Ian Cowley, Addenbrooke's NHS Trust, Cambridge 28/June/2017 - London time is 09:57 (BST)
Research
The Techie Bit
Valid XHTML 1.0!
Valid CSS!
My Research Projects
IMRT Fluence Maps and Film Images | Patient Dosimetry | Patient Positioning | Dosimetry Accuracy
IMRT Fluence Maps and Film Images

The IMRT page explains how intensity profiles are built up for use in IMRT. One of my more frivolous projects has been to create some fun intensity profiles and treat them on our Siemens Primus linear accelerators.

Siemens linear accelerators have MLC leaves that project to 1cm wide at the radiation isocentre, so we only have a 1cm lateral resolution with any profile we wish to create. To make things easier, our segmentation software, ImFAST uses 1cm x 1cm squares to construct our beam intensities. The MLC bank has 27 pairs of leaves and a maximum over-centre travel of 20 cm, thereby giving us a maximum resolution of 21 x 27 'pixels' to play with.

Siemens 27-leaf MLC bank
A Siemens Primus 27-leaf MLC bank

The first image I tried to irradiate was the Cambridge University crest. Having recently graduated from Cambridge, and with the aspect ratio of the crest being about right, it seemed an ideal choice. So, I took the University crest, shrank it to 21x27 pixels and converted it to greyscale. With IMRT, we can choose any intensity we wish for each 'pixel', but I chose to keep the crest to 16 levels of grey to keep the treatment simpler.

University Crest ... University Crest ... University Crest
The University Crest...resized to required pixel size...and reduced to greyscale

I converted the greayscale image to a matrix of values from 0 t 15, wrote it out into a .map file and imported into ImFAST. ImFAST worked its magic and segmented the map into individual beams, then exported the beams to our new Siemens Primus-K linac. To record the irradiation pattern, I used the electronic portal imager on the linac, a Siemens BeamviewTI unit. The new BeamView software has auto-sequencing for recording each segment of the treatment and automatically storing it. This is a life-saver, as this treatment has 69 segments! I set everything up and fired away.
The treatment took about 12 minutes and involved a total dose of 800 monitor units - 100 MU is calibrated to deliver 1 gray of radiation to water (body-equivalent) at isocentre. A whole-body dose of 8 gray would kill a human being! The dose was only so high because some of the segments were only receiving 10MU, and the minimum exposure of the EPID is about 5MU.
Beamview automatically collected each segment and then produced a composite version - the final image I was looking for. So here we have it, the Cambridge University crest irradiated on a linear accelerator!

Irradiated Cambridge University Crest
The final result of the Crest Irradiation Project!

After I had created this image, Neil Burnet, one of our Consultant Oncologists saw it and thought it was great. He shows the image to the University of Cambridge's Regis Professor of Physic when he came to tour the Oncology department, and the Professor was instantly able to understand what IMRT is all about. Sufficiently chuffed by this, Neil wanted more fun images.
So Neil furnished me with a picture of his face, already scaled down to 21 by 27 pixels:

Neil Burnet's face
Neil Burnet, waiitng to be IMRTed

This one was a lot easier to do, having worked out how all the components of the process fit together, and, securing some time on the same linac, I set the treatment going. Again, I used BeamView to auto-record each image, but this time I kept the individual images. Click on the composite image below to see the individual segements play out in all their glory (3MB animation)!

IMRT irradiation of Neil Burnet
Neil Burnet's face, irradiated onto film and a BeamView portal imager with a 50-segment treatment, using 510 monitor units.
Back to Top
This page updated on 18 December 2015 at 22:21