Introduction

Diffusion weighted imaging (DWI) is an important part of pediatric abdominal MRI. It is used to detect neoplasms and infection, differentiate benign and malignant tumors, and inflamed bowel in inflammatory bowel disease¹. As DWI acquisitions are inherently SNR limited, signal averages are typically employed to improve SNR at the cost of scan time. In children that may be unable to cooperate for prolonged exams, or sedated children in which minimizing time under anesthesia is critical, reducing scan times is desirable². Scan time for 2D acquisitions can be reduced by performing simultaneous multislice (SMS) imaging, in which excitation of multiple slices is performed concurrently, and a coil sensitivity profile in the slice direction is leveraged to properly reconstruct the simultaneously excited slices ³. However, coil geometry of modern phased array coils is typically designed for imaging of adults and does not confer the same advantages when imaging children⁴. Although SMS holds potential to reduce scan times for pediatric abdominal DWI, its performance in children has not been previously characterized in a routine clinical setting. In this work, we compare acquisitions of standard DWI (std-DWI) to SMS (SMS-DWI) with the goal of assessing its feasibility for use in routine clinical exams. DWI also produces quantitative data in the form of the apparent diffusion coefficient (ADC), which has shown promise in differentiating benign processes from malignancy⁵. Therefore, ability to generate comparable ADC values using SMS-DWI was also assessed.

Methods

Patients referred for clinically indicated MRI underwent examination on a 3T magnet (either GE SIGNA Premier or SIGNA Architect, GE Healthcare, Waukesha, WI). Patients were imaged with phased array coils that were matched to the patient’s size as best as possible in accordance with routine clinical practice; utilized coils all had at least 20 elements with a minimum of 3 rows of coils in the craniocaudal dimension over the range of the abdomen and pelvis. Imaging was obtained during free-breathing with a single-shot echo-planar DWI sequence both in standard and SMS mode with a SMS factor of 2. Remaining parameters in each acquisition were unchanged, including TR, number of slices, parallel imaging factor, spatial resolution, and b value. Two board-certified radiologists (A.B.S. and P.J., 8 and 6 years of experience respectively) were blinded to the acquisition method and graded comparative image quality in the liver, pancreas, kidneys, and adrenal glands on a 5-point scale. The blinded scores were subsequently un-blinded and translated such that scores of +2 referred to significant preference of SMS-DWI over std-DWI for evaluation of the organ in question, while +1 referred to mild preference of SMS-DWI over std-DWI for evaluation of the organ in question. A score of 0 indicated no preference. Similarly, -2 and -1 respectively referred to significant preference or mild preference of std-DWI over SMS-DWI. SNR and apparent artifacts were also scored on a comparative 5-point scale and analyzed similarly.

Non-inferiority in image quality for each organ, overall apparent SNR, and overall image artifacts was assessed by using a two one-sided test (TOST) method for equivalence based on Wilcoxon signed rank methodology.

ADC values were recorded from matched regions of interest in the liver (primarily in the right hepatic lobe) as well as the pancreatic head. Bland-Altman analysis of differences in ADC value was performed; clinical limits of agreement was set at ±10%. Total scan time for each acquisition was also recorded, and the mean scan time for SMS-DWI and std-DWI were compared using a one-sided paired t test to assess for significant reduction in scan time.

Holm-Bonferroni correction for multiple comparisons was applied to all p values.

Results

20 exams were obtained in 18 patients (12 male, 8 female) between January and September 2022. Patients ranged in age from 1 to 23 years of age. 45% of patients (n = 9) were scanned awake; remaining 55% (n = 11) were scanned under general anesthesia. SMS-DWI provided equivalent image quality for the liver (p < 0.001), pancreas (p < 0.001), kidneys (p < 0.001), and adrenal glands (p <0.001). SMS-DWI satisfied equivalence for SNR and presence of artifacts (p < 0.001 for both).

Bland Altman analysis revealed 95% limits of agreement from -7.9% to +8.4%, which was within the clinically significant limits of ±10%. A bias of 0.25% was present in for SMS-DWI.
SMS-DWI mean scan time was 65 seconds, compared to 119 seconds for std-DWI; this represents a 45% reduction in scan time which was statistically significant (p < 0.001).

Discussion

SMS-DWI and std-DWI demonstrated equivalence for evaluation of the liver, pancreas, kidneys, and adrenal glands in children spanning a wide range of ages and sizes, including sedated and awake children. Apparent overall SNR and artifacts were equivalent when evaluated by the two radiologists. ADC values were not significantly changed in SMS-DWI when compared to std-DWI. Scan time was significantly reduced in SMS-DWI. Together, these findings indicate that SMS-DWI may be equivalent to std-DWI for most organs interrogated in pediatric abdominal DWI exams while providing a method to significantly reduce acquisition times.

Conclusion

SMS-DWI of the pediatric abdomen significantly reduces scan time without impacting image quality compared to std-DWI and is effective over a range of patient ages and sizes.

Acknowledgements

No acknowledgement found.