Using an artificial neural network for fast mapping of the oxygen extraction fraction with combined QSM and quantitative BOLD.

PURPOSE: To apply an artificial neural network (ANN) for fast and robust quantification of the oxygen extraction fraction (OEF) from a combined QSM and quantitative BOLD analysis of gradient echo data and to compare the ANN to a traditional quasi-Newton (QN) method for numerical optimization. METHODS: Random combinations of OEF, deoxygenated blood volume ( ν ), R2 , and nonblood magnetic susceptibility ( χnb ) with each parameter following a Gaussian distribution that represented physiological gray matter and white matter values were used to simulate quantitative BOLD signals and QSM values. An ANN was trained with the simulated data with added Gaussian noise. The ANN was applied to multigradient echo brain data of 7 healthy subjects, and the reconstructed parameters and maps were compared to QN results using Student t test and Bland-Altman analysis. RESULTS: Intersubject means and SDs of gray matter were OEF =43.5±0.8 %, R2=13.5±0.3 Hz, ν=3.4±0.1 %, χnb=-25±5 ppb for ANN; and OEF = 43.8±5.2 %, R2=12.2±0.8 Hz, ν=4.2±0.6 %, χnb=-39±7 ppb for QN, with a significant difference ( P<0.05 ) for R2 , ν , and χnb . For white matter, they were OEF = 47.5±1.1 %, R2=17.1±0.4 Hz, ν=2.5±0.2 %, χnb=-38±5 ppb for ANN; and OEF =42.3±5.6 %, R2=16.7±0.7 Hz, ν=2.9±0.3 %, χnb=-45±9 ppb for QN, with a significant difference ( P<0.05 ) for OEF and ν . ANN revealed more gray-white matter contrast but less intersubject variation in OEF than QN. In contrast to QN, the ANN reconstruction did not need an additional sequence for parameter initialization and took approximately 1 s rather than roughly 1 h. CONCLUSION: ANNs allow faster and, with regard to initialization, more robust reconstruction of OEF maps with lower intersubject variation than QN approaches.

Comparison of gradient echo and gradient echo sampling of spin echo sequence for the quantification of the oxygen extraction fraction from a combined quantitative susceptibility mapping and quantitative BOLD (QSM+qBOLD) approach.

PURPOSE: To compare gradient echo (GRE) and gradient echo sampling of spin echo (GESSE) sequences for the quantification of the oxygen extraction fraction (OEF) from combined quantitative BOLD and quantitative susceptibility mapping (QSM) with regard to accuracy, precision and parameter initialization. METHODS: GRE and GESSE data were acquired from 7 healthy volunteers. QSM was applied to the GRE data and used as a regularization for the single-compartment quantitative BOLD fit to the GESSE and GRE data, respectively, to quantify OEF, deoxygenated blood volume (ν), R2 , and non-blood susceptibility (χnb ). Intersubject means within gray and white matter, respectively, were compared between GESSE and GRE (Student's t) and gray-white matter contrast was determined for each sequence separately. A single- and multi-compartment simulation was used to compare reconstruction accuracy. RESULTS: Intersubject means and SDs for gray and white matter were OEF = 32.4 ± 1.6%, ν = 2.9 ± 0.1%, R2 = 14.2 ± 0.5 Hz, χnb = -43 ± 5 ppb for GESSE and OEF = 43.0 ± 5.4%, ν = 3.5 ± 0.4%, R2 = 14.4 ± 0.7 Hz, χnb = -43 ± 8 ppb for GRE with a significant difference (P < 0.05) for OEF and ν. Gray-white matter contrast was significant (P < 0.05) in all parameters for GESSE but only in ν and R2 for GRE. All parameters reconstructed from GESSE had higher accuracy than from GRE in the single- but not multi-compartment simulation. CONCLUSION: GESSE yields higher parameter accuracy in simulated gray matter but produces unphysiological gray-white matter contrast in OEF in vivo. GRE produces uniform OEF maps in vivo and is more efficient, which could facilitate a clinical implementation, but revealed biases in simulation. The appropriate sequence should be chosen depending on application.

Influence of Gadolinium-Based Contrast Agents on Tissue Sodium Quantification in Sodium Magnetic Resonance Imaging.

OBJECTIVES: Sodium magnetic resonance (MR) imaging provides noninvasive insights to cellular processes by measuring tissue sodium concentration (TSC). Many clinical studies combine sodium MR imaging with clinical standard MR procedures, in which contrast media is frequently administered. This work investigates the influence of gadolinium-based contrast agents on quantification of TSC. Thus, either scan pauses between early and late contrast-enhanced acquisitions can be used efficiently or sodium imaging can be performed as the final scan after dynamic contrast-enhanced acquisition. MATERIALS AND METHODS: For this study, 2 gadolinium-based contrast agents, Dotarem and Gadovist, were diluted with saline solution covering contrast agent concentrations in a clinical range. In addition, agarose-based sample series were created to simulate tissue relaxation time behavior. In vivo, the influence of Dotarem on sodium acquisition and TSC quantification was investigated in 1 ischemic stroke patient. RESULTS: Proton relaxation times decreased for increasing contrast agent concentrations as hyperbolic functions. Sodium relaxation times displayed a negative slope in regression analysis in most cases. The largest influence (-1.52 milliseconds per mmol/L contrast agent) was measured for sodium T1. Worst case calculations in ultrashort echo time sequence signal analysis showed a signal drop of (1.21% ± 0.56%) on tissue sodium quantification. In vivo sodium brain acquisitions of a stroke patient before and after Dotarem injection resulted in statistically nonsignificant differences in TSC quantification of relevant tissues and stroke areas (P > 0.05). CONCLUSIONS: Our study showed a quantitative influence of Dotarem and Gadovist on sodium relaxation times. However, quantification of TSC was not impaired, which was proven by worst case calculations and nonsignificant differences in vivo in an ischemic stroke patient. We suggest performing sodium imaging in useful clinical positions in protocols regardless of included Dotarem or Gadovist administrations. Being flexible in the study protocol design will strengthen ongoing sodium imaging investigations for various pathologies.

Dose-dependent changes in renal (1)H-/(23)Na MRI after adjuvant radiochemotherapy for gastric cancer.

PURPOSE: Combined radiochemotherapy (RCT) for gastric cancer with three-dimensional conformal radiotherapy (3D-CRT) results in ablative doses to the upper left kidney, while image-guided intensity-modulated radiotherapy (IG-IMRT) allows kidney sparing despite improved target coverage. Renal function in long-term gastric cancer survivors was evaluated with 3T functional magnetic resonance imaging (MRI) including diffusion-weighted imaging (DWI) and (23)Na imaging. PATIENTS AND METHODS: Five healthy volunteers and 13 patients after radiotherapy were included: 11×IG-IMRT; 1×3D-CRT; 1× "positive control" with stereotactic body radiotherapy (SBRT) of a metastasis between the spleen/left kidney. Radiation doses were documented for the upper/middle/lower kidney subvolumes. Late toxicity was evaluated based on CTC criteria, questionnaire, and creatinine values. Morphological sequences, DWI images, and (23)Na images were acquired using a (1)H/(23)Na-tuned body-coil before/after intravenous water load (WL). Statistics for [(23)Na] (concentration) and apparent diffusion coefficient (ADC) values were calculated for upper/middle/lower renal subvolumes. Corticomedullary [(23)Na] gradients and [(23)Na] differences after WL were determined. RESULTS: No major morphological alteration was detected in any patient. Minor scars were observed in the cranial subvolume of the left kidney of the 3D-CRT and the whole kidney of the control SBRT patient. All participants presented a corticomedullary [(23)Na] gradient. After WL, a significant physiological [(23)Na] gradient decrease (p < 0.001) was observed in all HV and IG-IMRT patients. In the cranial left kidney of the 3D-CRT patient and the positive control SBRT patient, the decrease was nonsignificant (p = 0.01, p = 0.02). ADC values were altered nonsignificantly in all renal subvolumes (all participants). Renal subvolumes with doses ≥ 35 Gy showed a reduced change of the [(23)Na] gradient after WL (p = 0.043). No participants showed clinical renal impairment. CONCLUSIONS: Functional parameters of renal (23)Na MRI after gastric IG-IMRT are identical to those of healthy volunteers, in contrast to renal subvolumes after ablative doses in the control and 3D-CRT patient. While kidney doses to the cortex below 20-25 Gy in fractional doses of ~ 1 Gy in IG-IMRT (combined with intensified chemotherapy) do not seem to cause significant MRI morphological or functional alterations, doses of > 35 Gy in 1.5-2 Gy fractions clearly result in impairment.

Real time fMRI feedback of the anterior cingulate and posterior insular cortex in the processing of pain.

Self-regulation of brain activation using real-time functional magnetic resonance imaging has been used to train subjects to modulate activation in various brain areas and has been associated with behavioral changes such as altered pain perception. The aim of this study was to assess the comparability of upregulation versus downregulation of activation in the rostral anterior cingulate cortex (rACC) and left posterior insula (pInsL) and its effect on pain intensity and unpleasantness. In a first study, we trained 10 healthy subjects to separately upregulate and downregulate the blood oxygenation level-dependent response in the rACC or pInsL (six trials on 4 days) in response to painful electrical stimulation. The participants learned to significantly downregulate activation in pInsL and rACC and upregulate pInsL but not rACC. Success in the modulation of one region and direction of the modulation was not significantly correlated with success in another condition, indicating that the ability to control pain-related brain activation is site-specific. Less covariation between the areas in response to the nociceptive stimulus was positively correlated with learning success. Upregulation or downregulation of either region was unrelated to pain intensity or unpleasantness; however, our subjects did not learn rACC upregulation, which might be important for pain control. A significant increase in pain unpleasantness was found during upregulation of pInsL when covariation with the rACC was low. These initial results suggest that the state of the network involved in the processing of pain needs to be considered in the modulation of pain-evoked activation and its behavioral effects.

Artifact free T2*-weighted imaging at high spatial resolution using segmented EPI sequences.

The aim of this work was the development of novel measurement techniques that acquire high resolution T2*-weighted datasets in measurement times as short as possible without suffering from noticeable blurring and ghosting artifacts. Therefore, two new measurement techniques were developed that acquire a smoother k-space than generic multi shot echo planar imaging sequences. One is based on the principle of echo train shifting, the other on the reversed gradient method. Simulations and phantom measurements demonstrate that echo train shifting works properly and reduces artifacts in multi shot echo planar imaging. For maximum SNR-efficiency this technique was further improved by adding a second contrast. Both contrasts can be acquired within a prolongation in measurement time by a factor of 1.5, leading to an SNR increase by approximately 2. Furthermore it is demonstrated that the reversed gradient method remarkably reduces artifacts caused by a discontinuous k-space weighting. Assuming sequence parameters as feasible for fMRI experiments, artifact free T2*-weighted images with a matrix size of 256 × 256 leading to an in-plane resolution in the submillimeter range can be obtained in about 2s per slice.

How background noise shifts eigenvectors and increases eigenvalues in DTI.

INTRODUCTION: The signal-to-noise ratio of in vivo diffusion tensor imaging (DTI) is usually very limited, especially if high resolution data is acquired. In a variety of settings, the signal of diffusion weighted images can drop below the background noise level yielding an underestimated diffusion constant. In this work, we report two new artefacts in DTI that are important in this regime. METHODS: Both artifacts are described analytically and numerically and are demonstrated in DTI phantoms and in subjects in vivo. RESULTS: First, eigenvectors are systematically shifted towards distinct 'attractive' orientations of the gradient scheme. Second, certain eigenvalues can be overestimated due to the underestimation of the measured diffusion, which can result in the misordering of eigenvalues. DISCUSSION: We show that these effects are relevant for current clinical settings of DTI.

[Slice profile measurements of half pulse excitation for MR-imaging with ultra-short echo times (UTE)]. / Schichtprofilmessungen bei Anregung mit halben Pulsen für die MR-Bildgebung mit ultrakurzen Echozeiten (UTE).

Half pulse excitation plays an important role in imaging with ultra-short echo times (UTE imaging) of the order of TE < 100 micros. Based on half RF-pulses this method was theoretically modeled and experimentally applied in 1991 for the first time. Following this work, measurements of slice profiles produced by half pulses were performed and results were compared to the slice profiles produced by the original full pulse. Furthermore, the hypothesis was tested that short RF pulses may be of advantage in minimizing the relaxation effects during the pulse. Within the scope of these measurements no dependency of slice profile on the pulse duration could be found.

[Testing of model assumptions of MR oxygen extraction fraction mapping using a phantom study]. / Validierung von Modellannahmen einer MR-Sauerstoff- extraktionsbildgebung mittels einer Phantomstudie.

Oxygen supply is an important parameter for the evaluation of tissue viability and therefore of high interest in cancer diagnosis and therapy. One promising approach to extract relevant information from imaging data is the determination of oxygen saturation by means of the BOLD-effect. Using a simple model of tissue structure allows to evaluate the susceptibility difference between tissue and venous blood, from which the blood oxygen extraction fraction can be derived indirectly. The present study tested the validity of two model assumptions needed for exact quantification: the independence of the results of both capillary diameter d and the relative blood volume lambda. For this purpose a phantom was built, which allows the evaluation of susceptibility differences depending on d (27 microm-238 microm) and lambda (3%-12%). In agreement with model assumptions, delta chi(lambda) was widely constant and independent of ) lambda. In contradiction to the model, delta chi (d) showed a positive slope (delta chi range: 0.35-0.57 ppm). The present study suggests that the simple model investigated here has shortcomings in the quantification of oxygen extraction due to insufficient model assumptions.