Light-weight neural network for intra-voxel structure analysis.

We present a novel neural network-based method for analyzing intra-voxel structures, addressing critical challenges in diffusion-weighted MRI analysis for brain connectivity and development studies. The network architecture, called the Local Neighborhood Neural Network, is designed to use the spatial correlations of neighboring voxels for an enhanced inference while reducing parameter overhead. Our model exploits these relationships to improve the analysis of complex structures and noisy data environments. We adopt a self-supervised approach to address the lack of ground truth data, generating signals of voxel neighborhoods to integrate the training set. This eliminates the need for manual annotations and facilitates training under realistic conditions. Comparative analyses show that our method outperforms the constrained spherical deconvolution (CSD) method in quantitative and qualitative validations. Using phantom images that mimic in vivo data, our approach improves angular error, volume fraction estimation accuracy, and success rate. Furthermore, a qualitative comparison of the results in actual data shows a better spatial consistency of the proposed method in areas of real brain images. This approach demonstrates enhanced intra-voxel structure analysis capabilities and holds promise for broader application in various imaging scenarios.

Whole-body diffusion-weighted magnetic resonance imaging (WB-DW-MRI) vs choline-positron emission tomography-computed tomography (choline-PET/CT) for selecting treatments in recurrent prostate cancer.

OBJECTIVE: To determine the effectiveness of whole-body diffusion-weighted magnetic resonance imaging (WB-DW-MRI) in detecting metastases by comparing the results with those from choline-positron emission tomography-computed tomography (choline-PET/CT) in patients with biochemical relapse after primary treatment, and no metastases in bone scintigraphy, CT and/or pelvic MRI, or metastatic/oligometastatic prostate cancer (PCa). Patients with this disease profile who could benefit from treatment with stereotactic body radiation therapy (SBRT) were selected and their responses to these techniques were rated. MATERIALS AND METHODS: This was a prospective, controlled, unicentric study, involving 46 consecutive patients from our centre who presented biochemical relapse after adjuvant, salvage or radical treatment with external beam radiotherapy, or brachytherapy. After initial tests (bone scintigraphy, CT, pelvic MRI), 35 patients with oligometastases or without them were selected. 11 patients with multiple metastases were excluded from the study. WB-DW-MRI and choline-PET/CT was then performed on each patient within 1 week. The results were interpreted by specialists in nuclear medicine and MRI. If they were candidates for treatment with ablative SBRT (SABR), they were then evaluated every three months with both tests. RESULTS: Choline-PET/CT detected lesions in 16 patients that were not observable using WB-DW-MRI. The results were consistent in seven patients and in three cases, a lesion was observed using WB-DW-MRI that was not detected with choline-PET/CT. The Kappa value obtained was 0.133 (p = 0.089); the sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of WB-DW-MRI were estimated at 44.93, 64.29, 86.11, and 19.15%, respectively. For choline-PET/CT patients, the sensitivity, specificity, PPV, and NPV were 97.10, 58.33, 93.06, and 77.78%, respectively. CONCLUSIONS: Choline-PET/CT has a high global sensitivity while WB-DW-MRI has a high specificity, and so they are complementary techniques. Future studies with more enrolled patients and a longer follow-up period will be required to confirm these data. The initial data show that the best technique for evaluating response after SBRT is choline-PET/CT. Trial registration number NCT02858128.

Randomized comparison of distal and proximal cerebral protection during carotid artery stenting.

OBJECTIVES: This study sought to randomly compare cerebral protection with ANGIOGUARD (Cordis Corporation, Bridgewater, New Jersey) with Mo.Ma (Invatec/Medtronic Vascular Inc, Santa Rosa, California) during carotid artery stenting (CAS), using diffusion-weighted magnetic resonance imaging (DW-MRI) to detect new ischemic cerebral lesions. The number, size, and location of lesions were analyzed. BACKGROUND: The choice of the type of cerebral protection during CAS is controversial. METHODS: From July 2008 to July 2011, 60 patients undergoing CAS were randomized to ANGIOGUARD or Mo.Ma, distributed by chance, 30 patients for each group. All patients underwent DW-MRI before and after CAS. An independent neuroradiologist blinded to the cerebral protection used analyzed the images. Univariate and multivariate logistic models were fitted to analyze new ischemic lesions. Alternatively, a propensity score approach was used to reduce the bias due to differences between the groups. For the number of lesions, we used Poisson regression models. RESULTS: New ischemic lesions seen on DW-MRI were present in 63.3% of the ANGIOGUARD group versus 66.7% of the Mo.Ma cohort (p = 0.787). The number of ischemic cerebral lesions per patient, when present, was significantly lower in the Mo.Ma group (a median of 6 lesions per patient vs. a median of 10 in the ANGIOGUARD, p < 0.001). Most lesions were small (<0.5 mm) and localized in the ipsilateral territory. One patient in the ANGIOGUARD group had a minor stroke during CAS (1.66%). CONCLUSIONS: New ischemic lesions seen on DW-MRI were present in both groups in >60%, but the number of lesions per patient was greater in the ANGIOGUARD group. No death or disabling stroke occurred during at least 1 year of follow-up in both cohorts.