The same MRI. The same radiologist. Two different reports.
Spine MRI is the most variable read in radiology. We fixed it — not by hiring better radiologists, but by writing down how spine MRI should be read.
Variability is the unspoken crisis in spine MRI.
Two radiologists. Different reports.
Disagreement on stenosis grading, disc classification, and nerve root compression is one of the most-cited problems in spine MRI. The cause is not skill — it is standards that are too high-level and too ambiguous.
The same radiologist, a week later.
The same radiologist reading the same spine MRI a week apart produces a meaningfully different report 15–20%+ of the time. Not carelessness — too much room for variability.
Grades change treatment.
A “mild” vs. “moderate” stenosis read changes whether a patient gets conservative care or a surgical consult. The variability is a clinical and operational problem hiding in plain sight.
Every aRAD report is read to a single, published, peer-reviewed clinical standard.
The Spine MRI Pathology Identification Standards are a clinical reading specification co-developed by RemedyLogic with a panel of leading US neuroradiologists and spine surgeons. For each major lumbar, cervical, and thoracic spine pathology, the standard defines: which MRI projection to use, which peer-reviewed grading system to apply, and exactly what each grade means — with explicit visual criteria.
Every aRAD radiologist reads to this standard. Every aRAD AI model is trained against this standard. The Clarity Report™ is the artifact that results.
What the standard covers.
Spinal canal stenosis
Four-point scale on axial and sagittal T2.
Lee grading systemForaminal stenosis
Four-point scale with explicit perineural fat criteria on sagittal T1/T2.
Four-point grade · explicit criteriaSubarticular stenosis
Graded with explicit nerve root contact and displacement criteria.
Nerve root contact criteriaSpondylolisthesis
Measurement-based grading with millimeter slip thresholds.
mm slip thresholdsDisc bulge, protrusion, extrusion, sequestration
Classified per published nomenclature with consistent shape and migration definitions.
Standardized nomenclatureNerve root compression
Graded by location and severity with specific anatomic landmarks.
Location + severity gradeDisc and vertebral height loss
Quantified with explicit ratio thresholds.
Ratio thresholdsFacet joint arthrosis
Graded with reproducible criteria, not narrative descriptors.
Reproducible gradeModic endplate changes
Classified per Modic with signal-based criteria.
Modic types I–IIIIncidental findings
Transitional vertebrae, conus termination level, aortic aneurysm — identification and measurement standards every read should address.
Identification + measurementThe full standard is a 51-page proprietary document. The website publishes its existence, the pathologies covered, and the categories of grading systems used. Imaging-center buyers, attorneys, and credentialed clinicians can request the full standard under NDA.
Co-developed with senior US neuroradiologists and spine surgeons.
The Standards were co-developed by RemedyLogic in collaboration with a panel of leading US neuroradiologists and spine surgeons drawn from institutions including the Hospital for Special Surgery, Mount Sinai, and other major US academic centers. Each pathology grading approach is anchored to peer-reviewed scientific literature.
Want to see the full Standards?
Imaging centers, radiology groups, and credentialed clinicians can request the full Spine MRI Pathology Identification Standards under NDA. Radiologist applicants receive the Standards as part of onboarding.