Magnetic resonance imaging reveals distinct bone marrow patterns in indolent and advanced systemic mastocytosis.

Systemic mastocytosis (SM) is broadly subcategorized according to mast cell (MC) burden and organ involvement into indolent (ISM), smoldering (SSM), and advanced SM (AdvSM). However, the pattern and extent of bone involvement remains controversial. In this institutional review board (IRB)-approved study, 115 patients with different forms of SM (ISM (n = 37, 32%), SSM (n = 9, 8%), and AdvSM (n = 69, 60%)) underwent a whole-body magnetic resonance imaging including sagittal and coronal T1 and turbo inversion recovery magnitude (TIRM) sequences of the spine. The evaluation included the pattern and extent of pathologic bone marrow (BM) signals in the spine and extremities, osteolytic lesions, and vertebral fractures. A pathologic BM pattern was observed in 4/37 (11%), 8/9 (89%), and 66/69 (96%); affection of the appendicular skeleton in 3/37 (8%), 8/9 (89%), and 67/69 (97%); and vertebral fractures in 7/37 (19%), 0/9, and 13/69 (19%) patients with ISM, SSM, and AdvSM, respectively. In AdvSM, pathologic BM pattern included activated (62%), diffuse sclerotic (25%), and small-spotted BM (9%), respectively. Only activated/sclerotic BM was associated with significantly higher MC burden, organ damage, and inferior median survival (2.9 years, p = 0.04). Vertebral fractures resembled classical multi-segmental osteoporotic fractures in ISM but not in AdvSM in which they were only found in activated/sclerotic BM. Only one patient with AdvSM had a focal osteolytic lesion in the femur. Activated/sclerotic BM changes of the spine and affection of the appendicular skeleton are indicative for SSM or AdvSM. Osteolytic lesions, which are very rare, and osteoporotic fractures are ineligible for the diagnosis of AdvSM.

Nonenhanced ECG-gated quiescent-interval single shot MRA: image quality and stenosis assessment at 3 tesla compared with contrast-enhanced MRA and digital subtraction angiography.

PURPOSE: To evaluate the diagnostic accuracy of a nonenhanced electrocardiograph-gated quiescent-interval single shot MR-angiography (QISS-MRA) at 3 Tesla with contrast-enhanced MRA (CE-MRA) and digital subtraction angiography (DSA) serving as reference standard. MATERIALS AND METHODS: Following institutional review board approval, 16 consecutive patients with peripheral arterial disease underwent a combined peripheral MRA protocol consisting of a large field-of-view QISS-MRA, continuous table movement MRA, and an additional time-resolved MRA of the calves. DSA correlation was available in eight patients. Image quality and degree of stenosis was assessed. Sensitivity and specificity of QISS-MRA was evaluated with CE-MRA and DSA serving as the standards of reference and compared using the Fisher exact test. RESULTS: With the exception of the calf station, image quality with QISS-MRA was rated statistically significantly less than that of CE-MRA (P < 0.05, P = 0.17, and P = 0.6, respectively). A greater percentage of segments were not accessible with QISS-MRA (19.5-20.1%) in comparison to CE-MRA (10.9%). Relative to DSA, sensitivity for QISS-MRA was high (100% versus 91.2% for CE-MRA, P = 0.24) in the evaluated segments; however, specificity (76.5%) was substantially less than that of CE-MRA (94.6%, P = 0.003). CONCLUSION: Overall image quality and specificity of QISS-MRA at 3T are diminished relative to CE-MRA. However, when image quality is adequate, QISS-MRA has high sensitivity and, thus, has potential use in patients with contraindications to gadolinium.

In vivo sodium (23Na) imaging of the human kidneys at 7 T: preliminary results.

OBJECTIVE: To evaluate the feasibility of in vivo (23)Na imaging of the corticomedullary (23)Na gradient and to measure (23)Na transverse relaxation times (T2*) in human kidneys. METHODS: In this prospective, IRB-approved study, eight healthy volunteers (4 female, 4 male; mean age 29.4 ± 3.6 years) were examined on a 7-T whole-body MR system using a (23)Na-only spine-array coil. For morphological (23)Na-MRI, a 3D gradient echo (GRE) sequence with a variable echo time scheme (vTE) was used. T2* times were calculated using a multiecho 3D vTE-GRE approach. (23)Na signal-to-noise ratios (SNR) were given on a pixel-by-pixel basis for a 20-mm section from the cortex in the direction of the medulla. T2* maps were calculated by fitting the (23)Na signal decay monoexponentially on a pixel-by-pixel basis, using least squares fit. RESULTS: Mean corticomedullary (23)Na-SNR increased from the cortex (32.2 ± 5.6) towards the medulla (85.7 ± 16.0). The SNR increase ranged interindividually from 57.2% to 66.3%. Mean (23)Na-T2* relaxation times differed statistically significantly (P < 0.001) between the cortex (17.9 ± 0.8 ms) and medulla (20.6 ± 1.0 ms). CONCLUSION: The aim of this study was to evaluate the feasibility of in vivo (23)Na MRI of the corticomedullary (23)Na gradient and to measure the (23)Na T2* relaxation times of human kidneys at 7 T. KEY POINTS: • High field MR offers new insights into renal anatomy and physiology. • (23) Na MRI of healthy human kidneys is feasible at ultra-high field. • Renal (23) Na concentration increases from the cortex in the medullary pyramid direction. • In vivo measurements of renal (23) Na-T2* times are demonstrated at 7.0 T.

Assessment of acute intestinal graft versus host disease by abdominal magnetic resonance imaging at 3 Tesla.

OBJECTIVES: After allogeneic stem cell transplantation (SCT), a reliable diagnosis of acute graft versus host disease (aGvHD) is essential for an early and successful treatment. It is the aim of this analysis to assess intestinal aGvHD by magnetic resonance imaging (MRI). METHODS: Prior to allogeneic SCT, 64 consecutive patients underwent abdominal MRI examination on a 3 T MR system, including axial and coronal T2w sequences and a three-dimensional dynamic T1w, contrast enhanced sequence. After SCT, 20 patients with suspected aGvHD received a second MRI as well as an endoscopic examination. RESULTS: Nine patients suffered from histologically proven intestinal aGvHD. In eleven patients intestinal aGvHD was excluded. In all aGvHD patients typical MRI findings with long-segment bowel wall thickening--always involving the terminal ileum--with profound submucosal oedema, were detected. The bowel wall was significantly thickened in patients with intestinal aGvHD. Bowel contrast enhancement spared the submucosa while demonstrating strong mucosal hyperemia. CONCLUSIONS: In intestinal aGvHD, a characteristic MR-appearance can be detected. This MRI pattern might facilitate an early and non-invasive diagnosis of intestinal aGvHD. MRI might thus be used as a sensitive tool to rule out or support the clinical diagnosis of aGvHD. KEY POINTS: • Acute intestinal graft versus host disease (aGvHD) can be assessed by MRI. • The aGvHD of the bowel demonstrates a characteristic MR imaging pattern. • Bowel wall shows extensive long-segment wall thickening with profound submucosal oedema. • Terminal ileum seems invariably affected; other bowel segments show variable involvement. • Colonoscopy in suspected aGvHD should include inspection of terminal ileum.

Zoomed EPI-DWI of the head and neck with two-dimensional, spatially-selective radiofrequency excitation pulses.

OBJECTIVES: To evaluate the feasibility of zoomed diffusion-weighted EPI (z-EPI) in the head and neck in a healthy volunteer population and to compare to conventional single-shot EPI (c-EPI). MATERIAL AND METHODS: Nine volunteers were included in this prospective, IRB-approved study. Examinations were performed on a 3 T-MR system equipped with a two-channel, fully-dynamic parallel transmit array. The acquired sequences consisted of a T2w-TSE, a c-EPI, and two z-EPI acquisitions. For quantitative assessment of distortion artefacts, DW images were fused with T2-TSE images. Misregistration of DW images with T2-TSE images was assessed in the cervical spine. For qualitative assessment, two readers ranked c-EPI and z-EPI sequences in terms of susceptibility artefacts, image blur, and overall imaging preference. ADC values of several anatomical regions were calculated and compared between sequences. RESULTS: Mean maximum distortion with the c-EPI was 5.9 mm ± 1.6 mm versus 2.4 mm ± 1 mm (p < 0.05) with z-EPI. Both readers found more blur and susceptibility artefacts in every case with c-EPI. No statistically significant differences in calculated ADC values were observed. CONCLUSION: z-EPI of the head and neck leads to substantial image quality improvements relative to c-EPI due to a reduction in susceptibility artefacts and image blur. KEY POINTS: • Zoomed DWI is feasible in the head and neck. • Image quality improves substantially with zoomed DWI of the neck. • Zoomed DWI exhibits markedly reduced susceptibility artefacts.

3T renal (23)Na-MRI: effects of desmopressin in patients with central diabetes insipidus.

PURPOSE: The purpose of this prospective study was to assess physiologic changes in the renal corticomedullary (23)Na-concentration ([(23)Na]) gradient with (23)Na-MRI at 3.0T in patients with central diabetes insipidus (CDI) before and after intranasal administration of 20 µg desmopressin (DDAVP). METHODS AND MATERIALS: Four patients with CDI (all male, mean age 60.2 years) were included in this IRB-approved study. For (23)Na-imaging, a 3D density adapted, radial GRE-sequence (TE = 0.55 ms; TR = 120 ms; projections = 8,000; spatial resolution = 5 × 5 × 5 mm(3)) was used in combination with a dedicated (23)Na-coil and reference phantoms. The corticomedullary [(23)Na] gradient (in mmol/L/mm) was calculated pixel-by-pixel along a linear region-of-interest (ROI) spanning from the renal cortex in the direction of the medulla. Mean ± SDs of [(23)Na] were calculated for each patient as well as for the entire group. RESULTS: Mean [(23)Na] increased along the corticomedullary gradient from the cortex (pre-DDAVP 38.0 ± 6.3 mmol/L vs. post-DDAVP 30.7 ± 3.5 mmol/L) to the medulla (pre-DDAVP 71.6 ± 14.8 mmol/L vs. post-DDAVP 59.7 ± 10.8 mmol/L). The overall mean decrease of [(23)Na] after DDAVP administration was 17.1 ± 1.1 %. CONCLUSION: (23)Na-MRI with state-of-the-art techniques at 3T depicts the physiologic renal response to the administration of desmopressin in patients with central diabetes insipidus.

Assessment of glycosaminoglycan content in intervertebral discs using chemical exchange saturation transfer at 3.0 Tesla: preliminary results in patients with low-back pain.

OBJECTIVES: To evaluate glycosaminoglycan-dependent chemical exchange saturation transfer (gagCEST) imaging at 3-T magnetic resonance imaging (MRI) for quantification of glycosaminoglycan (GAG) content in intervertebral discs (IVDs) in patients with low-back pain (LBP). METHODS: Sixteen patients with LBP were examined in this Institutional Review Board-approved study using a clinical whole-body system. The MRI protocol included standard morphological imaging, sagittal T2-mapping and gagCEST imaging. IVD grading according to the Pfirrmann score and region-of-interest analysis of the annulus fibrosus (AF) and the nucleus pulposus (NP) in gagCEST and T2 maps were performed before data were statistically tested for correlations between imaging techniques and quantitative differences between different grades of IVD degeneration. RESULTS: GagCEST values of the NP were significantly (P < 0.001) lower in degenerative IVDs (Pfirrmann 3 + 4) compared with non-degenerative IVDs (Pfirrmann 1 + 2), but only a weak linear correlation (r = 0.299) with the T2 relaxation times was found. GagCEST values of the NP exhibited a moderate negative correlation with Pfirrmann grades (r = -0.449). CONCLUSIONS: The known loss of GAG in the NP with increasing grade of morphological degeneration can be assessed using gagCEST imaging at 3.0 T. The correlation with single Pfirrmann grades and T2 relaxation times only seems to be moderate, indicating a substantial difference in information provided by the techniques.

Steady state vascular imaging with extracellular gadobutrol: evaluation of the additional diagnostic benefit in patients who have undergone a peripheral magnetic resonance angiography protocol.

BACKGROUND: To evaluate the feasibility and additional diagnostic benefit of a high-resolution steady state 3D-volume interpolated breath-hold exam (VIBE) sequence between a continuous table movement (CTM) MR angiography of the entire runoff vasculature and a time-resolved (TWIST) MRA of the calves. METHODS: In this retrospective IRB approved study 224 patients (72 women, 152 men, mean age 67.29 ± 13.9) were included who had undergone a low-dose MR angiographic protocol at 3T (Siemens TimTrio) after injection of 0.1 mmol/kg gadobutrol including a CTM MRA, a time-resolved MRA of the calf station and a steady state 3D VIBE sequence prior to the time-resolved MRA. One board-certified radiologist rated the image quality of the steady state VIBE sequences on an ordinal three point scale (excellent, good, poor) and analyzed the images for additional diagnostic findings of and beyond the vascular system in comparison to the CTM MRA and the time-resolved MRA. Descriptive statistics and demographic patient data were used for further evaluation. RESULTS: The image quality of the steady state imaging of the pelvis, upper and lower leg was excellent in up to 88%, 84% and 47%, respectively, while poor image quality was only detected in the upper (2%) and lower leg (6%). An additional diagnostic benefit was found in 44% of the patients overall. The most common relevant pathologies included inflammatory processes of the soft tissues (26%), thrombi (14%), abscesses (13%) and tumors (11%). In subgroups of patients above the age of 60, 65, 70, 75 and 80 years an additional pathology was found in 50% 33%, 44%, 65% and 58%, respectively. There was no significant difference in terms of additional findings between men and women (46% and 39%, p > 0.05) and inpatients and outpatients (42% and 45%, p > 0.05). CONCLUSION: Steady state imaging is also feasible with extracellular contrast agents with good image quality yielding additional diagnostic findings in up to 44% and above in patients older than 60 years of age irrespective of gender or patient status. Given the short acquisition time of 4 minutes this sequence could be added to all peripheral MRA exams.

Impact of time-resolved MRA on diagnostic accuracy in patients with symptomatic peripheral artery disease of the calf station.

OBJECTIVE: The purpose of this article is to evaluate the added diagnostic accuracy of time-resolved MR angiography (MRA) of the calves compared with continuous-table-movement MRA in patients with symptomatic lower extremity peripheral artery disease (PAD) using digital subtraction angiography (DSA) correlation. MATERIALS AND METHODS: Eighty-four consecutive patients with symptomatic PAD underwent a low-dose 3-T MRA protocol, consisting of continuous-table-movement MRA, acquired from the diaphragm to the calves, and an additional time-resolved MRA of the calves; 0.1 mmol/kg body weight (bw) of contrast material was used (0.07 mmol/kg bw for continuous-table-movement MRA and 0.03 mmol/kg bw for time-resolved MRA). Two radiologists rated image quality on a 4-point scale and stenosis degree on a 3-point scale. An additional assessment determined the degree of venous contamination and whether time-resolved MRA improved diagnostic confidence. The accuracy of stenosis gradation with continuous-table-movement and time-resolved MRA was compared with that of DSA as a correlation. Overall diagnostic accuracy was calculated for continuous-table-movement and time-resolved MRA. RESULTS: Median image quality was rated as good for 578 vessel segments with continuous-table-movement MRA and as excellent for 565 vessel segments with time-resolved MRA. Interreader agreement was excellent (κ = 0.80-0.84). Venous contamination interfered with diagnosis in more than 60% of continuous-table-movement MRA examinations. The degree of stenosis was assessed for 340 vessel segments. The diagnostic accuracies (continuous-table-movement MRA/time-resolved MRA) combined for the readers were obtained for the tibioperoneal trunk (84%/93%), anterior tibial (69%/87%), posterior tibial (85%/91%), and peroneal (67%/81%) arteries. The addition of time-resolved MRA improved diagnostic confidence in 69% of examinations. CONCLUSION: The addition of time-resolved MRA at the calf station improves diagnostic accuracy over continuous-table-movement MRA alone in symptomatic patients with PAD.

Quantitative in vivo 23Na MR imaging of the healthy human kidney: determination of physiological ranges at 3.0T with comparison to DWI and BOLD.

OBJECTIVES: The purpose of this prospective study was to assess the normal physiologic ranges of the renal corticomedullary 23Na-concentration ([23Na]) gradient at 3.0T in healthy volunteers. The corticomedullary [23Na] gradient was correlated with other functional MR imaging parameters--blood oxygenation level dependent (BOLD) and diffusion-weighted imaging (DWI)--and to individual and physiologic parameters--age, gender, estimated glomerular filtration rate (eGFR), body mass index (BMI), and blood serum sodium concentration ([23Na]serum). METHODS AND MATERIALS: 50 healthy volunteers (30 m, 20 w; mean age: 29.2 years) were included in this IRB-approved study, without a specific a priori preparation in regard to water or food intake. For 23Na-imaging a 3D density adapted, radial gradient echo (GRE)-sequence (spatial resolution=5×5×5 mm3) was used in combination with a dedicated 23Na-coil and 23Na-reference phantoms. [23Na] values of the corticomedullary [23Na] gradient were measured by placement of a linear region of interest (20×1 mm2) from the renal cortex in the direction of the renal medulla. By using external standard reference phantoms, [23Na] was calculated in mmol/L of wet tissue volume (mmol/l WTV). Axial diffusion-weighted images (spatial resolution=1.7×1.7×5.0 mm3) and 2D GRE BOLD images (spatial resolution=1.2×1.2×4.0 mm3) were acquired. Mean values±standard deviations for [23Na], apparent diffusion coefficient (ADC) values, and R2* values were computed for each volunteer. The corticomedullary 23Na-concentration gradient (in mmol/l/mm) was calculated along the area of linear concentration increase from the cortex in the direction of the medulla. Correlations between the [23Na] and DWI, BOLD, and the physiologic parameters were assessed with Pearson correlation coefficients. RESULTS: The mean corticomedullary [23Na] for all healthy volunteers increased from the renal cortex (58±17 mmol/l WTV) in the direction of the medulla (99±18 mmol/l WTV). The inter-individual differences ranged from respective cortical and medullary values of 27 and 63 mmol/L WTV to 126 and 187 mmol/L WTV. No statistically significant differences in renal [23Na] were found based on differences in individual or physiologic parameters (age, gender, [23Na]serum, BMI, GFR). No ADC or R2* gradients were identified, and [23Na] did not correlate with these parameters. CONCLUSION: Renal corticomedullary [23Na] values increase from the cortex in the direction of the medullary pyramid, demonstrating wide inter-individual ranges and no significant correlations with age, gender, [23Na]serum, BMI, GFR, ADC, or R2* values. For future clinical evaluations, an approach relying on renal stimulation (e.g. pharmacologically induced diuresis) may be applicable to account for wide inter-individual ranges of normal [23Na].

Value of diffusion-weighted MR imaging for the detection of nephritis.

PURPOSE: To evaluate diffusion-weighted MR imaging (DWI-MRI) for the detection and assessment of infectious renal disease. MATERIALS AND METHODS: Twenty-one patients with suspicious increased signal intensity of the kidneys on DWI sequences and corresponding ADC decrease were identified. Sixty patients without clinical signs of renal infection served as a control group. All patients were examined with the following sequences: EPI-DWI (0/400/800 s/mm(2)), T2w HASTE, and T1w VIBE after intravenous injection of Gd-chelate. Confirmation of renal infection was established on the basis of clinical criteria. T1w and T2w images were assessed and compared to DWI for the presence of altered signal, and the degree of the visibility of pathology was graded on an ordinal three-point scale. RESULTS: In all 21 patients with positive DWI findings a renal infection could be confirmed. T2w imaging and contrast-enhanced T1w imaging displayed obvious pathologic signal in 3/21 (14%) and 11/19 (58%) patients and slightly pathologic signal in 17/21 (81%) and 7/19 (37%), respectively. The median visibility score of 2 for the DWI and the T1w images was significantly higher than the score of 1 for the T2w imaging, P = 0.0001 (DWI versus T2w) and P = 0.078 (T1w versus T2w). CONCLUSION: DWI of the kidneys seems to be highly sensitive for the detection of infections within the kidney.

Renal BOLD-MRI does not reflect renal function in chronic kidney disease.

Renal blood oxygen level-dependent magnetic resonance imaging (BOLD-MRI) is a noninvasive fast technique to characterize renal function. Here we evaluated the impact of renal function on the relaxation rate (R2(*)) in the cortex and medulla to provide baseline data for further use of renal BOLD-MRI. This parameter was evaluated in 400 patients scheduled for abdominal imaging who underwent transversal blood oxygen level-dependent measurements with a multi-echo gradient-echo sequence with 12 echo times. The loss of phase coherence (T2(*)) maps were generated in which kidney regions of interest were selected to differentiate the medulla and cortex, and R2(*) was equated to 1/T2(*). Individual R2(*) values were, in turn, correlated to the eGFR (MDRD formula of 280 patients with available serum creatinine measurements), age, and gender each for 1.5 and 3.0 T field-strength scans of 342 patients. At both the field strengths, no significant differences in R2(*) of the cortex and medulla were found between patient gender, age, eGFR, or between different stages of chronic kidney disease determined using the KDOQI system. Thus, BOLD-MRI of a non-specific patient population failed to discriminate between the patients with various stages of chronic kidney disease.

DCE-MRI of the human kidney using BLADE: a feasibility study in healthy volunteers.

PURPOSE: To evaluate the degree of motion compensation in the kidney using two different sampling methods, each in their optimized settings: A BLADE k-space acquisition technique and a routinely used kidney perfusion acquisition scheme (TurboFLASH). MATERIALS AND METHODS: Dynamic contrast enhanced magnetic resonance examinations were performed in 16 healthy volunteers on a 3 Tesla MR-system with two parameterizations of the BLADE sequence and the standard reference acquisition scheme. Signal intensity enhanced time curves were analyzed with a mathematical model and a widely published separable compartment model on cortex regions to assess robustness versus motion artifacts. RESULTS: BLADE-measurements with a strip-width of 32 lines constituted the smallest mean values for the sum of squared errors (6065 ± 4996) compared with the measurement with a strip-width of 64 lines (13849 ± 14079) or the standard TurboFLASH (11884 ± 8076). Calculations concerning goodness of the fit of the applied compartment model yielded an overall average of the Akaike Fit Error of 732 ± 141 for BLADE (646 ± 149 for a strip-width of 32 lines, 816 ± 53 for 64 lines) and 1626 ± 303 for the TurboFLASH (TFL) sequence. CONCLUSION: We demonstrated that renal dynamic contrast enhanced magnetic resonance imaging using BLADE k-space sampling with a strip-width of 32 is significantly less sensitive to motion than a widely published Turbo-Flash sequence with nearly similar parameters.

Comparison of 0.5 M gadoterate and 1.0 M gadobutrol in peripheral MRA: a prospective, single-center, randomized, crossover, double-blind study.

PURPOSE: To evaluate the diagnostic efficacy of macrocyclic paramagnetic gadolinium (Gd) chelates gadoterate (0.5 mmol/mL) and gadobutrol (1.0 mmol/mL) for the diagnosis of clinically significant abdominal/lower limb arterial diseases at 3.0T. MATERIALS AND METHODS: This study was conducted as a prospective, single-center, randomized, double-blind, intraindividual study comparing single dose (0.1 mmol/kg) gadoterate enhanced-MRA (magnetic resonance angiography) with gadobutrol enhanced-MRA at 3.0T for their diagnostic potential in patients with peripheral artery disease. A total of 20 patients were included in this trial. RESULTS: Fourteen patients were eligible for the final efficacy analysis. The overall image quality (excellent/more than adequate) was better rated with gadoterate than with gadobutrol (100% vs. 78.6%, 100% vs. 92.9%, 100% vs. 85.7%, 100% vs. 85.7% for readers 1, 2, 3, 4, respectively). Diagnostic confidence was rated high/excellent in 100% (readers 1, 2, and 3) and 92.9% (reader 4) with gadoterate compared to 92.9% (readers 1 and 2) and 85.7% (readers 3 and 4) with gadobutrol. Higher signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) values were obtained for gadobutrol compared to gadoterate (26.1/23.4, P = 0.01, and 22.7/20.2, P = 0.01). For the secondary criteria, no differences between groups were reported. No adverse events were reported. CONCLUSION: Gadobutrol yielded significantly higher SNR/CNR while gadoterate was better rated in terms of overall image quality and diagnostic confidence (P > 0.05).

Current Value of Biparametric Prostate MRI with Machine-Learning or Deep-Learning in the Detection, Grading, and Characterization of Prostate Cancer: A Systematic Review.

Prostate cancer detection with magnetic resonance imaging is based on a standardized MRI-protocol according to the PI-RADS guidelines including morphologic imaging, diffusion weighted imaging, and perfusion. To facilitate data acquisition and analysis the contrast-enhanced perfusion is often omitted resulting in a biparametric prostate MRI protocol. The intention of this review is to analyze the current value of biparametric prostate MRI in combination with methods of machine-learning and deep learning in the detection, grading, and characterization of prostate cancer; if available a direct comparison with human radiologist performance was performed. PubMed was systematically queried and 29 appropriate studies were identified and retrieved. The data show that detection of clinically significant prostate cancer and differentiation of prostate cancer from non-cancerous tissue using machine-learning and deep learning is feasible with promising results. Some techniques of machine-learning and deep-learning currently seem to be equally good as human radiologists in terms of classification of single lesion according to the PIRADS score.

Quantitative and qualitative (23)Na MR imaging of the human kidneys at 3 T: before and after a water load.

PURPOSE: To qualitatively and quantitatively assess the corticomedullary sodium 23 ((23)Na) concentration in human kidneys before and after oral administration of a water load by using 3-T magnetic resonance (MR) imaging. MATERIALS AND METHODS: Fourteen healthy volunteers (mean age, 28 years; range, 24-34 years) were included in this institutional review board-approved study between July and December 2009. For (23)Na MR imaging, a density-adapted three-dimensional radial gradient echo sequence (echo time, 0.55 msec; repetition time, 120 msec; spatial resolution, 5 × 5 × 5 mm) was used with a dedicated (23)Na-tuned coil. Beforehand, the coil profile was assessed by using phantom measurements, and the volunteer images were mathematically corrected accordingly. Images of the volunteers were obtained before and 30 minutes after oral ingestion of 1 L of water. As internal reference, (23)Na concentration of the cerebrospinal fluid (CSF) was calculated. Well-defined corticomedullary complexes in each kidney were assessed, with (23)Na concentrations in the cortex and medulla assessed at various standardized points. From these values, quantitative (23)Na concentrations were derived, and the slopes of the linear portion of the concentration gradient were calculated. Paired t tests were performed. RESULTS: Mean calculated (23)Na concentrations of CSF before (135.2 mmol/L ± 10.4) and after water load (135.5 mmol/L ± 11.0) fell within physiologic ranges (P = .95). An increase in average (23)Na concentration from 63.5 mmol/L ± 9.3 in the cortex to 108.0 mmol/L ± 10.9 in the medulla was identified. After the water load, this gradient was preserved, although (23)Na concentrations decreased significantly (P < .0001) to 48.6 mmol/L ± 5.3 in the cortex and 81.9 mmol/L ± 10.1 in the medulla-declines of 23.4% and 24.7%, respectively. CONCLUSION: This study demonstrates the physiologic evaluation of human kidneys with 3-T (23)Na MR imaging. The (23)Na imaging technique used allows the quantification of the corticomedullary (23)Na concentration and the assessment of its change with differing physiologic conditions.

Contrast-enhanced magnetic resonance angiography (MRA): evaluation of three different contrast agents at two different doses (0.05 and 0.1 mmol/kg) in pigs at 1.5 Tesla.

OBJECTIVE: To compare the image quality of contrast-enhanced magnetic resonance angiography (CE-MRA) of the supra-aortic vessels at 0.05 mmol/kg bw and 0.1 mmol/kg bw, between gadobutrol, Gd-DTPA and Gd-BOPTA quantitatively and qualitatively a total of eight pigs were evaluated intraindividually at 1.5 T. METHODS: Each pig was examined using 0.1 mmol/kg gadobutrol, Gd-DTPA and Gd-BOPTA on day one and 0.05 mmol/kg on day two. MRA datasets for the carotid artery and the infraorbital artery were qualitatively assessed regarding overall image quality on an ordinal four-point scale (4-excellent, 1-non-diagnostic). The signal-to noise-ratio (SNR) was measured. RESULTS: The qualitative assessment of the carotid artery showed a higher median image quality for the 0.1 mmol dose than for the 0.05 mmol dose for all three compounds. No difference was found for the infraorbital artery. Mean SNR of Gd-BOPTA, Gd-DTPA, gadobutrol at 0.05 mmol/kg were 36.0 ± 13.4/37.9 ± 16.3/43.7 ± 0.4 and at 0.1 mmol/kg they were 50.1 ± 12.4/46.6 ± 6.5 / 54.6 ± 10.2. Gd-BOPTA 0.05 revealed a significantly lower SNR than all other agents at normal dose. CONCLUSIONS: Full-dose gadolinium MRA results in higher image quality and significantly higher SNR compared with the half dose. Gadobutrol and Gd-BOPTA have similar enhancement properties at full dose but at half dose, gadobutrol appears superior.

Magnetic resonance angiography (MRA) of the calf station at 3.0 T: intraindividual comparison of non-enhanced ECG-gated flow-dependent MRA, continuous table movement MRA and time-resolved MRA.

OBJECTIVES: To compare 3D non-enhanced ECG-gated inflow-dependent MRA (NE-MRA) vs. continuous table movement (CTM) MR-angiography and time-resolved TWIST-MRA in the calf station at 3.0 T in a clinical patient collective. METHODS: 36 patients (27 male/9 female, 66.1 ± 14.4 years) with PAOD (stage II-IV) underwent during a single MRI: NE-MRA, contrast-enhanced CTM-MRA and TWIST-MRA with a single dose of a gadolinium-based contrast agent. The image quality (IQ) and the degree of stenoses were rated on a four-point scale. Positive (PPV) and negative predictive values (NPV), sensitivity (SS) and specificity (SP) for stenoses detection were calculated for NE-MRA vs. CTM-MRA and vs. TWIST-MRA. Values were obtained for overall graduation of wall changes and for severe stenoses (>70%). RESULTS: With NE-MRA 122/288 segments were not assessable. Compared with CTM-MRA and TWIST-MRA the IQ was significantly inferior (p < 0.0001 to p = 0.0426). CTM-MRA/TWIST-MRA detected stenoses in 44.9%/46.1% of the segments, NE-MRA in 53.5%. SS/NPV of the NE-MRA ranged from 97.8 to 100%. The SP and PPV ranged from 72.7 to 85.5% and 66.7 to 78.2%. CONCLUSION: Contrast-enhanced MRA techniques are superior to NE-MRA regarding IQ and correct identification of stenoses. If technically successful, NE-MRA is characterised by high NPV and overestimation of the degree of stenoses.

Assessment of the kidneys: magnetic resonance angiography, perfusion and diffusion.

Renal magnetic resonance (MR) imaging has undergone major improvements in the past several years. This review focuses on the technical basics and clinical applications of MR angiography (MRA) with the goal of enabling readers to acquire high-resolution, high quality renal artery MRA. The current role of contrast agents and their safe use in patients with renal impairment is discussed. In addition, an overview of promising techniques on the horizon for renal MR is provided. The clinical value and specific applications of renal MR are critically discussed.

Peripheral arterial occlusive disease: evaluation of a high spatial and temporal resolution 3-T MR protocol with a low total dose of gadolinium versus conventional angiography.

PURPOSE: To evaluate a peripheral magnetic resonance (MR) angiographic protocol combining continuous table movement (CTM) MR angiography of the entire runoff vasculature with time-resolved (TWIST) 3-T MR angiography of the calves with a total gadolinium dose of 0.1 mmol per kilogram of body weight. MATERIALS AND METHODS: In this retrospective institutional review board-approved study, 31 consecutive patients (22 men, nine women; mean age, 65 years ± 14 [standard deviation]) with peripheral arterial occlusive disease who had undergone a low-dose MR angiographic protocol that consisted of CTM MR angiography (repetition time msec/echo time msec, 2.4/1.0; 21° flip angle; voxel size, 1.2 mm(3); gadolinium dose, 0.07 mmol per kilogram of body weight) and TWIST MR angiography (2.8/1.1; 20° flip angle; voxel size, 1.1 mm(3); temporal resolution, 4.8-5.5 sec, gadolinium dose, 0.03 mmol/kg), as well as digital subtraction angiography (DSA), were included. Two radiologists rated image quality and stenosis degree on four-point scales. The accuracy of stenosis gradation and, specifically, the detection of high-grade stenoses (stenosis of 70%-99%) with CTM MR angiography alone and with the combined protocol were compared with accuracy of stenosis gradation and detection of high-grade stenoses with DSA. Means and standard deviations were calculated for all data. Interobserver agreement was determined with κ statistics. Positive and negative predictive values, sensitivity, specificity, and overall diagnostic accuracy were calculated for CTM MR angiography alone and for the combined protocol. RESULTS: For CTM MR angiography, image quality was good or excellent in 95.9% of vessel segments; for TWIST MR angiography, image quality was good or excellent in 94.3% and 97.8% of vessel segments for readers 1 and 2, respectively. The combined protocol resulted in high overall diagnostic accuracy of more than 80% for detection of stenosis and diagnostic accuracy of 93.5% for detection of high-grade vessel stenosis. Inclusion of TWIST MR angiography increased diagnostic value over that achieved with CTM MR angiography alone. CONCLUSION: A combined MR angiographic approach in which a low total gadolinium dose (0.1 mmol/kg) is used yields excellent image quality and is accurate in the diagnosis of peripheral arterial stenosis.