Radioprotection No 59-3 | Page 61

204 H . Sekkat et al .: Radioprotection 2024 , 59 ( 3 ), 203 – 210

precise dose estimation tailored to pediatric patients ( Frush et al ., 2003 ; Zacharias et al ., 2013 ).

Accurately evaluating radiation risk mandates precise estimation of absorbed radiation doses by pediatric patients . Common methods for quantifying radiation dose in CT exams , such as volumetric computed tomography dose index ( CTDIvol ) and dose length product ( DLP ), lack consideration for individual patient size variations critical for accurate dose estimation . CTDIvol and DLP , derived from measurements with cylindrical phantoms , fail to precisely reflect the actual radiation dose received by a patient , as highlighted in previous studies ( AAPM , 2011 , 2014 ).

To determine individual patient radiation doses , the American Association of Physicists in Medicine ( AAPM ) introduced Specific Size Dose Estimation ( SSDE ), involving normalizing CTDIvol with size-correction factors like effective diameter ( Deff ) or water-equivalent diameter ( Dw ). While effective diameter provides a straightforward measure of patient size , it overlooks patient composition and tissue attenuation . Dw , accounting for X-ray attenuation in patients , offers a more accurate method for estimating patient doses ( AAPM , 2011 , 2014 ; Amalaraj et al ., 2023 ). However , measuring Deff and Dw for each patient can be impractical in busy CT centers . As a pragmatic alternative , these values can be estimated using the patient ’ s weight or body mass index ( BMI ), as proposed in previous studies ( Menke , 2005 ; Boos et al ., 2018 ). Age-based estimation is another practical approach , though it may have some variance . Some studies have provided data tables correlating Deff with the patient ’ s age , such as ICRU 74 and AAPM 204 . Additionally , there are reports of a correlation between Deff and age for abdominal examinations , recognizing that the average patient diameter varies with age . However , predicting an individual patient ’ s size based solely on age may not yield consistently accurate results ( ICRU , 2005 ; AAPM , 2011 ; Cheng et al ., 2013 ).

To date , no information exists regarding the connection between patient size and age among pediatric individuals in Morocco for head CT . Our study aims to automatically compute Deff and Dw from pediatric head CT images and establish correlations with the patient ’ s age .

2 Materials and methods

The research encompassed patients who underwent head CT scans at a pediatric university hospital . The acquisition of head images was in helical mode , ensuring rapid scans and obviating the necessity for sedation in pediatric subjects . Exclusion criteria encompassed patients who underwent contrast-enhanced scans , those exhibiting poor cooperation leading to motion artifacts , and scans with insufficient diagnostic information .

Diameter measurements were executed on axial images obtained from 134 pediatric patients who underwent head CT examinations , utilizing the Siemens Emotion 16-multislice computed tomography ( MSCT ) scanner . Furthermore , all pertinent scan acquisition parameters and patient details were extracted from the Picture Archiving and Communication System ( PACS ) ( Tab . 1 ). Patients presenting torch infections such as hydrocephalus , microcephaly , and other congenital abnormalities affecting the brain and calvaria bones , thereby

Table 1 . Acquisition parameters used in pediatric head examinations .

causing abnormal head diameters , were purposefully excluded from the study .

Patient sizes ( Deff and Dw ) were determined through the measurement of lateral ( LAT ) and anterior-posterior ( AP ) dimensions extracted from patient images using the RadiAnt Dicom Viewer . Specifically , Deff was computed as the square root of the product of the LAT and AP diameters ( AAPM , 2011 ), as depicted in ( Fig . 1 ) and following ( Eq . ( 1 )): p

D ef f ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi AP LAT: ð1Þ

Notably , no established protocol currently outlines the requisite number of images for accurate Deff determination . In a pragmatic approach , we selected the middle slice from the patient ’ s image for analysis . While it remains feasible to calculate Deff from all images to enhance precision , this alternative entails a lengthier processing time .

The calculation of D w is based on the measurement of the water equivalent area ( A w ). The evaluation of ( A w ) is based on the CT numbers ( CT ( x , y )) of the pixels in this area , which are usually expressed in Hounsfield units ( HU ), and determined by the linear attenuation coefficient of each pixel as :

CTðx ; yÞ ¼ m ðx ; yÞ � m water m water

1000 ;

( AAPM , 2014 ) where m ( x , y ) is the linear attenuation coefficient of a tissue located in the pixel ( x , y ) at a given photon energy normalized to that of water m water at the same energy .

( A w ) is assessed by means of the CT numbers over the ( ROI ) as :

A W ¼ 1 1000 CTðx ; yÞ ROI

A ROI þ A ROI : ð3Þ

( AAPM , 2014 ) where A ROI is the sum of the pixel areas ( A ROI = Ʃ A pixel ) of the ROI . Subsequently , D w of the slice can be calculated as : rffiffiffiffiffiffi Aw

D w ¼ 2 : ð4aÞ p

( AAPM , 2014 ) ð2Þ