Abstract
Purpose: :
To evaluate the intraoperator reliability and interoperator reliability of measurements with the Simader-Sponer-Testeye (SISPOT) Vienna Reading Center (VRC)
Methods: :
The SISPOT was constructed roughly following the characteristics of a Gullstrand eye, it contains a focusing screen with a circle of known size crossed by 12 radial lines. Nine different operators took 5 independent pictures with one SISPOT sticking to a tight protocol concerning image taking and labelling, at 3 different cameras (FF450 Zeiss, HRAII and Spectralis Heidelberg Ing.). Pictures were evaluated digitally using a software to mark 24 points on each picture, PIXS was then calculated automatically. The mean and standard deviation (SD) was calculated for all 27 sets of measurements to evaluate repeatability and for all measurements at the same camera.
Results: :
Values for PIXS of FF450 pictures varied between 3,68 µm/pixel (µm/p) and 3,72 µm/p. SD ranged from 0,01 to 0,02 µm/p. 99,73% of the pictures differed less then 2% in PIXS for all operators and 99,73% of all pictures taken differed less then 2% in PIXS indicating a very high intraoperator and interoperator reliability. SD of PIXS of the HRAII picture sets varied between 0,01 and 0,12 µm/p, indicating that it was more difficult to take pictures of consistent quality. For 5 experienced operators 99,73% of the pictures differed less than 2% in PIXS. For 3 operators 68,27% of the pictures differed less than 2% in PIXS. 1 operator reached 95% < 5%. 95% of all HRAII measurements differed less than 5%. The difficulties for less experienced testeye users may be overcome by protocol improvements. For measurements with the Spectralis 95,45% of pictures of all 9 operators differed less than 3% in PIXS. SD range 0,1- 0,9 µm/p. 95,54% of the pictures differed less than 3% in PIXS between the operators.
Conclusions: :
The SISPOT is a reliable instrument to judge pixel size of pictures taken with various camera types, independent of the operator, when a tight protocol is followed. This allows to draw conclusions about real lesion size within pictures taken with tested cameras, leading to more safety of decisions concerning patient inclusion/exclusion on digital pictures by reading centers.
Keywords: imaging/image analysis: clinical • clinical research methodology • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)