CoMMpass Clinical Data Overview¶

The clinical dataset generated through the MMRF CoMMpass Study is one of the richest real-world and longitudinal clinical resources available for multiple myeloma. It integrates structured clinical assessments, laboratory data, treatment timelines, and long-term outcomes for more than 1,100 patients.

This page provides a high-level overview of the clinical data and how it was harmonized for use in the MMRF Virtual Lab.

Clinical Data Collection Framework¶

Clinical data were collected from:

Routine clinical care
Structured study visits
Patient-reported outcome instruments over time (EORTC QLQ-C30 and EORTC QLQ-MY20)

Because visit schedules varied across patients and centers, clinical timing was standardized using "days from baseline" relative to the diagnosis date to support reproducible analyses.

Major Categories of Clinical Data¶

The harmonized dataset includes variables across key clinical domains such as:

Demographic Information¶

Age at diagnosis
Sex
Race and ethnicity (when collected)
ECOG performance status

Diagnostic & Disease Characteristics¶

Laboratory values (hematology, chemistry, biomarkers)
Bone disease assessments
Imaging interpretations
Pathology impressions
Cytogenetics/FISH
- Local pathology data available for a subset
- Inferred cytogenetic abnormalities from all with tumor genomics (seqFISH)

Treatment Details¶

Initial therapy regimens
Drug-level administration data
Transplant information
Response and relapse documentation
Subsequent lines of therapy (observational cohort)

Clinical Outcomes¶

Response assessments based on IMWG criteria
Dates of progression, relapse, and treatment change
Progression Free Survival, Overall Survival, and time-to-event variables

Patient-Reported & Supportive Care Data¶

Where collected:

Quality-of-life questionnaire results
Supportive care medications and interventions

These data enable clinical insights into disease burden, treatment patterns, and real-world outcomes.

Harmonization of Clinical Data¶

Because CoMMpass combined data from two cohorts with different study designs, a multi-step harmonization process was required:

1. Variable Mapping¶

Clinical variables from the observational cohort and trial cohort were aligned to a unified data dictionary modeled after NCI GDC concepts.

2. Derived Variables¶

Standard logic was applied to produce:

Days-to-event variables
Therapy line assignments
Consolidated pathology and lab indicators
Normalized visit timing

3. Quality Control¶

All tables underwent:

Consistency checks
Logical sequencing validation
Cross-referencing with sequencing metadata
Harmonization across sites and timepoints

Use of Clinical Data in the Virtual Lab¶

The clinical dataset powers multiple tools:

Cohort Builder¶

Filter patients by disease features, treatments, outcomes, or demographics.

Cohort Comparison¶

Compare characteristics across user-defined cohorts with visual summaries.

Clinical Data Analysis¶

Explore distributions, survival, and statistics for any clinical variable.