Program Information
Amplitude-Based 4D-PET/MRI for Target Motion Characterization
J Patrick1*, J Thiessen2 , D Capaldi3 , R Thompson4 , R Stodilka5 , M Drangova6 , S Gaede7 , (1) London Regional Cancer Program, London, ON, (2) Lawson Health Research Institute, London, Ontario, (3) Robarts Research Institute, London, Ontario, (4) Lawson Health Research Institute, London, Ontario, (5) London Health Sciences Centre, London, Ontario, (6) Robarts Research Institute, London, Ontario, (7) London Regional Cancer Program, London, Ontario
Presentations
MO-F-CAMPUS-IT-1 (Monday, July 31, 2017) 4:30 PM - 5:30 PM Room: Imaging ePoster Theater
Purpose: The aim of this work is to evaluate a retrospective amplitude-based 4D-PET/MRI post-processing technique, which utilises a surrogate respiratory signal with commonly available MRI and PET acquisition methods to characterise target respiratory motion.
Methods: An MR-compatible motion stage moved four spherical lesion-compartments (28, 22, 17, 10 mm diameter) within lung-compartments of a torso phantom while being scanned with a 3T PET/MRI. Both the torso and lesion-compartments were filled with aqueous solutions of 18F, simulating a ~4:1 tumour-to-background ratio. Three respiratory motion scenarios were considered, including 2 irregular real patient profiles and a modified sine wave. For each scenario, dynamic 2D-MR images were acquired in each of 53 adjacent coronal planes using an ultrafast gradient echo pulse sequence. Simultaneously, dynamic PET was acquired in list-mode. To achieve our 4D-PET/MRI technique, all 2D-MR and PET were retrospectively sorted into 10 phases according to amplitude and slope of a surrogate respiratory trace. A 4D μ-map was generated by applying transformations, calculated from 4D-MRI to a single μ-map acquired while the phantom was static. Lesion-compartment centroids in PET, MRI, and μ-maps were compared with known positions. CNR measurements and compared with single-gate and ungated PET.
Results: 4D-PET CNR values were comparable to single-gate PET, improving on average by 10%. Centroid measurements in PET, MR, and μ-maps show excellent correlation (R2 = 0.82 to 1.00) with known positions. However, correlation coefficients were lowest for cross-plane motion in MR for smallest lesion-compartment (R2 = 0.82).
Conclusion: This work demonstrates the feasibility of a novel retrospective 4D-PET/MRI technique with attenuation correction in a PET/MR/CT compatible tumour respiratory motion phantom. Our technique has the advantage of using a clinically available pulse sequence, reducing the need for less available more exotic sequences or MR pulse sequence programming.
Contact Email: