How to convert an mri dicom scan to a pdf report
- Step 1Select the key slice on your workstation — MR sequences are many slices; this tool converts one. On your viewer, scroll to the representative slice (the one you'll cite in the report) and confirm the contrast looks right — the window you set is what should be baked into the exported instance.
- Step 2Set the window, then export that single instance — Apply your T1/T2/FLAIR window on the workstation and export the slice as an uncompressed single DICOM instance. The converter reads
WindowCenter/WindowWidthfrom the file, so a windowed export reproduces your view; compressed (JPEG2000/JPEG-LS) exports will not render correctly. - Step 3Drop the .dcm onto the converter — Upload the single
.dcm(or.dicom) file. It acceptsapplication/dicomand converts one file at a time — there is no series-upload mode. - Step 4Let it render automatically — There are no options to set; the page is fixed A4 portrait and rendering starts on drop. The grayscale slice is mapped through the stored window to 0–255 and placed on the page with the metadata caption beneath it.
- Step 5Verify the series caption — Check the caption shows the correct Series description and Study Date — this is what tells the reader the image is T2 rather than T1. Patient identifiers are also printed here, so the output is PHI.
- Step 6Insert into the report or share — Download the
.dicom-to-pdf.pdfand attach it to your report, or combine several key-slice PDFs into one figure sheet with PDF Merge. De-identify with the PII redactor if the recipient shouldn't see the patient name.
MR-specific rendering behaviour
How an MR instance is interpreted. All values come from the file; nothing here is adjustable in the UI.
| Aspect | What happens | Why it matters for MR |
|---|---|---|
| Slices per conversion | One slice → one page | An MR sequence is many slices; convert the key slice, not the stack |
| Contrast | Stored WindowCenter / WindowWidth applied | A windowed export reproduces your T1/T2/FLAIR view; an unwindowed one falls back to C 128 / W 256 |
| Rescale | RescaleSlope / RescaleIntercept applied before windowing | MR rescale is usually 1/0, so values pass through; harmless when absent |
| Grayscale polarity | MONOCHROME1 inverted, MONOCHROME2 normal | Both render dark-background as clinical viewers do |
| Series label | SeriesDescription printed in the caption | Lets the reader identify the sequence (T1 / T2 / FLAIR / DWI) |
| Bit depth | 16-bit read as Int16/Uint16 per PixelRepresentation | MR is typically 16-bit; rendered to 8-bit grayscale via the window |
Free vs Pro for MR files
MR instances vary in size; the per-file cap is the practical gate.
| Limit | Free | Pro |
|---|---|---|
| Max file size | 2 MB | 50 MB |
| Typical 256×256 16-bit MR slice | ~0.13 MB raw — fits free | Fits |
| Typical 512×512 16-bit MR slice | ~0.5 MB raw — fits free | Fits |
| Files per run | 1 | 1 |
| Output | One-page PDF | One-page PDF |
Cookbook
Representative MR reporting scenarios. Window values and identifiers are illustrative.
A T2 axial brain slice for the report figure
The radiologist windows a T2 axial on the workstation, exports the single slice, and converts it. The caption's Series description confirms it's T2.
Input: t2_ax_023.dcm 0.5 MB, MR, 512x512, MONOCHROME2
Action: drop on converter (auto-runs)
Output: t2_ax_023.dicom-to-pdf.pdf 1 page
caption: Modality MR, Series "AX T2 FSE"Building a multi-sequence figure sheet
To show T1, T2 and FLAIR side by side, convert each key slice separately, then merge. The tool never pages a whole series on its own.
Convert: t1_key.dcm -> t1_key.dicom-to-pdf.pdf
t2_key.dcm -> t2_key.dicom-to-pdf.pdf
flair_key.dcm -> flair_key.dicom-to-pdf.pdf
Merge: PDF Merge -> mri_figure_sheet.pdf (3 pages)Exported slice had no window baked in
If the export tool stripped the window tags, the converter falls back to a generic window and the slice can look flat. Re-export with the window applied.
Input: mr_nowin.dcm no WindowCenter/Width tags Output: rendered with default C 128 / W 256 Fix: set window on workstation, re-export the slice
Whole-sequence DICOMDIR or multi-frame mistake
Dropping a multi-frame MR object renders the first frame only. For a per-slice document, export the slices you need individually.
Input: mr_multiframe.dcm (enhanced MR, many frames) Output: 1 page = first frame only Wanted per-slice? Export each slice, convert, merge.
Compressed MR export renders as noise
Some PACS default to JPEG2000 for MR archives. That pixel data isn't decompressed here.
Input: mr_j2k.dcm transfer syntax JPEG2000 Result: noise instead of anatomy Fix: re-export as uncompressed Little Endian
Edge cases and what actually happens
Expecting a page per slice
By designAn MR sequence is many slices, but this tool converts one frame to one page. There is no series-to-pages mode. Convert each key slice and combine with PDF Merge for a multi-image figure.
Window not stored in the instance
Default appliedMR contrast is interactive on the workstation; if the export didn't write WindowCenter / WindowWidth, the converter uses C 128 / W 256, which often looks flat on 16-bit data. Apply the window before exporting so it's baked into the file.
Compressed transfer syntax
Renders incorrectlyJPEG2000, JPEG-LS, and RLE pixel data are not decompressed and render as noise. Re-export the slice as uncompressed (Implicit/Explicit VR Little Endian).
Multi-frame enhanced-MR object
First frame onlyEnhanced MR can store a whole sequence as one multi-frame instance. Only the first frame is rendered. Export individual single-frame instances for the slices you want.
File over the tier cap
BlockedFree allows up to 2 MB; Pro up to 50 MB. A single 16-bit MR slice is usually well under 2 MB, so most MR conversions fit free — but if a workstation exports a thick multi-frame or oversampled instance, it may need Pro.
Wrong series picked
Check captionBecause only the file you drop is converted, picking the wrong slice gives a clean-looking PDF of the wrong image. The caption's Series description and Study Date are your verification — confirm them before sharing.
Patient name printed on the figure
PHI exposureThe caption renders Patient and Patient ID onto the page. For a de-identified figure, redact with the PII redactor or de-identify the DICOM before converting — the metadata scrubber won't remove rendered caption text.
Not for diagnostic reading
Communication onlyThe output is a JPEG reproduction at one window on a fixed page. It supports the written report; it does not replace reviewing the full multi-sequence study on a calibrated display.
Invalid or renamed file
ErrorA non-DICOM file (e.g. a screenshot saved as .dcm) throws "Could not parse DICOM file". For ordinary images use Image to PDF.
Frequently asked questions
Will my workstation window/level be preserved?
The converter applies the WindowCenter / WindowWidth stored in the exported instance, not the interactive window you last set in the PACS session. If you apply the window and then export the slice, the file carries those tags and the PDF reproduces your view. If the export strips them, it falls back to center 128 / width 256.
Can I convert an entire MRI sequence at once?
No. It converts one slice to one page. For a multi-image figure, convert each key slice and merge them with PDF Merge.
How do I know which sequence the PDF shows?
The caption prints the Series description (e.g. "AX T2 FSE", "FLAIR", "DWI b1000") along with the Study Date and Modality. That's the field to check when assembling figures from several sequences.
Are there sliders or rendering options?
No. There is no window slider, no series picker, and no page-size option. Conversion auto-runs on drop with the file's own settings on a fixed A4 page. Control the appearance by windowing and choosing the slice before export.
What if my MR is JPEG2000-compressed in PACS?
It will render as noise — the parser doesn't decompress encapsulated pixel data. Export the slice as uncompressed Little Endian and convert that.
How large can an MR file be?
Up to 2 MB on free, 50 MB on Pro. A single 16-bit MR slice (256×256 or 512×512) is typically a fraction of a megabyte, so most MR conversions fit the free tier.
Is anything uploaded to a server?
No. Parsing and rendering happen in your browser with dicom-parser and pdf-lib; MR pixel data and identifiers never leave the device.
Can I annotate the figure with findings?
Add marks after conversion: the PDF Sign tool draws onto a page, or overlay a graphic with PDF Overlay / Stamp. The DICOM converter itself doesn't add annotations.
Why does my slice look washed out?
Almost always a missing or generic window. 16-bit MR data mapped through the default C 128 / W 256 looks flat. Set the proper window on the workstation and re-export so the tags travel with the file.
Is this PDF acceptable in a formal radiology report?
It's suitable as an embedded figure for communication. The diagnostic interpretation must still be made from the full study on a calibrated display — the PDF is a single windowed JPEG reproduction, not measurement-grade imaging.
How do I de-identify the figure?
Convert, then redact the patient name and ID with the PII redactor; or de-identify the DICOM at source. The caption is rendered into the page, so clearing PDF document properties won't hide it.
Can I script this for routine reporting?
Yes. dicom-to-pdf runs on the @jadapps/runner (Pro, runner-builtin): GET /api/v1/tools/dicom-to-pdf for the schema, pair the runner, then POST to 127.0.0.1:9789/v1/tools/dicom-to-pdf/run. The slice is processed locally.
Privacy first
All PDF processing runs locally in your browser using PDF-lib and pdf.js. No file is ever uploaded — only metadata counters are saved for signed-in dashboard stats.