• Pancreas
• RM

Background - Magnetic Resonance (MR) has acquired more and more weight in the presurgical evaluation of pancreatic cancer. The necessity to identify reliable biomarkersto perform a correct tumour staging and effective treatment planningled to an increaseof interest in quantitative imaging. However traditional quantitative sequences were shown to be too time-consuming to be used in vivo. In this scenario, MR fingerprinting (MRF) has recently been proposed for the acquisition of quantitative data in a shorter time.
Purpose - To assess the concordance between the T1-and T2-conventional maps with the MRF maps; To identify tumours with different degrees of aggressiveness among resectable pancreatic cancer by MRF quantitative data.
Materials and Methods - This is sub-analysis obtained from data collected in a prospective study entitled EXplanted LIver and ex-vivo Pancreatic Specimen Evaluation “EXLIPSE” (Protocol ID 13646; ClinicalTrials.gov ID NCT04012021). A total of 10 specimens, all obtained from pancreatic surgery, were included in the study. 7T MRI was performed on a Discovery MR 950 MR scanner (GE Healthcare) equipped with a 2‐channel transmit/8‐channel receive knee coil (autonomously customized). Imaging protocol included both morphological sequences (T2-weighted Fast Turbo Spin Echo -3D CUBE-acquisitions; T1-and T2-weighted water-only images obtained with IDEAL technique-Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares estimation) and quantitative imaging acquisitions (T1-map: MP2RAGE acquisition; T2-map: multi-echo Spin Echo-MSE-acquisition; multiparametric acquisition in the transient state -MRF).
Results - Statistical analysis confirmed a strong correlation between the conventional T1 maps and the 2D MRF T1 (p=0.0008). Good statistical correlation was also found between conventional IDEAL Water Fraction (WF) and MRF WF (p=0.0010) but not between conventional IDEAL Fat Fraction(FF) and FF MRF (p=0.0658). Only on conventional quantitative maps the difference between the normal tissue and tumour was statistically significant (T2 conventional map p=0,0122) (T1 conventional map p=0,0033). The correlation between the histological grading of the neoplasm and the values of the MRF maps was not statistical significant (p>0,05).
Conclusions - Our study confirmed the reliability of MRF maps in the quantitative analysis of the explanted specimens. Short acquisition time that characterizes the MRF sequences allows us to hypothesize a use in vivo as breath hold sequences. Collecting a larger sample may be necessary to evaluate the possibility of obtaining information on histological grading.