Comparative performance of fully-automated and semi-automated artificial intelligence methods for the detection of clinically significant prostate cancer on MRI: a systematic review.

OBJECTIVES: We systematically reviewed the current literature evaluating the ability of fully-automated deep learning (DL) and semi-automated traditional machine learning (TML) MRI-based artificial intelligence (AI) methods to differentiate clinically significant prostate cancer (csPCa) from indolent PCa (iPCa) and benign conditions. METHODS: We performed a computerised bibliographic search of studies indexed in MEDLINE/PubMed, arXiv, medRxiv, and bioRxiv between 1 January 2016 and 31 July 2021. Two reviewers performed the title/abstract and full-text screening. The remaining papers were screened by four reviewers using the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) for DL studies and Radiomics Quality Score (RQS) for TML studies. Papers that fulfilled the pre-defined screening requirements underwent full CLAIM/RQS evaluation alongside the risk of bias assessment using QUADAS-2, both conducted by the same four reviewers. Standard measures of discrimination were extracted for the developed predictive models. RESULTS: 17/28 papers (five DL and twelve TML) passed the quality screening and were subject to a full CLAIM/RQS/QUADAS-2 assessment, which revealed a substantial study heterogeneity that precluded us from performing quantitative analysis as part of this review. The mean RQS of TML papers was 11/36, and a total of five papers had a high risk of bias. AUCs of DL and TML papers with low risk of bias ranged between 0.80-0.89 and 0.75-0.88, respectively. CONCLUSION: We observed comparable performance of the two classes of AI methods and identified a number of common methodological limitations and biases that future studies will need to address to ensure the generalisability of the developed models.

Structural Brain Network Abnormalities and the Probability of Seizure Recurrence After Epilepsy Surgery.

OBJECTIVE: We assessed preoperative structural brain networks and clinical characteristics of patients with drug-resistant temporal lobe epilepsy (TLE) to identify correlates of postsurgical seizure recurrences. METHODS: We examined data from 51 patients with TLE who underwent anterior temporal lobe resection (ATLR) and 29 healthy controls. For each patient, using the preoperative structural, diffusion, and postoperative structural MRI, we generated 2 networks: presurgery network and surgically spared network. Standardizing these networks with respect to controls, we determined the number of abnormal nodes before surgery and expected to be spared by surgery. We incorporated these 2 abnormality measures and 13 commonly acquired clinical data from each patient into a robust machine learning framework to estimate patient-specific chances of seizures persisting after surgery. RESULTS: Patients with more abnormal nodes had a lower chance of complete seizure freedom at 1 year and, even if seizure-free at 1 year, were more likely to relapse within 5 years. The number of abnormal nodes was greater and their locations more widespread in the surgically spared networks of patients with poor outcome than in patients with good outcome. We achieved an area under the curve of 0.84 ± 0.06 and specificity of 0.89 ± 0.09 in predicting unsuccessful seizure outcomes (International League Against Epilepsy [ILAE] 3-5) as opposed to complete seizure freedom (ILAE 1) at 1 year. Moreover, the model-predicted likelihood of seizure relapse was significantly correlated with the grade of surgical outcome at year 1 and associated with relapses up to 5 years after surgery. CONCLUSION: Node abnormality offers a personalized, noninvasive marker that can be combined with clinical data to better estimate the chances of seizure freedom at 1 year and subsequent relapse up to 5 years after ATLR. CLASSIFICATION OF EVIDENCE: This study provides Class II evidence that node abnormality predicts postsurgical seizure recurrence.

Investigating Brain Network Changes and Their Association With Cognitive Recovery After Traumatic Brain Injury: A Longitudinal Analysis.

Traumatic brain injury (TBI) can result in acute cognitive deficits and diffuse axonal injury reflected in white matter brain network alterations, which may, or may not, later recover. Our objective is to first characterize the ways in which brain networks change after TBI and, second, investigate if those changes are associated with recovery of cognitive deficits. We aim to make initial progress in discerning the relationships between brain network changes, and their (dys)functional correlates. We analyze longitudinally acquired MRI from 23 TBI patients (two time points: 6 days, 12 months post-injury) and cross-sectional data from 28 controls to construct white matter brain networks. Cognitive assessment was also performed. Graph theory and regression analysis were applied to identify changed brain network metrics after injury that are associated with subsequent improvements in cognitive function. Sixteen brain network metrics were found to be discriminative of different post-injury phases. Eleven of those explain 90% (adjusted R 2) of the variability observed in cognitive recovery following TBI. Brain network metrics that had a high contribution to the explained variance were found in frontal and temporal cortex, additional to the anterior cingulate cortex. Our preliminary study suggests that network reorganization may be related to recovery of impaired cognitive function in the first year after a TBI.

Early deviation from normal structural connectivity: A novel intrinsic severity score for mild TBI.

OBJECTIVE: Studies of outcome after traumatic brain injury (TBI) are hampered by the lack of robust injury severity measures that can accommodate spatial-anatomical and mechanistic heterogeneity. In this study we introduce a Mahalanobis distance measure (M) as an intrinsic injury severity measure that combines in a single score the many ways a given injured brain's connectivity can vary from that of healthy controls. Our objective is to test the hypotheses that M is superior to univariate measures in (1) discriminating patients and controls and (2) correlating with cognitive assessment. METHODS: Sixty-five participants (34 with mild TBI, 31 controls) underwent diffusion tensor MRI and extensive neuropsychological testing. Structural connectivity was inferred for all participants for 22 major white matter connections. Twenty-two univariate measures (1 per connection) and 1 multivariate measure (M), capturing and summarizing all connectivity change in a single score, were computed. RESULTS: Our multivariate measure (M) was able to better discriminate between patients and controls (area under the curve 0.81) than any individual univariate measure. M significantly correlated with cognitive outcome (Spearman ρ = 0.31; p < 0.05). No univariate measure showed significant correlation after correction for multiple comparisons. CONCLUSIONS: Heterogeneity in the severity and distribution of injuries after TBI has traditionally complicated the understanding of outcomes after TBI. Our approach provides a single, continuous variable that can fully capture individual heterogeneity. M's ability to distinguish even mildly injured patients from controls and its correlation with cognitive assessment suggest utility as an imaging-based marker of intrinsic injury severity.

Automated volumetry of hippocampus is useful to confirm unilateral mesial temporal sclerosis in patients with radiologically positive findings.

Background and purpose We evaluated two methods to identify mesial temporal sclerosis (MTS): visual inspection by experienced epilepsy neuroradiologists based on structural magnetic resonance imaging sequences and automated hippocampal volumetry provided by a processing pipeline based on the FMRIB Software Library. Methods This retrospective study included patients from the epilepsy monitoring unit database of our institution. All patients underwent brain magnetic resonance imaging in 1.5T and 3T scanners with protocols that included thin coronal T2, T1 and fluid-attenuated inversion recovery and isometric T1 acquisitions. Two neuroradiologists with experience in epilepsy and blinded to clinical data evaluated magnetic resonance images for the diagnosis of MTS. The diagnosis of MTS based on an automated method included the calculation of a volumetric asymmetry index between the two hippocampi of each patient and a threshold value to define the presence of MTS obtained through statistical tests (receiver operating characteristics curve). Hippocampi were segmented for volumetric quantification using the FIRST tool and fslstats from the FMRIB Software Library. Results The final cohort included 19 patients with unilateral MTS (14 left side): 14 women and a mean age of 43.4 ± 10.4 years. Neuroradiologists had a sensitivity of 100% and specificity of 73.3% to detect MTS (gold standard, k = 0.755). Automated hippocampal volumetry had a sensitivity of 84.2% and specificity of 86.7% (k = 0.704). Combined, these methods had a sensitivity of 84.2% and a specificity of 100% (k = 0.825). Conclusions Automated volumetry of the hippocampus could play an important role in temporal lobe epilepsy evaluation, namely on confirmation of unilateral MTS diagnosis in patients with radiological suggestive findings.

Automated volumetry for unilateral hippocampal sclerosis detection in patients with temporal lobe epilepsy.

Hippocampal sclerosis (HS) is the most common cause of temporal lobe epilepsy (TLE) and can be identified in magnetic resonance imaging as hippocampal atrophy and subsequent volume loss. Detecting this kind of abnormalities through simple radiological assessment could be difficult, even for experienced radiologists. For that reason, hippocampal volumetry is generally used to support this kind of diagnosis. Manual volumetry is the traditional approach but it is time consuming and requires the physician to be familiar with neuroimaging software tools. In this paper, we propose an automated method, written as a script that uses FSL-FIRST, to perform hippocampal segmentation and compute an index to quantify hippocampi asymmetry (HAI). We compared the automated detection of HS (left or right) based on the HAI with the agreement of two experts in a group of 19 patients and 15 controls, achieving 84.2% sensitivity, 86.7% specificity and a Cohen's kappa coefficient of 0.704. The proposed method is integrated in the "Advanced Brain Imaging Lab" (ABrIL) cloud neurocomputing platform. The automated procedure is 77% (on average) faster to compute vs. the manual volumetry segmentation performed by an experienced physician.

ABrIL - Advanced Brain Imaging Lab : a cloud based computation environment for cooperative neuroimaging projects.

Neuroscience is an increasingly multidisciplinary and highly cooperative field where neuroimaging plays an important role. Neuroimaging rapid evolution is demanding for a growing number of computing resources and skills that need to be put in place at every lab. Typically each group tries to setup their own servers and workstations to support their neuroimaging needs, having to learn from Operating System management to specific neuroscience software tools details before any results can be obtained from each setup. This setup and learning process is replicated in every lab, even if a strong collaboration among several groups is going on. In this paper we present a new cloud service model - Brain Imaging Application as a Service (BiAaaS) - and one of its implementation - Advanced Brain Imaging Lab (ABrIL) - in the form of an ubiquitous virtual desktop remote infrastructure that offers a set of neuroimaging computational services in an interactive neuroscientist-friendly graphical user interface (GUI). This remote desktop has been used for several multi-institution cooperative projects with different neuroscience objectives that already achieved important results, such as the contribution to a high impact paper published in the January issue of the Neuroimage journal. The ABrIL system has shown its applicability in several neuroscience projects with a relatively low-cost, promoting truly collaborative actions and speeding up project results and their clinical applicability.