Overview

Smarter Diagnostics, Powered by AI

• PrediX delivers advanced, AI-driven solutions for radiology and pathology, helping health care professionals detect critical findings faster, reduce workload, and support more confident decision-making.

Modules

• Radiology – MRI Triage & ROI Guidance

  Our radiology module analyzes multi-sequence MRI scans to automatically highlight regions of interest, generate risk scores, and provide structured findings.

  Inputs: Classified images folders

  Outputs: Tumor probability of image and tumor classification

  Benefits:

  Accelerates triage and prioritization

  Reduces radiologist workload

  Enhances early and accurate decision-making

• Pathology – WSI Region Proposals & Second-Read

  AI-powered support for digital pathology.
  The pathology module identifies critical diagnostic regions in whole slide images and provides second-read assistance to pathologists.

  Inputs: Classified images folders

  Outputs: Tumor probability of image and tumor classification

  Benefits:

  Focuses attention on key regions

  Improves diagnostic accuracy and efficiency

  Fully compatible with leading digital pathology platforms

Intended Use

• Decision support for trained clinicians; not a replacement for clinical judgment.

Clinical Accuracy

• +90% accuracy; ongoing trials for lung, prostate, etc.

Integration Features

“PrediX enables Canadian hospitals to accelerate oncology diagnosis and treatment planning through AI-driven image classification. The platform is cloud-ready, PHIPA-compliant, and adaptable for integration into hospital IT systems.”

• Current functionality: accepts standard image files (JPG, BMP, PNG, TIFF) in classified folders to train and predict new cases.

• Optional integration modules (developed per site needs):DICOM / DICOMweb, HL7 v2, FHIR R4, SSO (SAML / OIDC), PACS / VNA connectors

  PrediX can be configured to connect with PACS, VNA, or EHR systems based on client requirements.

  Standard healthcare protocols (DICOM/DICOMweb, HL7 v2, FHIR R4, SSO via SAML/OIDC) are available as optional modules, to be developed per hospital platform.

  Current production release supports direct image file input (JPG, BMP, PNG, TIFF) via structured folders (e.g., pituitary, pineal, healthy).

  Flexible architecture allows customization for integration into each hospital’s IT environment.

Workflow Tools

• Real-time AI insights

• Automated reporting

• Customizable dashboards

• Training simulations

Safety

• Human-in-the-loop, confidence bands, OOD detection, versioned models, audit trails.

Key capabilities

• AI analysis for MRI and histopathology images

• Real-time region-of-interest feedback

• Adaptive learning with a simulation UI

• Simple desktop and REST API integration for hospitals and enterprises

Primary outcomes we target

• Fewer missed early-stage tumors

• More consistent interpretation between readers

• Faster turnaround under time pressure