Diagnosis and Localization of Prostate Cancer via Automated Multiparametric MRI Equipped with Artificial Intelligence

Prostate MRI scans for pre-biopsied patients are important. However, fewer radiologists are available for MRI diagnoses, which requires multi-sequential interpretations of multi-slice images. To reduce such a burden, artificial intelligence (AI)-based, computer-aided diagnosis is expected to be a critical technology. We present an AI-based method for pinpointing prostate cancer location and determining tumor morphology using multiparametric MRI. The study enrolled 15 patients who underwent radical prostatectomy between April 2008 and August 2017 at our institution. We labeled the cancer area on the peripheral zone on MR images, comparing MRI with histopathological mapping of radical prostatectomy specimens. Likelihood maps were drawn, and tumors were divided into morphologically distinct regions using the superpixel method. Likelihood maps consisted of pixels, which utilize the cancer likelihood value computed from the T2-weighted, apparent diffusion coefficient, and diffusion-weighted MRI-based texture features. Cancer location was determined based on the likelihood maps. We evaluated the diagnostic performance by the area under the receiver operating characteristic (ROC) curve according to the Chi-square test. The area under the ROC curve was 0.985. Sensitivity and specificity for our approach were 0.875 and 0.961 (p < 0.01), respectively. Our AI-based procedures were successfully applied to automated prostate cancer localization and shape estimation using multiparametric MRI.

Evaluating discrepancy rates of radiology resident provisional reports for cross-sectional body imaging studies at a tertiary hospital

Introduction Urgent radiological studies obtained during on-call hours are often preliminarily read by on-call residents before consultant radiologists finalise the reports at a later time. Such provisional radiology reports provide important information to guide initial patient management. This study aims to determine discrepancy rates between provisional reports and final interpretations, and to assess the clinical significance of such discrepancies. Methods This retrospective quality assurance project reviewed a total of 1218 cross-sectional imaging studies of the body (thorax, abdomen and pelvis) done between July 2015 and May 2016 during on-call hours. The studies included 1201 Computed tomography (CT) scans and 17 Magnetic Resonance Imaging (MRI) scans. Studies with incomplete or unavailable reports were excluded. Conclusions of both the provisional and final reports of each study were reviewed for concordance, with reference to the full report if needed. Discrepancies were graded according to the ACR 2016 RADPEER scoring system. Results There were 1210 studies with complete reports. Discrepant reports were noted in 183 (15.1%) studies. Of these, 89 (7.3%) were assessed to be clinically significant and the majority of these (55) were due to interpretations which should be made most of the time. CT of the abdomen and pelvis were the most prone to discrepant reports, accounting for 148 cases (80.9%). Conclusion The majority of preliminary reports for on-call body scans were concordant with final interpretations. The discrepancy rates for provisional body scan reports provided by residents while on call were comparable to those previously reported in literature.

Exploring AI-healthcare innovation: natural language processing-based patents analysis for technology-driven roadmapping

PurposeThe purposes of this paper are (1) to explore the overall development of AI technologies and applications that have been demonstrated to be fundamentally important in the healthcare industry, and their related commercialized products and (2) to identify technologies with promise as the basis of useful applications and profitable products in the AI-healthcare domain.Design/methodology/approachThis study adopts a technology-driven technology roadmap approach, combined with natural language processing (NLP)-based patents analysis, to identify promising and potentially profitable existing AI technologies and products in the domain of AI healthcare.FindingsRobotics technology exhibits huge potential in surgical and diagnostics applications. Intuitive Surgical Inc., manufacturer of the Da Vinci robotic system and Ion robotic lung-biopsy system, dominates the robotics-assisted surgical and diagnostic fields. Diagnostics and medical imaging are particularly active fields for the application of AI, not only for analysis of CT and MRI scans, but also for image archiving and communications.Originality/valueThis study is a pioneering attempt to clarify the interrelationships of particular promising technologies for application and related products in the AI-healthcare domain. Its findings provide critical information about the patent activities of key incumbent actors, and thus offer important insights into recent and current technological and product developments in the emergent AI-healthcare sector.

Magnetic Resonance Imaging to Investigate the Clinical Applicability of the Medial Femoral Trochlea Osteochondral Flap Within the Wrist

Background: The medial femoral trochlea (MFT) osteochondral flap is employed for reconstruction of unsalvageable scaphoid proximal pole nonunions. The convex surface of the cartilage-bearing proximal trochlea is used to replace the similarly contoured proximal scaphoid and articulate with the concave scaphoid fossa of the radius. A magnetic resonance imaging (MRI) comparison of the shape of the MFT as it relates to the native proximal scaphoid has not been previously performed. Our study aimed to quantifiably compare the shape of the MFT, proximal scaphoid, and scaphoid fossa. Methods: Using imaging processing software, we measured radius of curvature of the articular segments in MRI scans of 10 healthy subjects’ wrists and knees. Results: Compared with the scaphoid fossa, average ratio of the radius of circumference of the proximal scaphoid was 0.79 and 0.78 in the coronal and sagittal planes, respectively. Compared with the scaphoid fossa, average ratio of the radius of circumference of the MFT was 0.98 and 1.31 in the coronal and sagittal planes, respectively. The radius of curvature of the MFT was larger than the proximal scaphoid, in the coronal and sagittal planes. In the coronal plane, the MFT radius of curvature is nearly identical to the scaphoid fossa, a closer match than the scaphoid itself. In the sagittal plane, the radius of curvature of the MFT was larger than the radius of curvature of the scaphoid fossa. Conclusions: Our data suggest that the radius of curvature, in the sagittal and coronal planes, of the MFT and proximal scaphoid is disparate.

Exploratory Investigation of Brain MRI Lesions According to Whole Sample and Visual Function Subtyping in Children With Cerebral Visual Impairment

Background: There is limited research on brain lesions in children with cerebral visual impairment (CVI) of heterogeneous etiologies and according to associated subtyping and vision dysfunctions. This study was part of a larger project establishing data-driven subtypes of childhood CVI according to visual dysfunctions. Currently there is no consensus in relation to assessment, diagnosis and classification of CVI and more information about brain lesions may be of potential diagnostic value.Aim: This study aimed to investigate overall patterns of brain lesions and associations with level of visual dysfunction and to compare the patterns between the classification subgroups in children with CVI.Methods: School-aged children with CVI received ophthalmological and neuro-psychological/developmental assessments to establish CVI-related subtyping. Other pediatric information was collected from medical records. MRI scans were coded according to a semi-quantitative template including brain regions (right hemisphere, left hemisphere, visual pathways) and summed for total scores. Non-parametric analyses were conducted.Results: 28 children had clinical brain MRI scans available [44% of total sample, Group A (lower severity of visual dysfunctions) n = 16, Group B (higher severity) n = 12]. Total brain scores ranged between 0 and 18 (Group A mdn = 7, IQR = 0.8–10.0, Group B mdn = 10, IQR = 6.5–11.8) and were widespread across regions. 71 per cent had post-geniculate visual pathway damage. The median total brain and hemisphere scores of Group B were higher than subgroup A but differences did not reach statistical significance. No statistically significant associations were found between brain scores and vision variables (acuity, contrast sensitivity).Conclusion: This study found a spread of lesions across all regions on the brain scans in children with congenital CVI. The majority had damage in the postgeniculate visual pathways and visual cortex region suggesting this is an area of interest and potentially informative for diagnosis. However the subtyping classification did not show differences in number or region of lesions though the trend was higher toward Group B. This study confirms the complex diffuse and variable nature of brain lesions in children with congenital CVI, many of whom have other neurological impairments.

Confirming a Historical Diagnosis of Multiple Sclerosis: Challenges and Recommendations

Patients with a historical diagnosis of multiple sclerosis (MS) — a patient presenting with a diagnosis of MS made previously and by a different clinician — present specific diagnostic and therapeutic challenges in clinical practice. Application of the McDonald criteria is most straightforward when applied contemporaneously with a syndrome typical of an MS attack or relapse; however, retrospective application of the criteria in some patients with a historical diagnosis of MS can be problematic. Limited patient recollection of symptoms and evolution of neurological examination and MRI findings complicate confirmation of an earlier MS diagnosis and assessment of subsequent disease activity or clinical progression. Adequate records for review of prior clinical examinations, laboratory results, and/or MRI scans obtained at the time of diagnosis or during ensuing care may be inadequate or unavailable. This paper provides recommendations for a clinical approach to the evaluation of patients with a historical diagnosis of MS to aid diagnostic confirmation, avoid misdiagnosis, and inform therapeutic decision-making.

Deep learning-based pancreas volume assessment in individuals with type 1 diabetes

Pancreas volume is reduced in individuals with diabetes and in autoantibody positive individuals at high risk for developing type 1 diabetes (T1D). Studies investigating pancreas volume are underway to assess pancreas volume in large clinical databases and studies, but manual pancreas annotation is time-consuming and subjective, preventing extension to large studies and databases. This study develops deep learning for automated pancreas volume measurement in individuals with diabetes. A convolutional neural network was trained using manual pancreas annotation on 160 abdominal magnetic resonance imaging (MRI) scans from individuals with T1D, controls, or a combination thereof. Models trained using each cohort were then tested on scans of 25 individuals with T1D. Deep learning and manual segmentations of the pancreas displayed high overlap (Dice coefficient = 0.81) and excellent correlation of pancreas volume measurements (R2 = 0.94). Correlation was highest when training data included individuals both with and without T1D. The pancreas of individuals with T1D can be automatically segmented to measure pancreas volume. This algorithm can be applied to large imaging datasets to quantify the spectrum of human pancreas volume.

Temporoparietal Connectivity Within Default Mode Network Associates With Clinical Improvements in Schizophrenia Following Modified Electroconvulsive Therapy

Although modified electroconvulsive therapy (ECT) has been reported to be effective for the treatment of schizophrenia (SCZ), its action mechanism is unclear. To elucidate the underlying ECT mechanisms of SCZ, this study used a longitudinal cohort including 21 SCZ patients receiving only antipsychotics (DSZ group) and 21 SCZ patients receiving a regular course of ECT combining with antipsychotics (MSZ group) for 4 weeks. All patients underwent magnetic resonance imaging (MRI) scans at baseline (t1) and follow-up (t2) time points. A matched healthy control (HC) group included 23 individuals who were only scanned at baseline. Functional connectivity (FC) within the default mode network (DMN) was evaluated before and after ECT. Significant interaction of the group over time was found in FC between angular gyrus (AG) and middle temporal gyrus (MTG). Post-hoc analysis showed a significantly enhanced FC of left AG(AG.L) and right MTG (MTG.R) in the MSZ group relative to the DSZ group. In addition, the right AG (AG.R) showed significantly enhanced FC between MTG.R and left MTG (MTG.L) after ECT in the MSZ group, but no in the DSZ group. In particular, the FCs change in AG.L-MTG.R and AG.R-MTG.R were positively correlated with the Positive and Negative Syndrome Scale (PANSS) negative score reduction. Furthermore, the FC change in AG.L-MTG.R was also positively correlated with the PANSS general psychopathology score reduction. These findings confirmed a potential relationship between ECT inducing hyperconnectivity within DMN and improvements in symptomatology of SCZ, suggesting that ECT controls mental symptoms by regulating the temporoparietal connectivity within DMN.

Post-Irradiation Sinus Mucosa Disease in Nasopharyngeal Carcinoma Patients Treated with Intensity-Modulated Proton Therapy

In the past decade, patients with nasopharyngeal cancer (NPC) have been deemed candidates for proton radiotherapy, due to the large and comprehensive target volumes and the necessity for the retention of the surrounding healthy tissues. In this study, we aimed to compare the incidence and severity of post-irradiation sinusitis by detecting sinus mucosa diseases (SMDs) via the magnetic resonance imaging (MRI) of patients with NPC after intensity-modulated proton therapy (IMPT) and volume-modulated arc therapy (VMAT). A total of 53 patients in the IMPT group and 54 patients in the VMAT group were enrolled in this study. There were significantly lower endoscopic scores and Lund–Mackay staging scores determined from MRI scans in the IMPT group during different follow-up periods. For the most vulnerable sinuses, the incidence and severity of SMD were the highest during the third post-radiotherapy month in both groups. These decreased steadily, and there was no significant increase in the incidence and severity of SMD during the second post-radiotherapy year in the IMPT group. Our data show that NPC patients with IMPT have a significantly lower incidence and decreased severity of SMD than those with VMAT. A better and faster recovery of sinonasal function after radiotherapy in the IMPT group was also observed.

Hippocampal Volume, Neurite Density and Trace Eye-Blink Conditioning in Young Children

The current study examined the relations between hippocampal structure (e.g., volume and neurite density) and performance on a trace eye blink conditioning (EBC) task in young children. Our first aim assessed whether individual differences in hippocampal volume were associated with trace EBC performance, using both percent Conditioned Responses (% CR) and CR onset latency or the average latency (ms) at which the child started their blink, as measures of hippocampal-dependent associative learning. Our second aim evaluated whether individual differences in hippocampal neurite density were associated with EBC performance using the same outcome measures. Typically developing 4- to 6-year-olds (N = 31; 14 girls; Mage = 5.67; SDage = 0.89) completed T1 and diffusion-weighted MRI scans and a 15-minute trace eyeblink conditioning task outside of the scanner. % CR and CR onset latency were computed across all tone-puff and tone-alone trials. While hippocampal volume was not associated with any of our EBC measures, greater hippocampal neurite density bilaterally, was associated with later CR onset. In other words, children with greater left and right hippocampal neurite density blinked closer to the US (i.e., air puff) than children with less hippocampal neurite density, indicating that structural changes in the hippocampus assisted in the accurate timing of conditioned responses.