2D/2D portal vs DRR matching explained
Before volumetric imaging became routine, planar image matching was — and in many departments still is — the everyday way to verify a patient's setup. This guide explains what DRRs and portal images are, how a 2D/2D match works, why the correction is usually described as 5DOF, and when you'd reach for 2D/2D instead of CBCT.
Educational overview. General background for radiation-therapy learners — not clinical advice or protocol.
On this page
What is a DRR? What is a portal / setup image? Why "2D/2D"? Orthogonal pairs The 2D/2D matching workflow Why it's a 5DOF correction Which landmarks to match Tools that help 2D/2D vs CBCT How to practice 2D/2D matching FAQWhat is a DRR?
A DRR — digitally reconstructed radiograph — is a synthetic X-ray generated from the planning CT. The planning system casts virtual rays through the CT volume from a chosen beam direction and sums the densities along each ray, producing an image that looks like a conventional planar X-ray. Because it comes from the planning CT, the DRR represents the patient as planned. It is the reference image that the daily setup image is compared against.
What is a portal / setup image?
A setup image (historically a "portal" image) is a planar X-ray taken at the treatment machine. It can be acquired with the on-board kV imager (a diagnostic-energy source and flat panel) or with the MV treatment beam itself. kV images are now the most common for setup because they give better bony contrast at lower dose. Whatever the energy, the setup image captures the patient's position right now, on the couch.
Why "2D/2D"? Orthogonal pairs
A single planar image only tells you about position in the plane of that image — it can't resolve depth along the beam. To localise the patient in 3D you take two images from roughly perpendicular directions, classically an anterior–posterior (AP) and a lateral. Together this orthogonal pair pins down all three translation directions. Matching two setup images to two DRRs is where the name 2D/2D comes from.
The 2D/2D matching workflow
- Set up the patient to skin marks and lasers.
- Acquire the orthogonal pair (e.g. AP + lateral) at the machine.
- Overlay each setup image on its DRR. The software shows the live image against the reference.
- Align bony landmarks by translating and rotating until the anatomy in the setup image sits on the anatomy in the DRR. You read both images together, because a shift seen on one view is confirmed or refined on the other.
- Apply the resulting couch correction within tolerance, then treat.
Why it's a 5DOF correction
An orthogonal image pair reliably resolves five of the six degrees of freedom: the three translations plus two of the rotations.
The third rotation, yaw (couch rotation about the vertical axis), projects very little onto a front or side view, so it's difficult to measure from a planar pair. That's why a 2D/2D correction is normally expressed as 5DOF — three translations and two rotations.
Which landmarks to match
Planar images show bony anatomy best, so 2D/2D matching is fundamentally a bony match. Useful, high-contrast landmarks include:
- Skull / facial bones and sinuses for head & neck.
- Vertebral bodies, pedicles, and spinous processes for spine and many body sites.
- Pelvic ring, femoral heads, and pubic symphysis for pelvis.
- Carina and ribs as secondary references in the thorax.
Pick landmarks close to the target and stable from day to day, and match the same features you'd use to judge the plan.
Tip: read both views together. A translation that looks ambiguous on the AP image is often obvious on the lateral, and vice-versa — they constrain each other.
Tools that help
- Contrast / windowing on each image so bone stands out from soft tissue.
- Colour wash or overlay to see where the setup image and DRR agree and where they separate.
- A spyglass / split window to compare a local region of the two images edge-to-edge.
- Single-axis locks so you can adjust one direction at a time without disturbing the others.
2D/2D vs CBCT — when to use which
Neither is simply "better"; they trade off:
- 2D/2D is fast, low-dose, and excellent for routine bony verification. It can't show soft tissue well and doesn't resolve yaw.
- CBCT gives a full 3D volume, enabling soft-tissue matching and full 6DOF rotational assessment — at the cost of more imaging dose and time.
Many departments use 2D/2D for day-to-day bony checks and bring in CBCT when they need volumetric or soft-tissue information, or for higher-precision treatments. Knowing both is essential, which is why the trainer includes both workflows. For the volumetric side, see our companion guide on CBCT 6DOF registration.
How to practice 2D/2D matching
The fastest way to build the eye for a good match is repetition with instant feedback. The RT Image Matching Trainer lets you practice 2D/2D portal-to-DRR matching on brain, pelvis, thorax, and a monoisocentric breast case — dragging to translate and rotate an orthogonal pair, with single-axis "couch locks", contrast and overlay tools, and a live residual-error readout graded against tolerance. Generate a new random setup error and work it back to green.
Try a 2D/2D case nowOrthogonal AP + lateral matching, 5DOF correction, live error feedback.
Open the trainerFrequently asked questions
What is a DRR in radiation therapy?
A DRR (digitally reconstructed radiograph) is a synthetic X-ray computed from the planning CT by casting virtual rays through the CT volume from a beam direction. It looks like a planar X-ray and is the reference that setup images are matched against.
Why is 2D/2D matching usually a 5DOF correction?
An orthogonal pair (for example AP and lateral) reliably resolves the three translations plus pitch and roll. The third rotation, yaw (couch rotation), is hard to read from planar images, so 2D/2D corrections are typically expressed as 5DOF.
What is the difference between 2D/2D and CBCT?
2D/2D matches planar setup images to DRRs — fast and low-dose, but mostly bony anatomy in projection. CBCT produces a 3D volume allowing soft-tissue matching and full rotational assessment, at the cost of more dose and time. Departments often use 2D/2D for routine bony checks and CBCT when volumetric information is needed.
Educational use only. This guide is general background for learning, not clinical protocol — always follow your department's procedures. The trainer is not a medical device; its patient offsets and values are fictional, and its imaging derives from de-identified, openly licensed research datasets (see image credits).