Prostate fiducial marker matching explained
The prostate is a moving target. It drifts relative to the pelvic bones with bladder and rectal filling, so lining up the skeleton can leave the gland — and the dose — off position. Implanted gold markers solve this by giving a high-contrast surrogate you can match directly. This guide explains why seeds are used, how kV and CBCT fiducial matching works, and why three markers let you read a full 6DOF correction.
Educational overview. General background for radiation-therapy learners — not clinical advice or protocol.
On this page
The problem: the prostate moves off the bone What fiducial markers are Why three markers kV 2D/2D marker matching CBCT marker matching Reading a 6DOF correction Fiducial match vs bony match How to practice fiducial matching FAQThe problem: the prostate moves off the bone
Most setup verification aligns bony anatomy, which works well when the target is fixed to the skeleton. The prostate is not. It sits between the bladder and rectum and shifts by several millimetres — most often in the anterior–posterior and superior–inferior directions — as those organs fill and empty. Match the pelvis on a given day and the prostate can still be out of position. For a tightly conformal prostate plan, that residual error matters.
What fiducial markers are
A fiducial marker is a small, dense, radio-opaque object — typically a gold seed roughly a millimetre across and a few millimetres long — implanted into the prostate before planning, usually under transrectal ultrasound guidance. Because gold is far denser than tissue, each seed shows up as an unmistakable bright spot on any X-ray-based image. The markers move with the gland, so wherever the seeds are, the prostate is.
Why three markers
The number of markers determines how much of the patient's pose you can recover:
- One marker gives position only — translation in three directions.
- Three non-collinear markers define a rigid body: they pin down translation and rotation, and they add redundancy. If one seed is hard to see or has migrated slightly, the geometry of the other two still constrains the match.
Three is the common choice because it unlocks rotational information without over-complicating the implant.
kV 2D/2D marker matching
On a kV imaging system, two orthogonal radiographs (an AP and a lateral) are acquired and the three seeds appear as bright dots on each. The therapist aligns a planning marker triad — the seed positions from the planning CT — onto the imaged seeds in both views. Because the same three points are seen from two directions, their full 3D arrangement is constrained, and the system computes the shift that brings the planned markers onto today's seeds. This is fast and low-dose, and it is exactly the workflow recreated in the trainer's 2D/2D prostate case.
CBCT marker matching
With cone-beam CT, the seeds appear as bright voxels in the reconstructed volume. The registration is driven by the markers rather than the bony pelvis: you align the seeds across the three planes so the planned and imaged markers coincide. CBCT also lets you sanity-check the surrounding soft tissue and the rectum/bladder state at the same time. For the volumetric fundamentals, see CBCT 6DOF registration.
Reading a 6DOF correction
Three markers define a rigid body, so a marker match can resolve all six degrees of freedom:
Mathematically, the software finds the rigid transform that best maps the planned markers onto the imaged seeds (a least-squares point-set fit), then reports the couch correction. Whether all rotations are acted on depends on the couch and protocol, but the markers carry the information.
Fiducial match vs bony match
- Bony match: aligns the pelvis. Robust and simple, but it ignores prostate motion relative to the skeleton, so the target can be left off-position.
- Fiducial match: aligns the gland directly via its implanted seeds, following the actual target. This is why marker-based localisation is a standard of care for prostate IGRT.
Tip: if the markers and the bony pelvis disagree, trust the markers — that disagreement is the prostate motion you implanted the seeds to catch.
How to practice fiducial matching
The RT Image Matching Trainer includes prostate fiducial cases in both workflows: a kV 2D/2D case where you drag a three-marker triad onto the seeds in the AP and lateral views and the app solves a full 6DOF couch shift, and a CBCT prostate case with implanted gold seeds. Each applies a hidden offset you recover to within tolerance, with live feedback — the quickest way to build confidence reading a marker match before you're at the console.
Try the prostate fiducial caseDrag the marker triad onto the seeds; the trainer solves the 6DOF couch shift.
Open the trainerFrequently asked questions
Why are gold fiducial markers implanted in the prostate?
The prostate moves independently of the pelvic bones with bladder and rectal filling, so a bony match doesn't reliably localise it. Three gold seeds in the gland give a high-contrast, directly visible surrogate for the prostate on daily images, so the soft-tissue target can be matched precisely.
How does fiducial marker matching work?
On the daily kV 2D/2D pair or CBCT, the implanted seeds are aligned to the marker positions from the planning CT. Three non-collinear markers define a rigid body, so the software solves the translation and rotation mapping planned markers onto the imaged seeds and reads out the couch correction.
Why match the markers instead of the bones?
Because the prostate can shift several millimetres relative to the pelvis between fractions, aligning the bony pelvis can leave the target off-position. Matching the markers follows the gland itself, keeping the high-dose region on the prostate.
How many markers are used and why three?
Three is typical. One marker gives only translation; three non-collinear markers define a rigid body, so they also resolve rotation and add redundancy if one seed is hard to see or migrates.
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).