Reading a gyroscopic plan

Almost nothing transfers from a C-arm write-up. There is no gantry angle, no couch angle and no MLC, so the settings a linac notebook spends its paragraphs on do not exist here.

The geometry. A 3 MV linac rides on two coupled ("yoked") gimbals, which between them sweep a spherical workspace rather than a plane. The reachable solid angle is 2√2π steradian — about 71% of the surface around the head — with over a thousand distinct non-coplanar directions available in routine use.

The aperture is a wheel, not a bank of leaves. Eight circular collimators, 4 mm to 25 mm at a 45 cm reference distance. Conformality is bought by how many isocentres you place and where, not by shaping a field — which is why isocentre placement is the planning decision here, and the thing the automation literature attacks first.

Nodes, not a continuous arc. Unlike Gamma Knife, which fires from many positions at once, each isocentre is treated from a small number of discrete nodes with differing beam-on times. A "beam" in the plan is a node, and the meterset distribution across nodes is the shape of the plan.

Where the error budget sits. Entry-angle accuracy matters far more for organs distant from the target than for the target itself — near the isocentre the geometry is forgiving and stops being so as you move out. That is the opposite of the intuition a C-arm planner brings.

Reading MU. Calibration is quoted as 100 MU with the 25 mm field giving 1 Gy at 450 mm SAD and 7 mm depth in water. An MU total compared against a linac plan means nothing without that; compare within the platform or not at all.