Evaluation of MRI Sequences for Ultra-rapid Acquisition of Bile Ducts Images (BiliFast)

The Institute of Imaged-Guided Surgery (IHU Strasbourg) has two clinical Magnetic Resonance Imaging (MRI) scanners, one with a 3T (3 Teslas) magnetic field used for diagnosis, the other with a magnetic field of 1,5T (1,5 Teslas) used for the interventional (Pre / per / postoperative). The reference for the visualization of the biliary and pancreatic ducts is a relatively long sequence that needs a breathing-synchronized acquisition leading to artefacts on the images (blur effect). In order to reduce and/or standardize the acquisition time as well as to limit artefacts, accelerated sequences are developed. Such sequence is available in France recently in the form of WIP Siemens (Work In Progress: sequence in test phase at manufacturer to be marketed in the short or medium term on clinical machines). It incorporates a Compressed Sensing (CS) acquisition scheme allowing the acquisition of a 3D (3 dimensions) sequence similar to the usual sequence by drastically reducing the acquisition time, the sequence CS-SPACE. This sequence exists in two forms: An ultra-rapid sequence acquired in apnea An accelerated sequence but remaining synchronized with the breath. The study carried out here on a large number of patients, with two different magnetic fields, applied routinely for diagnosis or anticipation of surgery, could be used by the community of radiologists, hepatogastroenterologists and also digestive surgeons Hepatobiliary.

The Institute of Imaged-Guided Surgery (IHU Strasbourg) has two clinical MRIs, one with a 3T (Teslas) magnetic field used for diagnosis, the other with a magnetic field of 1,5T (Teslas) used for the interventional (Pre / per / postoperative). Cholangiopancreatography by MRI (or bili-MRI) accounts for about 25% of MRI examinations at the IHU and is used for the diagnosis of pathologies affecting the bile ducts (lithiasis pathology, cholangitis, tumoral pathology of the bile ducts) and the pancreas (search for a causal pathology to pancreatitis (calculus), ductal pathology, ductal study in tumor pathology). It can be done in preoperative context (cholecystectomy) for research of an associated calculus in the bile ducts, or for an anatomical analysis of the bile ducts (possible anatomical variants ...). The reference for the visualization of the biliary and pancreatic ducts is a very heavily weighted T2 3D SPACE sequence. A 3D volume is obtained in which they appear in white, and the projection of this volume makes it possible to visualize the entire tract. This sequence is relatively long and needs a breathing-synchronized acquisition. The time of acquisition is then very variable from one patient to another (4 to 8 minutes). This extended acquisition time in addition to the variability of rhythm and respiratory profile can lead to artefacts on the images (blur effect). These are the limits of the respiratory synchronization that triggers acquisition by referring to the position of the liver dome which is not perfectly reproducible from a breathing cycle to the other, potentially degrading the appearance of the bile ducts and the reliability of the diagnosis. A solution that makes it possible to reduce and / or standardize the acquisition time as well as to limit artefacts is to use an accelerated sequence. Such a sequence is available in France recently in the form of WIP Siemens (Work In Progress: sequence in test phase at manufacturer to be marketed in the short or medium term on clinical machines). It incorporates a Compressed Sensing (CS) acquisition scheme allowing the acquisition of a 3D sequence similar to the usual sequence by drastically reducing the acquisition time, the sequence CS-SPACE. This sequence exists in two forms: An ultra-rapid sequence acquired in apnea An accelerated sequence but remaining synchronized with the breath. This sequence has been shown to provide comparable, if not better, results to those obtained with the usual sequence in terms of image quality and visualization of structures, but these results were only obtained at 3T and for a reduced cohort of patients. A lot of work is done today in the field of MRI to reduce the time of sequence acquisition. Accelerated acquisition schemes like Compressed Sensing remain marginal but are set to become the norm in the long term. The use of the sequence CS-SPACE at the IHU therefore enters into this perspective of reducing examination times. The study carried out here on a large number of patients, with two different magnetic fields, applied routinely for diagnosis or anticipation of surgery, could be used by the community of radiologists, hepatogastroenterologists and also digestive surgeons Hepatobiliary.

For each patient who undergo a 1.5T MRI, 3 sequences will be done: (i) a conventional SPACE sequence, (ii) an ultra-rapid sequence (sequence CS-SPACE) acquired in apnoea and (iii) an accelerated sequence (sequence CS-SPACE) but remaining synchronized with the breath.

For each patient who undergo a 3T MRI, 3 sequences will be done: (i) a conventional SPACE sequence, (ii) an ultra-rapid sequence (sequence CS-SPACE) acquired in apnoea and (iii) an accelerated sequence (sequence CS-SPACE) but remaining synchronized with the breath.

On top of the conventional SPACE sequence, 2 supplementary sequences will be performed with the ultra-rapid sequence (sequence CS-SPACE) either acquired in apnoea or synchronized with the breath.

The global Likert score corresponds to the average of scores varying between 1 and 5 and based on the following criteria : i. diagnostic quality (feasibility of establishing a diagnosis when reading the image), including canal visibility (from score 1 = no diagnostic to score 5 = Excellent) ii. the sharpness of the image, especially the canal sharpness (from score 1 = too blurry / illegible to score 5 = sharp) iii. removal of the background signal (contrast quality) (from score 1 = too much background / illegible to score 5 = no background). The global Likert score will vary between 1 (worst) and 5 (best) as well.

Comparison of the global quality of the images evaluated with a Likert score between the three sequences performed with the 1.5T MRI scanner

The sequences will be compared in pairs based on the overall image quality to determine, where appropriate, the sequence providing the best performance. The global quality of the images will be evaluated with a Likert score as described in the primary outcome measure. The global Likert score corresponds to the average of scores varying between 1 and 5 and based on the following criteria : i. diagnostic quality (feasibility of establishing a diagnosis when reading the image), including canal visibility (from score 1 = no diagnostic to score 5 = Excellent) ii. the sharpness of the image, especially the canal sharpness (from score 1 = too blurry / illegible to score 5 = sharp) iii. removal of the background signal (contrast quality) (from score 1 = too much background / illegible to score 5 = no background). The global Likert score will vary between 1 (worst) and 5 (best) as well.

Comparison of the global quality of the images evaluated with a Likert score between the three sequences performed with the 3T MRI scanner

The sequences will be compared in pairs based on the overall image quality to determine, where appropriate, the sequence providing the best performance. The global quality of the images will be evaluated with a Likert score as described in the primary outcome measure. The global Likert score corresponds to the average of scores varying between 1 and 5 and based on the following criteria : i. diagnostic quality (feasibility of establishing a diagnosis when reading the image), including canal visibility (from score 1 = no diagnostic to score 5 = Excellent) ii. the sharpness of the image, especially the canal sharpness (from score 1 = too blurry / illegible to score 5 = sharp) iii. removal of the background signal (contrast quality) (from score 1 = too much background / illegible to score 5 = no background). The global Likert score will vary between 1 (worst) and 5 (best) as well.

Comparison of the sequences between 1.5T and 3T MRI scanners using the Likert score in order to determine the influence of the magnetic field on the global quality of the images. The global Likert score corresponds to the average of scores varying between 1 and 5 and based on the following criteria : i. diagnostic quality (feasibility of establishing a diagnosis when reading the image), including canal visibility (from score 1 = no diagnostic to score 5 = Excellent) ii. the sharpness of the image, especially the canal sharpness (from score 1 = too blurry / illegible to score 5 = sharp) iii. removal of the background signal (contrast quality) (from score 1 = too much background / illegible to score 5 = no background). The global Likert score will vary between 1 (worst) and 5 (best) as well.

Performance of each sequence in the visualization of the pathology and / or an anatomical variant in the 1.5T magnetic field group, evaluated in percentage

The performance is evaluated as the number of well-diagnosed patients out of the total number of cases seen at surgery in the 1.5T magnetic field group

Inclusion Criteria: Male or female over 18 years old Patient for whom MRI is required for the diagnosis or follow-up of a pathology Patient able to receive and understand information related to the study and give written informed consent Patient affiliated to the French social security system Exclusion Criteria: Patient with contraindications to MRI: pacemaker or automatic defibrillator, pump implanted auditory, anal, painkiller neurostimulator, etc ... ferromagnetic bodies in soft tissues, body intraocular foreigners, cerebrovascular clips claustrophobia Patient presenting, in the judgment of the investigator, an illness that may prevent participation in the procedures provided by the study Patient who has been operated urgently Pregnant or lactating patient Patient in exclusion period (determined by a previous or a current study) Patient under the protection of justice Patient under guardianship or trusteeship Patient subject to a legal protection measure or out of state to express their consent Patient in a situation of social fragility A patient may be excluded at the end of the examination if at least one of the three sequences of bili-MRI could not be obtained (apnea impossible, image reconstruction problem).

Bates DD, LeBedis CA, Soto JA, Gupta A. Use of Magnetic Resonance in Pancreaticobiliary Emergencies. Magn Reson Imaging Clin N Am. 2016 May;24(2):433-48. doi: 10.1016/j.mric.2015.11.010. Epub 2016 Feb 22.

Yoon JH, Lee SM, Kang HJ, Weiland E, Raithel E, Son Y, Kiefer B, Lee JM. Clinical Feasibility of 3-Dimensional Magnetic Resonance Cholangiopancreatography Using Compressed Sensing: Comparison of Image Quality and Diagnostic Performance. Invest Radiol. 2017 Oct;52(10):612-619. doi: 10.1097/RLI.0000000000000380.

Chandarana H, Doshi AM, Shanbhogue A, Babb JS, Bruno MT, Zhao T, Raithel E, Zenge MO, Li G, Otazo R. Three-dimensional MR Cholangiopancreatography in a Breath Hold with Sparsity-based Reconstruction of Highly Undersampled Data. Radiology. 2016 Aug;280(2):585-94. doi: 10.1148/radiol.2016151935. Epub 2016 Mar 16.