Free-breathing magnetic resonance fingerprinting for liver cancer diagnosis and treatment

Abstract

Purpose: This study aims to develop free-breathing MRF (FB-MRF) techniques for the diagnosis and treatment of liver cancer and evaluate its clinical efficacy using liver cancer patients. Specifically, we aim to develop a motion-robust three-dimensional MRF (MR­-3DMRF) technique for liver cancer diagnosis and develop a respiratory-correlated four-dimensional MRF technique (RC-4DMRF) for liver tumor motion management in radiation therapy. Methods and Materials: Thirteen liver cancer patients with hepatocellular carcinoma (HCC) were prospectively enrolled in this study. The k-space MRF signals of the liver were acquired during free-breathing. To develop MR-3DMRF, the k-space signals from the end-of-exhalation (EOE) state were extracted as the target phase based on the amplitude of respiratory surrogates. Hereafter, the rest k-space signals were evenly distributed to five moving phases and were deformed to the target phase using the non-rigid registration algorithm. A dynamics-weighting strategy was adapted in MR-3DMRF to reduce the influence of dynamics from moving phases while maintaining the dynamic integrity in pattern matching. To develop RC-4DMRF, the MRF signals were binned into eight respiratory phases based on respiratory surrogates, and inter-phase displacement vector fields (DVFs) were estimated using a phase-specific low-rank optimization method. Hereafter, the tissue property maps were reconstructed using a pyramid motion-compensated method that alternatively optimized inter-phase DVFs and subspace images. The evaluation of MR-3DMRF and RC-4DMRF was performed on digital phantom and in vivo patients. The mean absolute percentage errors (MAPEs) of the MRF-derived tissue maps were calculated to reveal tissue quantification accuracy. The full-width half maximum (FWHM) of the organ boundary was used to assess the image quality. Amplitude motion differences (AMDs) and Pearson correlation coefficients (PCCs) were determined to assess measurement agreement in tumor motion between RC-4DMRF and cine MRI. Paired Student’s t-test was used for statistical significance analysis with a p-value threshold of 0.05. Results: MR-3DMRF successfully reconstructs motion-resolved tissue property maps at the end of expiration (EOE) motion state with high image quality. The FWHM of organ boundaries in MR-3DMRF-derived tissue maps is 3.1mm ± 1.8mm, significantly lower than motion-blurred tissue maps (10.1mm ± 4.3mm, p-value=6.3×10-7). The linearity of test-retest MRF measurement (2 value) was 0.982, 0.894, 0.925 for T1, T2, and PD, respectively. RC-4DMRF achieved excellent performance in liver tumor motion management. First, RC-4DMRF has a high agreement in motion measurement with cine MRI, yielding the mean (± standard deviation) PCCs of 0.95 ± 0.05 and 0.93 ± 0.09 and AMDs of 1.48 ± 1.06 mm and 0.81 ± 0.64 mm in the superior-inferior and anterior-posterior directions, respectively. Moreover, RC-4DMRF achieved high accuracy in tissue property quantification, with MAPEs of 8.8%, 9.6%, and 5.0% for T1, T2, and PD, respectively. Conclusion: FB-MRF was successfully developed and validated in a prospective cohort of liver cancer patients. MR-3DMRF provided accurate and repeatable tissue property maps from the free-breathing MRF acquisition scheme, demonstrating an excellent diagnostic performance for liver cancer. RC-4DMRF presented high accuracy in simultaneous respiratory motion measurement and tissue property quantification, holding great promise in improving the motion management of liver cancer radiation therapy.