University Hospital Ljubljana Research Pilot

AI-Orchestrated
Hematologic
Diagnostics

Reducing bone marrow morphology review from weeks to hours through intelligent workflow orchestration.

Peer-Reviewed Publication
Academic Partnership
-95%
2-6 weeks → <24h
Time to diagnostic result
+200%
3-5×
Expert productivity increase
10-21 days
Faster treatment initiation
Critical Bottleneck

The Diagnostic
Capacity Crisis

Bone marrow morphology review is the gold standard for hematologic diagnosis, but manual microscopy creates severe capacity constraints.

40-60% Expert Time on Routine Triage

Hematopathologists spend half their time on straightforward cases that could be pre-screened.

Limited Expert Capacity

Global shortage of trained hematopathologists constrains diagnostic throughput.

2-6 Week Diagnostic Delays

Patients wait weeks for results, delaying critical treatment decisions.

Expert Time Allocation

Routine Triage 40-60%
Complex Cases 25-35%
Research & Training 15-20%

"Half of expert time is spent on routine work that AI could pre-screen, freeing capacity for complex diagnostics."

Integrated Workflow

Four-Module Diagnostic Platform

Orchestrated AI pipeline that augments expert capacity while maintaining clinical rigor

UScan

Digitizes bone marrow slides at clinical-grade resolution, creating standardized image dataset for AI processing.

Whole Slide Imaging 40× Magnification

Annotation Module

Expert-validated cell type labeling tool that builds ground truth dataset for model training and validation.

Expert Validation Quality Control
v1 v2 main

Nexus

YOLO/ResNet AI model that detects and classifies bone marrow cells, pre-screening routine cases for expert review.

YOLO Detection ResNet Classification

Compliance

Audit trail and regulatory framework ensuring clinical-grade quality and traceability for medical device compliance.

Audit Trail Regulatory Ready

Seamless Integration: From Sample to Secure Diagnosis

Four modules orchestrate end-to-end AI-assisted diagnostic workflow with expert validation at every critical decision point

UScan

Sample Digitization

Annotation Module

Expert Validation

Nexus

AI Classification

Compliance

Audit & Compliance

Peer-Reviewed Publication

Published Research

Academic validation of our AI-orchestrated diagnostic environment through peer-reviewed publication.

Publisher: PubMed Central (PMC)
Published: 2023
Institution: UH Ljubljana

P1352: Bone Marrow Morphology Diagnostic Environment Using YOLO/ResNet Artificial Intelligence Model

PMC10430733 • National Center for Biotechnology Information

Abstract

"We developed an AI-based diagnostic environment for bone marrow morphology evaluation using YOLO and ResNet architectures. The system demonstrates significant potential for reducing manual review burden while maintaining clinical accuracy."

Technical Architecture

YOLO
Object Detection
ResNet
Feature Extraction
RT DETR
Real-Time
Read Full Publication on PubMed
Measurable Impact

Clinical Impact Metrics

Validated improvements in diagnostic workflow efficiency and patient care timelines

<24h
Time to Result
vs 2-6 weeks traditional
3-5×
Productivity Gain
Expert capacity multiplier
10-21
Days Saved
To treatment initiation
40-60%
Routine Cases
Pre-screened by AI

Patient Journey Timeline

Traditional Workflow

Day 0

Sample Collection

Bone marrow biopsy performed

Day 7-14

Manual Review

Expert microscopy analysis

Day 14-42

Final Report

Diagnostic conclusion

AI-Orchestrated Workflow

Day 0

Sample Collection + Scan

UScan digitization

Day 0

AI Pre-Screening

Nexus automated triage

<24h

Expert Validation + Report

Focused review of flagged cases

Result: 95% reduction in time to diagnosis, enabling earlier treatment and improved patient outcomes

Join the Research

Research Partnership Opportunities

We're seeking academic and clinical collaborators to advance AI-orchestrated diagnostics

Academic Institutions

Universities and research labs advancing AI in clinical diagnostics

  • Co-authorship opportunities
  • Dataset access for research
  • Model improvement collaboration

Clinical Centers

Hospitals and labs implementing AI-assisted workflows

  • Platform deployment support
  • Clinical validation studies
  • Expert panel participation

Technology Partners

AI/ML teams and medical device companies

  • Model architecture innovation
  • Integration partnerships
  • Regulatory compliance support

Express Your Interest

Reach out to discuss research collaboration opportunities