• Quantitative metric for assessment of pancreatic ductal adenocarcinoma treatment response in T1-weighted gadolinium-enhanced magnetic resonance imaging

      Liau, J.; Vedantham, S.; Babiker, H.M.; McGlothin, T.; Martin, D.R.; Department of Medical Imaging, University of Arizona; Biomedical Engineering, University of Arizona; Early Phase Clinical Trials Program, University of Arizona Cancer Center (AME Publishing Company, 2020)
      Background: We theoretically derived a new quantitative metric reflecting the product of T1 signal intensity and contrast media concentration (T1C) using first principles for the signal provided by the gradient echo sequence. This metric can be used with conventional gadolinium contrast-enhanced magnetic resonance imaging (CE-MRI) exams. We used this metric to test our hypothesis that gadolinium enhancement changes with pancreatic ductal adenocarcinoma (PDA) treatment response, and that this metric may differentiate responders from non-responders. Methods: Out of 264 initially identified patients, a final total of 35 patients with PDA were included in a retrospective study of responders (n=24) and non-responders (n=11), which used changes in cancer antigen 19-9 (CA 19-9) and tumor size as reference standards. T1C was computed for the pancreatic mass in the arterial, portal venous, and delayed phases in pre-treatment and post-treatment MRIs. Changes in measurements and correlations with treatment response were assessed by repeated measures analysis of variance and paired t-tests. Results: In the treatment responder group, T1C significantly increased in the arterial, portal venous, and delayed phases (P=7.57e-5, P=3.25e-4, P=1.75e-4). In the non-responder group, T1C did not significantly change in any phase (P>0.58). Post-treatment T1C significantly differed between responders and non-responders (P=0.044) by repeated measures analysis of variance. Conclusions: T1C significantly increases in all phases of CE-MRI in responders to treatment, but does not change in non-responders. T1C correlates with treatment response, can be computed from clinical MRI exams, and may be useful as an additional metric to stratify patients undergoing treatment. © 2021 Annals of Pancreatic Cancer. All rights reserved.
    • Slice-stacking T2-weighted MRI for fast determination of internal target volume for liver tumor

      Han, S.; Liang, X.; Li, T.; Yin, F.-F.; Cai, J.; Department of Biomedical Engineering, University of Arizona (AME Publishing Company, 2021)
      Background: To investigate the feasibility of generating maximum intensity projection (MIP) images to determine internal target volume (ITV) using slice-stacking MRI (SS-MRI) technique. Methods: Slice-stacking is a technique which applies a multi-slice MRI acquisition to generate a 3D MIP for ITV contouring, without reconstructing 4D-MRI. 4D digital extended cardiac-torso (XCAT) phantom was used to generate MIP images with sequential 2D HASTE sequence, with different tumor diameters (10, 30 and 50 mm) and with simulated regular and irregular (patient) breathing motions. A reference MIP was generated using all acquisition images. Consecutive repetitions were then used to generate MIP to analyze the relationship between Dice's similarity coefficient (DSC) and the number of repetitions, and the relationship between the relative ITV volume difference and the number of repetitions. Images from XCAT phantom and from three hepatic carcinoma patients were collected in this study to demonstrate the feasibility of this technique. Results: For both regular and irregular breathing motion, the average DSC of ITV is >0.94 and the average relative ITV volume difference is <10% (approximately 0.15 cm3) when using 5 repeated scanning images to reconstruct MIP for tumor diameter of 10 mm. As tumor diameter increases, the DSC of ITV is >0.97 and the relative ITV volume difference is <5% for regular breathing motion, and the DSC of ITV is >0.97 and the relative ITV volume difference is <5.5% for irregular breathing motion when using 5 repeated scanning images to reconstruct MIP. In patient image study, the mean relative ITV volume difference is <3% and the mean DSC is 0.99 when using 5 repeated scanning images to reconstruct MIP. Conclusions: The number of scans required to generate tumor ITV for slice-stacking method (5-7 repetition) is 3-4 times less than that of 4D-MRI (15-20 repetitions). It is feasible to generate a fast clinically acceptable ITV using slice-stacking method with sequential 2D MR images. © Quantitative Imaging in Medicine and Surgery. All rights reserved.