Voxel-based morphometry（VBM）は，3D-T₁WIを用いて脳容積を評価する手法である．3D-T₁WIの信号値が解析結果に強く影響するため，信号対雑音比（SNR）が低下すると測定精度が低下する，空間的歪みが大きいと空間的正規化精度が低下する，静磁場強度の相違により解析結果は影響される，信号不均一性の影響で測定精度が低下するなどの技術的な問題点があるため，これらの特性を理解し，結果を解釈する必要がある．

　VBM解析ソフトウェアとして代表的なものにstatistical parametric mapping（SPM）がある．Voxel-based specific regional analysis system for Alzheimer's disease（VSRAD）ソフトウェアはSPMソフトウェアの解析アルゴリズムを一部利用しVBM解析の汎用性を高めたものであり，VBM解析の臨床普及にも貢献した．VBM解析では3D-T₁WIを灰白質画像や白質画像に分画（セグメンテーション）した後，ボクセルごとに統計解析するために，個々の脳に対して空間的正規化を行う．この空間的正規化では，線形と非線形の形態変形が行われるため，分画エラー（ミスセグメンテーション）の存在により正規化エラー（ミスレジストレーション）が発生し，ミスセグメンテーションが生じた領域のみならず，離れた場所でも解析エラーを観察することがある．また，T₁WI以外のコントラスト画像を用いたVBM解析の可能性も示されている．

　VBMで算出される脳容積は真値との比較が困難であるため，解析パラメータや前処理方法，撮像条件や解析する画像コントラスト最適化の検討において，なにが正しいのかを結論付けるには，多くの修飾因子が影響し合う点（例えば，静磁場強度の相違により，最適な信号不均一補正法や空間的歪み補正は変化する可能性がある．）に注意する必要がある．また，VBM解析法はミスレジストレーションやミスセグメンテーションを含んだ結果を示している．しかし，全脳を探索的に解析する手法としては非常に汎用性が高くそのメリットは大きい．使用する側が探索的ツールであることを理解することで，ミスリードされるリスクも大幅に減ると考える．

　When patients with implanted devices are referred to MRI examination, radiologists have to check if it is safe to perform the examination along with these medical devices. Most of the latest implants consist of non-magnetic materials, which do not create a contra-indication for the MRI. However, in the case of the patients with “MR conditional” implants, a specific MRI environment, such as a magnetic field and a scan parameter, is required. In addition, the specified condition for MRI varies with each device. Therefore, this makes it difficult for the radiologists to modify the MRI environment for each patient.

　In addition to this check of the implanted device at the timing of referral to MRI examination, the radiologist who is responsible for the examination has to confirm the scan safety before the examination.

　MRI scan for patients with implanted medical devices that are contraindicated for the examination should be avoided. However, unnecessary cancellation of the examination owing to insufficient information regarding the medical devices should also be avoided.

　For larger blood vessels, such as cervical arteries and aortic arteries, 4D-Flow imaging with high signal to noise ratio (SNR) can be used to collect accurate measurements. When the SNR is sufficient and the voxel size is less than 30% of the vessel diameter, the error rate for the cross-sectional average flow velocity obtained by 4D-Flow is less than 10%. When the SNR is sufficient and the voxel size is less than 10% of the vessel diameter, error rate for the maximum flow velocity is also less than 10%. However, for smaller vessels, such as intracranial arteries, 4D-Flow imaging underestimates the flow velocities owing to the low spatial resolution or low SNR. Meanwhile, because of the partial volume phenomenon, the velocity of each voxel is underestimated within the vessel and overestimated near the vessel wall. Thus, the spatial resolution affects the velocity profile in the blood vessels. Higher spatial resolution leads to more accurate velocity profile and more accurate wall shear stress (WSS). However, it should be noted that the WSS determined by 4D-Flow is smaller compared to the true value.

　We can obtain the 3D velocity vector fields, maximum flow velocity, spatially averaged flow velocity, volume flow rate, streamlines, pathlines, streak lines, and WSS and its derivatives using a flow analysis software.

　The spatial resolution and SNR of 4D-Flow affects the accuracy of each voxel, velocity profile in blood vessels, and ultimately, the calculated WSS. However, there is a trade-off between the spatial resolution and SNR and hence there are limitations to increase the spatial resolution. Artificial intelligence (AI) may be able to interpolate lower spatial resolution data, and therefore, address this problem in the future. AI may also help us to obtain flow related biomarkers like WSS and its derivatives more easily and quickly in clinical practice. Development of the magnetic resonance fluid dynamics is ongoing and can provide a promising solution.

　The quantitative evaluation of liver function is important, not only for monitoring, but also for the preoperative evaluation of the functional hepatic reserve. The indocyanine green (ICG) clearance has been used as a reliable quantitative liver function test; however, ICG clearance cannot evaluate segmental liver function. Gd-EOB-DTPA is a bolus-injectable paramagnetic contrast media taken up by hepatocytes via organic anion transporting peptide, specifically expressed on cell membrane of hepatocytes; therefore quantitative evaluation of segmental liver function with use of Gd-EOBD-TPA-enhanced MR imaging is feasible. Methods for quantitative evaluation of liver function using Gd-EOB-DTPA-enhanced MR imaging can be divided into the following two major categories : dynamic and static methods. In the dynamic method, the time-intensity curve obtained from dynamic contrast enhanced MR imaging is analyzed using various pharmacokinetic models. The compartment model analysis is a representative method, categorized as a dynamic model-dependent analysis. In the static method, post contrast enhanced images, usually during the hepatobiliary phase, are analyzed using relaxation time or signal intensity of the liver. The static methods are easy to apply in clinical practice. The hepatocellular uptake index (HUI) is a static signal intensity-based method, which is obtained from the signal intensity and volume of the liver and spleen on Gd-EOB-DTPA-enhanced MR images during the hepatobiliary phase. HUI has been accepted as the most reliable predictor of segmental liver function. In this article, pharmacokinetics of Gd-EOB-DTPA is first described using the compartment model. Subsequently, derivation of 2-in-liner-1-uptake-1-out-2-compartment model and HUI are shown as examples of quantitative evaluation of the liver function with the use of Gd-EOB-DTPA-enhanced MR imaging.