Genomic Analysis & Tumor Board

Once genomic variants are uploaded, annotated, and imported into the OMOP CDM, Parthenon provides three population-level analysis types and an individual patient tumor board dashboard. These tools connect molecular findings to clinical outcomes, enabling translational research and precision medicine workflows.

---

Genomic Analysis Suite

The analysis suite is accessible from Genomics > Analysis in the navigation. All analyses query the OMOP CDM directly, correlating genomic measurement data with clinical outcomes, treatments, and demographics.

Survival Analysis

Kaplan-Meier survival analysis compares outcomes between patients who carry a specific variant and those who do not (wild-type).

How to run a survival analysis:

1. Navigate to Genomics > Analysis.
2. Select the Survival tab.
3. Configure:
   • Data Source --- the OMOP source containing both genomic and clinical data
   • Gene --- the gene of interest (e.g., EGFR, KRAS, TP53)
   • HGVS Variant --- optionally restrict to a specific variant (e.g., L858R). Leave blank to compare any mutation in the gene vs. wild-type.
4. Click Run Analysis.

Results include:

• Survival curves --- Kaplan-Meier curves for mutated vs. wild-type groups, plotted over time from the index date (variant detection or cohort entry) to the event (death) or censoring (end of observation)
• Log-rank p-value --- statistical significance of the survival difference between groups
• Median survival --- median survival time for each group (with confidence intervals)
• Patient counts --- number of patients in each group, number of events, and number censored

Choosing the right gene

Start with genes known to have prognostic significance in your disease area. For lung cancer, consider EGFR, KRAS, ALK, and TP53. For breast cancer, consider BRCA1, BRCA2, and HER2. The survival analysis is most informative when you have sufficient patients in both the mutated and wild-type groups.

Treatment-Variant Response Matrix

The treatment-variant response matrix cross-tabulates drug exposures against genomic variants to identify treatment-variant combinations associated with different outcomes.

How to run a treatment-variant analysis:

1. Navigate to Genomics > Analysis.
2. Select the Treatment Matrix tab.
3. Configure:
   • Data Source --- the OMOP source
   • Genes --- select up to 10 genes to include in the matrix
4. Click Run Analysis.

Results display:

• A matrix with drugs (from the CDM drug_exposure table) on one axis and gene variants on the other
• Each cell shows the outcome metric (e.g., time to progression, response rate) for patients who received that drug and carry that variant
• Color coding highlights favorable (teal) and unfavorable (red) treatment-variant combinations
• Cell counts show the number of patients in each combination

This analysis helps answer questions like: "Do patients with EGFR L858R who receive osimertinib have better outcomes than those who receive standard chemotherapy?"

Genomic Characterization

Population-level genomic profiling provides an overview of the variant landscape in your data source.

How to run characterization:

1. Navigate to Genomics > Analysis.
2. Select the Characterization tab.
3. Select a Data Source.
4. Click Run.

Results include four visualizations:

Chart	Description
Top Mutated Genes	Horizontal bar chart showing the most frequently mutated genes across all patients, ranked by variant count
TMB Distribution	Histogram of tumor mutational burden across the patient population, showing the distribution shape and identifying patients with high TMB
Variant Type Breakdown	Pie or bar chart showing the proportion of SNVs (single nucleotide variants), insertions, deletions, and MNVs (multi-nucleotide variants)
ClinVar Significance	Distribution of ClinVar clinical significance classifications across all annotated variants

These characterization results are useful for understanding the molecular profile of your patient population before designing targeted analyses or clinical trials.

---

Genomic Cohort Criteria

The genomics module extends Parthenon's cohort builder with six genomic criteria types, allowing you to define patient cohorts based on molecular characteristics.

Available Criteria Types

Type	Description	Example Use Case
Gene Mutation	Patients with a specific mutation in a specific gene	EGFR L858R carriers for lung cancer therapy selection
TMB	Patients above or below a tumor mutational burden threshold	TMB > 10 mutations/Mb for immunotherapy eligibility
MSI	Patients with a specific microsatellite instability status	MSI-High patients for pembrolizumab candidacy
Fusion	Patients with a specific gene fusion event	ALK-EML4 fusion carriers for ALK inhibitor studies
Pathogenicity	Patients with variants of a specific ClinVar classification in a gene	Any pathogenic variant in BRCA1 or BRCA2 for hereditary cancer screening
Treatment Episode	Patients who received a specific treatment following a genomic finding	Patients started on osimertinib after EGFR mutation detection

Creating a Genomic Criterion

1. Navigate to Genomics > Criteria (or access from the Cohort Builder's genomic criteria panel).
2. Click New Criterion.
3. Select a criteria type from the list above.
4. Configure the criterion:
   • For gene mutation: specify the gene, and optionally the specific HGVS variant
   • For TMB: specify the operator (>, >=, <, <=) and threshold value
   • For MSI: select the status (MSS, MSI-L, MSI-H)
   • For fusion: specify the two fusion partner genes
   • For pathogenicity: select one or more ClinVar classes and optionally restrict to specific genes
   • For treatment episode: select the treatment and the temporal relationship to the genomic finding
5. Optionally check Shared to make the criterion available to all researchers in the organization.

Genomic criteria integrate seamlessly with standard clinical criteria in the cohort builder. You can combine genomic filters with condition, drug, procedure, and demographic criteria to build precision medicine cohorts. For example: "Patients aged 18+ with stage IIIB NSCLC who carry EGFR L858R and have at least 365 days of follow-up."

---

Tumor Board Dashboard

The tumor board dashboard provides a comprehensive molecular evidence panel for an individual patient, consolidating genomic findings with clinical history in a single view. It is designed to support molecular tumor board discussions.

Accessing the Tumor Board

1. Navigate to Genomics > Tumor Board.
2. Enter a Person ID for the patient.
3. Select a Data Source.
4. The dashboard loads the patient's complete molecular and clinical profile.

Dashboard Sections

The tumor board display is organized into five sections:

1. Genomic Variants

A table of all variants for the patient, sorted by clinical significance (pathogenic first). Each variant shows:

• Gene symbol and HGVS notation (coding and protein)
• Variant type and functional consequence
• ClinVar significance with a link to the NCBI ClinVar entry
• Allele frequency and read depth
• Quality score

Pathogenic and likely pathogenic variants are highlighted for immediate attention.

2. ClinVar Annotations

For each annotated variant, the panel displays:

• ClinVar variation ID (hyperlinked to NCBI)
• Clinical significance classification
• Review status (e.g., reviewed by expert panel, criteria provided)
• Associated conditions from ClinVar
• Last evaluated date

3. Drug Exposures

A timeline of the patient's medication history from the OMOP CDM, showing:

• Drug name and class
• Start and end dates
• Duration of exposure

This helps the tumor board understand what treatments the patient has already received and how they relate to genomic findings.

4. Condition History

The patient's condition history from the CDM, including:

• Diagnosis codes and descriptions
• Condition start and end dates
• Condition type (primary, secondary, billing)

This provides clinical context for interpreting the relevance of genomic variants.

5. Treatment Timeline with Genomic Overlay

A visual timeline that overlays:

• Drug exposure periods (horizontal bars)
• Condition events (markers)
• Genomic variant detection dates (vertical lines)

This integrated view shows the temporal relationship between molecular findings and clinical events, which is critical for understanding whether treatments were selected based on genomic results and how the disease responded.

When to use the Tumor Board

The tumor board dashboard is most useful when reviewing individual patients with complex molecular profiles --- for example, patients with multiple actionable variants, patients who have progressed through multiple lines of therapy, or patients being considered for clinical trial enrollment based on molecular markers.

Research use only

The genomics module is designed for research and quality improvement purposes. It is not a validated clinical diagnostic tool. All clinical decisions should be made in consultation with qualified medical geneticists and using validated clinical-grade sequencing and interpretation pipelines.