Identification And Thermometry Of Brown Adipose Tissue Using Magnetic Resonance Imaging

Lots of studies have demonstrated that the brown adipose tissue(BAT)in mammal is an important heat production organ.Being activated,BAT burns fat and dissipates energy in heat.Thus,BAT can be considered as the next target for treating obesity and other metabolic diseases.Temperature change in BAT can characterize the activity of BAT,which provides a useful tool for the research on the relationship between activity of BAT and metabolic diseases.This relationship may have great clinical significance and provides a new evaluation tool for obesity drugs development.Targeting on the identification and thermometry in BAT based on magnetic resonance imaging(MRI),the whole study includes three parts: robust chemical shift encoding-based fat-water separation algorithm,correction of bipolar readout sequence induced eddy current and development of a fat-referenced proton resonance frequency shift(PRFS)thermometry with high accuracy.Finally a preliminary experimental study in rodents for BAT identification and activity assessment using the proposed methods.First,due to the different fat content in adipocytes of BAT and white adipose tissue(WAT),the fat-water separation method based on chemical shift encoded imaging is a powerful technique to idenfy BAT from WAT.What is more,the separated results provide reliable initial values for thermometry in fat-containing tissue which is a key step for accurate temperature map in the latter study.The biggest challenge in fat-water separation is the fat water ambiguity problem.This study proposed a novel region growing algorithm with self-feeding phasor estimation for robust fat-water separation.The proposed method was compared to traditional region growing methods,multiresolution methods,and graph-cut methods using data from the ISMRM 2012 Challenge.This Challenge included 17 multi-echo datasets from different institutions with known separation results as reference.We tested all 17 datasets with the proposed method and compared the results with the reference results.All methods were scored on a scale of 0 to 10,000.Larger score means better performance.The average score of all 17 datasets from the ISMRM 2012 Challenge was 9928 with 13 of the 17 scores surpassing 9900.There was no apparent fat-water swap observed throughout these datasets of the proposed method.The self-feeding mechanism of phasor estimation ensures the reliability of seed pixel selection at the finest resolution.Compared to traditional multiple resolution methods and region growing methods,the proposed method is more robust when applied to disjoint areas and to regions with strong field inhomogeneity.Second,multi-echo GRE sequence with bipolar readout gradients can reduce the echo spacing and increase the image SNR efficiency,which is significant for the increase of the accuracy of BAT idenfication and precision of temperature map.However,bipolar readout gradients with opposite polarity would introduce an additional eddy current induced-phase(EC-phase)between odd and even echoes.This study proposed a novel EC-phase estimation method considering the high order terms in EC-phase.The nonlinearity of the EC-phase is precisely approached by a hierarchical iterative linear-fitting algorithm(HILA)through the whole image.The accuracy of FF from the proposed HILA method was evaluated using the FF from unipolar readout dataset as reference through fat-water phantom experiment and in vivo human experiment.The phantom study shows that the maximum mean FF error after EC-phase correction using the proposed HILA method is smaller than 2%,implying HILA can approximate the high-order term of EC-phase through the step-wise linear fitting.In conclusion,the proposed HILA method provides a simple and efficient EC-phase correction method for bipolar acquisition without acquiring additional data.Third,assessing the BAT activity is critical to the BAT related studies.The noninvasive thermometry in activated BAT based on MRI may play an important role in BAT characterization.Since the proton frequency in lipids is insensitive to temperature changes,the traditional proton resonance frequency shift-based(PRFS)thermometry fails in BAT thermometry.In this study,a novel fat-referenced PRFS MR thermometry was proposed,considering several confounding factors in previously proposed method,such as the multiple fat peak model,temperature dependency on T2*.To increase the measurement precision,a dual-step iterative temperature estimation algorithm was proposed.The performance of the DITE was evaluated with a Monte Carlo simulation,a fat-water phantom and ex vivo brown adipose tissue(BAT)experiments and then compared with the performance of previously proposed methods.Our experiments demonstrate that the temperatures estimated by the DITE are highly consistent with thermometer results.Besides,DITE shows better accuracy and precision than that of previous multi-echo fat-referenced PRFS methods.In the BAT heating experiment,the average mean error(ME)/standard deviation of the error(SD)/root mean squared error(RMSE)are-0.08/0.46/0.56 °C over all the measurements within the region of interest(ROI),respectively.As a result,our proposed DITE method can improve both the accuracy and precision of temperature measurements in fat-containing tissues.Finally,the DITE method was applied to acquire the temperature distribution and FF image in activated BAT in vivo.Similar patterns of the temperature and FF change in BAT was observed in all four rats.Further studies include the validation of the proposed method in BAT using gene expression measurements.The study may provide a powerful tool for BAT activity assessment and related anti-obesity therapy development.