Patent Pending · MIMIC-IV Validated

Clinical AI that detects deterioration before traditional systems can.

DiviScan analyzes dozens of organ-system endpoints simultaneously through a 9-layer diagnostic engine, identifying patient deterioration significantly earlier than conventional early warning scores. Built on multi-knowledge-base inference, validated against MIMIC-IV critical care data.

0.815–0.890
MIMIC-IV AUC Range
1M+
Phenotype-Gene Associations
5
Trained Disease Models
$134M
Trial Rescue Simulation

9-Layer Diagnostic Engine

Not a single-variable scoring system. DiviScan layers biomarker ingestion, multi-knowledge-base querying, phenotype mapping, ML inference, and utility-weighted decision logic into a unified clinical intelligence pipeline.

Layer 01

Biomarker Ingestion

Multi-source lab and vitals intake with normalization across organ-system panels. Structured for downstream inference.

Layer 02

Panel Analysis

Organ-system panel decomposition — hepatic, renal, cardiac, metabolic, hematologic — with cross-panel correlation mapping.

Layer 03

Knowledge Base Query

Parallel lookup across HPO, OMIM, and clinical ontologies. 1M+ phenotype-gene associations for differential enrichment.

Layer 04

Phenotype Mapping

Translates abnormal results into standardized phenotype terms. Maps biomarker deviations to known clinical phenotypes.

Layer 05

ML Inference

Five trained disease models (CKD, liver disease, diabetes, sepsis, heart failure) generate probability scores from feature vectors.

Layer 06

Utility Gate

Proprietary decision layer applying utility-weighted thresholds — not raw probability cutoffs. Context-sensitive alerting that reduces alarm fatigue.

Layer 07

Cross-Organ Synthesis

Integrates signals across organ systems to detect multi-system deterioration patterns invisible to single-variable monitors.

Layer 08

Risk Stratification

Composite acuity scoring combining ML outputs, utility gates, and temporal biomarker trajectories into actionable risk tiers.

Layer 09

Clinical Reporting

Structured output with organ-level breakdowns, confidence intervals, phenotype-gene evidence trails, and suggested next-step diagnostics.

Beyond Single-Variable Scoring

Traditional early warning systems rely on a handful of vital signs. EHR-embedded tools apply raw probability thresholds. DiviScan operates on a fundamentally different architecture.

Capability NEWS / MEWS Epic Sepsis Model DiviScan
Input Variables 6–7 vital signs EHR data fields Dozens of organ-system endpoints
Scoring Method Aggregate threshold Single-model probability Utility-weighted multi-model
Knowledge Bases None None HPO, OMIM, clinical ontologies
Disease Coverage General deterioration Sepsis only 5 models + cross-organ synthesis
Alert Logic Fixed threshold Probability cutoff Utility Gate (context-sensitive)
Phenotype Inference No No 1M+ phenotype-gene associations
Alarm Fatigue High High Reduced via utility weighting

Two Markets, One Engine

DiviScan's multi-layer architecture serves distinct but complementary clinical intelligence needs.

Pharmaceutical

Trial Rescue Intelligence

Identify why trials fail before they fail. DiviScan's simulation engine models biomarker trajectories across trial arms, detecting subpopulation signals and endpoint risks that traditional DSMB reviews miss.

  • Simulate biomarker-driven enrollment stratification
  • Detect subpopulation responders in failing Phase II/III trials
  • Model endpoint sensitivity across organ-system panels
  • $134M value demonstrated in trial rescue simulation
Hospital / Health System

Early Deterioration Detection

Move beyond NEWS/MEWS. DiviScan processes routine lab panels and vitals through its 9-layer engine to surface multi-organ deterioration patterns before they become critical events.

  • Continuous multi-organ-system monitoring from existing lab data
  • Utility-weighted alerts that reduce alarm fatigue
  • Phenotype-backed evidence trails for clinical decision support
  • Validated against MIMIC-IV critical care dataset (AUC 0.815–0.890)

Simulation-Backed Validation

DiviScan's performance is validated against real-world critical care data. We are transparent about our stage — these are demonstration-level results with clear clinical signal.

0.815–0.890
MIMIC-IV AUC Performance
Area under the curve across five disease models validated against the MIMIC-IV critical care database — a widely used benchmark in clinical ML research.
$134M
Trial Rescue Simulation
Modeled value recovered by identifying salvageable subpopulations in simulated Phase III trial failure scenarios using biomarker-driven stratification.
1,000,000+
Phenotype-Gene Associations
Queried from HPO and OMIM knowledge bases during inference, enabling genotype-informed differential diagnosis enrichment.
USPTO PROVISIONAL

DiviScan's utility-weighted diagnostic methodology is protected under USPTO Provisional Patent #63/706,309. The core innovation — utility gate scoring applied to multi-knowledge-base clinical inference — is proprietary to RESSS Global Holdings LLC.

Software Demonstration with Simulation Validation

We believe in transparency. DiviScan's HyperCore OS is operational software with demonstrated capabilities — not a concept deck. Here is exactly where we stand.

  • HyperCore Diagnostic OS — Operational. 9-layer engine processing biomarker inputs through multi-knowledge-base inference.
  • 5 Disease Models — Trained and validated against MIMIC-IV. CKD, liver disease, diabetes, sepsis, heart failure.
  • Trial Rescue Module — Demonstrated. $134M simulation with biomarker-driven subpopulation stratification.
  • Clinical Reporting — Generating structured organ-level reports with evidence trails.
  • Clinical Pilot — Seeking hospital and pharma pilot partners for prospective validation.
  • DiviScan Device (Phase 2) — Hardware platform planned. Point-of-care saliva-based diagnostics with NV-Diamond sensing.
Technology Stack
Engine HyperCore OS
Models 5 Disease + Utility
Knowledge Bases HPO, OMIM, ClinVar
Validation MIMIC-IV
IP Protection USPTO #63/706,309
Target Partners
Pharma VP Clinical Dev / CMO
Hospital CMIO / VP Informatics
Stage Pilot Ready

Explore a DiviScan Pilot

Whether you're evaluating trial rescue capabilities or early warning infrastructure, we'd like to show you what DiviScan can do with your data.

Or email us directly at diviscan-gh-os@polsia.app

Inquiry Received

We'll review your submission and respond within 2 business days.