RPS 301a - Pancreas and biliary disease

Purpose:

To evaluate the diagnostic accuracy of dynamic magnetic resonance (MR) with diffusion weighted imaging (DWI) in the identification of mural nodules of intraductal papillary mucinous neoplasms (IPMN) of the pancreas.

91 preoperative MRI scans with histopathological diagnosis of IPMN were reviewed by two radiologists. They evaluated the presence, number, and size of mural nodule, the SI of the nodule on T1-WI after i.v. contrast medium administration and on DWI, and the size of the cystic lesion and dilation of the main pancreatic duct. Two pathologists reviewed each case, evaluating the presence, number, and size of papillary structures, and the degree of epithelial dysplasia. Qualitative and quantitative analysis were performed. Interobserver agreement was calculated.

Statistically significant differences were found for the following correlations: presence of nodules >5mm on MRI and confirmed by pathological review, size and number of mural nodule on pathological review and degree of dysplasia, size and number of mural nodule evaluated on MRI and tumoral dysplasia, presence of nodule >5mm with post-contrast enhancement and hyperintensity on DWI and degree of dysplasia, dilation of MPD and tumoral dysplasia. Interobserver agreement was moderate for the presence of mural nodule (K=0.56), the presence of high signal intensity on DWI (K=0.60), and enhancement of mural nodule (K= 0.58). Histogram analysis of ADC map showed a correlation between entropy of the entire cystic lesion and the degree of dysplasia.

Dynamic MRI with DW imaging has a good accuracy in the detection of IPMN mural nodule >5mm. DWI and T1-WI after contrast medium administration can be useful in the evaluation of malignancy of IPMN. Entropy could be used as a predictive parameter of malignancy of IPMN.

None

n/a

Purpose:

To assess the safety of a surveillance MRI protocol through a 10 year follow-up of patients with branch-duct intraductal papillary mucinous neoplasms (BD-IPMNs) of the pancreas.

All patients with “presumed” MR diagnosis of BD-IPMNs observed from 2008 to 2009 were retrospectively recruited from our radiological database. Only patients without worrisome features (WF) and/or high-risk stigmata (HRS) at diagnosis were enrolled. MRI protocol included axial T1w/T2w sequences, MRCP, and diffusion-weighted MRI with multiple b values. Contrast-enhanced T1-weighted sequences were obtained only at the baseline and if there was a suspicion of degeneration during the follow-up. The policy for follow-up consisted of abdominal MRI/MRCP every 6 months in the first 2 years from diagnosis and then each year in absence of clinical/radiological signs of progression. Cyst growth rate analysis and development of WF, HRS, and pancreatic cancer (PC) during follow-up were evaluated. Patients with missing data and/or without 10 years follow-up were excluded from analysis.

69 patients fulfilling all the inclusion criteria represented our study group. During surveillance, five out of 69 patients (7%) underwent surgery (PC, n=2; IPMN with low grade dysplasia, n=3). WF and HRS developed in 10/69 (14,5%) and 3/69 (4%) cases, respectively. The cysts were dimensionally unchanged or slightly reduced in size in 18/69 (26%) and 3/69 (4,3%) patients, respectively. In the remaining 48/69 (69,5%) cases, cyst enlargement was appreciable with a median annual growth rate of 0,97±0,87 mm/yr (range: 0,13-5,0 mm).

In our series, pancreatic BD-IPMNs slowly grow over the time. Worrisome features, high-risk stigmata, and pancreatic cancer developed in a low percentage of patients, supporting the safety of our surveillance MRI protocol.

Retrospective study.

n/a

Purpose:

To evaluate whether pancreatic adenocarcinoma (PAC) could be accurately differentiated from the surrounding pancreatic tissue (PT) using the shortened MR scanning protocol by quantitative analysis of ADC and T2W signal intensity including the histogram-derived analysis.

We retrospectively analysed 31 cases of PAC who underwent shortened 1.5T-MRI scanning protocol, which consisted of T2W FSSE and diffusion-weighted imaging (DWI) sequence with the apparent diffusion coefficient (ADC) by using two b values (0 and 800 mm²/s). Tumour and surrounding pancreatic tissue were segmented on three consecutive the-largest-tumour-area sections in T2W series and the corresponding ADC map. Histogram analysis of T2W signal of PAC and PT was performed using MAZDA software. Mean values of ADC and T2W signal intensity and histogram-derived quantitative parameters (variance, skewness, kurtosis, minimum and maximum normalised values) of PAC and PT were compared using the paired-samples-t or Wilcoxon test depending on normality of parameter distribution, which was explored using the Shapiro-Wilk test.

Mean ADC of PAC (1.537±0.221x10^-3 mm²/s) did not significantly differ from PT (1.555±0.196x10^-3 mm²/s) (p=0.644). Mean value of T2W signal intensity of PAC (median, 639.46) was non-significantly higher (p=0.410) than PT (589.76). Of histogram-derived parameters, variance of T2W signal intensity was significantly lower (p=0.030), kurtosis of histogram was higher (p=0.050), and minimum normalised signal intensity was significantly higher (p<0.001) in PAC than in PT.

Histogram-derived quantitative analysis of T2W signal intensity improves accuracy in differentiation of PAC from surrounding pancreatic tissue on shortened MRI scanning protocol.

There were no obvious limitations.

n/a

No funding was received for this work.

Purpose:

To determine the diagnostic interval of patients with pancreatic ductal adenocarcinoma (PDAC) in our region’s population and which factors, including imaging-related delays, may impact survival.

This study evaluated 257 patients (129 men, 128 women; mean 71.8 years) diagnosed with PDAC from 2014-2015. Patient data was retrieved from the Cancer Registry database, including demographics, stage, and dates of initial presentation/diagnosis and death.

The imaging archive was searched for US, CT, and MRI examinations performed during the diagnostic interval; these were independently rated by two radiologists according to the American College of Radiology RADPEER system. Univariate Kaplan-Meier analysis was performed for age (<65, 65-79, 80), sex, tumour location, stage, surgery, and RADPEER score. Multivariate Cox proportion hazards model was performed to assess for association between these covariates and survival.

RADPEER 1-2 and 3 scores were assigned to 219 and 38 patients, respectively. The overall diagnostic interval was 77 days; this was much longer in the RADPEER 3 group (190 vs. 58 days, p=0.018). However, K-M analysis showed worse survival in the RADPEER 1-2 group (p=0.007).

Survival was also worse with distal PDAC (p=0.016), stage (p<0.0001), and no surgery (p<0.001). Survival was similar amongst the sexes (p=0.083). The Cox proportional hazards model showed better survival in women (p=0.008) and surgical patients (p<0.001), and worse survival in patients aged >80 years (p=0.016) and with stage IV disease (p=0.006).

Imaging-related delays are associated with much longer diagnostic intervals. However, this did not impact survival in our population.

Retrospective study, registry data, other factors associated with survival: patient’s comorbidities.

Research ethics board approval and PHIA adherence.

No funding was received for this work.

Purpose:

To evaluate prevalence and stability predictors of pancreatic lesions (PLs) using non-contrast whole-body MRI (WBMRI) screening.

3236 consecutive healthy subjects (1765 male, 1471 female, mean: 55±13 years, range: 20-92 years) receiving non-contrast WBMRI using 1.5T or 3T MRI scanners covering brain, neck, thorax, abdomen, pelvis, and spine from 2008-2018 were retrospectively analysed. Abdomen protocol included T2w-TSE, T1w-FS, and DWI. Images were evaluated by radiologists (10+ years of experience). Subjects with PLs were identified and necessary follow-ups (interval: 1-9 years) were conducted to confirm the PLs' nature. Prevalence was counted as the proportion of PL reported to the total of WBMRI cases. PLs were divided into 2 groups (I: need treatment/close-monitor, II: simple/benign cysts). Pearson-correlation was conducted to evaluate the relationship between prevalence of PL and age. Chi-square and t-test were used to evaluate PL stability regarding sex and PL size.

In this study, 74 (2.3%) subjects (40 males, 34 females) with PL were observed using WBMRI. Group I included 2 pancreatic carcinoma, 9 intraductal papillary mucinous neoplasms (IPMN), 1 serous cystadenoma, and 4 pancreatic cysts with changes in size/number/density. Group II consisted of 58 stable pancreatic cysts. PLs’ size ranged from 0.1-4 cm (mean:0.6±0.6mm). Group I had significantly larger PLs (size:1±0.95mm) compared to group II (size:0.52±0.5mm)(p<0.05), with no significant difference in gender (p=0.066) or age (p=0.42) between groups.

Prevalence of PL was age-dependent (Pearson-coef.=0.35;p<0.005), with no pancreatic lesion found in subjects under 30 years old, and gender-independent (p=0.8;2.3%(male),2.3%(female)).

Prevalance of PL screened by WBMRI is 2.3%, and is age-dependent but gender-independent. PL size might be a potential predictor of whether further treatment/close-monitoring is needed.

This is a retrospective study, hence, an inadvertent bias could not be excluded.

Ethical approval has been obtained.

No funding was received for this work.

Purpose:

Traditional techniques for characterising pancreatic tissue such as endoscopy and computed tomography are invasive, limited by interobserver variability and lack sensitivity. Magnetic resonance imaging (MRI) successfully predicts liver outcomes, correlates with histology, and does not use ionizing radiation. MRI is emerging as a promising technique for monitoring the pancreas in patients with T2DM. Here we explore the utility of multiparametric MRI of the pancreas within a cohort of T2DM participants compared to healthy participants.

459 participants underwent a pancreatic MRI scan as part of the UK Biobank imaging study. 3 regions of interest (10mm) were placed on proton density fat fraction (PDFF) and T1 (a fibroinflammation biomarker) maps in the head, body, and tail of the pancreas. PDFF and T1 values were successfully acquired from 90.8% and 90.4% of scans, respectively. Participants were excluded if pancreatic T2* measured < 23.1 ms (high iron). Data from 126 participants (20 T2DM/106 healthy) were included for analysis.

Two-sample Kolmogorov-Smirnov tests revealed T2DM participants had significantly greater PDFF than healthy participants (6.79% vs 3.09%, p < 0.0001), and there was a trend towards increased T1 in T2DM patients (639ms vs 603ms, p = 0.2). PDFF was not uniformly distributed within the pancreas with a trend towards increased fat in the body compared to the head and tail of the pancreas (T2DM: head 4.94%, body 7.06%, tail 4.57%; healthy: head 2.69%, body 3.28%, tail 3.07%). Data is presented as median.

Multi-parametric MRI had a high success rate for quantitative characterisation of pancreatic tissue, even in this high-throughput epidemiological study. Pancreatic PDFF, but not T1, was significantly higher in patients with T2DM vs heathy participants.

Single analyst.

n/a

No funding was received for this work.

Click here to add a description...

Purpose:

To evaluate the added value of gadoxetic acid-enhanced magnetic resonance imaging (MRI) in the monitoring of disease progression and liver function in patients with primary sclerosing cholangitis (PSC).

We retrospectively identified 120 patients (87 male, 33 female; mean age 42.58±12.3 years) with confirmed diagnosis of PSC who underwent 259 gadoxetic acid-enhanced MRI examinations between January 2008 and May 2019. Relative enhancement (RE) in hepatobiliary phase (HBP) was calculated in each Couinaud liver segment and in the whole liver and correlated with liver function tests (LFTs) and different prognostic clinical scoring systems, including model for end-stage liver disease (MELD) score, Mayo risk score, and Amsterdam-Oxford-PSC score. Furthermore, RE was correlated with qualitative MRI parameters including lobar atrophy, T2 hyperintensity, significant bile duct stenosis, and bile duct irregularities (pruning and beading).

There was a significant negative correlation between RE in HBP and Mayo risk score (r = -0.716), MELD score (r = -0.554), Amsterdam-Oxford-PSC score (r = -0.465), LFTs including bilirubin (r=-0.521), aspartate aminotransferase (r=-0.449), alkaline phosphatase (r=-0.428), INR (r = -0.552), Alanine aminotransferase (r=-0.286) and Gamma-glutamyl transferase (r=-0.202) and a significant positive correlation with albumin (r=0.552). Furthermore, correlation between RE in HBP and qualitative MRI findings revealed a significant negative correlation with biliary duct dilatation (r=-0.286), significant bile duct stenosis (r=-0.204), lobar atrophy (r=-0.122), T2 hyperintensity (r=-0.096), and bile duct caliber irregularity (r=-0.112).

Gadoxetic acid-enhanced MRI can be used as a surrogate parameter for the evaluation of liver function and monitoring of disease progression in patients with PSC.

No correlation with ERCP or clinical outcome.

The study was approved by the institutional review board.

No funding was received for this work.

Purpose:

To evaluate the utilisation of gadoxetic acid excretion in the hepatobiliary phase of gadoxetic acid-enhanced magnetic resonance imaging (MRI) as a prognostic parameter for disease progression and liver function in patients with primary sclerosing cholangitis (PSC).

We retrospectively identified 120 patients (87 male, 33 female; mean age 42.58±12.3 years) with confirmed diagnosis of PSC who underwent 259 gadoxetic acid-enhanced MRI examinations between January 2008 and May 2019. Patients were divided into excreting (181 patients) and non-excreting (78 patients) groups depending on the presence or absence of gadoxetic acid excretion in the common bile duct (CBD) in hepatobiliary phase (HBP) respectively. Both groups were compared regarding relative liver enhancement (RE) in HBP, liver function tests (LFTs), and different prognostic clinical scoring systems including model for end-stage liver disease (MELD) score, Mayo risk score, and Amsterdam-Oxford-PSC score.

Relative liver enhancement (RE) in HBP was significantly higher (P < 0.005) in the excreting group while clinical scoring systems (Mayo risk score, MELD score, and Amsterdam-Oxford-PSC score) were significantly higher (P < 0.005) in the non-excreting group. LFTs comparison between both groups demonstrated a significantly higher levels of bilirubin (P < 0.005), aspartate aminotransferase (P = 0.001), alkaline phosphatase (P < 0.005), and INR (P = 0.001) in the non-excreting group and a significantly higher level of serum albumin (P < 0.005) in the excreting group.

Gadoxetic acid excretion in the hepatobiliary phase of gadoxetic acid-enhanced MRI can be used as a prognostic parameter for monitoring of disease progression and evaluation of liver function in patients with PSC.

No comparison regarding clinical outcomes or ERCP findings.

The study was approved by the institutional review board.

No funding was received for this work.