Image matching cheat sheet

CT

The planning reference — high bone contrast, geometric accuracy

On CT, bone is bright white, air is black, and soft tissues are intermediate shades of gray. Hounsfield units make CT the dosimetry reference: the plan is calculated on the CT and all other images are fused to it.

Identify CT by crisp bony detail and true geometry — if bone is bright and anatomy looks "textbook," you are probably looking at a planning CT.

MRI

Unmatched soft-tissue contrast; bone cortex is dark

T1-weighted images: fat bright, CSF dark, anatomy bright and detailed. T2-weighted: fluid, edema, and most tumors appear bright. Cortical bone gives no signal (dark), which distinguishes MRI from CT instantly.

See soft tissue detail that CT cannot show plus a dark skull or dark vertebral cortex → MRI. Check for T1 vs T2 before choosing a target definition.

PET / PET-CT

Metabolic imaging — hot spots where the tumor is active

FDG-avid tissue appears as bright "hot" areas on the PET component, always fused with a CT for anatomy. Resolution is coarse; anatomy detail comes from the CT, and intensity is judged by SUV, not brightness alone.

Color-washed hot spots over a gray CT = PET-CT. Great for staging and target volume definition, never used alone for dose calculation.

CBCT

Daily linac imaging — volumetric, gray-scale, slightly noisy

Cone-beam CT is acquired on the treatment machine with a flat panel: gray-scale volumes similar to CT but with more scatter, lower soft-tissue contrast, and ring or streak artifacts. Used for daily setup verification and soft-tissue matching.

A gray volume with the patient in treatment position, mild noise, and artifacts around hips or shoulders → CBCT.

kV vs MV imaging

Planar checks — high contrast (kV) vs metal-tolerant (MV)

kV orthogonal pairs (on-board imager) give high bone contrast at low dose — the workhorse for daily bony setup. MV port films penetrate metal and show anatomy through hardware but carry a higher dose and lower contrast.

Choose the match structure by modality: kV for bony anatomy, MV when hip prostheses or hardware obscure kV images.

Ultrasound alignment

Non-ionizing daily soft-tissue matching for pelvis

Systems like Clarity capture 3D ultrasound volumes of prostate or gynecologic targets and register them to the planning CT. No radiation dose, ideal for daily soft-tissue alignment where organ motion dominates.

Probe pressure changes anatomy — keep probe placement consistent with the reference volume, and verify against periodic CBCT.

Bony vs soft-tissue matching

Which structure you register decides where the dose lands

Bony matching aligns anatomy landmarks; soft-tissue matching aligns the target itself (prostate, liver, lung). When the target moves relative to bone, a bony match can miss the tumor — fiducials, ultrasound, or direct soft-tissue match are preferred.

Always match the structure the plan was optimized on; document the registration structure so every therapist matches the same way.

Overlay alignment technique

Translate and rotate until the target overlaps, not the whole image

In overlay practice, apply shifts one axis at a time (left–right, superior–inferior, anterior–posterior), then rotation. Prioritize the target region over peripheral anatomy, and stay within the plan's rotational tolerance before accepting a match.

Fix the biggest offset axis first, then fine-tune; chasing global alignment while the target is off is the classic overlay error.

CBCT landmarks by body region

The structures therapists use to orient a scan in seconds

Head: skull contour, mask, airway, cervical vertebrae. Thorax: vertebral bodies, ribs, carina, heart. Abdomen: vertebrae, ribs, fiducials. Pelvis: pelvic ring, femoral heads, prostate fiducials, bladder-rectum interface.

Orient with large symmetric structures first (skull, pelvis, vertebral column), then refine with site-specific targets and fiducials.

Common matching pitfalls

Errors that quietly misalign a treatment

Rotational offsets left uncorrected, matching mobile anatomy to bone, couch vs gantry coordinate mix-ups, metal artifacts around prostheses, weight-loss anatomy changes in head & neck, and accepting an auto-match without visual verification.

Eyeball every auto-match against the target region and apply the documented action level before treating.