Friday, 12 February 2016

A Snapshot of Radiotherapy Physics: 4DCT

The centre I work at is in the process of commissioning 4DCT for some radiotherapy planning, so this seemed like a good moment to try to explain what 4DCT is, and why we use it.

Let's start with the basics - 'standard' CT [computed tomography]. Sometimes called a 'CAT' scan [computed axial tomography], CT is used extensively in diagnostic radiology. It's the doughnut-shaped one in which an x-ray source and detector array spin around (usually around one rotation per second) while the couch on which the patient lies moves through the bore of the scanner. (Not to be confused with the doughnut-shaped one which is very loud, uses a magnet, and which many people find terribly claustrophobic - that's MRI [magnetic resonance imaging]).

The signal received by the detectors varies depending on what's between the x-ray source and the detectors - more or denser tissue between the two means less of the x-ray beam gets through the patient to the detector. The signal from each projection - every angle at which x-rays pass through the patient - is analysed and used to reconstruct regions of different density within the patient, forming axial (slice) images from head to toe - or whichever bits are required!

In radiotherapy, we use CT scans to plan treatment. The way radiation dose is deposited in tissue is determined largely by electron density (at least for megavoltage photon beams, which are used for the majority of radiotherapy), which is related to the density and composition of the tissue. That means that if we know the relationship between electron density and CT number (how bright or dark a pixel is on the CT image) for a CT scanner, we can use the information from the CT scan to accurately model the therapeutic radiation dose we give each patient.

So, that's a standard CT scan, and that's all that's required for the majority of patients. Some parts of the human body are constantly moving1 - particularly in the thorax, where all sorts of mobile things are located, including the oesophagus, diaphragm and lungs - and this can pose a problem when planning treatment of tumours in these areas. As I mentioned, CT tube rotation for standard protocols is usually around one second per rotation, and a ballpark figure for couch speed (speed of motion through the x-ray beam) is perhaps 1 mm/s - that will only give a snapshot of a lung tumour for a normal breathing rate of around 15-20 breaths per minute. All the detailed information the CT scan provides is specific to that point in the patient's breathing cycle, and breathing patterns - speed and range of motion - vary from person to person.

So how do we know where the tumour will be, if it's moving and we only have a snapshot of its position? Well, with a standard CT scan, we don't. One way of getting around this is to add a margin to the treatment volume to allow for the uncertainty in position - but that has to be done using information about the 'average' breathing pattern, and not all patients are average.

Enter 4DCT! The fourth dimension, of course, is time, added to the three dimensional CT scan we normally acquire. The couch speed is slowed down so that we can acquire additional information for each image slice - each position within the patient - information which covers more of the breathing cycle. Information about the patient's breathing during the scan is also acquired (by various methods, e.g. reflective markers on the patient's abdomen or using an elasticated belt), and used by the CT scanner to separate the images acquired according to the part of the breathing cycle.

Usually the breathing cycle is split into ten phases, giving us ten complete sets of CT images which cover the whole range of tumour motion. So instead of adding an 'average' margin to account for tumour motion, we can actually look at how the tumour moves, and plan treatment to cover the tumour over the whole breathing cycle.


1. OK, on some level, everything in the human body is constantly moving; in practical terms, though, it's a matter of how large, how much and how fast.


(An appalling absence of three years in updating this blog! One of my new year's resolutions for 2016 was to write more; specifically, to finish the Medical Physics series - if I could ever consider it finished, with the variety of things we do!)