The first nuclear magnetic resonance (NMR) signals from a living animal were acquired from an anesthetized rat in 1968 . The capability of NMR to differentiate tumors from normal tissue was reported by Damadian in 1971 . Magnetic resonance imaging (MRI) was developed by Lauterbur  based on the encoding spatial information of NMR signals with magnetic field gradients. The first cross-sectional image of a living mouse was published in 1974 by Lauterbur . The echo-planar imaging (EPI) technique was developed by Mansfield . The first MRI body scan of a human being was performed by Damadian in 1977 . For their efforts, Paul C. Lauterbur and Peter Mansfield were awarded the Nobel Prize in Physiology and Medicine in 2003 in justification of the fundamental importance and applicability of MRI in medicine. The k-space was patented by Richard S. Likes in U.S.A (#U.S. Patent 4,307,343). The first clinical MRI system was installed in the early 1980s. The first whole-body MRI scanner was developed in Korea in 1982 and was commercialized in 1984 by Goldstar (Seoul, Korea). At same time, another whole-body MRI system was built at the University of Aberdeen and was used at St. Bartholomew’s Hospital (London, UK) in 1983. Superconducting MRI was developed and it was clinically used in 1985 to 1988 at Siemens (Erlangen, Germany), General Electric (GE, Boston, MA, USA), and Goldstar. Fig. 1 shows the timeline of MRI developments and a summary of the major contributions achieved internationally and domestically.
The first MRI system was developed by the Korean Advanced Institute of Science and Technology (Daejeon, Korea), and was installed in Shin Hwa Hospital (Shin Hwa Nursing Hospital, Seoul, Korea) in 1984. This MRI scanner used a 0.2 T permanent magnet. Therefore, its use for clinical studies was limited considerably at that time. The first commercial MRI (Spectro-20000; Goldstar) was installed in Seoul University Hospital (Seoul, Korea) in 1987. This MRI system was a 2.0 Tesla superconductivity magnet. During this period, GE, and Siemens also developed the 1.5 T superconductive MRI systems. In 1990, the first animal 4.7 T MRI system (Biospec, Bruker, Switzerland) was installed in Seoul’s Asan Joong Ang Hospital (Asan Medical Center, Seoul, Korea). Among the many incremental developments in MRI in Korea, one of the notable progresses made during the early 21st century was the development of ultrahigh field (UHF) 7.0 T MRI and its applications [7 -11]. Another notable worldwide research activity was the development of the hybrid positron emission tomography–MRI (PET–MRI) system. Since the introduction of the UHF (7.0 T) MRI system in Korea, an effort was initiated for the construction of a hybrid system with 7.0 T MRI and PET that led to the development of one of the world’s most advanced PET–MRI systems in 2007. With these developments, a number of neurotransmitter studies was initiated to study the serotonergic distribution in the brainstem in vivo [7,8]. Another new and interesting development with UHF 7.0 T MRI was the super-resolution tractography initiative with resolution down to 200 µm .
In 1950, spin echoes and free induction decay were detected by Hahn [13,14]. Therefore, the spin echo (SE) is commonly referred to as Hahn’s echo. MR angiography was developed by Charles L. Dumoulin and Howard R. Hart at the General Electric in 1986 (“Blood-flow checker”. Popular Science: 12. 1987). The fluid attenuation inversion recovery (FLAIR) pulse sequence that yields high-signal regions in normal white matter, was demonstrated by Hajnal et al. in 1992 . Blood Oxygen Level Dependent signal was recognized by Ogawa  at AT&T Bell labs in 1990. Susceptibility-weighted imaging (SWI) was developed by Reichenbach et al.  at Washington University in 1997.
Fig. 2 shows patient cases that demonstrate the use of imaging contrast agents acquired from a 83-year-old female, 71-year-old male, and from a 26-year-old male with a 3 T MRI system.
Speed is always important to clinical practice. Fast MRI techniques were introduced and were based on the use of multiple refocusing pulses, commonly referred to as turbo spin-echo (TSE) imaging or turbo gradient-echo imaging. Rapid Acquisition with Refocused Echoes (RARE) technique was commercially implemented after Hennig  described it in 1986. In this sequence, the echo train length (ETL), also known as the turbo factor, denotes the number of echoes acquired at a given repetition time (TR). Spiral sequences were introduced in 1986 by Ahn et al. .
Several pulse sequences for fast imaging were developed, such as TSE or turbo field echo, half-Fourier, single-shot turbo spin echo (HASTE), gradient and spin echo (GRASE), balanced steady-state free-precession (bSSFP), EPI, and spiral . First, the principle of operation of TSE or RARE sequences was based on the filling of many k-space lines after every radiofrequency (RF) excitation that reduced the MR scan time by reducing the number of RF excitations required for each image . Second, HASTE was a single-shot version of the TSE technique  wherein slightly more than half the total number of required phase encoding lines of the k-space were acquired after a single RF excitation. Third, turbo gradient spin echo or GRASE, is a turbo SE sequence with additional phase encoding gradient echoes between successive 180° refocusing pulses . Fourth, bSSFP is a gradient-echo sequence that accomplishes fast acquisitions by eliminating idle times . In the most common case, the data are mapped line-by-line from the top to bottom parts of k-space in a lexicographic manner, commonly referred to as Cartesian sampling. Fifth, EPI allows the collection of all the data required to reconstruct an image after a single RF excitation  and spiral imaging acquires data using two oscillating gradients that create a spiral trajectory in k-space . The single-shot approach of EPI or spiral is the fastest with a subsecond acquisition time. Therefore, these techniques are usually used in functional MRI (fMRI). Furthermore, radial trajectories have also been introduced .
Several reconstruction methods were also developed to achieve higher signal-to-noise ratios (SNR) and reduced scan times. Parallel imaging techniques in conjunction with the use of phased array coils have been developed to reduce scan times by acquiring a reduced amount of k-space data with an array of receiver coils. These techniques include the simultaneous acquisition of spatial harmonics (SMASH) , sensitivity encoding (SENSE) , and generalized auto-calibrating partially parallel acquisitions (GRAPPA) . SENSE is based on reconstruction in the imaging domain, but GRAPPA is based on k-space domain reconstruction. These parallel imaging techniques are being further developed nowadays . Recently, multiband excitation imaging and finger printing imaging techniques have been developed to reduce the scan times even further.
The simultaneous multislice (SMS) pulse sequence  applies a multiband composite RF pulse with a slice-selective gradient to simultaneously excite multiple slice planes . The multiband technique is simultaneous image refocused (SIR) EPI or simultaneous echo refocused (SER) EPI that does not use parallel imaging and coil sensitivity for image separation . Another recently developed fast imaging technique is magnetic resonance fingerprinting (MRF) . This technique allows simultaneous and efficient measurements of multiple tissue properties with one acquisition . In MRF, the acquisition parameters, such as the RF excitation angle, phase, repetition time, and k-space sampling trajectory, are varied throughout the acquisition. When implemented properly, this acquisition could generate a unique signal time course for each tissue. For every MRF sequence, the dictionary of signal evolutions can be generated on a computer using mathematical algorithms to predict spin behavior and signal evolution during the acquisition.
Ultrafast imaging is used to eliminate the effects of physiological motion, thus capturing dynamic events in real time or shortening the total scan time. Shortening the scan time helps improve patient comfort and compliance, thereby minimizing motion during scans . At higher imaging speeds, it becomes feasible to examine a wide range of relevant physiological processes or to freeze induced motion that may otherwise lead to artifacts. Relevant physiological processes include, among others, respiration , cardiac rhythm , and hemodynamics from neuronal activation . For example, ultrafast imaging is often used in cardiovascular imaging [40,41].
SENSE or GRAPPA are commonly used in the clinical practice nowadays. Multiband technique is used in fMRI. However, there is always an inherent trade-off between imaging speed and quality. MRF was initially evaluated in brain relaxometry , prostate , liver , cardiac , musculoskeletal imaging , arterial spin labeling (ASL) perfusion measurement , and microvascular properties .
A diffusion process is described by Fick’s law. Einstein described the relationship between the mean-squared displacement and diffusion coefficient in Brownian motion . Diffusion is dependent on concentration, pressure, and temperature (https://en.wikipedia.org/wiki/Diffusion). Diffusion MRI is currently a well-established technique that is used in routine clinical practice to identify lesions and to characterize them. Diffusion MRI was developed by Le Bihan in 1989 . Diffusion tensor imaging (DTI) and fiber tractography were developed by Peter Basser and Le Bihan in the early 1990s [49 -51]. The diffusion coefficient for water at 37°C is approximately D=3×10−9 m2/s.
Diffusion-weighted imaging (DWI) was developed to investigate microstructural properties by evaluating the proton diffusion process. The technique is used to characterize the microscopic behaviors of protons noninvasively. The diffusion coefficient or apparent diffusion coefficient (ADC) is measured on the basis of the use of diffusion gradients around the refocusing pulse in the SE sequence, or the use of bipolar gradients in the gradient-echo sequence within the echo time (TE) time period, and is fitted by the mono-exponential decay curve .
where b is the b-factor that can be calculated using the strength of the applied diffusion gradient, the duration of the diffusion gradient, and the duration of the period between two diffusion gradients.
A two-compartment model was applied to measure diffusion from intracellular and extracellular compartments, respectively referred to as slow (Dslow) and fast (Dfast) diffusion with the corresponding volume fractions Fslow and Ffast .
The intravoxel incoherent motion (IVIM) method was introduced to separate diffusion from flow effects in a voxel . In this technique, the flowing vascular volume fraction of incoherently flowing blood in the tissue (F), pseudo-diffusion coefficient (D*) associated to the IVIM effect, and true diffusion coefficient (D) can be measured by applying multiple b-values.
Because in vivo proton movement is not isotropic, anisotropic diffusion was introduced by assuming ellipsoidal diffusion, known as DTI . The diffusion coefficient in the DTI model is a 3×3 symmetric matrix. Eigenvalues and eigenvectors can be extracted from the tensor matrix to calculate the isotropic diffusion effects by using the mean diffusivity (MD=(Dxx+Dyy+Dzz)/3), and anisotropic diffusion effects by using fractional anisotropy (FA) index that represents the variance of the three eigenvalues divided by norm of the eigenvalues.
Tensor models can be applied by assuming Gaussian distributions of proton movements. However, in biological structures, water molecular diffusion is hindered in the extra-axonal space and restricted in the intra-axonal space . To take this effect into account, other models were introduced, such as the ball-and-stick model , Q-ball imaging (QBI) , diffusion spectrum imaging (DSI) , spherical deconvolution model , and neurite orientation dispersion and density imaging (NODDI) . Furthermore, non-Gaussian diffusion kurtosis imaging, was applied in clinical studies .
Finally, tractography techniques were introduced to detect white matter fiber bundles by measuring the eigenvectors along the fibers [60,61]. Improvements of fiber tracking were further achieved by using probabilistic models . Tractography is currently used to investigate white matter connectivity in healthy brains and in several pathologic conditions .
Diffusion MRI techniques, including DWI, DTI, and tractography, are currently extensively used in clinical settings. DWI is routinely applied in stroke and tumor patients. Diffusion indices of MD and FA are altered in the presence of ischemic injury  and in neoplasms . In brain neoplasms, ADC values have been shown to be decreased in highly cellular tumors, such as central nervous system (CNS) lymphoma, medulloblastoma, and high-grade glioma . Diffusion imaging techniques were applied in Alzheimer’s disease [67 -71]. Furthermore, malignant lesions have lower ADC values compared with surrounding normal tissue, edema, and benign tumors in the brain, head and neck malignancies, prostate, and liver cancer . Diffusion-weighted whole-body imaging with background suppression (DWIBS) was performed with the short tau inversion recovery (STIR) EPI sequence with a high b value for background suppression to evaluate metastatic lesions in the body . DTI was used to evaluate dislocation, disruption, infiltration, and edema. Tractography was used to evaluate corticospinal tract fibers, optic radiation fibers, and language fibers in patients to perform presurgical planning in neoplastic brain tumor cases.
Perfusion refers to the delivery of blood to a capillary bed in tissue. This is different compared with bulk flow motion (https://en.wikipedia.org/wiki/Perfusion). August Krogh first described the mechanism of regulation of capillaries in muscle, and was awarded the Nobel Prize in Physiology or Medicine . Perfusion MRI is categorized according to the use of a contrast agent or not. Dynamic-susceptibility-contrast (DSC)-based perfusion MRI was developed by Villringer et al. in 1988 . The ASL technique was developed by Koretsky in 2012 .
Details of perfusion MRI were summarized in a previous paper . Three important techniques are currently used in clinical practice to obtain perfusion-related parameters. The first-pass DSC-enhanced MR perfusion is based on the susceptibility effects of gadolinium-based contrast agents on the signal echo. Therefore, T2- or T2*-weighted imaging sequence, usually a gradient-echo or an EPI sequence, is used to obtain the signal attenuation of the time-series images. Cerebral blood volume (CBV) and flow (CBF) values as well as time-related parameters, such as the mean transit time (MTT) and time-to-peak can be mapped in each pixel. Convolution theory is used to evaluate the measured concentrations of the hemodynamic changes of the contrast agent as follows ,
where Ca(t) is the measured arterial input function (AIF) that describes the shape of the tracer that enters a voxel, and R(t-t’) is the unknown residue function that describes the probability of the tracer entering a voxel. Therefore,
, and CBF=CBV/MTT.
The dynamic contrast-enhanced (DCE) MR perfusion is based on the relaxivity effects of gadolinium-based contrast agents on the signal echo. Therefore, a transverse relaxation (T1)-weighted imaging sequence (usually a three-dimensional sequence), is used to obtain signal increments of time-series images. The area under-the-curve can be mapped. Furthermore, a pharmacokinetic model is used to map permeability-related parameters such as Ktrans and Kep and the corresponding volume fractions such as vp and ve. The general equation used to express the hemodynamic event after injecting the contrast agent is expressed with the extended Tofts model as follows ,
where Cp(t) is the concentration of contrast agent in blood, Ktrans is the permeability–surface area constant from the vascular to the extracellular space, and ve is volume fraction in the extravascular and extracellular space.
The ASL MR perfusion is based on an endogenous contrast agent using magnetically labeled arterial blood water as a diffusible flow tracer. Therefore, the proton-density-weighted sequence is used to obtain signal changes with and without the use of magnetically labeled blood by either continuous or pulsed RF pulses. CBF value can be quantified. A general kinetic model for the evaluation of the difference of the longitudinal magnetization in the tissue owing to the labeled blood can be expressed to quantify the blood flow as follows ,
where M0b is the equilibrium magnetization in a blood filled arterial voxel, Ca(t) is the delivery function, such as AIF, R(t-t’) is the residue function that describes the washout of tagged spins from a voxel, and m(t-t’) includes the longitudinal magnetization relaxation effects. The pseudo-continuous ASL (pCASL) technique was introduced to improve the ASL signals , and is currently used in clinical practice.
Perfusion MRI is a promising tool used to assess stroke, tumors, and neurodegenerative diseases. DSC perfusion MRI is the standard technique used to evaluate brain diseases, such as stroke  and tumors . A combination of perfusion and DWI is used to evaluate a mismatch between the size of a perfusion defect and the diffusion abnormality that is referred to as the ischemic penumbra . Tumor grade, recurrence, postoperative changes, or radiation effects can be established with DSC perfusion imaging .
DCE perfusion MRI is often applied in brain diseases  and in patients with breast, prostate, pelvic, and muscle diseases, and can be useful in differentiating between tumor recurrence and radiation necrosis. The DCE parameters reflect a more extensive BBB disruption and a higher tumor grade , evaluate the treatment prediction , and differentiate pseudo-progression from true progressive disease in GBM patients .
ASL perfusion MRI is mainly applied in brain diseases, such as neurodegenerative [90,91], renal, and cardiovascular diseases. ASL perfusion MRI has been used to evaluate pseudo-progression in brain tumor . This method is particularly useful for patients with poor intravenous access, infants and children, and pregnant women .
MR spectroscopy (MRS) is used to determine the molecular structure of compounds, or to detect their presence. Proton MRS is based on the proton’s magnetic moment and its interaction with magnetic fields. MRS is thus sensitive to certain aspects of tissue metabolism. Proton MRS can detect N-acetyl aspartate (NAA), creatine/phosphocreatine (Cr/PCr), choline (Cho), glucose (Glu), myoinositol (ml or mlns), lactate (Lac), alanine (Ala), glutamate and glutamine (GIx), citrate, and ethanol in the human body. MRS can detect other nuclei in compounds of biological interest, such as phosphorus-31 (found in PCr) or carbon-13 (found in glycogen). However, proton MRS is more routinely performed in clinical practice compared with 31P or 13C MRS. Therefore, in this review, we only discuss proton MRS.
Chemical exchange saturation transfer (CEST) is a novel MR technique that enables molecular imaging to obtain certain compounds at concentrations that are too low to impact the contrast of standard MRI and too low to be directly detected in MRS at typical water imaging resolutions.
A single voxel spectroscopy (SVS) study is performed with short or long TE values. The SVS pulse sequences are the following: point resolved spectroscopy (PRESS) , stimulated echo acquisition mode (STEAM) , image selected in vivo spectroscopy (ISIS) , and depth-resolved spectroscopy (DRESS) . Longer TE results in the signal decrease as a result of the transverse relaxation (T2) that leads to the alteration of the phase of multiplet signals because of J-coupling . TE values in the range of 135–144 ms are typically used, as this leads to the production of a spectrum in which the doublet signal of Lac with a J-coupling constant of nearly 7 Hz is entirely reversed owing to the short TE and long TR .
A spectroscopic imaging is an area of interest. Chemical shift imaging (CSI) is used for multiple-voxel spectroscopic acquisitions. In spectroscopic imaging, phase encoding gradients can be applied in all three dimensions to sample k-space to select a volume that resembles methods used in MRI .2) Water suppression methods
The water resonance must be suppressed to detect the millimolar concentrations of other molecules/moieties. The most common suppression method is based on the use of a chemical shift selected (CHESS) pulse sequence .3) Spectral quantification
The typical postprocessing techniques used include Fourier Transform (FT), baseline correction, zero filling, and phasing. Quantification of the MRS signal amplitude can provide a means for estimating the tissue concentration of the signal generating molecules. While MRS signals are usually acquired in the time domain as free induction decays or echoes, they are usually viewed and analyzed in the frequency domain. The frequency domain representation is derived from the acquired time domain data by the FT. Signal averaging is used in virtually all MRS studies to increase the SNR by averaging the signals obtained in repeated measurements. To quantify the proton spectrum in most of the clinical studies, the internal reference signal is typically used that is either the Cr signal at 3.05 ppm or the water signal at 4.69 ppm. When Cr is used as a reference signal, it is more common to report results as signal amplitude ratios, such as NAA/Cr, or Cho/Cr. One weak aspect related to the use of the Cr signal as a reference is attributed to the fact that the Cr signal is not as uniform throughout the normal brain . Another weak aspect pertains to the fact that the assumption of the Cr levels do not change with disease and other physiological characteristics may be erroneous. If the water signal is used as a reference signal, its amplitude must be measured by performing a separate measurement in the same brain region without using water suppression.
A linear least-squares optimization procedure has been established and used in spectral fitting techniques. The most popular spectral fitting software is currently the LCModel [103,104]. Furthermore, spectrum fitting software is available either in the time  or frequency domains [106,107].
CEST can be used to apply molecular imaging . This technique is more appropriate compared with CSI or MR spectroscopy imaging because it provides relatively high resolution. The principle of CEST is based on the use of the magnetization transfer effects from other molecules to water molecules. Therefore, the requirement of CEST is that the chemical species must have in their structures a 1H proton that is exchangeable with those of water. Known endogenous diamagnetic CEST agents are involved with exchangeable groups of amide proton (-NH), amine proton (-NH2), and hydroxyl proton (-OH), whose chemical shifts are ~3.5, ~1.8–3.0, and ~0.5–1.5 parts per million (ppm), respectively . Amide CEST is usually referred to as amide proton transfer (APT). CEST techniques have been applied to map glutamate (amine proton), creatine (amine proton), glycosaminoglycan (Gag) (hydroxyl proton), myoinositol (MI) (hydroxyl proton), and glucose (hydroxyl proton). The detailed principle of CEST technique has been described in numerous previous papers [110 -116].
The goal of clinical spectroscopy is to provide physicians with biochemical information that will assist in the differential diagnosis when standard clinical and radiologic tests fail or are too invasive. Proton spectroscopy has attained clinical value in that it can monitor the evolution of diseases and associated therapies. Disease can sometimes lead to large changes in metabolite levels.
MR spectroscopy has also been shown to be of diagnostic value for evaluating and monitoring the progression of certain brain diseases, such as stroke , epilepsy, multiple sclerosis , Alzheimer’s disease [119,120], brain tumors  and other tumors, such as prostate cancer  and breast cancer . Lactate signal levels are elevated in ischemic brain tissue. Choline signal levels are elevated in some neoplastic or inflamed tissues.
The main applications for amide CEST or APT are the detection of cancer and ischemic stroke. In tumor regions, the concentration of proteins are elevated compared with surrounding tissues, and thus lead to increased APT levels . This method was applied to classify tumor progression from radiation necrosis . The CEST technique is applied in stroke because reduced pH in the ischemic region leads to lowered APT exchange rate, and results in decreased CEST values . CEST was applied in other-than-brain pathologies, such as breast , prostate , and knee .
PET–MRI is an imaging system that incorporates MRI and PET to gain from the benefits of soft tissue morphological imaging (MRI) and metabolic imaging (PET). This hybrid system is mainly used in the fields of oncology and neurology for clinical and preclinical studies. Some systems operate in totally separate rooms, but other systems do operate in the same room with separate machines. In these cases, a bed is shared to transfer the subjects from MRI to PET or to fully integrated systems. The first whole-body PET–MRI systems were produced by Philips (Amsterdam, The Netherlands) and were installed in the US (Mount Sinai Medical Center, New York, NY) and in Switzerland (Geneva University Hospital, Geneva) in 2010. The system featured a PET and MRI scanner separated by a revolving bed. The simultaneous PET–MR acquisition system (Siemens) was installed in 2010. The fully integrated whole-body systems were provided in 2011 by Siemens and in 2014 by GE. In Korea, the first PET–MRI system was installed at the Gachon University Hospital (Inchen, Korea) [7,8].1) Technical developments
In this part, we only discussed the issues of the fully integrated PET–MRI system. Placing PET detectors in the MR magnet can alter the local magnetic field strength causing the protons to spin at wrong frequencies, thus leading to the formation of severe image distortions and artifacts, such as susceptibility artifacts. The presence of PET hardware within the gradient coil significantly alters the MR eddy current characteristics of the system possibly leading to degraded spatial linearity. The gradient imperfection would impact imaging spatial resolution and image homogeneity. The lutetium-based scintillation crystals have acceptable magnetic properties . The avalanche photodiode for PET can be used in a 7 T MRI system without inducing major effects to magnetic fields .
Attenuation correction describes a method to account for the self-absorption of the emitted annihilation photons, and is a prerequisite for accurate quantification of the PET data . It is not possible to directly derive the attenuation properties of tissues from MRI measurements. MRI-based attenuation correction methods have been introduced : segmentation-based  that are usually used in T1-weighted images  or Dixon-sequence-based images , atlas-based , and reconstruction-based  methods. Attenuation correction of bone is calculated by using ultrashort echo time images .
The simultaneous acquisition of MRI and PET data in the fully integrated PET–MRI systems has major advantages compared with sequential acquisitions. Two-dimensional or three-dimensional (3D) navigator MRI is used to correct respiratory motion in PET images . 3D Cine MRI has been used to correct cardiac motion in PET images . A high resolution MRI has been used for the correction of the partial volume effect in PET images . MR-guided PET reconstruction techniques have been developed by incorporating “a priori” anatomical information from the MRI . Finally, MR anatomical images have been used for aligning functional information obtained from PET .2) Clinical applications
While PET provides a) the high sensitivity required to detect minute amounts of radiotracers and b) the ability to quantify radiotracer activity throughout the body in absolute terms, MRI provides excellent soft-tissue contrast according to multiple contrast mechanisms at high-spatial resolution. Therefore, PET/MRI imaging combines the diagnostic breadth and information content of both PET and MRI. The PET–MRI system has been used for the study of patients with hepatobiliary cancer , neuroendocrine tumors , pancreatic adenocarcinoma , prostate cancer , primary brain tumors , dementia , epilepsy , musculoskeletal tumor , and coronary artery disease .
High-intensity focused ultrasound (HIFU) is a noninvasive therapeutic technique that uses nonionizing ultrasonic waves to heat tissue . HIFU has been combined with MRI to enable guidance of the treatment and monitoring. It is referred to as MR-guided focused ultrasound surgery (MRgFUS). It is a 3D imaging technique that features high-soft-tissue contrast and provides information about temperature, thus allowing the monitoring of ablation. In 1992, reports were published that described MRgFUS on ex vivo muscle tissue , and the following year on in vivo tissue . MRgFUS developed by Hynynen  was later transferred to InsighTec in Haifa, Israel in 1998. The InsighTec ExAblate 2000 was the first MRgFUS system used to obtain Food and Drug Administration market approval in the US in 2004.1) Technical developments
In MRgFUS, MR is used for both target localization and in vivo real time monitoring of temperature based on a technique referred to as MR thermometry , and for verifying tissue destruction using a postprocessing tool. The thermometric technique of temperature-dependent phase changes in gradient-recalled echo pulse sequences are commonly used to determine the temperature change . The change in temperature is represented as
where Δϕ is the phase change, γ is the gyromagnetic ratio, c is the proton-resonance frequency shift constant (–0.01 ppm/°C) , Bo is the main magnetic field strength, and TE is the echo time. For the bone tumors, the calcification issues can be resolved be using susceptibility-weighted MRI to identify calcifications rather than computer tomography. An ultrashort TE sequence can be used to improve thermometry in bone.2) Clinical applications
Clinical applications of MRgFUS are still limited. MRgFUS has been used in clinical studies in patients with uterine fibroids , bone metastases , prostate cancer , breast cancer , and brain diseases, such as brain tumors , intractable essential tremor , Parkinson’s disease , obsessive-compulsive disorder , major depressive disorders , and neuropathic pain .
MRI-guided linear accelerator (MRI–LINAC) is a recently developed and advanced radiation treatment system. As indicating the name, the MRI–LINAC is fully integrated with the MRI for imaging soft-tissue tumors together with LINAC for the radiotherapy to treat cancers throughout the body. The advantage of MRI-based imaging on a linear accelerator has superior high-definition image quality, especially for some soft-tissue cancers compared with CT-based imaging in the traditional linear accelerators to visualize the target area and adjacent anatomy for treatment setup and delivery. The first technical prototype MR–Linac was developed and installed in the University Medical Center Utrecht in Utrecht, The Netherlands. Similar types of measurements have been performed on a hybrid MRI Cobalt-60 device . The first clinically active MRI-guided radiation therapy machine (ViewRay) was installed at the Alvin J. Siteman Cancer Center at Barnes-Jewish Hospital at the Washington University School of Medicine (St. Louis, MO, USA). The treatment of the first patients was announced in February 2014.1) Technical developments
LINAC is affected by MRI. Two main configurations of MRI–LINAC that are being pursued with the radiotherapy beam are either parallel or perpendicular to the main magnetic field. This configuration is affected by the interference between the delivery of the radiation beam of LINAC and MRI. The operation of the multileaf collimator in the strong magnetic field can be a problem in the shaping of the X-ray beam . Both configurations have this problem, and vendors lowered magnetic field to minimize this issue. Another issue in the MRI–LINAC is that the accelerated electrons used to produce the X-ray beam can be deviated or defocused, thus causing a loss of the beam current. Previous studies showed that the perpendicular configuration is dominant to the total beam loss compared with the parallel one [170,171]. Skin dose can be increased by secondary electrons owing to the influence of the magnetic field . In this case, the perpendicular configuration should be advantageous, although the electron return effect can still appear . Receiver coils can cause attenuation of the primary beam and can increase the skin dose. Detailed explanation can be found in another paper .
MRI quality is also affected by the LINAC. Any RF noise generated by the LINAC can cause artifacts or noise in images. In addition, LINAC components, such as the accelerator or multileaf collimators cause inhomogeneity of the main magnetic field, thus worsening the imaging quality . Finally, the radiation beam can affect conductors or electronics in the coil, causing imaging artifacts [176,177].2) Clinical applications
The MRI-LINAC can adapt the radiation treatment plan based on movement of the organs or tumor, and also track the motion of the tumor. This system reduces complications after radiation treatments. The MRI–LINAC can be used to improve the personalization of the radiation therapy using existing contrast imaging mechanisms, such as diffusion, perfusion, functional, and metabolic, to evaluate treatment effects. The hybrid system has been focused on daily plan changes based on geometric changes in the organs-at-risk [178,179]. Furthermore, MRI has been used to evaluate radiation treatment effects . Currently, this hybrid system is used to treat patients with prostate cancer , pelvic lymph nodes , and the esophagus .
The research was supported by the National Research Foundation of Korea grant funded by Ministry of Science and ICT (No. 2020R1A2C1004749, GHJ), Republic of Korea.
The authors have nothing to disclose.
The authors confirm that the data supporting the findings of this study are available within the article.
Conceptualization: Geon-Ho Jahng. Data curation: Geon-Ho Jahng, Soonchan Park, Chang-Woo Ryu, and Zang-Hee Cho. Writing–original drafting: Geon-Ho Jahng. Writing–review & editing: Geon-Ho Jahng, Soonchan Park, Chang-Woo Ryu, and Zang-Hee Cho.