Fast Imaging Employing Steady-state Acquisition Images Research Articles

Multi-view convolutional neural networks for automated ocular structure and tumor segmentation in retinoblastoma

In retinoblastoma, accurate segmentation of ocular structure and tumor tissue is important when working towards personalized treatment. This retrospective study serves to evaluate the performance of multi-view convolutional neural networks (MV-CNNs) for automated eye and tumor segmentation on MRI in retinoblastoma patients. Forty retinoblastoma and 20 healthy-eyes from 30 patients were included in a train/test (N = 29 retinoblastoma-, 17 healthy-eyes) and independent validation (N = 11 retinoblastoma-, 3 healthy-eyes) set. Imaging was done using 3.0 T Fast Imaging Employing Steady-state Acquisition (FIESTA), T2-weighted and contrast-enhanced T1-weighted sequences. Sclera, vitreous humour, lens, retinal detachment and tumor were manually delineated on FIESTA images to serve as a reference standard. Volumetric and spatial performance were assessed by calculating intra-class correlation (ICC) and dice similarity coefficient (DSC). Additionally, the effects of multi-scale, sequences and data augmentation were explored. Optimal performance was obtained by using a three-level pyramid MV-CNN with FIESTA, T2 and T1c sequences and data augmentation. Eye and tumor volumetric ICC were 0.997 and 0.996, respectively. Median [Interquartile range] DSC for eye, sclera, vitreous, lens, retinal detachment and tumor were 0.965 [0.950–0.975], 0.847 [0.782–0.893], 0.975 [0.930–0.986], 0.909 [0.847–0.951], 0.828 [0.458–0.962] and 0.914 [0.852–0.958], respectively. MV-CNN can be used to obtain accurate ocular structure and tumor segmentations in retinoblastoma.

Definitions and MR imaging of duropathies

Kumar defined duropathies as disorders with dural defects-related cerebrospinal fluid leaks, particularly of spinal anterior dura mater, Superficial (hemo) siderosis (SS) and multisegemental amyotrophy (MSAM) were included in duropathies. Dural defects of SS had two types; one was incomplete closure of the dura in the spinal and cranial operations, the other was a spontaneous defect in the spinal anterior dura mater. In a majority of SS patients, spontaneous dural defects were detected at the levels of C7/Th1 to Th2/Th3 on axial FIESTA (fast imaging employing steady state acquisition) images. Similarly, dural defects in our 6 cases with MSAM were showed at C7/Th1 to Th2/Th3. Axial FIESTA images were crucial on MR imaging. T2 weighted images demonstrated abnormal hyperintense lesions in both anterior horns at the level of C3 spinal cord in all of 7 patients with MSAM. A dural defect in one case was not found.

The Role of Fast Imaging Employing Steady-State Acquisition (FIESTA) Magnetic Resonance Imaging for Assessment of Delayed Enhancement of Fat Graft Packing on Postoperative Imaging After Vestibular Schwannoma Surgery.

To investigate the role and efficacy of fast imaging employing steady-state acquisition (FIESTA) imaging in distinguishing fat graft enhancement from residual or recurrent tumor after vestibular schwannoma (VS) surgery. A retrospective study of 33 patients who underwent VS resection via the retrosigmoid or translabyrinthine approach with fat graft reconstruction was performed. Magnetic resonance imaging (MRI) was collected at different time points: preoperative, immediate postoperative (24-48 hours), delayed postoperative (3-6 months after surgery), and yearly postoperative. The image sets contained T1, T2, fat-suppressed T1-weighted with gadolinium, and FIESTA. The radiographs were analyzed for tumor recurrence by the primary neurosurgeon and an independent blinded neuroradiologist. If fat-suppressed T1-weighted images demonstrated postoperative enhancement in the resection bed, a comparison was made with FIESTA imaging. At 3-6 months postoperatively and at 1 year and beyond, 28 (84.8%) and 33 (100%) of patients, respectively, displayed delayed enhancement of the fat graft on postgadolinium fat-suppressed T1-weighted MRI. The enhancement seen on postgadolinium, fat-suppressed, T1-weighted MRI consistently correlated with the characteristic fat graft signal on FIESTA imaging and not tumor recurrence. FIESTA imaging was able to distinguish residual tumor from enhancing fat graft compared with postgadolinium, fat-suppressed, T1-weighted MRI (P < 0.0001) due to distinctive signaling patterns. FIESTA is an effective tool in discerning fat graft enhancement from residual or recurrent tumor on delayed postoperative imaging after VS resection. Fat graft used in reconstruction consistently enhances on delayed postoperative postgadolinium, fat-suppressed, T1-weighted imaging, which correlates with the fat graft signal seen on FIESTA images.

Visualization of the Abducens Nerve in its Petroclival Segment Using Contrast-Enhanced FIESTA MRI: The Size of the Petroclival Venous Confluence Affects Detectability.

Visualization of the abducens nerve in its petroclival segment still remains challenging. We aimed to investigate the detectability of the petroclival segment of the abducens nerve and to evaluate the role of the size of the petroclival venous confluence (PVC) in the visibility of the nerve using contrast-enhanced fast imaging employing steady-state acquisition (FIESTA) magnetic resonance imaging (MRI). We retrospectively evaluated the contrast-enhanced FIESTA images of 237 patients (female/male: 127/110; mean age: 49.0 ± 14.7). Two radiologists divided the imaging findings of the petroclival segments of the abducens nerves into three groups: 0 (not visualized), 1 (partially visualized), and 2 (completely visualized). Another radiologist measured the anteroposterior diameter of the PVC from the clival bony surface to the inner layer of dura at the dural entrance level of the abducens nerve. One-way analysis of variance, Tukey's test, and receiving operating curve analysis were performed. Among 474 abducens nerves, 76 were classified as group 0 (76/474, 16.03 %), 100 were classified as group 1 (100/474, 21.10 %), and 298 were classified as group 2 (298/474, 62.87 %). There was significant difference in mean anteroposterior diameters of the PVC for each group (group 0, 0.95mm; group 1, 1.80mm; group 2, 2.51mm). The cut-off values for the differentiation of group 0 from group 1 and group 1 from group 2 were found to be 1.35 and 2.09mm, respectively. Abducens nerve in its petroclival segment can be reliably identified using contrast-enhanced FIESTA MRI especially in those with a greater anteroposterior diameter of the PVC.

Optimal setting of image bounding box can improve registration accuracy of diffusion tensor tractography.

When we register diffusion tensor tractography (DTT) to anatomical images such as fast imaging employing steady-state acquisition (FIESTA), we register the B0 image to FIESTA. Precise registration of the DTT B0 image to FIESTA is possible with non-rigid registration compared to rigid registration, although the non-rigid methods lack convenience. We report the effect of image data bounding box settings on registration accuracy using a normalized mutual information (NMI) method METHODS: MRI scans of 10 patients were used in this study. Registration was performed without modification of the bounding box in the control group, and the results were compared with groups re-registered using multiple bounding boxes limited to the region of interest (ROI). The distance of misalignment after registration at 3 anatomical characteristic points that are common to both FIESTA and B0 images was used as an index of accuracy. Mean ([Formula: see text]SD) misalignment at the 3 anatomical points decreased significantly from [Formula: see text] to [Formula: see text]mm, [Formula: see text]), [Formula: see text] to [Formula: see text]mm, ([Formula: see text], and [Formula: see text] to [Formula: see text]mm, ([Formula: see text], each showing improvement compared to the control group CONCLUSION: Narrowing the image data bounding box to the ROI improves the accuracy of registering B0 images to FIESTA by NMI method. With our proposed methodology, accuracy can be improved in extremely easy steps, and this methodology may prove useful for DTT registration to anatomical image.

Delineation of Optic Nerves and Chiasm in Close Proximity to Large Suprasellar Tumors with Contrast-enhanced FIESTA MR Imaging

To evaluate whether contrast material-enhanced (CE) fast imaging employing steady-state acquisition (FIESTA) can depict the anterior optic pathways in patients with large suprasellar tumors. Institutional review board approval was obtained. Twenty-eight patients with large suprasellar tumors undergoing surgical treatment (19 pituitary adenomas, six meningiomas, and three additional miscellaneous tumors) underwent preoperative magnetic resonance (MR) imaging, including CE FIESTA, at 3.0 T. Two radiologists assessed the visibility of five segments of the optic pathways (bilateral optic nerves and optic tracts, optic chiasm) with CE FIESTA and conventional MR imaging, including thin-section coronal T2-weighted imaging and CE T1-weighted imaging, by using a three-point scale. Moreover, the preoperative signal intensity of the optic pathways was correlated with pre- and postoperative visual impairment to determine whether high signal intensity at CE FIESTA is predictive of persistence of visual impairment after surgery. The χ(2) test was used to compare scores assigned to CE FIESTA and conventional MR images. The percentage of anterior optic pathways rated as visible was significantly higher with CE FIESTA than with conventional MR imaging (100% [140 of 140 segments] vs 78% [109 of 140 segments], P < .05). Thirty-one of the 140 segments (22%) were not depicted with conventional MR imaging; all of these 31 segments were visualized with CE FIESTA. For 12 patients who underwent transcranial surgery, the anatomic locations of the optic pathways at CE FIESTA were compatible with the surgical findings. CE FIESTA helped predict persistent visual impairment after surgical treatment with a sensitivity of 75% (three of four patients) and a specificity of 96% (23 of 24 patients). The accuracy of CT FIESTA in the prediction of visual loss was significantly higher than that of T2-weighted imaging (93% [26 of 28 patients] vs 50% [14 of 28 patients], P < .05). CE FIESTA is useful for the preoperative localization of the anterior optic pathways in patients with large suprasellar tumors and offers the potential to predict persistent visual impairment after decompression.

Patients with idiopathic trigeminal neuralgia have a sharper-than-normal trigeminal-pontine angle and trigeminal nerve atrophy

Trigeminal neuralgia (TN) is primarily diagnosed by symptoms and patient history. Magnetic resonance (MR) imaging can be helpful in visualizing the neurovascular compression of the trigeminal nerve in TN patients, but the current parameters used as diagnostic markers for TN are less than optimal. The aim of this study is to assess whether the angle between the trigeminal nerve and the pons (the trigeminal-pontine angle) on the affected side of patients with idiopathic TN differs from that of the unaffected side and that found in controls without TN. A case-control study of 30 clinically diagnosed idiopathic TN patients aged 30 to 79 years and 30 age- and sex-matched controls was conducted. We compared the trigeminal-pontine angle and trigeminal nerve atrophy via fast-imaging employing steady-state acquisition (FIESTA) MR imaging. A sharp trigeminal-pontine angle was observed in 25 patients (25/30) on the affected side. As such, the mean angle of the trigeminal nerve on the affected side (40.17) was significantly smaller than that on the unaffected side (48.91, p = 0.001) and that in the control group (52.02, p < 0.001). A sharp trigeminal-pontine angle on the affected side was found in idiopathic TN patients by FIESTA imaging. This suggests that a sharp trigeminal-pontine angle increases the chance of neurovascular compression on the medial side of the trigeminal nerve.

SU‐E‐T‐281: Quantitative Evaluation of the MRI Image Distortion in Gamma Knife Radiosurgery

Purpose: To quantify the spatial distortion of the MRI images in the Elekta Leksell GammaPlan version 9 planning system using different registration methods. Methods: The RPC SRS phantom was imaged using a GE Signa HDxt 1.5Tesla MR scanner and a GE LightSpeed VCT CT scanner. Both the T1 weighted MRI images and the fast imaging employing steady state acquisition (FIESTA) MRI images were acquired, along with the axial and helical CT images. The skull shell and the target in the phantom were contoured in all the four series of images acquired. The MRI images were registered in the planning system in three different ways: 1) using the fiduciary marks in the imaging box; 2) using a global co‐registration with FOV just covering skull shell; 3) using a local co‐registration to max 8.5cm FOV at image center. The target positions in all the four series of images were compared using the software tool in the planning system. Results: The target positions as obtained from the helical and the axial CT images agree within 0.1mm. For the registration method the fiduciary marks, the T1 weighted MRI images are shifted from the CT images 0.9mm in the anterior‐posterior direction, and 0.5mm in both the superior‐inferior and the left‐right directions. The corresponding displacements for the FIESTA images are 0.6mm, 0.5mm and 0.5mm respectively. The shifts of the target positions are significantly reduced in the global co‐registration approach, and all less than 0.2mm in the local co‐registration approach. Conclusions: MRI image distortion in GammaPlan version 9 is in the sub‐millimeter range for the central region of the MRI image. The spatial distortion in the FIESTA images is smaller than that in the T1 weighted images. Co‐registration to the central region of images can reduce MRI imaging distortion effect significantly.

Initial experience with combined endoscopic third ventriculostomy and choroid plexus cauterization for post-hemorrhagic hydrocephalus of prematurity: the importance of prepontine cistern status and the predictive value of FIESTA MRI imaging

Post-hemorrhagic hydrocephalus of prematurity (PHHP) is among the most common causes of infant hydrocephalus in developed nations. This population has a high incidence of shunt failure, infection, and slit ventricle syndrome. Although effective for other etiologies of infant hydrocephalus, the efficacy of combined endoscopic third ventriculostomy and choroid plexus cauterization (ETV/CPC) in PHHP has not been investigated. This pilot study reports the initial experience. Ten patients (four grade III and six grade IV intraventricular hemorrhage) requiring definitive treatment for PHHP underwent ETV/CPC within 6 months of birth. Seven had a prior ventriculo-subgaleal shunt. Mean age at birth was -12.8 weeks, or 25.2 weeks gestation (24-28 weeks), and at surgery was -1.6 weeks (-11 to +11 weeks). Mean weight at surgery was 3.3 (1.0-5.5 kg). Each patient had preoperative magnetic resonance imaging (MRI) with fast imaging employing steady-state acquisition (FIESTA). Four of ten (40%) required no further operations related to hydrocephalus (mean follow-up, 29.7 months). Six required another procedure (five ultimately shunted). Prepontine cistern status correlated with outcome (p = 0.033). Procedures in all infants with unobstructed cisterns were successful but failed in six of seven with cisternal obstruction, with the one success having an alternative lamina terminalis endoscopic third ventriculostomy. Preoperative MRI FIESTA images correlated well with intraoperative assessment of the cistern. Results from this small homogenous cohort suggest cistern status is an important determinant of outcome. FIESTA imaging correlated with endoscopic observation. Preliminary analysis suggests ETV/CPC as an effective treatment for PHHP, but only when the cistern is unscarred. This information should guide patient selection for future study protocols.

Preoperative Evaluation of Unruptured Cerebral Aneurysms by Fast Imaging Employing Steady-State Acquisition Image

In aneurysm surgery, understanding the microanatomy around the aneurysm such as perforating arteries and cranial nerves is mandatory. To assess the usefulness in determining the microanatomy around the cerebral aneurysms by the use of fast imaging employing steady-state acquisition (FIESTA) images of magnetic resonance imaging preoperatively, in addition to computed tomography and digital subtraction angiography. Between October 2006 and June 2009, 123 patients with 140 unruptured cerebral aneurysms were treated in our institution. Eighty-two patients were assessed with FIESTA by the operators on the workstation of the magnetic resonance image before surgical clipping of the aneurysms. The small vessels and cranial nerves were confirmed intraoperatively before or after obliteration of the aneurysms. Sensitivities and specificities of FIESTA imaging were 100% in detecting hypothalamic artery around anterior communicating artery aneurysms, oculomotor nerve attachment to the posterior communicating artery aneurysm domes, and anterior choroidal artery adhesion to the posterior communicating artery aneurysms. This technique was also useful for predicting adhesion between the aneurysm and adjacent main trunks or perforators. Although the specificity was 100%, sensitivity was 56% in detecting vessel adhesion around the middle cerebral aneurysms. This technique can provide limited information in large aneurysms or aneurysms located in minimal cerebrospinal fluid space. The overall outcomes of the patients included 120 excellent recoveries, 1 moderate deficit, 1 severe deficit, and 1 persistent vegetative state according to the Glasgow Outcome Scale. By giving information on the minute anatomical structure around the aneurysm, FIESTA can contribute to thorough preoperative evaluations of cerebral aneurysms.

Evaluation of the Optic Nerve Complex in the Orbit Using Coronal Fast Magnetic Resonance Imaging

Recently available coronal fast magnetic resonance imaging (MRI) has very high spatial resolution with good contrast between the optic nerves and cerebrospinal fluid (CSF). The aim of this study was to evaluate the diagnostic value of coronal fast imaging in optic nerve diseases. Thirty-five patients with various Neuro-ophthalmic conditions including 9 with optic neuritis, 6 with optic atrophy, 5 with glaucoma, 4 with segmental optic nerve hypoplasia and 11 with other optic neuropathies including orbital apex syndrome were evaluated with the three-dimensional fast imaging employing steady-state acquisition (FIESTA) sequence in addition to standard MRI protocols. The optic nerve complexes were evaluated on coronal images of the orbits. Detailed demonstration of the optic nerve complex—the optic nerve, the perineural CSF space and dural sheath—could be readily obtained with FIESTA sequence. The acute phase of both optic neuritis and perineuritis showed enlargement of the perineural CSF space; the optic nerve was swollen in optic neuritis but not in perineuritis. Cases of optic atrophy and glaucoma showed perineural CSF space enlargement with normal optic sheath circumference and a thinner optic nerve, while optic nerve hypoplasia showed a smaller dural sheath circumference without perineural CSF space enlargement. In the cases of orbital apex syndrome optic nerve compression by the extraocular muscles was clearly shown. Coronal FIESTA imaging of the orbit is capable of delineating detailed structural changes in the optic nerve complex and is of diagnostic value for the differentiation of optic nerve diseases.

Two‐dimensional fast imaging employing steady‐state acquisition (FIESTA) cine acquisition of fetal non–central nervous system abnormalities

To evaluate the value of two-dimensional fast imaging employing steady-state acquisition (2D FIESTA) cine MR with parallel imaging techniques in the diagnosis of fetal non-central nervous system (CNS) anomalies. A total of 28 pregnant women were referred for further MR evaluation on fetuses after abnormal sonographic results. A total of 33 fetal MR examinations were performed by a 1.5 T MR scanner with eight-channel phase-arrayed body coils. Single-shot fast spin-echo (SSFSE(R), GE) of three orthogonal planes and 2D FIESTA for cine fetal MR of three sagittal planes (midsagittal and 10 mm off midline on left and right) were routinely acquired. Additional planes on target organs with variable imaging frames were added if indicated. Nine of the 33 examinations (9/33; 27.3%) had motion artifacts obscuring the detail in SSFSE imaging; 2D FIESTA imaging provided motion-artifact-free imaging in all of them. Cine 2D FIESTA imaging provided additional information on the visceral peristalsis. The information helped in differentiating dilated gastrointestinal (GI) tract from other intraabdominal cystic lesions and in confirming the nature and level of GI tract obstruction. With sub-half-second temporal resolution of the 2D FIESTA sequences, fetal movement is no longer problematic. In addition to the anatomical information also provided by conventional SSFSE sequences, 2D FIESTA demonstrates information on motility and peristalsis of hollow organs and helps the diagnosis of fetal visceral anomalies.

Comparison of FIESTA and gadolinium‐enhanced T1‐weighted sequences in magnetic resonance of acoustic schwannoma

AbstractAcoustic schwannomas (AS) are a treatable cause of sensorineural hearing loss (SNHL). Currently, the gold standard magnetic resonance (MR) examination for exclusion of acoustic schwannoma is gadolinium‐enhanced T1‐weighted images of the cerebropontine angles (CPA). The FIESTA (fast imaging employing steady‐state acquisition) sequence is a true‐FISP (free induction steady‐state precession) sequence that provides high‐resolution fluid‐bright images of the CPAs and basal cisterns. The T1‐weighted sequence adds cost and time to the examination, and possibility of contrast medium reaction exists. This study of 50 patients compared the gadolinium‐enhanced T1‐weighted images and FIESTA images. The hypothesis was that the FIESTA sequence could repIACe the gadolinium‐enhanced T1‐weighted images for the diagnosis of AS. The results showed that in 98% of cases, this was possible. The conclusion was that when the FIESTA sequence does not demonstrate AS, this sequence may be employed in isolation. However, gadolinium‐enhanced images should be employed when pathology is seen and to follow‐up post‐surgical patients.

MR colonoscopy at 3.0 T: comparison with 1.5 T in vivo and a colon model

Retrospectively, magnetic resonance (MR) colonography images obtained from a colon model and in routine examinations of patients screened for polyps were compared in terms of whether, and to what degree, image quality improved at a higher field strength of 3.0 T compared to 1.5 T. One hundred twenty-eight MR colonography images from 40 patients, of whom 20 had each been scanned at 1.5 and 3.0 T, respectively, using a four-element phased-array torso coil, were compared. At both field strengths, imaging included T1-weighted fat-suppressed spoiled gradient-echo (T1-fs-GE), T2/T1-weighted fast imaging employing steady-state acquisition (FIESTA), and T2-weighted single-shot fast spin-echo (T2-SSFSE), with breath-hold technique. Using receiver operating characteristic analysis performed by seven readers, the three types of images from the colon model and from 20 patients each at 1.5 and 3.0 T were compared. While a time window of 20 s was allowed for picture assessment in a chance-generated succession of images on a monitor, image quality was rated with a score of 1-5 (1=very good; 5=very bad). Statistical significance was calculated with Mann-Whitney U test. At both field strengths, T2-SSFSE images received the best ratings, followed by FIESTA images (P=.001). Although, overall, the 3.0-T images obtained scores worse than those of the 1.5-T images, a better detection of phantom polyps was noted in the colon model (P=.001). Although MR colonography with the breath-hold technique using the same four-element phased-array coil at 3.0 and 1.5 T does not perform better at a higher field strength in general, an improved detection of small polyps may be obtained.

Cranial nerve assessment in posterior fossa tumors with fast imaging employing steady-state acquisition (FIESTA)

Steady-state free precession is widely used for ultra-fast cardiac or abdominal imaging. The purpose of this work was to assess fast imaging employing steady-state acquisition (FIESTA) and to evaluate its efficacy for depiction of the cranial nerve affected by the tumor. Twenty-three consecutive patients with posterior fossa tumors underwent FIESTA sequence after contrast agent administration, and then displacement of the cranial nerve was evaluated. The 23 patients with posterior fossa tumor consisted of 12 schwannomas, eight meningiomas, and three cases of epidermoid. Except in the cases of epidermoid, intensity of all tumors increased on FIESTA imaging of the contrast enhancement. In the schwannoma cases, visualization of the nerve became poorer as the tumor increased in size. In cases of encapsulated meningioma, all the cranial nerves of the posterior fossa were depicted regardless of location. The ability to depict the nerves was also significantly higher in meningioma patients than in schwannoma patients (P<0.05). In cases of epidermoid, extension of the tumors was depicted clearly. Although the FIESTA sequence offers similar contrast to other heavily T2-weighted sequences, it facilitated a superior assessment of the effect of tumors on cranial nerve anatomy. FIESTA sequence was useful for preoperative simulations of posterior fossa tumors.

Cellular Imaging of Inflammation after Experimental Spinal Cord Injury

The ability to visualize the cellular inflammatory responses after experimental spinal cord injury (SCI) was investigated using a clinical 1.5-T magnetic resonance imaging scanner, a custom-built, high-strength gradient coil insert, a 3-D fast imaging employing steady-state acquisition (FIESTA) imaging sequence and a superparamagnetic iron oxide (SPIO) contrast agent. An "active labeling" approach was used, with SPIO administered intravenously at different time points following SCI. Our results show that this strategy can be used to visualize clusters of iron-labeled cells associated with the inflammatory response in SCI. Of particular importance for this application was the finding that in FIESTA images hemorrhage does not cause signal loss. In T2-weighted spin echo or T2*-weighted gradient-echo images, which are more commonly used to detect signal loss associated with SPIO, the signal loss associated with hemorrhage interferes with the detection of iron-induced signal loss. FIESTA, therefore, allowed us to discriminate between iron associated with blood products in hemorrhage that occurs in acute SCI and the iron associated with SPIO-labeled cells accumulating in the injured cord.