Influence of anatomical structure on antiscatter grid performance in 2D: application to x-ray angiography and a...

Objective.The use of uniform phantoms to assess the influence of x-ray scatter and antiscatter grids on x-ray angiography and fluoroscopy image quality disregards the influence of spatially variable x-ray attenuation of patients. The purpose of this work was to measure scatter to primary ratio (SPR) and antiscatter grid SNR improvement factor (K_SNR) using experimental conditions which better mimic patient imaging conditions.Approach.Three adult-sized anthropomorphic phantoms were used. AP and lateral projection images of the thorax and abdomen were acquired with and without an antiscatter grid. Grids with ratio 15:1 and 29:1 (r15, r29) and x-ray fields of view 20, 25 (thorax) and 32, 42 cm (abdomen) were tested. Combined with a-priori measurements of grid scatter and primary transmission fractions, these images were used to calculate 2D SPR andK_SNRmaps.Main results.Results demonstrated that measurements by uniform phantom do not describe the complex 2D SPR andK_SNRdistributions associated with anthropomorphic phantoms. The regions of the images with the lowest primary x-ray intensity (greatest attenuation) had the highest SPR and the highestK_SNRattributable to the grids. Considering all conditions, the 95th percentile of the SPR maps was in the range 42%-185% greater than the median values and that of theK_SNRmaps was 4%-20% higher than the median values. The combined influences of SID 120 vs. 107 cm and r29 vs. r15 grid resulted inK_SNRin the range 1.05-1.49.Significance.Performance of anti-scatter grids using anatomically complex phantoms highlights the substantial variation of SPR andK_SNRwithin 2D images. Also, this work demonstrates the benefit of the prototype r29 grid for thoracic and abdominal angiography imaging conditions is substantial, especially for large patients and radiodense image regions.