Los combos y bandas como producto de la cultura

5.3.1 In tro d u ctio n

MRI relies on the presence o f hydrogen nuclei in the tissue being examined. A

powerful magnet aligns the hydrogen nuclei and a radio-frequency pulse is applied

deflecting the net magnetization o f the hydrogen nuclei in the tissue. The excited

nuclei spin around the axis o f the magnetic field and produce an electric signal

detected in a receiver coil. The decay o f the signal as the nuclei return to equilibrium

is monitored. The return to equilibrium is called magnetic relaxation and is a unique

property o f each tissue, described by T1 and T2 relaxation times. These are important

determinants o f image contrast and signal intensity in MRI.

MRI shows increased signal intensity with the inflammatory changes associated with

pelvic RT. A characteristic pattern o f varying oedema post-radiation has been

described [Blomlie et al., 1996]. Irradiated soft tissues also show enhancement with

intravenous gadolinium contrast[Fletcher et al., 1990]. Dynamic gadolinium enhanced

MRI has been used to show changes in the vascularity o f pelvic tumours (cervical

cancer) [Gong et al., 1999] and the subsequent response to RT but there are no reports

o f the utilisation o f dynamic MRI in the quantitative assessment o f post-RT changes

in the anorectum.

The MR scanner available for this work was a Siemens™ Magnetom™ open magnet

scanner. Fig. 5.3. This utilises a relatively low field strength magnet (0.2 Tesla) and

conventional higher field strength scanners (1.5 Tesla). However, the open scanner

has the advantages o f easy access and is less likely to make patients feel

claustrophobic.

MR scanning cannot produce an absolute density value for a specific tissue density

such as a Hounsfield number that can be applied in CT scanning. However, dynamic

gadolinium enhanced MR .scanning can be used to quantify the degree o f

enhancement o f a specific anatomical site. Enhancement is dependant on the rate o f

delivery o f the contrast, blood flow and vascular permeability. Following injury and

inflammation there is increased blood flow and vascular permeability resulting in

increased uptake o f contrast. Irradiation o f the anal canal resulting in inflammation

might be detectable by increased enhancement on dynamic gadolinium enhanced T1

weighted MRI.

Any detectable enhancement in the rectum following dynamic gadolinium enhanced

MRI would also be o f relevance and interest. However this was considered unreliable

because the resolution o f the MRI images obtainable with the MRI scanner available

for this work and the thickness and variable position o f the rectal wall impact on the

Fig. 5.3 O pen m agn et M RI scan n er

5.3.2 Methods

D ynam ic gadolinium enhanced MRI o f the p elv is w as performed in m en with carcinom a o f the prostate (25 ) and bladder (3 ) b efore, 6 w eek s and in 10 patients 6 m onths after RT. W ith the patient lying supine on the table, a 4 5cm diam eter b ody coil w as p osition ed around the p elvis, with the inferior border o f the coil at the upper

level o f the sym p h y sis pubis. A Vasculon^^ 20 G auge IV cannula (B ecton D ick inson )

w as sited in the ante-cubital fossa and attached to a 3 -w a y tap. A fter an initial scout scan axial, coronal and saggital im ages w ere taken and these were centred over the anal canal.

The dynamic sequences were aligned over the anal canal from the previous axial and

coronal images. The dynamic sequence involved 20 acquisitions, each consisting o f

three saggital slices, centred on the anal canal. Each acquisition o f three slices took 11

seconds to complete. Immediately after the end o f the second acquisition lOmls o f

DOTAREM ™ (gadoteric acid 279.3mg/ml Guerbet Laboratories Ltd) was injected as

rapidly as possible via the ante-cubital fossa cannula. After completion o f each scan

images were saved and stored on an optical disk. On completion o f all the scans the

anal canal slices that were most anatomically central were collated into individual

series and sent to a computer via the DICOM ( Digital Imaging and Communications

in Medicine) server. DICOM is the universal data language for imaging technology.

Two elliptical regions o f interest (ROI) each o f I94.4mm^, were identified in the anal

canal in each series. One ROI was situated in the lower anal canal and one in the

J

Fig. 5.4 Saggital M R I sh o w in g up p er and low er anal ROI (oran ge ellipses)

N um eric signal data throughout each scan w ere created from each ROI and the data collated on spread sheets. The data were plotted and the degree o f enhancem ent (calculated as final enhancem ent-enhancem ent at tim e zero) and the rate o f enhancem ent over the first 55 second s (calculated as the gradient o f the enhancem ent curve using M icrosoft E xcel^^) were determ ined for each ROI on each scan. The results (d ifferen ces b etw een before and after RT) w ere tested for norm ality using a Shapiro-W ilk test. The statistical sign ifican ce o f d ifferen ces betw een m easurem ents w as tested using either paired t-tests i f the data were parametric and norm ally distributed or a W ilcox on signed-rank test i f the data w ere non-parametric.

The data were further utilised to identify evidence o f a dose related response. As

discussed in Chapter 3 there was a large variance in the estimated anal canal dose

between patients (from 4.2 to 83.7% o f the total treatment dose - see Patient Dose

Data Set -Table 3.6). Anal canal dose (Appendix 5.f) was therefore plotted against the subsequent change in the rate and degree o f anal canal enhancement after

radiotherapy to identify any correlation between dose and enhancement.

The set-up for each scan in this study was strictly consistent. Positioning o f the patient

and coil, orientation o f acquisitions, injection o f gadolinium and outlining o f the

regions o f interest were all personally performed/supervised by the author. An

automated injection o f the gadolinium was not available. This would have allowed

precise timing o f the delivery o f the contrast, although the variability in delivery time

would be negligible. This with other minor set-up variations should contribute only

random and not cumulative error to the results.

5.3.3 Results

An example o f the enhancement curves produced for the lower anal canal in one

30 0