MRI reconstruction with enhanced self-similarity using graph convolutional network.

BACKGROUND: Recent Convolutional Neural Networks (CNNs) perform low-error reconstruction in fast Magnetic Resonance Imaging (MRI). Most of them convolve the image with kernels and successfully explore the local information. Nonetheless, the non-local image information, which is embedded among image patches relatively far from each other, may be lost due to the limitation of the receptive field of the convolution kernel. We aim to incorporate a graph to represent non-local information and improve the reconstructed images by using the Graph Convolutional Enhanced Self-Similarity (GCESS) network. METHODS: First, the image is reconstructed into the graph to extract the non-local self-similarity in the image. Second, GCESS uses spatial convolution and graph convolution to process the information in the image, so that local and non-local information can be effectively utilized. The network strengthens the non-local similarity between similar image patches while reconstructing images, making the reconstruction of structure more reliable. RESULTS: Experimental results on in vivo knee and brain data demonstrate that the proposed method achieves better artifact suppression and detail preservation than state-of-the-art methods, both visually and quantitatively. Under 1D Cartesian sampling with 4 × acceleration (AF = 4), the PSNR of knee data reached 34.19 dB, 1.05 dB higher than that of the compared methods; the SSIM achieved 0.8994, 2% higher than the compared methods. Similar results were obtained for the reconstructed images under other sampling templates as demonstrated in our experiment. CONCLUSIONS: The proposed method successfully constructs a hybrid graph convolution and spatial convolution network to reconstruct images. This method, through its training process, amplifies the non-local self-similarities, significantly benefiting the structural integrity of the reconstructed images. Experiments demonstrate that the proposed method outperforms the state-of-the-art reconstruction method in suppressing artifacts, as well as in preserving image details.

An Evaluation of Screening Pathways Using a Combination of Magnetic Resonance Imaging and Prostate-specific Antigen: Results from the IP1-PROSTAGRAM Study.

BACKGROUND: The use of prostate-specific antigen (PSA) testing to screen for prostate cancer has been fraught with under- and overdiagnosis. Short, noncontrast magnetic resonance imaging (MRI) might detect more grade group ≥2 cancers with similar rates of biopsy. OBJECTIVE: To evaluate strategies that combined PSA and MRI to select men based in the community for a prostate biopsy. DESIGN, SETTING, AND PARTICIPANTS: IP1-PROSTAGRAM was a prospective, population-based, paired cohort study of 408 men aged 50-69 yr conducted at seven UK primary care practice and two imaging centres (from October 10, 2018 to May 15, 2019). INTERVENTION: All participants underwent screening with a PSA test, MRI (T2-weighted and diffusion), and transrectal ultrasound (b-mode and elastography). If any test was screen positive, a systematic 12-core biopsy was performed. Additional image-fusion targeted biopsies were taken if the MRI or ultrasound was positive. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: We conducted an analysis, set out in the statistical plan a priori, comparing 13 different pathways including PSA-alone, MRI-alone, and a range of PSA thresholds and MRI scores. The performance of each pathway was evaluated focusing on the trade-offs between biopsy referral rates and detection of grade group ≥2 cancers. A targeted biopsy was performed only where the PROSTAGRAM MRI showed a lesion score of 3, 4, or 5. RESULTS AND LIMITATIONS: The standard PSA pathway (PSA ≥3 ng/ml + systematic biopsy) would lead to 10% of men being referred for a biopsy and a 1.0% detection rate of grade group ≥2 cancers. Pathways that relied on MRI alone set at a threshold score of 3 for a biopsy led to higher biopsy rates, but with benefit of high cancer detection rates. The pathway that combined an initial low PSA threshold (≥1.0 ng/ml) and MRI score ≥4 accurately identified a high rate of grade group ≥2 cancers (2.5%, 95% confidence interval 1.3-4.6) while recommending fewer patients for a biopsy (7.1%, 95% confidence interval 4.9-10.2). The results are pertinent to only one screening round, the impact of repeat screening rounds is not evaluated, and the required MRI capacity is currently lacking. CONCLUSIONS: Our results highlight the trade-off that exists between reducing excessive numbers of biopsies and maintaining grade group ≥2 cancer detection rates. A pathway that combines PSA ≥1 ng/ml and MRI score ≥4 maintains the detection of grade group ≥2 cancers while recommending fewer men for biopsies and would be the preferred strategy to evaluate in future studies at the first screening round. PATIENT SUMMARY: The IP1-PROSTAGRAM study shows that PROSTAGRAM magnetic resonance imaging in men with a prostate-specific antigen level of ≥1.0 ng/ml could be a promising pathway to evaluate in future screening trials.

X-ray-free protocol for pectus deformities based on magnetic resonance imaging and a low-cost portable three-dimensional scanning device: a preliminary study.

OBJECTIVES: To compare a standard protocol using chest computed tomography (CT) to a non-irradiant protocol involving a low-cost portable 3D scanner and magnetic resonance imaging (MRI) for all pectus deformities based on the Haller index (HI). METHODS: From April 2019 to March 2020, all children treated for pectus excavatum or carinatum at our institution were evaluated by chest CT, 3D scanning (iPad with Structure Sensor and Captevia-Rodin4D) and MRI. The main objectives were to compare the HI determined by CT or MRI to a derived index evaluated with 3D scanning, the external Haller index (EHI). The secondary objectives were to assess the inter-rater variability and the concordance between CT and MRI for the HI and the correction index. RESULTS: Eleven patients were evaluated. We identified a strong correlation between the HI with MRI and the EHI (Pearson correlation coefficient = 0.900; P < 0.001), with a strong concordance between a radiologist and a non-radiologist using intra-class correlation for the HI with MRI (intra-class correlation coefficient = 0.995; [0.983; 0.999]) and the EHI (intra-class correlation coefficient = 0.978; [0.823; 0.995]). We also identified a marked correlation between the HI with CT and the EHI (Pearson coefficient = 0.855; P = 0.002), with a strong inter-rater concordance (intra-class correlation coefficient = 0.975; [0.901; 0.993]), a reliable concordance between CT and MRI for the HI and the correction index (Pearson coefficient = 0.886; P = 0.033). CONCLUSIONS: Non-irradiant pectus deformity assessment is possible in clinical practice, replacing CT with MRI and 3D scanning as a possible readily-accessible monitoring tool.

Diagnostic Accuracy of Single-plane Biparametric and Multiparametric Magnetic Resonance Imaging in Prostate Cancer: A Randomized Noninferiority Trial in Biopsy-naïve Men.

BACKGROUND: Urological guidelines recommend multiparametric magnetic resonance imaging (mpMRI) in men with a suspicion of prostate cancer (PCa). The resulting increase in MRI demand might place health care systems under substantial stress. OBJECTIVE: To determine whether single-plane biparametric MRI (fast MRI) workup could represent an alternative to mpMRI in the detection of clinically significant (cs) PCa. DESIGN, SETTING, AND PARTICIPANTS: Between April 2018 and February 2020, 311 biopsy-naïve men aged ≤75 yr with PSA ≤15 ng/ml and negative digital rectal examination were randomly assigned to 1.5-T fast MRI (n = 213) or mpMRI (n = 98). INTERVENTION: All MRI examinations were classified according to Prostate Imaging-Reporting and Data System (PI-RADS) version 2. Men scored PI-RADS 1-2 underwent 12-core standard biopsy (SBx) and those with PI-RADS 4-5 on fast MRI or PI-RADS 3-5 on mpMRI underwent targeted biopsy in combination with SBx. Equivocal cases on fast MRI (PI-RADS 3) underwent mpMRI and then biopsy according to the findings. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: The primary outcome was to compare the detection rate of csPCa in both study arms, setting a 10% difference for noninferiority. The secondary outcome was to assess the role of prostate-specific antigen density (PSAD) in ruling out men who could avoid biopsy among those with equivocal findings on fast MRI. RESULTS AND LIMITATIONS: The overall MRI detection rate for csPCa was 23.5% (50/213; 95% confidence interval [CI] 18.0-29.8%) with fast MRI and 32.7% (32/98; 95% CI 23.6-42.9%) with mpMRI (difference 9.2%; p = 0.09). The reproducibility of the study could have been affected by its single-center nature. CONCLUSIONS: Fast MRI followed by mpMRI in equivocal cases is not inferior to mpMRI in the detection of csPCa among biopsy-naïve men aged ≤75 yr with PSA ≤15 ng/ml and negative digital rectal examination. These findings could pave the way to broader use of MRI for PCa diagnosis. PATIENT SUMMARY: A faster MRI (magnetic resonance imaging) protocol with no contrast agent and fewer scan sequences for examination of the prostate is not inferior to the typical MRI approach in the detection of clinically significant prostate cancer. If our findings are confirmed in other studies, fast MRI could represent a time-saving and less invasive examination for men with suspicion of prostate cancer. This trial is registered at ClinicalTrials.gov as NCT03693703.

Non-iterative image reconstruction from sparse magnetic resonance imaging radial data without priors.

The state-of-the-art approaches for image reconstruction using under-sampled k-space data are compressed sensing based. They are iterative algorithms that optimize objective functions with spatial and/or temporal constraints. This paper proposes a non-iterative algorithm to estimate the un-measured data and then to reconstruct the image with the efficient filtered backprojection algorithm. The feasibility of the proposed method is demonstrated with a patient magnetic resonance imaging study. The proposed method is also compared with the state-of-the-art iterative compressed-sensing image reconstruction method using the total-variation optimization norm.

Rapid-sequence brain magnetic resonance imaging for Chiari I abnormality.

OBJECTIVE Fast magnetic resonance imaging (fsMRI) sequences are single-shot spin echo images with fast acquisition times that have replaced CT scans for many conditions. Introduced as a means of evaluating children with hydrocephalus and macrocephaly, these sequences reduce the need for anesthesia and can be more cost-effective, especially for children who require multiple surveillance scans. However, the role of fsMRI has yet to be investigated in evaluating the posterior fossa in patients with Chiari I abnormality (CM-I). The goal of this study was to examine the diagnostic performance of fsMRI in evaluating the cerebellar tonsils in comparison to conventional MRI. METHODS The authors performed a retrospective analysis of 18 pediatric patients with a confirmed diagnosis of CM-I based on gold-standard conventional brain MRI and 30 controls without CM-I who had presented with various neurosurgical conditions. The CM-I patients were included if fsMRI studies had been obtained within 1 year of conventional MRI with no surgical intervention between the studies. Two neuroradiologists reviewed the studies in a blinded fashion to determine the diagnostic performance of fsMRI in detecting CM-I. For the CM-I cohort, the fsMRI and T2-weighted MRI exams were randomized, and the blinded reviewers performed tonsillar measurements on both scans. RESULTS The mean age of the CM-I cohort was 7.39 years, and 50% of these subjects were male. The mean time interval between fsMRI and conventional T2-weighted MRI was 97.8 days. Forty-four percent of the subjects had undergone imaging after posterior fossa decompression. The sensitivity and specificity of fsMRI in detecting CM-I was 100% (95% CI 71.51%-100%) and 92.11% (95% CI 78.62%-98.34%), respectively. If only preoperative patients are considered, both sensitivity and specificity increase to 100%. The authors also performed a cost analysis and determined that fsMRI was significantly cost-effective compared to T2-weighted MRI or CT. CONCLUSIONS Despite known limitations, fsMRI may serve as a useful diagnostic and surveillance tool for CM-I. It is more cost-effective than full conventional brain MRI and decreases the need for sedation in young children.

T1-weighted dual-echo MRI for fat quantification in pediatric nonalcoholic fatty liver disease.

AIM: To determine in obese children with nonalcoholic fatty liver disease (NAFLD) the accuracy of magnetic resonance imaging (MRI) in assessing liver fat concentration. METHODS: A case-control study was performed. Cases were 25 obese children with biopsy-proven NAFLD. Controls were 25 obese children matched for age and gender, without NAFLD at ultrasonography and with normal levels of aminotransferases and insulin. Hepatic fat fraction (HFF) by MRI was obtained using a modification of the Dixon method. RESULTS: HFF ranged from 2% to 44% [mean, 19.0% (95% CI, 15.1-27.4)] in children with NAFLD, while in the controls this value ranged from 0.08% to 4.69% [2.0% (1.3-2.5), P < 0.0001]. HFF was highly correlated with histological steatosis (r = 0.883, P < 0.0001) in the NAFLD children. According to the histological grade of steatosis, the mean HFF was 8.7% (95% CI, 6.0-11.6) for mild, 21.6% (15.3-27.0) for moderate, and 39.7% (34.4-45.0) for severe fatty liver infiltration. With a cutoff of 4.85%, HFF had a sensitivity of 95.8% for the diagnosis of histological steatosis ≥ 5%. All control children had HFF lower than 4.85%; thus, the specificity was 100%. After 12 mo, children with weight loss displayed a significant decrease in HFF. CONCLUSION: MRI is an accurate methodology for liver fat quantification in pediatric NAFLD.