T1-weighted Gradient-echo MRI Research Articles

Computer-aided Detection of Brain Metastases in T1-weighted MRI for Stereotactic Radiosurgery Using Deep Learning Single-Shot Detectors.

Background Brain metastases are manually identified during stereotactic radiosurgery (SRS) treatment planning, which is time consuming and potentially challenging. Purpose To develop and investigate deep learning (DL) methods for detecting brain metastasis with MRI to aid in treatment planning for SRS. Materials and Methods In this retrospective study, contrast material-enhanced three-dimensional T1-weighted gradient-echo MRI scans from patients who underwent gamma knife SRS from January 2011 to August 2018 were analyzed. Brain metastases were manually identified and contoured by neuroradiologists and treating radiation oncologists. DL single-shot detector (SSD) algorithms were constructed and trained to map axial MRI slices to a set of bounding box predictions encompassing metastases and associated detection confidences. Performances of different DL SSDs were compared for per-lesion metastasis-based detection sensitivity and positive predictive value (PPV) at a 50% confidence threshold. For the highest-performing model, detection performance was analyzed by using free-response receiver operating characteristic analysis. Results Two hundred sixty-six patients (mean age, 60 years ± 14 [standard deviation]; 148 women) were randomly split into 80% training and 20% testing groups (212 and 54 patients, respectively). For the testing group, sensitivity of the highest-performing (baseline) SSD was 81% (95% confidence interval [CI]: 80%, 82%; 190 of 234) and PPV was 36% (95% CI: 35%, 37%; 190 of 530). For metastases measuring at least 6 mm, sensitivity was 98% (95% CI: 97%, 99%; 130 of 132) and PPV was 36% (95% CI: 35%, 37%; 130 of 366). Other models (SSD with a ResNet50 backbone, SSD with focal loss, and RetinaNet) yielded lower sensitivities of 73% (95% CI: 72%, 74%; 171 of 234), 77% (95% CI: 76%, 78%; 180 of 234), and 79% (95% CI: 77%, 81%; 184 of 234), respectively, and lower PPVs of 29% (95% CI: 28%, 30%; 171 of 581), 26% (95% CI: 26%, 26%; 180 of 681), and 13% (95% CI: 12%, 14%; 184 of 1412). Conclusion Deep-learning single-shot detector models detected nearly all brain metastases that were 6 mm or larger with limited false-positive findings using postcontrast T1-weighted MRI. © RSNA, 2020 See also the editorial by Kikinis and Wells in this issue.

The opacity of mineral ion-loaded bead (DC beads®) on low-keV monochromatic images from dual energy CT and T1-weighted gradient-echo MRI.

To evaluate the opacity of DC beads® (DCB) loaded with mineral ions on low-keV monochromatic images from dual energy computed tomography (DECT) and T1-weighted gradient-echo (T1-GRE) MRI. Fe2+ or Ca2+-loaded DCBs were prepared by mixing DCBs in 100mM FeSO4 or CaSO4 solution and scanned by DECT from 10min to 27h after mixing. The Hounsfield units (HUs) of sedimented DCBs on 40-keV monochromatic images were measured. Next, we mixed DCBs in 100, 10, 5 and 1mM FeSO4 solutions, and scanned these solutions from 15 to 120min after mixing using a 3T MR scanner. The signal-noise ratios (SNRs) of sedimented DCBs on T1-GRE were measured. Venous blood was scanned to compare with DCBs. The CT values of DCBs in FeSO4 and CaCl2 solutions gradually increased, and were 113.3 and 43.1HU at 27h, respectively; that of blood was 17.8HU. The SNR of DCB in 1mM FeSO4 solution increased and achieved equilibrium at 120min, and was 120.5 and higher than in the other FeSO4 solutions. The SNR of blood was 49.7. Optimally Fe2+-loaded DCBs can be discriminated from venous blood on 40-keV monochromatic images from DECT and T1-GRE.

Macrophage Infiltration in the Saccular Intracranial Aneurysm Wall as a Response to Locally Lysed Erythrocytes That Promote Degeneration.

Saccular intracranial aneurysm (sIA) rupture is often fatal. Rupture-prone sIA walls are infiltrated by macrophages expressing hemoglobin-receptor CD163, suggesting a role for erythrocyte lysis in the degenerative remodeling predisposing to rupture. We therefore studied erythrocyte remnants in 16 unruptured and 20 ruptured sIA walls using histology and immunohistochemistry. Glycophorin A (GPA), an erythrocyte membrane protein, was present in 34/36 (94%) sIA walls and correlated with loss of αSMA+ cells, reflecting loss of mural smooth muscle cells ([SMCs]; r = -0.592, p < 0.001), wall degeneration (p = 0.008), and rupture (p = 0.005). GPA correlated with high numbers of CD163+ and CD68+ phagocytes (r = 0.65 and r = 0.54, p ≤ 0.001 for both). CD163+ phagocytes were mostly HLA-DR-. Interestingly, single SMCs expressed HLA-DR and also CD163 was expressed in sporadic SMCs, which may reflect their response to hemoglobin accumulation. GPA associated with iron (p = 0.014) was detectable by MRI. An additional 11 sIAs were therefore imaged ex vivo with a 4.7 T MRI prior to histology. In the sIA walls, high GPA and iron accumulation associated with signal intensity in T1-weighted gradient echo MRI. We conclude that accumulation of lysed erythrocytes is a potential driver of inflammatory response in the sIA walls and is associated with the degenerative wall remodeling, thereby predisposing to rupture.

Superiority of constructive interference in steady-state MRI sequencing over T1-weighted MRI sequencing for evaluating cavernous sinus invasion by pituitary macroadenomas.

Preoperatively determining the extent of parasellar invasion of pituitary macroadenomas is useful for surgical planning and patient counseling. Here, the authors compared constructive interference in steady state (CISS), a T2-weighted gradient-echo MRI sequence, to volume-interpolated breath-hold examination (VIBE), a T1-weighted gradient-echo MRI sequence, for evaluation of cavernous sinus invasion (CSI) by pituitary macroadenomas. VIBE and CISS images of 98 patients with pituitary macroadenoma were retrospectively analyzed and graded using the modified Knosp classification. The Knosp grades were correlated to surgical findings of CSI, which were determined intraoperatively using 0° and 30° endoscopes. The predictive accuracies for CSI according to the Knosp grades derived from the CISS and VIBE images were compared using receiver operating characteristic (ROC) curves. Postoperative MRI was used to evaluate the gross-total resection (GTR) rates. The CSI rate by pituitary macroadenomas was 27.6% (27 of 98 cases). Of 196 assessments (left and right sides of 98 macroadenomas), 45 (23.0%) had different Knosp grades when scored using VIBE versus CISS images. For the VIBE images, 0% of Knosp grade 0, 4.5% of grade 1, 23.8% of grade 2, 42.1% of grade 3A, 100% of grade 3B, and 83.3% of grade 4 macroadenomas were found to have CSI intraoperatively. For the CISS images, 0% of Knosp grade 0, 2.1% of grade 1, 31.3% of grade 2, 56.3% of grade 3A, 100% of grade 3B, and 100% of grade 4 macroadenomas were found to have CSI intraoperatively. Two pituitary macroadenomas were classified as grade 4 on VIBE sequences but grades 3A and 2 on CISS sequences; CSI was not observed intraoperatively in both cases. The GTR rate was 64.3% and 60.0% for high-grade (3A, 3B, and 4) macroadenomas classified using VIBE and CISS sequences, respectively. The areas under the ROC curves were 0.94 and 0.97 for VIBE- and CISS-derived Knosp grades (p = 0.007), respectively. Knosp grades determined using CISS sequence images are better correlated with intraoperative CSI than those determined using VIBE sequence images. CISS sequences may be valuable for the preoperative assessment of pituitary macroadenomas.

Renal imaging at 7 Tesla: preliminary results

To investigate the feasibility of 7T MR imaging of the kidneys utilising a custom-built 8-channel transmit/receive radiofrequency body coil. In vivo unenhanced MR was performed in 8 healthy volunteers on a 7T whole-body MR system. After B(0) shimming the following sequences were obtained: 1) 2D and 3D spoiled gradient-echo sequences (FLASH, VIBE), 2) T1-weighted 2D in and opposed phase 3) True-FISP imaging and 4) a T2-weighted turbo spin echo (TSE) sequence. Visual evaluation of the overall image quality was performed by two radiologists. Renal MRI at 7T was feasible in all eight subjects. Best image quality was found using T1-weighted gradient echo MRI, providing high anatomical details and excellent conspicuity of the non-enhanced vasculature. With successful shimming, B(1) signal voids could be effectively reduced and/or shifted out of the region of interest in most sequence types. However, T2-weighted TSE imaging remained challenging and strongly impaired because of signal heterogeneities in three volunteers. The results demonstrate the feasibility and diagnostic potential of dedicated 7T renal imaging. Further optimisation of imaging sequences and dedicated RF coil concepts are expected to improve the acquisition quality and ultimately provide high clinical diagnostic value.

IC-P-080: Effects of aging on whole brain gray and white matter volumes and hippocampal subfields

Although memory is commonly affected in elderly healthy subjects, neuroimaging studies seeking for age related hippocampal volume loss have been controversial. One reason for this might be that the hippocampus is not homogenous but consists of histologically and functionally different subfields which might be differently affected by age. 1. To test for global age related gray and white matter volume changes. 2. To test if hippocampal age effects are global or subfield specific. 45 cognitively normal controls (aged 21–85, mean 54.5±17.4, m/f: 27/18) were studied on a 4T magnet with the following sequences: 1. Volumetric T1–weighted gradient echo MRI (MPrage) resolution 1 × 1 × 1 mm. 2. High resolution T2–weighted fast spin echo 0.4 × 0.5 × 2 mm resolution, 2, angulated perpendicular to the long axis of the hippocampus. The MPrage was segmented into gray (GM) and white maps (WM) which were used for optimized voxel–based–morphometry (VBM) in SPM2. The entorhinal cortex (ERC), subiculum, CA1, CA2 and CA3/4/dentate–compound were marked on the high resolution image on 5 consecutive slices using anatomical landmarks. The influence of age, gender and total intracranial volume on hippocampal subfields was tested using linear regression and regionally unbiased tests for age related GM/WM reductions were performed using SPM2. VBM showed age related GM reductions bilaterally in hippocampus, putamen, thalamus, cerebellum, operculum, anterior and posterior cingulate and parieto–occipital and frontal cortex while WM reductions were found in corona radiata, corpus callosum, and periventricular, temporal and frontal WM. Significant negative age effects were found for ERC (p>0.005) and CA1 (p<0.00001) and significant age and gender effects (p<0.025) for CA2 which was primarily driven by volume loss in men. Neither CA1 nor ERC volume loss was correlated with GM losses after correction for age effects. Healthy aging is associated with regionally restricted volume loss that involves cortical regions incl. ERC. In the hippocampus age effects are restricted to specific subfields, notably CA1 while others, e.g. subiculum, are less affected. This regionally distinct pattern of atrophic changes, particularly also in the hippocampus, may have diagnostic value for early detection of Alzheimer's disease.

Scaphoid blood flow and acute fracture healing

We have investigated whether assessment of blood flow to the proximal scaphoid can be used to predict nonunion in acute fractures of the scaphoid. We studied 32 fractures of the scaphoid one to two weeks after injury, by dynamic fat-suppressed T1-weighted gradient-echo MRI after the intravenous administration of gadopentetate dimeglumine (0.1 mmol/kg body-weight). Steepest slope values (SSV) and percentage enhancement values (%E) were calculated for the distal and proximal fragments and poles. All the fractures were treated by immobilisation in a cast, and union was assessed by CT at 12 weeks. Nonunion occurred in four fractures (12%), and there was no statistically significant difference between the proximal fragment SSV and %E values for the fractures which united and those with nonunion. The difference between the proximal pole SSV and %E values for the union and nonunion groups reached statistical significance (p &lt; 0.05), but with higher enhancement parameters for the nonunion group. Our results suggest that poor proximal vascularity is not an important determinant of union in fractures of the scaphoid.

Scaphoid blood flow and acute fracture healing. A dynamic MRI study with enhancement with gadolinium.

We have investigated whether assessment of blood flow to the proximal scaphoid can be used to predict nonunion in acute fractures of the scaphoid. We studied 32 fractures of the scaphoid one to two weeks after injury, by dynamic fat-suppressed T1-weighted gradient-echo MRI after the intravenous administration of gadopentetate dimeglumine (0.1 mmol/kg body-weight). Steepest slope values (SSV) and percentage enhancement values (%E) were calculated for the distal and proximal fragments and poles. All the fractures were treated by immobilisation in a cast, and union was assessed by CT at 12 weeks. Nonunion occurred in four fractures (12%), and there was no statistically significant difference between the proximal fragment SSV and %E values for the fractures which united and those with nonunion. The difference between the proximal pole SSV and %E values for the union and nonunion groups reached statistical significance (p < 0.05), but with higher enhancement parameters for the nonunion group. Our results suggest that poor proximal vascularity is not an important determinant of union in fractures of the scaphoid.

Dynamic MRI sensitized to cerebral blood oxygenation and flow during sustained activation of human visual cortex.

Changes in cerebral blood oxygenation and flow during prolonged activation of human visual cortex (6-min video projection) were monitored using high-resolution T2*- and T1-weighted gradient-echo MRI in identical sessions. Oxygenation-sensitive recordings displayed an initial signal increase (oxygenation "overshoot"), a subsequent signal decrease extending over 4-5 min (relative deoxygenation), and a signal drop after the end of stimulation that mirrored the initial response (oxygenation "undershoot"). Flow-sensitive MRI demonstrated that the inflow effect remained elevated during the entire period of stimulation. The observation of gradually decreasing cerebral blood oxygenation, despite persisting elevation of blood flow, may be understood to be an accumulation of deoxyhemoglobin due to the progressive up-regulation of oxidative phosphorylation. The present findings support a concept in which transitions between functional states lead to an uncoupling of perfusion (oxygen delivery) from oxidative metabolism (oxygen consumption) whereas steady-state activity achieves their recoupling.