Angelo Don Ii Grasparil1, Hemali Gupta1, Elizabeth Sheybani1, Govind B Chavhan2,3. 1. Department of Diagnostic Imaging, The Hospital for Sick Children, 555 University Ave., Toronto, ON, M5G 1X8, Canada. 2. Department of Diagnostic Imaging, The Hospital for Sick Children, 555 University Ave., Toronto, ON, M5G 1X8, Canada. drgovindchavhan@yahoo.com. 3. Medical Imaging Department, University of Toronto, Toronto, ON, Canada. drgovindchavhan@yahoo.com.
Abstract
BACKGROUND: Low b-value (50-100 s/mm2) diffusion-weighted images (low b-value DWI) have been shown to be superior to T2-weighted fast spin echo sequence (T2-W) in detecting liver lesions in adults. There are no such studies assessing this difference in children. OBJECTIVE: The purpose of the study was to compare the sensitivity of low b-value DWI images and T2-W images in detecting focal liver lesions in children. MATERIALS AND METHODS: A retrospective review of liver magnetic resonance imaging (MRI) to assess focal liver lesions in 50 children (22 males, 28 females; age: 2 months to 17 years [mean: 10.9 years]) was performed. Two radiologists reviewed both low b-value DWI and T2-W sequences independently on different occasions to note lesions, the smallest lesion size and the location. A consensus reading of the entire MRI examination and a correlation with follow-up, other imaging modalities and pathology in available cases were used to determine the final number of lesions as a reference standard. The average number of lesions per patient detected by both readers on each sequence was compared with each other and with the reference standard using the signed-rank test. The smallest lesions detected by each sequence were compared using the paired t-test. RESULTS: One hundred seventy hyperintense lesions were identified on consensus review to serve as the reference standard. The average number of lesions identified by both readers on low b-value DWI was 134 (79%) and on T2-W was 95 (56%). There was excellent interobserver agreement for detection of lesions on low b-value DWI (intraclass correlation coefficient [ICC]=0.96 [0.93-0.98]) and T2-W (ICC=0.85 [0.75-0.91]), with slightly better agreement on low b-value DWI. Both readers identified significantly more number of lesions on low b-value DWI compared to T2-W (Reader 1: P=0.0036, Reader 2: P=0.0001). Compared to the reference standard (mean number of lesions: 3.45), T2-W detected significantly fewer lesions (mean number of lesions: 1.91; P=0.0001) while there was no significant difference in lesion detection on low b-value DWI (mean number of lesions: 2.68; P=0.1527). Low b-value DWI and T2-W were not significantly different in identifying the smallest lesion size (Reader 1: P=0.19, Reader 2: P=0.47). CONCLUSION: Low b-value DWI images are more sensitive than T2-W sequences in detecting hyperintense focal liver lesions in children.
BACKGROUND: Low b-value (50-100 s/mm2) diffusion-weighted images (low b-value DWI) have been shown to be superior to T2-weighted fast spin echo sequence (T2-W) in detecting liver lesions in adults. There are no such studies assessing this difference in children. OBJECTIVE: The purpose of the study was to compare the sensitivity of low b-value DWI images and T2-W images in detecting focal liver lesions in children. MATERIALS AND METHODS: A retrospective review of liver magnetic resonance imaging (MRI) to assess focal liver lesions in 50 children (22 males, 28 females; age: 2 months to 17 years [mean: 10.9 years]) was performed. Two radiologists reviewed both low b-value DWI and T2-W sequences independently on different occasions to note lesions, the smallest lesion size and the location. A consensus reading of the entire MRI examination and a correlation with follow-up, other imaging modalities and pathology in available cases were used to determine the final number of lesions as a reference standard. The average number of lesions per patient detected by both readers on each sequence was compared with each other and with the reference standard using the signed-rank test. The smallest lesions detected by each sequence were compared using the paired t-test. RESULTS: One hundred seventy hyperintense lesions were identified on consensus review to serve as the reference standard. The average number of lesions identified by both readers on low b-value DWI was 134 (79%) and on T2-W was 95 (56%). There was excellent interobserver agreement for detection of lesions on low b-value DWI (intraclass correlation coefficient [ICC]=0.96 [0.93-0.98]) and T2-W (ICC=0.85 [0.75-0.91]), with slightly better agreement on low b-value DWI. Both readers identified significantly more number of lesions on low b-value DWI compared to T2-W (Reader 1: P=0.0036, Reader 2: P=0.0001). Compared to the reference standard (mean number of lesions: 3.45), T2-W detected significantly fewer lesions (mean number of lesions: 1.91; P=0.0001) while there was no significant difference in lesion detection on low b-value DWI (mean number of lesions: 2.68; P=0.1527). Low b-value DWI and T2-W were not significantly different in identifying the smallest lesion size (Reader 1: P=0.19, Reader 2: P=0.47). CONCLUSION: Low b-value DWI images are more sensitive than T2-W sequences in detecting hyperintense focal liver lesions in children.
Entities:
Keywords:
B-value; Children; Diffusion-weighted imaging; Liver; Magnetic resonance imaging; Neoplasm; T2-weighted fast spin echo
Authors: Christoph J Zech; Karin A Herrmann; Olaf Dietrich; Wilhelm Horger; Maximilian F Reiser; Stefan O Schoenberg Journal: Invest Radiol Date: 2008-04 Impact factor: 6.016
Authors: Tejas Parikh; Stephen J Drew; Vivian S Lee; Samson Wong; Elizabeth M Hecht; James S Babb; Bachir Taouli Journal: Radiology Date: 2008-01-25 Impact factor: 11.105