Tomotherapy: Optimized planning and delivery of radiation therapy

Abstract

Abstract In contemporary radiotherapy dose optimization, radiation beams and beam modifiers are iteratively selected until the dose distribution is acceptable. Another approach, referred to as the “inverse problem,” is: Given the dose prescription, compute the optimal set of photon beams while preventing unphysical solutions such as negative beam weights, and iterate to achieve the prescription as closely as possible. This solution to this inverse problem, which uses image reconstruction mathematics, entails the delivery of large numbers of nonuniform beam intensities to produce uniform dose distributions. These dose distributions can be arranged to conform very closely to even complex target volumes, yet spare surrounding sensitive tissue. Alternatively, the dose distributions can be arranged to generously treat a regional field and “conformally avoid” overtreating sensitive volumes within the field. Multiple dose prescriptions can be delivered without additional effort. We propose that a practical way of delivering optimized dose distributions would be to intensity modulate a photon beam, using collimator leaves intersecting a slit field of radiation. Modulation is achieved by varying the time that the leaves are blocking the field. A practical geometry to deliver such a beam is a computed tomography‐like gantry configuration, which also lends itself to tomographic setup verification of dose delivered to the patient. We refer to such a delivery method as “tomotherapy.” Several types of tomotherapy simulations have been conducted. A fully three dimensional optimized treatment planning system using iterative filtered back‐projection have been developed. We will present examples of conformal plans for breast and prostate radiotherapy. We have constructed an experimental apparatus for simulating helical tomotherapy delivery by simultaneously rotating and longitudinally translating a phantom past an intensity‐modulated fan beam. A comparison between a computation and and experimentally realized plan is presented. A Monte Carlo simulation of the angular distribution and energy fluence spectrum of 10‐MV photons produced by a tungsten target have been used to estimate the optimized shape and mass of a primary shielding required to meet regulatory standards.