Quantitative lung perfusion evaluation using Fourier decomposition perfusion MRI.

PURPOSE: To quantitatively evaluate lung perfusion using Fourier decomposition perfusion MRI. The Fourier decomposition (FD) method is a noninvasive method for assessing ventilation- and perfusion-related information in the lungs, where the perfusion maps in particular have shown promise for clinical use. However, the perfusion maps are nonquantitative and dimensionless, making follow-ups and direct comparisons between patients difficult. We present an approach to obtain physically meaningful and quantifiable perfusion maps using the FD method. METHODS: The standard FD perfusion images are quantified by comparing the partially blood-filled pixels in the lung parenchyma with the fully blood-filled pixels in the aorta. The percentage of blood in a pixel is then combined with the temporal information, yielding quantitative blood flow values. The values of 10 healthy volunteers are compared with SEEPAGE measurements which have shown high consistency with dynamic contrast enhanced-MRI. RESULTS: All pulmonary blood flow (PBF) values are within the expected range. The two methods are in good agreement (mean difference = 0.2 mL/min/100 mL, mean absolute difference = 11 mL/min/100 mL, mean PBF-FD = 150 mL/min/100 mL, mean PBF-SEEPAGE = 151 mL/min/100 mL). The Bland-Altman plot shows a good spread of values, indicating no systematic bias between the methods. CONCLUSION: Quantitative lung perfusion can be obtained using the Fourier Decomposition method combined with a small amount of postprocessing.

Quantitative lung ventilation using Fourier decomposition MRI; comparison and initial study.

OBJECTIVE: The Fourier decomposition (FD) method is a noninvasive method for assessing ventilation and perfusion-related information in the lungs, but the lack of quantifiable values is a drawback. We demonstrate a novel technique for quantification of the FD ventilation maps, compare it to two published methods, and show results from both healthy volunteers and patients diagnosed with lung cancer. MATERIALS AND METHODS: We quantified the standard FD ventilation images by utilizing additional information, i.e., the zero-frequency component image, which is also obtained from the Fourier analysis. This image acts as a baseline for the changes recorded in the FD ventilation image and can therefore be used to calculate the ventilation. Using this technique, we compared the ventilation values from ten healthy volunteers and ten patients to two previously published methods for quantitative ventilation assessment. RESULTS: All methods showed good overall agreement (mean difference between the methods was 14-38 ml/min). The mean minute ventilation for the FD method was calculated to be 693 ml/min for a 2D slice, which is in the expected range. CONCLUSION: The zero-frequency component image can be used as a baseline to quantify the FD ventilation maps. Our initial study showed good agreement with published methods in healthy volunteers, but less so in patients with lung cancer.

Non-invasive quantitative pulmonary V/Q imaging using Fourier decomposition MRI at 1.5T.

OBJECTIVES: Techniques for quantitative pulmonary perfusion and ventilation using the Fourier Decomposition method were recently demonstrated. We combine these two techniques and show that ventilation-perfusion (V/Q) imaging is possible using only a single MR acquisition of less than thirty seconds. METHODS: The Fourier Decomposition method is used in combination with two quantification techniques, which extract baselines from within the images themselves and thus allows quantification. For the perfusion, a region assumed to consist of 100% blood is utilized, while for the ventilation the zero-frequency component is used. V/Q-imaging is then done by dividing the quantified ventilation map with the quantified perfusion map. The techniques were used on ten healthy volunteers and fifteen patients diagnosed with lung cancer. RESULTS: A mean V/Q-ratio of 1.15 ± 0.22 was found for the healthy volunteers and a mean V/Q-ratio of 1.93 ± 0.83 for the non-afflicted lung in the patients. Mean V/Q-ratio in the afflicted (tumor-bearing) lung was found to be 1.61 ± 1.06. Functional defects were clearly visible in many of the patient images, but 5 of 15 patient images had to be excluded due to artifacts or low SNR, indicating a lack of robustness. CONCLUSION: Non-invasive, quantitative V/Q-imaging is possible using Fourier Decomposition MRI. The method requires only a single acquisition of less than 30 seconds, but robustness in patients remains an issue.

Whole body sodium MRI at 3T using an asymmetric birdcage resonator and short echo time sequence: first images of a male volunteer.

Sodium magnetic resonance imaging (²³Na MRI) is a non-invasive technique which allows spatial resolution of the tissue sodium concentration (TSC) in the human body. TSC measurements could potentially serve to monitor early treatment success of chemotherapy on patients who suffer from whole body metastases. Yet, the acquisition of whole body sodium (²³Na) images has been hampered so far by the lack of large resonators and the extremely low signal-to-noise ratio (SNR) achieved with existing resonator systems. In this study, a ²³Na resonator was constructed for whole body ²³Na MRI at 3T comprising of a 16-leg, asymmetrical birdcage structure with 34 cm height, 47.5 cm width and 50 cm length. The resonator was driven in quadrature mode and could be used either as a transceiver resonator or, since active decoupling was included, as a transmit-only resonator in conjunction with a receive-only (RO) surface resonator. The relative B1-field profile was simulated and measured on phantoms, and 3D whole body ²³Na MRI data of a healthy male volunteer were acquired in five segments with a nominal isotropic resolution of (6 × 6 × 6) mm³ and a 10 min acquisition time per scan. The measured SNR values in the ²³Na-MR images varied from 9 ± 2 in calf muscle, 15 ± 2 in brain tissue, 23 ± 2 in the prostate and up to 42 ± 5 in the vertebral discs. Arms, legs, knees and hands could also be resolved with applied resonator and short time-to-echo (TE) (0.5 ms) radial sequence. Up to fivefold SNR improvement was achieved through combining the birdcage with local RO surface coil. In conclusion, ²³Na MRI of the entire human body provides sub-cm spatial resolution, which allows resolution of all major human body parts with a scan time of less than 60 min.