MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

A Multi-Phase Cross-Modality Deformable Bio-Tissue Phantom for Deformable Image Registration Validation

A Qin*, M Snyder, M Liu, X Ding, W Zheng, Q Liu, S Chen, J Liang, D Yan, Beaumont Health System, Royal Oak, MI

Presentations

(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose:
To develop a deformable phantom with bio-tissue material to validate deformable image registration (DIR) accuracy.

Methods:
A phantom was developed with pork liver embedded in the deformable foam. Fresh liver samples were purchased from food market. Two intact liver lobes were collected for phantom development. Four 3cm-length 0.5mm-Visicoils were cut into gold marks of length about 1mm. The markers were then implanted into the liver samples with a needle, trying to get an even distribution inside. Silicone casting foam was used to wrap the liver sample and construct the final deformable phantom. Four phases of deformation were simulated with the phantom1 under different external compression(10, 20, 20-RL, 25mm). Phantom2 was steamed to simulate tissue shrinkage. For each phase, the phantoms were scanned with CT and MR(T1, T2,BTFE) under the exact same deformation state. The gold marks were manually located throughout all phases on CTs and are acting as the ground truth for DIR evaluation. Two commercial DIR were evaluated between un-deformed and deformed phases on the masked-CT (gold marks digitally removed) and MRs, using target registration error(TRE) to quantify the DIR accuracy. In addition, an in-house biomechanical-based tool was evaluated on phantom2.

Results:
62 and 43 landmarks were identified on phantom1 and phantom2, respectively. The TRE of two DIR tools are close on CT, but significantly different on MRs. For example, (the mean and 90th percentile TRE) of Phantom_1_phase_4 is(1.9mm and 4.2mm) for DIR1, (2.0mm and 3.2mm)for DIR2 on CT; while on MR_T1 is (4.6mm and 9.0mm) for DIR1, (1.3mm and 2.4mm)for DIR2. On phantom2(shrunk to56.5%) , better mean TRE was achieved by DIR2(5.3, 4.6mm for CT_masked, MR_T1) compared to DIR1, and was improved to 4.1mm with our biomechanical DIR tool.


Conclusion:
A bio-tissue phantom was developed and demonstrated to be effective in evaluating DIR accuracy multiple imaging modalities.

Funding Support, Disclosures, and Conflict of Interest: This research is partially supported by Elekta Research grant

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email