A lumbar or cervical spine MRI is rarely one uniform 3D block — it usually arrives as a sagittal stack that runs the length of the segment, plus several short axial stacks, each angled to a single disc. This viewer opens every one of those series straight from your DICOM folder or ZIP export, keeps the pixel data on your device, and lets you scroll, window, and measure the discs, canal, and nerve roots the way they were acquired. When you want a second set of eyes on a pattern you're seeing, request AI support built from the rendered images.
Upload Your Spine MRIA standard spine study is built from a handful of purpose-built stacks rather than one uniform volume. A sagittal T2 sequence runs the full length of the imaged segment in a single stack — it is what you use to judge disc hydration (a bright, water-rich nucleus versus a dark, desiccated one), the width of the spinal canal, and whether one vertebral body has slipped against the next. A sagittal T1 stack covers the same range and is read mainly for vertebral marrow — normal fatty marrow reads bright, while a fracture, infection, or metastasis replaces that brightness with darker signal. A STIR, or fat-saturated T2, sagittal stack suppresses marrow fat so that oedema and an acute fracture line stand out as bright regions that would otherwise be masked. Layered on top of those three long sagittal stacks are several short axial stacks, usually one per disc level, each individually angled parallel to that one disc space rather than continuing straight through the spine.
Spine Conditions on MRIDrop your MRI ZIP file or individual .dcm files. Supports all standard DICOM transfer syntaxes including JPEG Lossless.
Navigate through slices with keyboard arrows or the slider. Adjust windowing, zoom, pan, and measure distances.
4 AI models (GPT, Gemini, Claude, Grok) analyze independently. Claude synthesizes a unified report with findings and confidence levels.
Radiology reports describe spine findings by disc level, not by slice number, so the fastest way to get oriented is to work the same way the report does. Start on the midline sagittal T2 slice — the one where the spinous processes line up in profile — and count vertebral bodies down from a fixed landmark, such as the sacrum for a lumbar study or C2 for a cervical one. Each disc space between two numbered bodies has a name: L4-L5, L5-S1, C5-C6, and so on. Once you know which disc the report is discussing, jump from the sagittal stack into the short axial stack acquired for that specific level — the series description or slice location in the DICOM header usually identifies which axial stack belongs to which disc — and step through it left and right to see the same finding the report flagged from the transverse view.
A CT of the spine is usually acquired as a single isotropic volume, so it reconstructs cleanly into any plane after the fact. A spine MRI does not work that way. The sagittal stacks run straight down the segment, but the axial stacks are acquired level by level, each one individually tilted to sit parallel to its own disc space rather than perpendicular to one straight axis through the spine.
That per-level angulation is what lets each axial image sit flat against the disc it targets, but it also means the axial stacks do not line up into a continuous 3D block — there are gaps between levels, and neighboring axial stacks are not parallel to each other. The viewer still lets you page through every stack that was acquired and cross-reference the sagittal and axial views for a given level, but true multiplanar reconstruction across the whole spine needs an isotropic acquisition, which most routine spine MRI protocols do not use.
See how MPR and volume rendering work for isotropic scansBecause a spine study is really a set of stacks rather than one volume, the viewer treats each stack as its own scrollable series. You can move slice by slice through the long sagittal T2 or T1 stack with the arrow keys, then switch directly into whichever short axial stack corresponds to the disc level you are examining, without losing your place in the sagittal view. Windowing and leveling let you push contrast toward bone detail for a suspected fracture or toward soft tissue for disc and nerve assessment. The measurement tool works directly in millimetres, so you can mark the anteroposterior width of the canal or the height of a neural foramen instead of estimating it by eye. If you have an older spine MRI on file, you can open both studies and step through the same disc level in each to see whether a stenosis or a disc bulge has changed.
On an axial slice through a lumbar disc level, the thecal sac sits centrally inside the spinal canal, bounded by the vertebral body in front and the lamina and ligamentum flavum behind. Just off to each side of the sac, in the space called the lateral recess, the traversing nerve root separates from the sac before it curves outward. Farther laterally, between the pedicles of the two vertebrae that form the level, sits the neural foramen — the exit tunnel where the nerve root leaves the canal, and the structure that narrows first when a disc bulges posterolaterally or a facet joint hypertrophies. Reading these three landmarks together — central canal, lateral recess, and foramen — is what lets a single axial slice tell you whether a finding is central, subarticular, or foraminal, which is exactly the language a spine radiology report uses to localize a herniation or stenosis.
AI support in this app reads the rendered images and any text you provide and explains what terms like central stenosis, foraminal narrowing, or a Modic change generally mean — it is not a radiologist's read and does not replace one. Where it helps most is turning a dense spine report into questions you can bring to your surgeon or physiatrist: which level is driving your symptoms, whether the finding matches your pain distribution, and whether the next step is conservative care or further imaging.
Read how AI explanation differs from a radiology second opinionRaw MRI files stay on your device for DICOM parsing, rendering, and viewing. If you request AI analysis, the app uses rendered images and minimal context.
Works in any modern browser — Chrome, Firefox, Safari, Edge. No plugins, no Java, no desktop software to download.
Four independent AI models can explain rendered MRI images. Each model provides its own findings, then Claude synthesizes a consensus report.
Multi-plane navigation, windowing presets (bone, soft tissue, brain), zoom up to 4x, pan, and distance measurement in millimeters.
T1, T2, PD, STIR, FLAIR, gradient echo, and fat-suppressed sequences. Sagittal, coronal, and axial planes.
Generate a professional report with structured findings, severity grading, model agreement, and recommendations for your physician.
DICOM (.dcm) — individual slice files
ZIP archives — containing DICOM series folders
JPEG Lossless (1.2.840.10008.1.2.4.70) transfer syntax
Explicit VR Little Endian, Implicit VR Little Endian
8-bit, 12-bit, and 16-bit pixel depth
MONOCHROME1 and MONOCHROME2 photometric interpretation
Yes. Viewing is free with no cap on how many spine studies you open, and there are no hidden fees for scrolling, windowing, or measuring. AI analysis of a spine study draws from a paid credit ($10 per session) because it calls out to several AI models.
No. Your DICOM files are parsed and rendered locally with JavaScript and WebAssembly, so the sagittal and axial stacks never leave your device just to be viewed. Only the rendered images you choose to send are shared with AI models, and only when you request analysis.
A sagittal T2 stack for disc and canal detail, a STIR or fat-saturated sequence for oedema, and the axial stacks for each disc level together give the fullest picture. The viewer reads whatever sequences your study includes and does not require a specific protocol.
Yes. The viewer does not assume which segment of the spine was imaged — it opens whatever sagittal and axial stacks are in your DICOM series, whether that is a cervical study built around C2 through C7 or a lumbar study built around L1 through S1.
Four independent AI models review the rendered images separately, and Claude combines their output into one report along with a confidence level and a note on where the models agreed or disagreed. It is meant to support a conversation with a clinician, not to stand in for one.
Yes. You can open two studies and step through the same disc level in each to see whether a stenosis, herniation, or marrow change has progressed, stayed stable, or resolved since the prior scan.
Most routine spine MRI protocols acquire short axial stacks angled to each disc rather than one continuous isotropic volume, so there isn't a single 3D block to reconstruct. Isotropic joint scans used for 3D rendering are a different acquisition entirely.
Yes. The DICOM parser handles files from Siemens, GE, Philips, Canon, and Hitachi scanners at both 1.5T and 3T field strength, so the manufacturer of your spine MRI scanner should not matter.
Not sure your spine MRI export will open here? Check the supported formats reference for the full list of transfer syntaxes and bit depths.
On Analyze My Spine, the viewer helps you review spine imaging and understand disc, nerve, canal, alignment, and fracture language in reports.
Review herniated discs, stenosis, degenerative disc disease, sciatica-related nerve compression, facet-joint changes, ankylosing spondylitis, compression fractures, scoliosis, and spondylolisthesis.
Spine reports can sound alarming. These pages help distinguish anatomy descriptions from urgent symptoms that need clinician review.
© 2026 Analyze My Spine