A .dcm file will not open in Windows Photos, macOS Preview, or a browser drag-and-drop, and that trips up more people than it should. This page is the straight answer: what is actually inside a DICOM file, the methods that convert it to JPEG or PDF, which of those methods are safe to run on real patient studies, and where lossy conversion stops being appropriate.
Why DICOM will not just open like a photo
DICOM (Digital Imaging and Communications in Medicine) is not an image format the way JPEG or PNG are. A single .dcm file bundles the pixel data together with a header of metadata: patient name and ID, study and series identifiers, the modality that acquired it, and display parameters like window center and width that decide how the pixel values should be mapped to what you see on screen. Some files hold one image; others hold a full stack of slices or a multi-frame cine loop in a single file. A general-purpose image viewer has no idea what to do with any of that, which is exactly why it fails silently or shows a blank thumbnail instead of a picture.
Opening a DICOM file for what it actually is takes DICOM-aware software: a desktop or web DICOM viewer, or the archive that sits behind one. We cover that split in detail in the DICOM viewer guide; this page is about the narrower, more mechanical question of getting a DICOM image turned into an ordinary JPEG or a PDF someone without DICOM software can open.
Four ways to convert DICOM to JPEG or PDF
1. A free desktop DICOM viewer, exported by hand
The lowest-friction route for a one-off file is a free desktop viewer with a built-in export or save-as function. Open the study, set the window and level the way you want the image to look, then export the current frame or the whole series as JPEG or PNG. Three viewers cover most of what people already have installed, and we have an honest comparison page for each:
- Horos, free, Mac only, with a straightforward export panel per series.
- MicroDICOM, free, Windows, exports single images or whole series to JPEG, PNG, BMP or TIFF.
- RadiAnt, paid, Windows, exports images and can also print a study straight to PDF through the OS print dialog.
This route is fine for a handful of files. It does not scale to converting hundreds of studies, and the exact export options differ per app, so check what window and level got baked into the output before you send it anywhere.
2. Command line: dcmtk and ImageMagick, for batches
For more than a few files, or anything you want to script,dcmtk is the standard free open-source toolkit. Its image-export tool (called dcm2img in current dcmtk releases, still often referenced by its older name dcmj2pnm) converts one DICOM instance to a bitmap file:
dcm2img +oj +Wm study.dcm study.jpg
+oj writes an 8-bit lossy JPEG; +Wm tells it to compute a sensible window automatically from the pixel data when the file has no window baked in. Loop it over a directory to convert a whole folder of instances in one pass:
for f in *.dcm; do
dcm2img +oj +Wm "$f" "${f%.dcm}.jpg"
doneImageMagick can also read DICOM directly through its own DICOM coder, which is convenient for a quick one-liner but gives you far less control over windowing than dcmtk:
magick study.dcm study.jpg
To get a PDF out of a batch of converted images, ImageMagick combines JPEGs into a single multi-page PDF directly:
magick *.jpg study.pdf
This is the right route when you are converting many files regularly, or need the process to be repeatable rather than a series of manual clicks in a viewer.
3. Online DICOM to JPEG or PDF converters
Free online converters exist, and they work fine on a file with no real patient in it. They are the wrong tool for an actual clinical study. A DICOM file's metadata carries patient name, date of birth and other identifiers regardless of what the pixels show, and a free web converter is an unvetted third party you do not have a HIPAA business associate agreement with. Uploading a real patient's scan to one is a genuine compliance exposure, not a hypothetical one. Save this route for de-identified test files or teaching material, and use a local method instead for anything with a real patient behind it.
4. From inside a PACS: skip the conversion step entirely
If the study already lives in a PACS, the fastest path is usually not converting a file at all, it is using the export the archive already has built in. In MiniPACS, every series in the browser viewer carries its own download menu with two choices side by side: DICOM for physicians, or JPEG Images for patients, so a patient- ready JPEG copy of one series is one click away rather than a file you run through a converter. A whole study can also come out as a ZIP or a burned CD/USB image, with an option to include or exclude the report, when the receiving side specifically needs the original DICOM rather than JPEG.
Reports go the other direction: build one in the report editor and it previews and downloads as a real PDF at any stage, draft or signed, on the clinic's own letterhead, plus a plain-text copy for pasting straight into another system, with nothing to render or convert by hand.
And often the actual goal is not a converted file at all, it is getting a colleague or a patient to see the images. A share link with a QR code opens the study directly in a browser, images and report both viewable inline, with DICOM, JPEG and report-PDF download buttons already sitting on the other end, no conversion step for anyone.
Lossy vs lossless: when JPEG is the wrong answer
Every method above that writes a standard JPEG is a lossy conversion, and that matters for more than file size. Diagnostic DICOM pixel data is commonly 12 or 16 bits per pixel, with window and level values that decide what range of that data gets mapped to visible gray. A baseline JPEG export flattens that down to 8 bits baked to one fixed window, and a multi-frame or multi-slice study collapses into single flat pictures with no ability to scroll or rewindow afterward. None of the information lost in that flattening comes back.
That is a fine tradeoff for a referral letter, a patient handout, an email attachment, or a teaching slide, all cases where the image only needs to be seen once, at whatever window it was exported at. It is not an acceptable tradeoff for the actual diagnostic read. That reading has to happen in a real DICOM viewer against the original pixel data, where window and level can still be adjusted and every slice or frame is still there. Converting a study to JPEG for someone to look at is normal. Reading a case off the JPEG instead of the DICOM is not.
A PDF built from a study's images inherits the same limits as the JPEGs it was built from; a PDF of a written report is a different document made of text, not flattened pixel data, and does not carry that tradeoff.
FAQ
How do I convert DICOM to JPEG?
Four honest routes, in order of how often you will actually want each one. For a single study you already have open, a free desktop DICOM viewer like Horos, MicroDICOM or RadiAnt has an export or save-as function that writes a JPEG per image or per series. For a batch of files, or to script the job, the command-line tool dcmtk converts one file at a time (dcm2img +oj +Wm study.dcm study.jpg) and a short loop handles a whole folder. Avoid uploading real patient studies to a random online converter; see the HIPAA question below. If the study already lives in a PACS, check whether it can export JPEG directly instead of converting by hand, which is what MiniPACS does per series.
Is there a free DICOM to PDF converter?
For turning DICOM pixel data into a PDF of images, dcmtk plus ImageMagick is the free command-line route: convert each instance to JPEG with dcm2img, then combine the JPEGs into one PDF with ImageMagick's convert. There is no single dedicated free DICOM-to-PDF utility we would point you to instead. If what you actually want is a study's diagnostic report as a PDF, that is a different document from the images, and in MiniPACS every report has a PDF preview and download built in, no conversion step needed.
Can I convert DICOM online safely?
Only if the study has no real patient identifiers in it. A DICOM file carries name, date of birth, and other identifying fields inside its metadata even when the pixels look anonymous, and a free online converter is an unvetted third party with no signed HIPAA business associate agreement. Uploading an actual patient's study to one is a real compliance problem, not a theoretical one. Use an online converter for de-identified test files or teaching images only. For real patient data, convert locally with a desktop viewer or dcmtk, or use the export tools inside your PACS.
Is DICOM to JPEG conversion lossy, and is it safe for diagnosis?
Yes it is lossy, and no, a converted JPEG should never be read for diagnosis. A DICOM instance can carry 12 or 16 bits of pixel data plus the window and level settings that decide what that data should look like on screen. A standard JPEG export flattens all of that into an 8-bit image baked to one fixed window, and multi-frame or multi-slice studies collapse to single flat pictures. That is fine for a referral letter, a patient handout, or a teaching slide. It is not fine for the actual reading, which has to happen in a real DICOM viewer against the original pixel data.
Does MiniPACS convert DICOM to JPEG or PDF?
Yes, without a separate conversion step. Every series in the browser viewer has its own download menu with two choices: DICOM for physicians, or JPEG Images for patients, so a JPEG copy exists one click away, not a file you convert by hand. Reports come out the other way already: build one in the report editor and it previews and downloads as a real PDF at any stage, draft or signed, plus a plain-text copy for pasting into another system. For getting images to someone else entirely, sharing a link is usually faster than converting anything: the guest portal linked to a study lets the recipient view the images and the report inline in a browser, and download DICOM, JPEG, or the report PDF themselves, with no conversion on your end at all.