Virtual magnetic resonance lumbar spine images generated from computed tomography images using conditional generative adversarial networks.

INTRODUCTION: The aim of this study was to generate virtual Magnetic resonance (MR) from computed tomography (CT) using conditional generative adversarial networks (cGAN). METHODS: We selected examinations from 22 adults who obtained their CT and MR lumbar spine examinations. Overall, 4 examinations were used as test data, and 18 examinations were used as training data. A cGAN was trained to generate virtual MR images from the CT images using the corresponding MR images as targets. After training, the generated virtual MR images from test data in epochs 1, 10, 50, 100, 500, and 1000 were compared with the original ones using the mean square error (MSE) and structural similarity index (SSIM). Additionally, two radiologists also performed qualitative assessments. RESULTS: The MSE of the virtual MR images decreased as the epoch of the cGANs increased from the original CT images: 8876.7 ± 1192.9 (original CT), 1567.5 ± 433.9 (Epoch 1), 1242.4 ± 442.0 (Epoch 10), 1065.8 ± 478.1 (Epoch 50), 1276.1 ± 718.9 (Epoch 100), 1046.7 ± 488.2 (Epoch 500), and 1031.7 ± 400.0 (Epoch 1000). No considerable differences were observed in the qualitative evaluation between the virtual MR images and the original ones, except in the structure of the spinal canal. CONCLUSION: Virtual MR lumbar spine images using cGANs could be a feasible technique to generate near-MR images from CT without MR examinations for evaluation of the vertebral body and intervertebral disc. IMPLICATIONS FOR PRACTICE: Virtual MR lumbar spine images using cGANs can offer virtual CT images with sufficient quality for attenuation correction for PET or dose planning in radiotherapy.

Usefulness of Contrast-Enhanced 3D-FLAIR MR Imaging for Differentiating Rathke Cleft Cyst from Cystic Craniopharyngioma.

BACKGROUND AND PURPOSE: Because it can be difficult to discriminate between a Rathke cleft cyst and cystic craniopharyngioma by conventional MR imaging alone, we investigated whether contrast-enhanced 3D T2-FLAIR MR imaging at 3T helps to distinguish a Rathke cleft cyst from a cystic craniopharyngioma. MATERIALS AND METHODS: We evaluated pre- and postcontrast T1-weighted and 3D T2-FLAIR images of 17 patients with pathologically confirmed Rathke cleft cyst (n = 10) or cystic craniopharyngioma (n = 7). All underwent 3T MR imaging studies before surgery. Two neuroradiologists independently recorded the enhancement grade of the lesion wall as grade 2 (most of the wall enhanced), grade 1 (some of the wall enhanced), and grade 0 (none of the wall enhanced). One neuroradiologist performed a blinded reading study of conventional MR images with/without 3D T2-FLAIR images. Interobserver agreement was determined by calculating the κ coefficient. Statistical analyses, including receiver operating characteristic curve analysis were performed. RESULTS: Interobserver agreement for postcontrast T1WI and 3D T2-FLAIR images was excellent (κ = 0.824 and κ = 0.867, respectively). Although the difference in the mean enhancement grade of Rathke cleft cysts and cystic craniopharyngiomas was not significant on postcontrast T1WIs, it was significant on postcontrast 3D T2-FLAIR images (P = .0011). The area under the receiver operating characteristic curve of the conventional MR alone and conventional MR with 3D T2-FLAIR readings was 0.879 and 1.0, respectively, though there was no significant difference in the area under the curve between the 2 readings. CONCLUSIONS: Contrast-enhanced 3D T2-FLAIR imaging at 3T helps to distinguish a Rathke cleft cyst from cystic craniopharyngioma.

Distinguishing imaging features between spinal hyperplastic hematopoietic bone marrow and bone metastasis.

BACKGROUND AND PURPOSE: Systematic investigations of the distinguishing imaging features between spinal hyperplastic hematopoietic bone marrow and bone metastasis have not been reported, to our knowledge. The purpose of this study was to determine the distinguishing imaging features of the 2 entities. MATERIALS AND METHODS: We retrospectively reviewed the radiologic images of 8 consecutive male patients (age range, 52-78 years; mean, 64 years) with suspected spinal metastasis on MR imaging and FDG-PET, which was later confirmed as hyperplastic hematopoietic bone marrow. MR imaging, FDG-PET, CT, and bone scintigraphy images were qualitatively and/or quantitatively evaluated. Imaging findings in 24 patients with spinal metastasis were compared, and differences were statistically analyzed. RESULTS: All 8 vertebral hyperplastic hematopoietic bone marrow lesions were hypointense on T1- and T2-weighted images; lesions contiguous with the adjacent vertebra were significantly more often seen in hyperplastic hematopoietic bone marrow than in metastasis (P = .035). T2 signal intensity of the lesion was significantly different between the 2 entities (P = .033). FDG-PET showed slightly higher uptake in all hyperplastic hematopoietic bone marrow lesions; their maximum standard uptake value was significantly lower than that of metastatic lesions (P = .037). CT attenuation of hyperplastic hematopoietic bone marrow was equal to or slightly higher than that of adjacent normal-appearing vertebra; the CT appearances of hyperplastic hematopoietic bone marrow and metastasis were significantly different (P < .01). Bone scintigraphy showed normal uptake for all vertebrae with hyperplastic hematopoietic bone marrow; the uptake was significantly different from that of metastasis (P < .01). CONCLUSIONS: If a lesion was isointense to hyperintense to normal-appearing marrow on MR imaging or had a maximum standard uptake value of >3.6, the lesion was considered metastatic. A normal appearance on CT or bone scintigraphy excluded metastasis.

Can 3T MR angiography replace DSA for the identification of arteries feeding intracranial meningiomas?

BACKGROUND AND PURPOSE: For identifying the arterial feeders of meningiomas, the usefulness of 3D TOF MRA at 3T has not been systematically investigated. This study was intended to assess whether unenhanced 3D TOF MRA at 3T can replace DSA for the identification of arteries feeding intracranial meningiomas and whether it is useful for assessing their dural attachment. MATERIALS AND METHODS: Twenty-one consecutive patients with intracranial meningiomas (18 women, 3 men; aged 42-77 years, mean 57 years) underwent DSA, conventional MR imaging, and 3D TOF MRA. Two neuroradiologists independently evaluated the primary and secondary feeders of each tumor on maximum-intensity-projection and source MRA images. They also identified the location of dural attachments based on information from MR imaging/MRA images. Interobserver and intermodality agreement was determined by calculating the κ coefficient. RESULTS: For the identification of primary and secondary feeders on MRA images, interobserver agreement was very good (κ=0.83; 95% CI, 0.66-1.00) and moderate (κ=0.58; 95% CI, 0.34-0.82) and intermodality agreement (consensus reading of MRA versus DSA findings) was excellent (κ=0.94; 95% CI, 0.84-1.00) and good (κ=0.72; 95% CI, 0.51-0.93), respectively. With respect to the dural attachment of meningiomas, interobserver agreement was very good (κ=0.95; 95% CI, 0.84-1.00). The agreement in the diagnosis between MR imaging/MRA and surgery was excellent (κ=1.00). CONCLUSIONS: Unenhanced 3D TOF MRA at 3T cannot at present supplant DSA for the identification of the feeding arteries of intracranial meningiomas. This information may be useful for evaluating their dural attachment.

Evaluation of brain and head and neck tumors with 4D contrast-enhanced MR angiography at 3T.

BACKGROUND AND PURPOSE: Systematic assessment of brain and head and neck tumors with 4D-CE-MRA at 3T has not been investigated. The purpose of this study was to test the hypothesis that 4D-CE-MRA at 3T can replace DSA in the identification of feeding arteries and tumor stain to plan interventional procedures in hypervascular brain and head and neck tumors. MATERIALS AND METHODS: Fifteen consecutive patients with brain and head and neck tumors underwent 4D-CE-MRA at 3T and DSA. 4D-CE-MRA combined randomly segmented central k-space ordering, keyhole imaging, SENSE, and half-Fourier imaging. We obtained 30 dynamic scans every 1.9 seconds at an acquired spatial resolution of 0.9 × 0.9 × 1.5 mm; the matrix was 256 × 256. Two independent observers inspected the 4D-CE-MRA images for the main arterial feeders and tumor stain. Interobserver and intermodality agreement was assessed by κ statistics. RESULTS: For 4D-CE-MRA, the interobserver agreement was fair with respect to the main arterial feeders and very good for the degree of tumor stain (κ = 0.28 and 0.87, respectively). Intermodality agreement was moderate for the main arterial feeders (κ = 0.45) and good for the tumor stain (κ = 0.74). CONCLUSIONS: Although 4D-CE-MRA may be useful for evaluating tumor stain in hypervascular brain and head and neck tumors, it is not able to replace DSA in planning interventional procedures.

Quantitative blood flow measurements in gliomas using arterial spin-labeling at 3T: intermodality agreement and inter- and intraobserver reproducibility study.

BACKGROUND AND PURPOSE: QUASAR is a particular application of the ASL method and facilitates the user-independent quantification of brain perfusion. The purpose of this study was to assess the intermodality agreement of TBF measurements obtained with ASL and DSC MR imaging and the inter- and intraobserver reproducibility of glioma TBF measurements acquired by ASL at 3T. MATERIALS AND METHODS: Two observers independently measured TBF in 24 patients with histologically proved glioma. ASL MR imaging with QUASAR and DSC MR imaging were performed on 3T scanners. The observers placed 5 regions of interest in the solid tumor on rCBF maps derived from ASL and DSC MR images and 1 region of interest in the contralateral brain and recorded the measured values. Maximum and average sTBF values were calculated. Intermodality and intra- and interobsever agreement were determined by using 95% Bland-Altman limits of agreement and ICCs. RESULTS: The intermodality agreement for maximum sTBF was good to excellent on DSC and ASL images; ICCs ranged from 0.718 to 0.884. The 95% limits of agreement ranged from 59.2% to 65.4% of the mean. ICCs for intra- and interobserver agreement for maximum sTBF ranged from 0.843 to 0.850 and from 0.626 to 0.665, respectively. The reproducibility of maximum sTBF measurements obtained by methods was similar. CONCLUSIONS: In the evaluation of sTBF in gliomas, ASL with QUASAR at 3T yielded measurements and reproducibility similar to those of DSC perfusion MR imaging.