fMRI, and magnetic resonance spectroscopy (MRS).
- Electrophysiology: 32-channel EEG (including P300 potentials) and Transcranial Magnetic Stimulation (TMS) to assess cortical excitability.
- Retinal Imaging: Optical Coherence Tomography (OCT) and electroretinography (ERG) serve as non-invasive "windows to the brain" to study CNS microvasculature and neurodegeneration.
- Biobanking: Collection of blood and CSF for transcriptomics and proteomics, alongside the generation of human-induced pluripotent stem cells (hiPSCs).
Cohort and Data Integration
The initiation phase of the CDP cohort includes 381 participants: 194 individuals with SMI (primarily schizophrenia spectrum and bipolar disorders) and 187 age- and gender-matched healthy controls (HC). A key strength of the CDP study is its integration with previous cohorts, including the MIMICSS imaging study and a dedicated hiPSC cohort.
This integration allows for robust longitudinal and translational analysis, connecting clinical phenotypes with cellular models.
Key Insights: From Biotypes to "Clinical Trials in a Dish"
The CDP study introduces several innovative concepts to psychiatric research:
- Transdiagnostic Biotypes: Identifying biological patterns (e.g., circuit dysfunctions) that appear across traditional diagnostic boundaries.
- The Retina as a Biomarker: Preliminary results suggest thinner inner retinal layers in schizophrenia spectrum disorders, highlighting the retina's potential as a high-resolution biomarker.
- Patient-Specific Cell Models: By biobanking hiPSCs from deeply phenotyped patients, researchers can create "clinical trials in a dish," testing molecular pathways and screening drugs in vitro before clinical application.
- Global Compatibility: By aligning with HCP protocols and RDoC units, the CDP data is compatible with international efforts like ENIGMA and the UK Biobank.
Feasibility and Participation
A primary achievement of the CDP study is the demonstration of feasibility. Over the first two years, the team successfully recruited nearly 400 participants and maintained high participation rates across demanding modalities.
Data indicates no statistically significant difference in the willingness of patients versus healthy controls to participate in most examinations, including EEG, OCT, and blood sampling. While MRI participation was slightly lower in patients (p=0.06), the overall data collection remains robust.
Limitations and Future Directions
While the current sample of 381 is significant, larger datasets are required for high-dimensional "big data" analytics to identify rare biotypes. Currently, much of the data is cross-sectional, though longitudinal follow-ups (six months to two years) are underway to capture biological changes over time. Additionally, disentangling the effects of long-term medication from disease pathology remains a challenge, which the study aims to address by recruiting drug-naïve and first-episode patients. The next phase involves integrating "omics" data with clinical and imaging phenotypes for AI-supported patient stratification.
Summary updated