Room: Karl Dean Ballroom B1
Purpose: To construct and validate phantoms with tissue-matching NMR properties (T1 and T2) for optimization of MR images for cervical cancer brachytherapy planning.
Methods: Three carrageenan-based MRI phantoms were fabricated with relaxation times simulating healthy endometrium, muscle tumors, and cervical cancer. Relevant phantom materials included carrageenan (gelling agent), GdCl₃ (T1 modifier), agarose (T2 modifier), and NaN₃ (antiseptic). Empirical equations relating NMR properties to phantom material concentrations established by previous studies were used to construct these phantoms for a 1.5-T clinical MRI scanner. Target phantom T1 and T2 values obtained from previously published data were 1881ms (T1) and 101ms (T2) for endometrium, 1083ms (T1) and 87ms (T2) for muscle tumor, and 1881ms (T1) and 113ms (T2) for cervical cancer. Phantom NMR properties were measured for validation via standard inversion recovery (TR=5000ms, TE=6.3ms, and TI range 150-3000ms) and spin-echo techniques (TR=9600ms, TE range 6-150ms) along with open-source and in-house software.
Results: Measured T1 times (mean±SD) for the endometrium, muscle tumor, and cervical cancer phantoms were 2101±104, 1380±44, and 2100±91ms respectively. Corresponding measured T2 times were 126±3, 107±2, and 139±4ms respectively. While the measured T2 times differed from the target values (mean 23.85% difference) more so than the T1 times (mean 16.90% difference), the phantom NMR properties were deemed acceptable for purposes of MR image optimization.
Conclusion: Carrageenan-based MR phantoms were constructed to mimic NMR properties of healthy endometrium, muscle tumors, and cervical cancer. Validation tests of these phantoms showed they produced relaxation times similar to the human tissues for which they were originally designed. In the future, the phantoms will be used to optimize MR image contrast between cancerous and healthy tissue for MR-guided brachytherapy of cervical cancer.