How I learned to stop worrying and love the bias

Within a therapeutic beam of radiation, you can expect a homogeneous dose of absorbed energy until you approach the beam edge. By convention, at the beam edge, 50% of the absorbed dose is delivered. A few millimetres from the edge towards the central axis of the beam, it’s close to 100%. A few millimetres away, it’s close to 0%. The beam penumbra is that region of rapidly changing absorbed dose that extends from the central axis of the beam. In most situations, you’d want a sharp demarcation between what you want to treat and what you don’t, so a small penumbra is good. But what if you have to treat with two beams, side by side? You don’t want a gap, which would leave the junction between the two beams under-dosed. You don’t want overlap, which would give the tissues at the junction twice the prescribed dose. Aligned perfectly, with 50% at the beam edge for each, there would be neither gap nor overlap. But the slightest error, or patient movement, courts catastrophe.