Radiographic calibration for pubic diastasis assessment in bladder exstrophy-epispadias complex: a phantom study

Abstract

Abstract Background The assessment of pubic diastasis is important for the surgical planning of patients with bladder exstrophy-epispadias complex. Understanding how the diastasis changes during surgical follow-up may help predict patient morbidity. Radiography can follow diastasis but may be affected by patient and technical imaging factors including body size, imaging protocol, and equipment. Using imaging calibration and anatomic ratios may mitigate differences due to these aspects. Objective Use imaging phantoms to assess the effect of radiographic calibration on measurements of pubic diastasis and an internal anatomic ratio as a child grows. Materials and methods Radiographic images were obtained of three different sizes of computed tomography phantoms (older child, child, and infant) using three imaging techniques that include the osseous pelvis in children. All phantoms were imaged with abdomen and pelvis techniques. The infant phantom was additionally imaged using a thoracoabdominal technique. These exposures were all repeated with systems from three manufacturers. Linear measurements were made between radiographic markers placed to simulate pubic diastasis and sacral width. A ratio was also created between these distances. Measurements with and without image calibration were made by two pediatric radiologists using rulers placed at the time of image acquisition. Results There was excellent interrater agreement for measurements, ICC >0.99. Anterior distances were more affected by magnification than posterior ones with a significant difference between uncalibrated versus calibrated anterior distances (p=0.04) and not for posterior ones (p=0.65). There was no difference between radiographic equipment manufacturers without or with calibration (p values 0.66 to 0.99). There was a significant difference in simulated pubic distance between thoracoabdominal and abdomen (p=0.04) as well as pelvic (p=0.04) techniques which resolved with calibration, each p=0.6. The ratio between the simulated pubic diastasis and sacral width differed by phantom size (all p<0.01) and imaging technique (p values 0.01 to 0.03) with or without calibration. However, the numerical differences may not be clinically significant. Conclusion Image calibration results in more uniform measurements that are more accurate than uncalibrated ones across patient size, imaging techniques, and equipment. Image calibration is necessary for accurate measurement of inter-pubic distances on all projection imaging. Small differences in the pelvic ratio likely are not clinically significant, but until there is a better understanding, image calibration may be prudent. Graphical abstract