24.5. RT Plans

Table of Contents

Treatment planning information

In clinical radiotherapy, an RT Plan is a structured description of how a treatment machine should deliver radiation to a patient. It is created by a treatment planning system, stored in DICOM RT Plan format, and then sent to the treatment machine. When you use GATE for dose calculations or research on treatment techniques, the RT Plan becomes the bridge between the clinical world and your simulation setup.

This chapter focuses on what information an RT Plan typically contains and how that information is conceptually used when you build a GATE simulation. The practical details of parsing DICOM files are covered elsewhere, so here you only need to understand the structure and purpose of the planning data.

What an RT Plan represents

An RT Plan describes the full treatment prescription in machine coordinates and times. It links together the patient geometry and contours (from CT and RT Structure Set) with the technical parameters of the beams that will deliver dose.

For a single external beam plan, the RT Plan tells you which machine is used, how many beams exist, their directions, shapes, and weights, and how these parameters may change over time. For complex clinical treatments like IMRT or VMAT, the RT Plan contains a sequence of control points that define the dynamic behavior of gantry, collimators, and multi‑leaf collimators, as well as the dose rate and cumulative monitor units.

In GATE, you typically translate these elements into source configurations, beam geometry, and time‑dependent motions so that the Monte Carlo simulation mimics the clinical plan as closely as possible.

Links between RT Plan and other DICOM objects

The RT Plan does not stand alone. It is designed to be used together with DICOM CT images and RT Structure Sets.

The RT Plan references the CT series used in planning, so you know which patient geometry the plan belongs to. It references the RT Structure Set, which contains the contours of organs, targets, and other regions of interest. It also includes the dose prescription to one or more target volumes, often using the same region names as in the structure set.

When you build a GATE simulation, you combine these: the CT provides voxelized geometry, the structure set gives semantic meaning to regions (such as tumor and organs at risk), and the RT Plan provides how beams are oriented and modulated in time. Only together do they fully reproduce the clinical scenario.

Core RT Plan content

An RT Plan file is organized into nested sequences. At a high level, the most relevant groups for GATE users relate to:

Plan identification and prescription.
Fractionation and total dose.
Beam definitions.
Control points for each beam.
Machine and technique information.

You rarely need every detail for Monte Carlo. Instead, you identify the parameters that must map to GATE configuration, such as beam direction, field size, jaw and MLC positions, isocenter coordinates, and total monitor units.

A useful way to think about the RT Plan is as a list of beams, where each beam has its own list of time‑ordered control points. Each control point describes how the machine looks at a specific moment. In static beams there is usually a single control point pair that fully describes the beam, while in modulated techniques there can be dozens or hundreds.

Beam geometry and isocenter

The beam geometry section of an RT Plan defines how each beam is oriented relative to the patient and the machine isocenter. This information is essential for reproducing clinical beams in GATE.

For each beam you typically find:

The isocenter position, given in the patient coordinate system. This is the point about which the gantry and couch rotate and is usually the reference point for field shaping and dose prescription. In GATE, you use this as a reference for placing the source and collimation system relative to the patient geometry.

The gantry angle, which defines the rotation of the X‑ray or particle beam around the patient. This tells you from which direction the beam enters the patient. In many plans, the gantry angle may vary over the course of delivery.

The collimator angle, which defines the rotation of the collimator system about the beam central axis. This determines the orientation of jaws and multi‑leaf collimator leaves with respect to the patient. Mapping this angle correctly in GATE is important whenever you model detailed collimator geometry.

The couch (or patient support) angle, which defines the rotation of the patient table around its vertical axis. When non‑zero, the patient is effectively rotated relative to the incoming beam. In Monte Carlo simulations you can implement this as a rotation of either the patient geometry or the beam orientation, but you must be consistent with your coordinate system.

Together, these angles and the isocenter location define the beam central axis in 3D space. In GATE, the beam direction vector is derived from the gantry and couch rotations, and the source position is defined relative to the isocenter along this axis.

Field size and collimation

An RT Plan describes how the beam is shaped before it reaches the patient. This shaping is achieved with jaws and optional multi‑leaf collimators.

Jaws are large collimator blocks that define a rectangular field. In the RT Plan, jaw positions are usually given in millimeters at isocenter distance, referenced to the beam central axis. You can interpret these as the half‑widths of the rectangular field in the beam coordinate system. If you only approximate the field in GATE, these jaw positions alone are often sufficient to define a simple rectangular field.

A multi‑leaf collimator uses many leaves that can move independently to create complex field shapes. In the RT Plan, each control point stores leaf positions, again in millimeters at isocenter distance. Leaves are grouped into banks, usually labeled A and B, which move from opposite sides. These positions define which parts of the beam aperture are open or closed.

When you create detailed collimator geometry in GATE, jaw and leaf positions from the RT Plan are mapped to the translation or extent of geometry objects representing jaws and MLC leaves. For static fields, this mapping is constant. For dynamic techniques, it is time‑dependent and controlled by the sequence of control points.

In some clinical plans there are also additional collimation devices, such as custom blocks or compensators. These are sometimes referenced in the RT Plan but may not be fully described in a standardized way. For Monte Carlo, you may need to approximate them or model them separately as custom geometry that matches clinical documentation.

Beam weights, monitor units, and fractionation

RT Plans link the geometric description of beams to the intended dose through beam weights and monitor unit values.

Each beam has a prescribed number of monitor units, which is the quantity used clinically to control the linac output. Monitor units are proportional to the delivered dose at a reference point in a reference geometry. For a Monte Carlo simulation, the absolute value of monitor units can be translated into an effective number of primary particles, scaling factor for fluence, or simulation time, depending on your approach.

Beams can also have relative weights to indicate how the total plan dose is distributed among them. For example, in a plan with two beams, one might carry 60 percent of the dose and the other 40 percent. This is encoded by their monitor units or explicit weight fields. When you simulate the plan, you can reflect this by assigning proportionally more primary histories to the more heavily weighted beams.

Fraction information describes how many treatment fractions will be delivered and how the total dose is split across them. An RT Plan typically contains the number of fractions and the dose per fraction for each target. In GATE, this fractionation does not usually change the geometry or physics, but it affects how you interpret the simulated dose. If you simulate one fraction, you scale the results by the number of fractions to estimate the total course dose, or vice versa.

In dose simulations, monitor units and beam weights must be consistently converted into source intensity or number of primary particles. Incorrect scaling leads to wrong absolute dose values, even if the geometry and physics are perfectly modeled.

Control points and dynamic delivery

Modern techniques like IMRT and VMAT rely on dynamic delivery where gantry angle, collimator positions, and leaf positions change continuously during irradiation. In an RT Plan this motion is captured as a sequence of control points for each beam.

Each control point represents the machine state at a specific cumulative fraction of the monitor units. You typically see:

A cumulative meter set value, which tells you what fraction of the total monitor units has been delivered up to this control point.

The gantry, collimator, and couch angles at that point. For static beams these angles do not change, but in VMAT they usually vary smoothly across control points.

Jaw positions and MLC leaf positions at that point. In step‑and‑shoot IMRT the leaves change only between static segments. In sliding window and VMAT, leaves move continuously, and their positions at control points are used to approximate the motion.

When you translate this into a GATE simulation, you usually discretize the delivery. You can model each control point, or groups of control points, as a separate simulation segment with fixed geometry and a beam weight proportional to the monitor unit increment between control points. For time‑dependent simulations, you can also represent the evolution explicitly as a series of time intervals using GATE motion capabilities.

The more control points you include, the more accurately you reproduce the dynamic delivery, but the more complex the simulation becomes. A balance is often needed between accuracy and computational cost.

Doses and reference points in the RT Plan

RT Plans contain prescription information that links the geometric plan to clinical objectives. Common elements include:

Total prescribed dose for a course of therapy.

Dose per fraction.

Names of target volumes that should receive the prescription dose. These typically correspond to structures in the RT Structure Set, such as PTV or CTV.

Reference points or dose calculation points where the treatment planning system has calculated expected doses. These points can be used to perform quick checks of Monte Carlo results by comparing simulated dose to planned dose at specific locations.

In GATE dose simulations, you usually use this information to check if your Monte Carlo reproduction of the plan is correct. For example, after running a simulation with beams and fractionation derived from the RT Plan, you can compare the dose at the plan reference point or within the PTV volume with the values in the clinical plan.

When validating simulations against clinical RT Plans, compare doses using consistent definitions, such as dose per fraction or total course dose, and ensure that coordinate systems for reference points and structures are correctly aligned before drawing conclusions.

Coordinate systems and transformations

The RT Plan uses the same patient coordinate system as the CT image and RT Structure Set. However, beam geometry is often described in a beam‑centric coordinate system. When you build a Monte Carlo simulation, you must correctly transform between these coordinate systems.

The key transformation concepts are:

Patient coordinates, typically defined as x from patient right to left, y from anterior to posterior, and z from inferior to superior, with known orientation tags in the CT series.

Machine coordinates for the gantry and collimator, defined relative to the treatment machine. The RT Plan provides angles that express how the beam coordinate system is rotated relative to the patient coordinate system.

Beam coordinates, where the z‑axis is along the beam central axis, and x and y define the lateral and vertical directions in the field.

In GATE, you align the patient geometry using the CT orientation, then you place the beam using the RT Plan angles and isocenter position. Correct interpretation of these rotations is essential, particularly in treatments with non‑zero couch angles, because any mismatch will change where dose is deposited in the patient geometry.

Using RT Plans in GATE simulations

Once you understand the structure of RT Plans, the workflow to use them conceptually in GATE is:

Interpret the CT and RT Structure Set to build the patient geometry and define organs and targets.

Extract from the RT Plan the beam list, including beam types, gantry, collimator, and couch angles, isocenter coordinates, field sizes, jaw positions, and MLC leaf positions.

For each beam, interpret monitor units, beam weights, and control point cumulative meter set values to define how many particles to simulate for each segment or time interval.

Translate each static or dynamic segment into GATE geometry and motion definitions, ensuring that field shaping, beam direction, and source parameters correspond to the clinical plan.

Map prescription and reference doses into quantities that can be compared against Monte Carlo dose actor outputs to validate that the simulation reproduces the clinical intent.

By treating the RT Plan as a precise technical description of the treatment machine behavior, you can build GATE simulations that closely mirror clinical delivery and support reliable dose verification, research on advanced techniques, and development of new treatment strategies.