The AI-native operating model

A radiology workflow built from scratch for the AI era.

Most teleradiology groups bolt AI onto legacy systems. We did the opposite. We started with the AI and built every layer of the workflow around it — DICOM viewer, PACS, RIS, reporting. The result is a radiologist experience that is faster, more accurate, and built for the role radiology is becoming.

The thesis

The radiologist’s job is changing. So is the workflow.

For three decades, the radiologist has been the producer of the report. They identified findings, measured them, dictated them, and signed them. Every layer of the imaging stack — PACS, DICOM viewer, reporting software — was built for that producer model.

AI changes the role. The radiologist is becoming the reviewer: confirming, refining, and signing findings that a purpose-built foundation model has identified and measured first. That shift is not a feature you can add to a 25-year-old workflow. It requires a different operating model.

aRAD was built for that future. Every screen, every keystroke, every report template assumes the AI saw the case first.

Four layers, all purpose-built

One stack, engineered around the AI.

Layer 01 — DICOM Viewer

A viewer designed for AI-first reads.

Most viewers are designed for the radiologist to find findings. Ours is designed to verify findings the AI has already identified. Annotations are surfaced, not searched for. Measurements are pre-populated, not drawn by hand.

Layer 02 — PACS

Cloud-native, AI-aware archive.

Engineered to ingest, route, and version both the original DICOM and the AI-annotated derivatives. Every study has a complete audit trail of what the AI saw, what the radiologist changed, and what was signed.

Layer 03 — RIS

Scheduling built around subspecialty.

Spine MRI studies route automatically to a spine-focused radiologist on shift. No subspecialty queue confusion. No misrouting to a generalist.

Layer 04 — Reporting

Structured reports, signed in minutes.

The Clarity Report™ schema is enforced at the template level. The AI populates findings, measurements, and key images. The radiologist reviews, edits, and signs.

The workflow

From study upload to signed report.

Step 01 — Upload

Studies route in.

Spine MRIs route to aRAD through your existing PACS or DICOM connection.

Step 02 — AI annotates

The model measures first.

The foundation model identifies, segments, and measures every relevant structure and grades pathology to a published standard before the radiologist opens the case.

Step 03 — Review & sign

A spine specialist signs.

A board-certified, spine-focused radiologist confirms findings and signs the Clarity Report™. Delivered in hours, not days.

The engine

A purpose-built spine MRI foundation model. The first FDA-cleared.

aRAD is built on RemedyLogic’s spine MRI foundation model — a purpose-built vision-language model trained specifically on spine MRI, and the first FDA-cleared AI in the United States for spine MRI interpretation. It identifies vertebrae, segments discs, measures displacements, detects stenosis, grades pathology to published clinical standards, and benchmarks anatomical aging against a reference population.

It is not a generic radiology AI retrofitted to one body part. It was engineered for spine MRI from the first line of code.

Powered by RemedyLogic AI — the first FDA-cleared AI for spine MRI interpretation in the US.

Foundation model · spine MRI
✓ Vertebrae identification & labeling
✓ Disc segmentation & displacement (mm)
✓ Canal & foraminal stenosis grading
✓ Listhesis measurement
✓ Pathology graded to published standard
✓ Anatomical SpineAge™ benchmark
HIPAA

HIPAA-compliant infrastructure end-to-end.

Encryption

Encrypted in transit and at rest.

Access

Role-based access controls.

Audit

Full audit trail on every study.

FDA

FDA-cleared AI components.

See the workflow on one of your own cases.

A 20-minute call. We walk through the workflow and read one of your spine MRIs so you can compare it to your current vendor.