Implementing a fast MR screening protocol for pancreas in patients with mutations of BRCA1 and BRCA2: preliminary data on Diffusion-Weighted Images.
Andrea Agostini1, Mitchell Raeside1, Richard Do1, Amita Shukla-Dave2, David Aramburu Nunez2, Ramesh Paudyal2, Olga Smelianskaia1, Maggie Fung3, Monika Khan1, David Kelsen4, Gabriella Carollo5, and Lorenzo Mannelli1

1Department of Radiology, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, 2Department of Medical Physics, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, 3Global MR Applications and Workflow, GE Healthcare, New York, NY, United States, 4Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, United States, 5Saint Vincent Ferrer High School, New York, NY, United States


The aim of the study is the qualitative and quantitative evaluation of Diffusion-weighted images (DWI) with reduced field-of-view (rFOV), that are incorporated into a fast rapid MR pancreas screening protocol (MRpsp). We are developing a fast rapid MRpsp to be performed in carriers of BRCA mutations after a breast MRI, using the built-in body coil of a 3T magnet without repositioning a patient who is in the prone position. 15 patients were scanned. For b-values <100 s/mm2, b=20 s/mm2 DWI had the best contrast-to-noise and signal-to-noise ratios. Among b-values >100 s/mm2, b=250 s/mm2 DWI showed the highest diagnostic information.


Patients with mutations of BRCA have higher risk of pancreatic cancer (3-10 fold for BRCA2; 2.6-3.6 fold for BRCA11), and thus should be considered for pancreatic screening programs2. At present there is no accepted screening protocol, due to uncertainties in diagnostic performance, potential costs, and invasiveness of some diagnostic tests. Taking advantage of the fact that carriers of BRCA mutations undergo MRI screening for breast cancer, we are developing and optimizing a < 10 minutes fast MR pancreas screening protocol (MRpsp) to be performed at the conclusion of a breast MRI without patient repositioning and using the built-in-body-coil. The protocol includes: axial and coronal T2, axial T1, and reduced Field of View (rFOV) diffusion weighted images (DWI). The aims of this study are: 1) To compare the quality of different b values in rFOV DWI, and 2) to evaluate the inter-reader agreement in normal pancreas ADC measurements.


This prospective study was approved by the local IRB.

Inclusion criteria were: female patients who are BRCA1 or BRCA2 mutation carriers undergoing breast MR screening. MRpsp was performed with a 3T magnet (MR 750, GE Healthcare, United States) after a contrast enhanced breast MR (Multihance, 0.2 ml/kg, Bracco Diagnostics, United States) with patients in the prone position and using the built-in-body-coil. DWI sequence parameters are reported in table 1. DWI of the pancreatic head, body, and tail was evaluated in consensus by two radiologists with a 5 points semi-quantitative score: 1= full diagnostic information is present; 2= most of diagnostic information is present; 3= some diagnostic information is present; 4= minimal diagnostic information is present; 5= non diagnostic. Quantitative analysis was performed by two radiologists with a commercially available post-processing workstation (Advantage Workstation 4.6, GE Healthcare, United States) placing circular, 5 mm diameter Region-of-Interest (ROI) in pancreatic head, body, and tail, and in peripancreatic fat. ROIs were automatically propagated through all the evaluated series and average signal intensity (SI) and standard deviation (SD) per ROI were recorded. Signal-to-Noise Ratio (SNR) was calculated as ROISI,Pancreas/ROISD,Pancreas. Contrast-to-Noise Ratio (CNR) was calculated as (ROISI,Pancreas–ROISI,Fat)/ROISD,Fat. Inter-reader agreement for ADC measurements was calculated with Intraclass Correlation Coefficient. For DWI, preliminary statistical analysis was conducted separately for lower b-value images (b<100 s/mm2) and for higher b-value images (b>100 s/mm2), considering average of measures of both readers without correction for the NEX variability for different b-values. Non-parametric tests were used, and p values <0.05 were considered significant.


Fifteen patients were imaged (Figure 1).

All low b-value (<100 s/mm2) DWI were ranked as fully diagnostic with the exception of one pancreatic tail (subject 2, rank 3= some diagnostic information, due to inconsistent breathing). As expected, DWI with b-values >100 s/mm2 showed higher variability in diagnostic image quality: b=250 s/mm2 DWI was ranked significantly better than b=500 s/mm2 and b= 800 s/mm2 DWI (Wilcoxon, p<0.05). No other significant differences were recorded (Table 2). For DWI with b<100 s/mm2, SNR was significantly higher in the pancreatic head, body and tail when using b=20 s/mm2. For DWI with b>100 s/mm2, no relevant differences were recorded in terms of SNR (Table 3). For DWI with b<100 s/mm2 images CNR was significantly higher in the pancreatic head, body and tail, when using b=20 s/mm2. For b>100 s/mm2 images, no significant differences in CNR were demonstrated (Table 3). ADC values (x10-3 mm2/s) in pancreatic head, body and tail were respectively for reader 1: 1.35; 1.69; and 1.65; for reader 2: 1.47; 1.55; and 1.59. Values were comparable with the ranges reported in the literature3. Intra-class correlation coefficients between the two readers measuring ADC in different pancreatic regions were 0.6860 (head), 0.8800 (body), and 0.7143 (tail) (Table 4).


All rFOV DWI obtained with built-in-body coil and patient in prone position were of diagnostic quality. At b<100 s/mm2, all images were of high diagnostic quality. Among b<100 s/mm2 the highest SNR and CNR were obtained when using, b=20 s/mm2. Among b>100 s/mm2 qualitatively better images were obtained using b=250 s/mm2. Limited inter-reader agreement was demonstrated for ADC measurements of the pancreas.


No acknowledgement found.


1. Chang MC, Wong JM, Chang YT, et al. Screening and early detection of pancreatic cancer in high risk population. World J Gastroenterol 2014; 20(9): 2358-2364.

2. Canto MI, Harinck F, Hruban RH, et al. International Cancer of the Pancreas Screening (CAPS) Consortium summit on the management of patients with increased risk for familial pancreatic cancer. Gut 2013;62:339–347.

3. Barral M, Taouli, Guiu B, et al. Diffusion-weighted MR Imaging of the Pancreas: Current Status and Recommendations. Radiology 2015; 274(1): 45-63.


Table 1. Technical parameters of DWI acquisition. NEX in different b-values have been modified to optimize image quality. 1: parameters are expressed as median and range when indicated. 2: Navigator Trigger.

Figure 1. DWI and ADC map, with ROIs placed on pancreatic head and locoregional fat. Signal Intensity and Standard Deviation have been recorded. a) b=0 s/mm2, b) b=20 s/mm2; c) b=50 s/mm2; d) b=80 s/mm2; e) b=250 s/mm2; f) b=500 s/mm2; g) b=800 s/mm2; h) ADC map.

Table 2. Qualitative assessment with a 5 points score (1: full diagnostic information; 5: non diagnostic, see text). Median values and ranges are reported. Among high b-values, b=250 s/mm2 DWI have better score than b=500 s/mm2 and b=800 s/mm2 (Wilcoxon, p<0.05).

Table 3. Table showing SNR (ROISI,Pancreas/ROISD,Pancreas) and CNR ((ROISI,Pancreas–ROISI,Fat)/ROISD,Fat). Median and 25 – 75 percentile are reported. Among DWI with b<100 s/mm2, b=20 has the best CNR and SNR (Wilcoxon p<0.05). No substantial differences in CNR and SNR among b>100 s/mm2 (Wilcoxon, p>0.05).

Table 4. Table showing ADC measurements (x10-3 mm2/sec) of both readers for pancreatic head, body and tail. Mean values, Standard Deviation (SD) and ranges are reported. Reproducibility of measurements is tested with Intraclass Correlation Coefficient.

Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)