ESTRO 36 Abstract Book

S932

ESTRO 36 2017

_______________________________________________________________________________________________

of agreement was noted for all tests (Table 1), of which

for geometric accuracy (inside length=0.4mm;

diameter=1.0mm), slice thickness accuracy (T1=0.3mm;

T2=0.5mm), slice position accuracy (T1 slice 1=1.4mm; T1

slice 11=0.5mm; T2 slice 1=0.9mm; T2 slice 11=0.6mm),

intensity uniformity (T1=0.0%; T2=0.1%), percent-signal

ghosting (0.0003) and low-contrast object detectability

(T1=2.9; T2=3.4) were all much smaller than the

corresponding ACR criterion. As illustrated in Fig. 1, all

data points fell within the 95% limit of agreement except

for diameter accuracy, for which 3 out of 54 (~5.6%) data

points fell outside the 95% limit of agreement.

Furthermore, small measurement bias close to zero was

also obtained for most tests. In terms of ICC, excellent

inter-observer agreement (ICC>0.75) was achieved in

geometric accuracy (ICC>0.99), spatial resolution (ICC =

1), slice position accuracy (ICC = 0.81), image intensity

uniformity (ICC = 0.80), percent ghosting ratio (ICC = 0.85)

and low-contrast object detectability (ICC = 0.89). A fair

inter-observer agreement was seen in the slice thickness

accuracy (ICC = 0.42). Based on both BA-analysis and ICC,

excellent inter-observer agreement could be achieved in

the ACR MRI phantom test under RT-setting.

Conclusion

Our results showed that ACR MRI phantom test under RT-

setting was highly reproducible and subject very little to

inter-observer

disagreement.

EP-1723 Optimisation of an Elekta XVI (R.5.0.2)

system for clinical protocols – image quality vs dose.

D. Oborska-Kumaszynska

, D. Northover

Royal Wolverhampton NHS Trust, MPCE Department,

Wolverhampton, United Kingdom

Purpose or Objective

The use of CBCT in radiotherapy has significantly

increased in recent times, which has led to an increase in

the concomitant dose received by some patients. Often,

the generic preset protocols provided by the

manufacturers are not optimised for a particular

department. This work aimed to optimise image quality

and dose for CBCT clinical protocols using an Elekta XVI

(R.5.0.2) machine for all clinically relevant treatment

sites.

Material and Methods

The Elekta XVI system was fully calibrated and Acceptance

Tests (AT) were performed for all FOVs before the

optimisation procedure. Three different phantoms were

used to complete the optimisation: CATPHAN 600, Phillips

WEP Phantom Set (PWEPPS) (5 circular objects 15, 20,

26.5, 36.5, 50cm diameter – Fig.1.) and Rando phantom

(RP). The optimisation methodology was designed to

assess dose vs the following image quality parameters:

spatial resolution (SR), uniformity (UN), contrast (CON),

CNR, SD, SNR and artefacts. These parameters were

evaluated as absolute values and compared to the

“standard” image results. These “standard” images were

taken for AT presets. The optimisation process was

performed by setting the exposure parameters: mA per

frame, ms per frame and gantry start/stop angles. The

first step involved taking CATPHAN images using varying

mA and ms settings. SR, UN, CON, CNR, SD and SNR values

were recorded. Final mA and ms settings were chosen

based on SNR and UN results, and were no worse than 20%

and 5% respectively in relation to the “standard image”.

Images were also compared using the same mAs but

different mA and ms values. The second step involved

taking PWEPPS images using the final mA and ms settings

for each protocol. SNR and UN were evaluated for

phantom diameters relevant to the treatment site in

question. The RP was used to assess image quality for the

finalised clinical protocols

Results

The optimisation process resulted in better image quality

in relation to the original presets and “standard images”.

Dose was reduced by a factor ranging from 2-4 times. For

a given mAs, superior image quality was seen for a higher

mA and lower ms, indicating that the detector response

was better for a higher dose rate. Saturation artefacts

(Fig.2) were visible for 64mA and 10ms when the images

included the intersection between the test object and air.

The worse UN was seen for LFOV. This was affected by

“cutting” from the reconstruction 40 pixel rows at the

edge of the panel. It was done because the bad pixel map

correction algorithm could not effectively correct the bad

pixels. Additionally, for 2D kV images, bad pixel artefacts

were visible using the TOR18FDG phantom. The kV

detector panel was replaced and the new one was

calibrated to get similar gains so the optimisation process