Hepatectomy Risk Assessment With Functional Magnetic Resonance Imaging

Overview

Liver resection remains the only curative option for primary or metastatic liver cancer, but a more accurate prediction of post-hepatectomy liver failure (PHLF) is needed to further reduce morbidity and mortality and to extend the indication to a wider patient population. Magnetic resonance Imaging (MRI) is a promising new source of liver function tests as it can provide segmental function alongside measurements of perfusion, tissue structure and standard morphological assessment. The primary aim of HEPARIM is to determine if quantitative MRI biomarkers of liver function and perfusion can improve predictions of post-hepatectomy liver function, as measured by an indocyanine green (ICG) liver function test. Secondary aims is to validate the MRI measurements of liver function against ICG. HEPARIM is an observational cohort study recruiting patients referred locally for a one- or two-stage liver resection of 2 segments or more. Before surgery, all participants will undergo an ICG liver function test and a Dynamic Gadoxetate-enhanced (DGE) MRI scan of the liver. The ICG test will be repeated at one day after surgery. The Gadoxetate Clearance (GC) of the future liver remnant (FLR-GC) will be determined from the DGE-MRI data and correlated to the post-operative ICG R15 as primary outcome measure. Preoperative ICG R15 will be correlated against GC of the whole liver (WL-GC) to address the secondary objective. In patients that undergo a staged hepatectomy, an additional MRI and ICG test will be performed before the first stage to assess its effect on volumetric and functional growth of the FLR. Additional pre- and postoperative data will be collected from medical records including demographics and medical histories, biochemistry, pathology and radiology reports, and any long-term outcome data collected in the 90-day follow-up visit. These data will be used in a multi-variate analysis to determine which preoperative biomarkers are most predictive of immediate and long-term outcomes, to identify the added value of functional MRI over routine clinical markers, and to derive a multi-variate prediction model that can be validated in future studies.

OBJECTIVES (1) To determine if DGE-MRI can improve predictions of post-hepatectomy liver function; (2) To validate DGE-MRI based measurements of liver function against gold-standard liver function tests; (3a) to determine if preoperative DGE-MRI can improve predictions of post-operative PHLF; (3b) to develop a multi-parametric prediction model for 90-day post-operative outcome combining MRI, clinical chemistry and patient history; (3c) To determine the relationship between volumetric and functional liver growth after portal vein embolisation. RECRUITMENT Patients considered for major liver resection will be identified at the Hepatobiliary multi-disciplinary team (MDT) meeting at St James University Hospital Leeds, and contacted by phone. Patients will have at least 3 days to consider this information, and after that they will be contacted by the member of their clinical care team to schedule the baseline visit. Formal written consent will be obtained on the day of the visit. If they don't consent to participation, then only the MRI scans will be collected and stored in their medical records. BASELINE VISIT An initial dataset will be captured incorporating the patient's medical history and up-to-date investigations. This will be obtained through a combination of patient interview, review of paper notes and review of the patient's electronic medical record. An MRI safety screening form will be administered. If a contraindication to MRI is found at this point, the visit will be cancelled, and the patient will be excluded from the study. An ICG safety questionnaire will be used to exclude patients with contraindications to ICG use. Patients excluded from ICG testing at this point will not be excluded from the study as other measures of liver function will still be available for analysis of tertiary outcomes. Patients will be cannulated and ICG will be administered intravenously as a bolus dose of 0.5 mg/kg through the cannula. The ICG will be reconstituted with 5 ml of water for injection for every 25 mg of ICG, leading to a dilution of 5mg ICG per 1 ml. Serial percutaneous ICG measurements will then be taken using an ICG clearance meter using a finger clip device (LiMON; Maquet, Germany). Patients will lie in the supine position and calibration will be performed before testing starts. In order to derive absolute ICG concentrations, two separate blood samples of 2.5mL each will be taken, both after ICG injection (pre-operative). The samples will be processed and stored in a dedicated facility in the Leeds Institute for Cardiovascular and Metabolic Medicine (LICAMM) and analyzed as a batch at the end of the study. The remaining samples will be stored for a maximum of 5 years to measure metabolomics data for secondary research. After 5 years any remaining samples will be destroyed according to standard LICAMM procedure for sample disposal. The MRI will be performed straight after ICG measurement in the 3 T MRI by radiographers of the department. The scan will last max 1.5hrs and will include the injection of a standard dose of MRI contrast agent Gadoxetate (Primovist, Bayer AG). The same cannula will be used to administer the MRI contrast as for the ICG test. Other scans that will be taken in the same session include T1- and T2-weighted anatomical imaging, T2\*-mapping, T1-mapping, proton-density fat fraction. SURGERY AND FOLLOW-UP After the baseline visit some patients will require re-discussion in the MDT and all patients will attend for a preoperative clinic appointment. A proportion of patients will not progress to surgery due to clinician or patient choice; these patients will not be followed up further. The data that has already been collected will be used for addressing the secondary objective. Intraoperative measurements of the resected liver segments will be taken, including weight and volume (measured using fluid displacement). After surgery, data obtained as part of the patient's standard intra and post-operative care will be collected from their medical records. This includes any blood tests during admission, the characteristics of the operation, complications and length of stay. Pathology results will be collected. The ICG test with the finger-clip device will be repeated postoperatively at bedside to determine a reference value for postoperative liver function. This will be performed on day 1 post-surgery if feasible. A PHLF score will be determined according the International Study Group of Liver Surgery (ISGLS) definition. All patients who underwent surgery will attend the hospital for a routine post-operative clinic appointment after 90 days. Data from this appointment will be collected by the clinical research fellow, including complications, readmissions and mortality. At the end of the study patients who have requested to do so will be sent a summary of the study's findings and implications. PVE/ALPPS ARM Patients requiring portal vein embolisation (PVE) or Associating Liver Partition and Portal vein Ligation for Staged hepatectomy (ALPPS) will be recruited in the same way as participants in the resection arm. They will attend the same baseline visit and the routine clinic visit. Patients may decline the PVE/ALPPS at this point, or the clinician may decide it is inappropriate at the clinic appointment. If this occurs and the patient proceeds to surgery, they will remain in the study. If they do not proceed to surgery their data will be used to address the secondary objectives. A patient who does proceed to PVE/ALPPS will attend the hospital for a routine computed tomography (CT) scan after the first stage, and a second research visit will be scheduled at that point. This visit will proceed in the same manner as the baseline visit. They will be re-discussed in the MDT and come for a preoperative clinic appointment. Patients may be deemed ineligible for surgery, or make the decision not to proceed with surgery, in which case their data already collected can feed into the secondary objectives. If they do proceed to surgery, the same intra-operative and post-operative data will be taken as for the resection arm. SAMPLE SIZE The aim is to recruit 134 participants into the study, which will include an estimated 122 in the primary surgery arm and 12 in the PVE arm. Of the 134 recruited participants an estimated 112 will feed into the primary objective. The power calculation assessed the impact of study size by simulating a hypothetical population of 1 million patients and evaluating model robustness to estimate the model beta coefficient of FLR-GC in relation to ICG clearance. Modelling was informed by pilot data in 29 cases. 10k samples of 122 patients are drawn with replacement, generating empirical distributions of the parameters. Results demonstrated that even at the poorest precision simulated, the fixed sample of 122 is sufficient to obtain reliable beta coefficients. DATA ANALYSIS Statistical analysis shall be performed by Professor Gilthorpe at the University of Leeds using the anonymized data sheet and the RStudio software package (RStudio, Boston, USA). A univariable analysis will be performed comparing FLR-GC against post-operative ICG-R15 (primary objective) and WL-GC against preoperative ICG-R15 (secondary objective). Second the correlations will be investigated in a multivariable analysis to determine if the MRI adds value to existing preoperative biomarkers, which MRI biomarkers are most predictive, and to derive a model for predicting PHLF. It will then be tested whether this combined score is significantly worse when DGE-MRI is removed from the prediction. In the PVE/ALPPS group, changes in volume will be correlated to changes in function of the FLR between pre PVE/pre-ALPPS and post-PVE/post-ALPPS. DATA MANAGEMENT Identifiable data will not be accessed by researchers outside of the clinical care team and are only retained for quality control, backup, and to follow-up on any incidental findings. Anonymised research data will be recorded using the participant's study ID. Paper forms used for data capture will be stored in a secure location in St James University Hospital. Electronic research data will be stored in a participant study record in a secure database in St James University Hospital, and in password protected network drives set up and maintained by University IT. Additional databases will be set up to retain the source data for further processing, including MRI in DICOM (Digital Imaging and Communications in Medicine) format, raw ICG data, and digitized histopathology slides. MRI data will be recorded using the patient ID in the secure Picture Archiving and Communication System (PACS) of the radiology department. Routine radiology reports will be retrieved and included in the participant's anonymous study file. The MRI images will be transferred anonymously from the scanner to the university's DICOM database, where they will be analyzed by a research fellow to extract the relevant imaging biomarkers. All excel sheets with anonymised study data will be transferred to a long-term data repository where they will be maintained for a minimum of 25 years. Source data for MRI and ICG will be retained indefinitely in a dedicated imaging data management system hosted by the University to enable secondary research, educational use, and data sharing. INCIDENTAL FINDINGS AND BLINDING All data collected for the study, except calculated MRI biomarkers and ICG measurements, are acquired as part of the routine workup and will therefore be available to the clinical care team through the patient's medical records. All ICG liver function measurements from finger clip device will be reviewed separately by the hepatology consultant. Any unusual findings will be flagged up to the clinical care team as incidental findings, and dealt with according to usual practices. For those patients that have an extra MRI scan as part of their participation in the study, these MRI's will be reviewed by a radiologist. Any unusual findings will be flagged up to the clinical care team as incidental findings, and dealt with according to usual practices. Unvalidated MRI biomarkers will not be reviewed for incidental findings or otherwise communicated to the clinical care team. The non-clinical research fellows working on analysis of MRI and ICG data will be blinded to the clinical history of the patient and other function tests. WITHDRAWAL OF PARTICIPANTS If a patient decides to withdraw their consent to inclusion in the study at any point then no further data regarding their care will be collected. Should the patient wish, data collected up until withdrawal of consent will be destroyed using the hospital's confidential waste bins. Patients who lose their capacity to consent after initial inclusion in the study for reasons that are not related to their surgery will be treated in the same manner as if they had withdrawn consent. Previously collected data will not be destroyed. If a patient is unable to complete their MRI study for any reason they will automatically be withdrawn from the study. Data collected until the point of withdrawal will be retained.