đź’ŠLevel 2: Therapies & Trials

What platinum/PARPi/DDR therapy should we give next?

Top 5 drug recommendations ranked by S/P/E framework. Mechanism-based matching with transparent scoring. Validated AUROC 0.70 (n=149).

Once you have CSI score, unlock drug recommendations and clinical trial matching. S/P/E framework (validated AUROC 0.70, n=149) ranks therapies by mechanism fit for DDR-targeted treatments.

The Problem: Generic Drug Ranking for DDR-Targeted Therapy

For patients with advanced, heavily pretreated cancer, clinicians need to know: ...

Problem

No Mechanism Understanding

Generic mutation lists don't explain why a DDR-targeted drug fits

Problem

No Transparent Reasoning

No clear scoring to understand why PARPi ranks higher than platinum for a specif

Problem

No Clinical Trial Matching

Missing mechanism-based trial matching means patients miss DDR-targeted combinat

The Solution: CSI-Powered Drug Recommendations with S/P/E Framework

Once you have CSI score, unlock drug recommendations and clinical trial matching

âś“ Validated

S/P/E Framework

Sequence (30%) + Pathway (40%) + Evidence (30%) for comprehensive drug ranking

âś“ Validated

Confidence & Tiers

Evidence tiers (Supported/Consider/Insufficient) with confidence scores and badg

âś“ Validated

Insights Integration

4 insights chips provide confidence lifts when biological signals align

How Level 2 Works

Four steps to unlock this capability

1

Take CSI Score from Level 1

Input: CSI score from Level 1
2

Add Genomic Test Results (NGS)

Input: Genomic test results (NGS)
3

Calculate S/P/E Scores

Method: S/P/E framework: 30% Sequence, 40% Pathway, 30% Evidence
4

Rank Drugs and Match Clinical Trials

Formula: efficacy_score = 0.3Ă—S + 0.4Ă—P + 0.3Ă—E + ClinVar_prior
Complete Level 2 to Unlock:
8-pathway transcriptomic scoring (EXHAUSTION, TIL, T_EFFECTOR, ANGIOGENESIS, etc.)
3x responder enrichment (10-15% → 30-50%)
KEYNOTE-158 proxy validated: delta +0.358 (3.5x threshold)
STK11/KEAP1-loss identification for IO-refractory routing

CSI in Action: Continuous Monitoring

See how CSI updates automatically as tumor evolves. Track chemosensitivity across treatment lines with real-time alerts when CSI drops below threshold.

2

Level 2: Genomic Monitoring

Add genomic testing to detect resistance mutations and find clinical trials

What You Can See:
CSI Score, CA-125, Imaging, ctDNA, Active Trials
Data Required:
+ Genomic test results (NGS)

CSI Continuous Monitoring

Track CSI score updates as tumor evolves. Never miss a chemosensitivity change. • Patient AK

LIVE• Level 2 of 5 Unlocked
CSI Score
72/100
Decreased from 78

ChemoSensitivity Index

3/6/2026

CA-125
900
Plateau detected

Tumor Marker

2/28/2026

Imaging
No progression
Last scan: 3 weeks ago

CT Scans

2/14/2026

ctDNA
KRAS G12D: 0.8% VAF
New mutation detected

Liquid Biopsy

3/6/2026

Active Trials
3 matches
New trial added

Clinical Trials

3/5/2026

Resistance Alerts
Locked
Unlock at Level 3
Requires resistance prediction

Active Alerts & Insights

3 Active(1 locked at Level 3)

CSI Score Decreased

Your CSI score dropped, meaning chemo is becoming less effective

CA-125 Plateau

Your CA-125 stopped declining and is staying flat

New Trial Match

A clinical trial that matches your specific tumor biology just opened

Unlock at Level 3: Resistance Prediction
Resistance Mutation will be available
Unlock at Level 5: Complete Care Plan
Full monitoring timeline showing progression across all treatment lines will be available
3

Next: Level 3 - Resistance Prediction

Get early warnings 3-6 weeks before imaging shows treatment failure

Unlocks:
Resistance Alerts
Level 2 Validation

Level 2 Performance Metrics

Validated performance for this capability

Retrospective Tested

Validation Status

Validated performance for this capability

Ready for Level 3?

What resistance class is active right now? Monitors 10 resistance classes validated across 680 patients from 6 independent datasets with temporal ctDNA modeling.

Unlocks:10 resistance class detection (BRCA reversion, ABCB1 efflux, SLFN11, lineage plasticity), 6 independent datasets (ARIEL, Patch, Christie, TCGA-OV, Abbott, MSK-SPECTRUM)
Continue to Level 3