The role of Magnetic Resonance Cholangiopancreatography (MRCP) in assessment of childhood hepatobiliary disorders, both congenital (choledochal cyst, anomalous pancreaticobiliary junction) as well as acquired (pancreatitis, choledocholithiasis, sclerosing cholangitis), is well established. Achieving an adequate quality examination in this cohort of patients presents unique challenges including varying body habitus, lack of cooperation and breath hold, and small calibre ducts. The possible predictors for a non-diagnostic examination have been previously highlighted in research by Glenn et al., who stressed the importance of sedation and general anesthesia for good image quality [1].

The paper by Sodhi et al. published in this issue of IJP represents an important continuum to this research [2]. The authors explore the Compressed SENSE (CS) MRI technique, which has previously shown good results in pediatric brain as well as cardiac imaging for reduction of acquisition time with good image quality. Similar results were found in this study as well, the imaging time for the MRCP sequence was reduced to nearly half, from 310+/-98 s for conventional MRCP (C-MRCP) to ~148+/-61 for CS-MRCP. On comparison of image quality, no significant difference was found between the incidence of artefacts (p = 1), as well as overall image quality (p = 0.18). These results present a leap in knowledge in the realm of pediatric hepatobiliary imaging.

It is of note however, that the authors used a 3 Tesla MRI for all examinations and have not attempted to discuss the effect of changing magnetic strength upon image quality. It has been reported by Blaise et al. that grading of image quality for CS-MRCP is inferior to C-MRCP at 1.5 Tesla, and the difference is statistically significant [3]. In fact, in the study by Glenn et al., it was consistently found in multivariate logistic regression as well as stepwise linear regression models that imaging at 1.5 Tesla is associated with higher image quality and lesser number of non-diagnostic scans [1]. Therefore, this aspect must be explored in future research on the subject to ascertain what would be the best combination of field strength and acquisition parameters.

Another point of contention remains the acquisition mode of MRCP; whether acquired in free breathing, navigator triggered (NT) mode or during breath hold (BH). It is the BH mode which offers the most drastic reduction in acquisition time (17 s BH-CS vs. 127.5 s NT-CS) [4], which may allow for truly non-sedated scans in a single breath hold duration. The caveat however, is that the BH-CS sequence has been found inferior to the NT mode, particularly for depiction of small ductal structures, which are most often encountered in children. Optimization of the BH sequence to improve image quality in adults has been suggested, and similar parameter adjustment must be investigated in children in future work on the subject. The use of deep learning algorithms post image acquisition, in the image reconstruction domain has also shown promise in improving signal-to-noise and contrast-to-noise ratio in BH-MRCP [5].

It is also worth mentioning that MRCP (conventional or CS) is a single sequence which is only a part of the MRI performed for evaluation of hepatobiliary and pancreatic pathology, which also includes other sequences. Indeed, it has been found that an average of 24+/-12 min are spent by a child in the MR gantry prior to the MRCP sequence being performed [1]. As pediatric radiologists, we must understand the implication of total scan time, and therefore anesthesia time, and endeavour to comprehensively optimise the entire scan and not just a single sequence.

To conclude, while much ground has been gained in improving outcomes of pediatric MRCP, the effort to revise, optimize and evaluate MR protocols must be continuous, till we achieve the best level of comfort and diagnostic accuracy for our patients.