Magnetic Resonance Imaging Interpretation Research Articles

Anatomic Study of the Elbow Joint in a Bengal Tiger (Panthera tigris tigris) Using Magnetic Resonance Imaging and Gross Dissections.

Simple SummaryWe present a report describing the normal elbow joint anatomy in a Bengal tiger (Panthera tigris tigris) using magnetic resonance imaging (MRI) and gross dissections of this region. Anatomical findings detected using MRI of the different bony and soft tissues were evaluated according to the characteristics of signal intensity and compared with the corresponding gross anatomical dissections. The main anatomical structures were labelled and identified. This study provides a valuable resource for veterinarians, biologists, and researchers involved in Bengal tiger (Panthera tigris tigris) conservation.The objective of our research was to describe the normal appearance of the bony and soft tissue structures of the elbow joint in a cadaver of a male mature Bengal tiger (Panthera tigris tigris) scanned via MRI. Using a 0.2 Tesla magnet, Spin-echo (SE) T1-weighting, and Gradient-echo short tau inversion recovery (GE-STIR), T2-weighting pulse sequences were selected to generate sagittal, transverse, and dorsal planes. In addition, gross dissections of the forelimb and its elbow joint were made. On anatomic dissections, all bony, articular, and muscular structures could be identified. The MRI images allowed us to observe the bony and many soft tissues of the tiger elbow joint. The SE T1-weighted MR images provided good anatomic detail of this joint, whereas the GE-STIR T2-weighted MR pulse sequence was best for synovial cavities. Detailed information is provided that may be used as initial anatomic reference for interpretation of MR images of the Bengal tiger (Panthera tigris tigris) elbow joint and in the diagnosis of disorders of this region.

Sympathetic activity contributes to the fMRI signal

The interpretation of functional magnetic resonance imaging (fMRI) studies of brain activity is often hampered by the presence of brain-wide signal variations that may arise from a variety of neuronal and non-neuronal sources. Recent work suggests a contribution from the sympathetic vascular innervation, which may affect the fMRI signal through its putative and poorly understood role in cerebral blood flow (CBF) regulation. By analyzing fMRI and (electro-) physiological signals concurrently acquired during sleep, we found that widespread fMRI signal changes often co-occur with electroencephalography (EEG) K-complexes, signatures of sub-cortical arousal, and episodic drops in finger skin vascular tone; phenomena that have been associated with intermittent sympathetic activity. These findings support the notion that the extrinsic sympathetic innervation of the cerebral vasculature contributes to CBF regulation and the fMRI signal. Accounting for this mechanism could help separate systemic from local signal contributions and improve interpretation of fMRI studies.

Isolated Neurosarcoidosis Mimicking Demyelination and Metastases

Involvement of the central nervous system (CNS) is commonly seen in patients with systemic sarcoidosis; however, an isolated case of neurosarcoidosis is a substantially infrequent scenario. Neurosarcoidosis may present with numerous nonspecific clinical manifestations and radiological features, which makes it difficult to diagnose. Correct interpretation of magnetic resonance imaging (MRI) findings incorporating sarcoidosis as a possible diagnosis is important for timely management of the disease. We report the unusual case of a 47-year-old female with brain lesions and a sudden appearance of spinal lesions mimicking CNS demyelination and metastases. Biopsy provided a definitive diagnosis of neurosarcoidosis.

Errors in Interpretation of Magnetic Resonance Imaging for Thymic Lesions.

The objectives of this study were to evaluate the magnetic resonance imaging (MRI) features of thymic lesions and to review those features that may result in diagnostic uncertainty or error. This multicenter retrospective study included patients with pathologically proven thymic epithelial neoplasms (TEN) and thymic cysts who underwent preoperative MRI. Clinical reports were evaluated for the radiologist's first-choice diagnosis, if not already known before the MRI. MR images were reviewed by thoracic radiologists to identify cystic components on T2-weighted imaging (T2WI) as well as the presence of enhancement on postcontrast series. Through a search of electronic medical records of 3 tertiary academic medical centers, we identified 41 TEN and 13 cysts. Of 35 evaluable clinical reports, 9 (26%) gave an incorrect diagnosis. Of these, 5 cases were misdiagnosed related to contrast enhancement, and 3 cases of cysts were misdiagnosed because of lack of fluid signal on T2WI. Upon rereview of images, of the 41 TENs, 4 (10%) did not show qualitative enhancement on the first postcontrast phase (although all enhanced eventually), and 4 cysts (31%) did not show fluid signal on T2WI. In addition, 4 cases of cysts (31%) showed mural or septal enhancement. Although atypical or misleading imaging features are uncommon in TEN and thymic cysts, they do occur and may lead to diagnostic error. We suggest that lesions with a T2 signal less than simple fluid and without enhancement on postcontrast images undergo follow-up imaging to confirm a benign diagnosis.

Liver MR Elastography Technique and Image Interpretation: Pearls and Pitfalls.

Liver MR elastography is an imaging technique used to measure liver stiffness in the evaluation for possible fibrosis or cirrhosis. Liver stiffness measurement (LSM) is useful for predicting the stage of liver fibrosis. However, obtaining and reporting accurate and reliable LSMs with MR elastography require an understanding of the three core components of liver MR elastography: optimization of imaging technique, prompt quality control of images, and proper interpretation and reporting of elastogram findings. When performing MR elastography, six important technical parameters that should be optimized are patient fasting before the examination, proper passive driver placement, proper MR elastography section positioning over the largest area of the liver, use of MR elastography-related sequences at end expiration, choosing the best timing of the MR elastography sequence, and optimization of several essential pulse sequence parameters. As soon as the MR elastography examination is performed, the elastograms should be reviewed to ensure that they are of diagnostic quality so that corrective steps can be taken, if needed, and MR elastography can be repeated before the diagnostic portion of the examination concludes. Finally, the interpreting radiologist needs to understand and be able to perform the proper technique for LSMs, including determining which areas of the liver to include or avoid in the measurements; knowing which conditions, other than fibrosis or cirrhosis, can increase liver stiffness; and understanding how to report elastography results. This article reviews the proper technique for performing liver MR elastography and subsequent quality control assessment, as well as the principles for interpreting and reporting studies. This review may be helpful for implementing and operating a clinical liver MR elastography service.Online supplemental material is available for this article.©RSNA, 2019.

MMP-2 AND NGAL AS BIOMARKERS IN GLIOBLASTOMA: A PILOT STUDY

Pseudoprogression and radiation necrosis prevent confident interpretation of magnetic resonance imaging (MRI) results and lead to challenges in the management of glioblastoma (GBM). We examined the utility of matrix metalloproteinase-2 (MMP-2) and neutrophil gelatinase-associated lipocalin (NGAL) in the serum and urine as biomarkers of tumor burden in patients with GBM, with the goal of improving the interpretation of MRI and predicting overall survival (OS) and progression free survival (PFS). Expression of MMP-2 and NGAL was analyzed by immunohistochemistry in GBM and non-tumor epileptic control tissues. Serum and urine samples were collected pre and postoperatively and at each MRI. MMP-2 and NGAL levels in the serum and urine were measured by enzyme-linked immunosorbent assay. MRI results, tumor volumes, survival, symptoms, and quality of life were assessed using repeated measures and autoregressive models correlation structure. Although the staining intensity of NGAL was indistinguishable, GBM tissues exhibited significantly higher number of NGAL positive cells as compared to control epileptic brain tissue. Serum MMP-2 was significantly higher in GBM patients than control subjects (p=0.0112) and was highest in patients who underwent biopsy compared to maximal resection (p=0.0038). Elevated levels of both NGAL and MMP-2 in the preoperative samples indicated a trend towards shorter PFS and OS, while preoperative urine NGAL levels were marginally predictive of PFS. Trajectory of symptoms and quality of life results were too variable to adequately correlate with the biomarker levels. Although biomarkers did not aid in differentiating between pseudoprogression, radiation necrosis, and tumor growth, preoperative levels correlated with survival.

Nonoperative Treatment of Elbow Ulnar Collateral Ligament Injuries With and Without Platelet-Rich Plasma in Professional Baseball Players: A Comparative and Matched Cohort Analysis

Background: Recent studies evaluating nonoperative treatment of elbow ulnar collateral ligament (UCL) injuries augmented with platelet-rich plasma (PRP) have shown promising results. To date, no comparative studies have been performed on professional baseball players who have undergone nonoperative treatment with or without PRP injections for UCL injuries. Hypothesis: Players who received PRP injections would have better outcomes than those who did not receive PRP. Study Design: Cohort study; Level of evidence, 3. Methods: The Major League Baseball (MLB) Health and Injury Tracking System identified 544 professional baseball players who were treated nonoperatively for elbow UCL injuries between 2011 and 2015. Of these, 133 received PRP injections (PRP group) before starting their nonoperative treatment program, and 411 did not (no-PRP group). Player outcomes and a Kaplan-Meier survival analysis were compared between groups. In addition, to reduce selection bias, a 1:1 matched comparison of the PRP group versus the no-PRP group was performed. Players were matched by age, position, throwing side, and league status: major (MLB) and minor (Minor League Baseball [MiLB]). A single radiologist with extensive experience in magnetic resonance imaging (MRI) interpretation of elbow injuries in elite athletes analyzed 243 MRI scans for which images were accessible for tear location and grade interpretation. Results: Nonoperative treatment of UCL injuries resulted in an overall 54% rate of return to play (RTP). Players who received PRP had a significantly longer delay in return to throwing (P < .001) and RTP (P = .012). The matched cohort analysis showed that MLB and MiLB pitchers in the no-PRP group had a significantly faster return to throwing (P < .05) and the MiLB pitchers in the no-PRP group had a significantly faster RTP (P = .045). The survival analysis did not reveal significant differences between groups over time. The use of PRP, MRI grade, and tear location were not statistically significant predictors for RTP or progression to surgery. Conclusion: In this retrospective matched comparison of MLB and MiLB pitchers and position players treated nonoperatively for a UCL tear, PRP did not improve RTP outcomes or ligament survivorship, although there was variability with respect to PRP preparations, injection protocols, time from injury to injection, and rehabilitation programs. MRI grade and tear location also did not significantly affect RTP outcomes or progression to surgery.

Performance of On-Call Radiology Residents in Interpreting Total Spine MRI Studies for the Detection of Spinal Cord Compression or Cauda Equina Compression.

OBJECTIVE. Accurate diagnosis of spinal cord compression (SCC) or cauda equina compression (CEC) is important in the emergency setting so management decisions may be made promptly. The purpose of this study is to evaluate the performance of on-call radiology residents in interpreting total spine MRI studies for the detection of SCC or CEC. MATERIALS AND METHODS. On-call radiology residents' preliminary interpretation of total spine MRI studies performed over a period of two and half years were evaluated. Agreements and disagreements were determined for SCC (defined as severe spinal stenosis with compression of the spinal cord and lack of surrounding CSF), CEC (defined as > 75% narrowing of the lumbar canal and lack of CSF in the thecal sac), and the presence of other significant findings that might impact clinical management. Studies with true-positive, true-negative, false-positive, and false-negative findings were identified. RESULTS. SCC or CEC was present in 72 of the 295 studies with preliminary interpretations done by on-call residents. The sensitivity and specificity for detecting SCC or CEC were 93.1% and 98.2%, respectively. Sensitivity was higher for 3rd- and 4th-year residents (97.7%) than for 2nd-year residents (86.2%). Other significant findings were present in 34 of the 295 studies. The overall sensitivity and specificity for the detection of other significant findings were 85.3% and 100.0%, respectively. The level of training did not affect the detection of other significant findings. CONCLUSION. The overall sensitivity of on-call radiology residents' interpretation of total spine MR images for the detection of SCC or CEC was high, improving with their level of training. However, residents' sensitivity was slightly less for the detection of other significant findings. Resident performance can be further improved with focused training and the use of a preliminary interpretation template.

The accuracy of magnetic resonance imaging diagnosis of non-osseous knee injury at Steve Biko Academic Hospital.

BackgroundPreoperative magnetic resonance imaging (MRI) has internationally been proven to reduce unnecessary knee arthroscopies and assist with surgical planning. This has the advantage of avoiding unnecessary surgery and the associated anaesthetic risk, as well as reducing costs. No data were found in the recently published literature assessing the accuracy of MRI interpretation of knee ligament injury in the public sector locally.ObjectivesThis pilot study aimed to determine the accuracy of MRI in detecting non-osseous knee injury in a resource-limited tertiary-level academic hospital in Pretoria, South Africa, compared to the gold standard arthroscopy findings.MethodThis was an exploratory retrospective analysis of 39 patients who had MRI and arthroscopy at Steve Biko Academic Hospital (SBAH). True positive, true negative, false positive and false negative results were extrapolated from findings in both modalities and translated into sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for each structure.ResultsNegative predictive values were recorded as 97%, 81%, 90% and 100% (anterior cruciate ligament [ACL], medial meniscus [MM], lateral meniscus [LM] and posterior cruciate ligament [PCL], respectively), which were comparative to recently published international literature. The PPV results were lower than those previously evaluated at 55%, 58%, 55% and not applicable. The sensitivities and specificities of the ligaments were 83%, 58%, 83% and not applicable; and 87%, 81%, 70% and not applicable, respectively.ConclusionMagnetic resonance imaging was found to be sensitive and specific, with a high NPV noted in all structures evaluated. Negative results can therefore be used to avoid unnecessary surgery to the benefit of the patient and state. The study reiterates that high accuracy can be obtained from MRI on a 1.5-tesla non-dedicated scanner, with interpretation by generalist radiologists.

Interobserver Reliability When Classifying MR Imaging of the Lumbar Spine: Written Instructions Alone Do Not Suffice.

Purpose:Numerous classification systems have been proposed to analyze lumbar spine MRI scans. When evaluating these systems, most studies draw their conclusions from measurements of experienced clinicians. The aim of this study was to evaluate the impact of specific measurement training on interobserver reliability in MRI classification of the lumbar spine.Methods:Various measurement and classification systems were assessed for their interobserver reliability in 30 MRIs from patients with chronic lumbar back and sciatic pain. Two observers were experienced spine surgeons. The third observer was an inexperienced medical student who, prior to the study measurements, in addition to being given the detailed written instructions also given to the surgeons, obtained a list of 20 reference measurements in MRI scans from other patients to practice with.Results:Excellent agreement was observed between the medical student and the spine surgeon who had also created the reference measurements. Between the two spine surgeons, agreement was markedly lower in all systems investigated (e.g., antero-posterior spinal canal diameter intraclass correlation coefficient [ICC] [3.1] = 0.979 vs. ICC [3.1] = 0.857).Conclusion:These data warrant the creation of publicly available standardised measurement examples of accepted classification systems to increase reliability of the interpretation of MR images.

Urology Residents’ Experience With Simulation: Initial Evaluation of MRI/US Fusion Biopsy Workshop

To evaluate our initial experience with a multi-institutional workshop model of MRI/US fusion biopsy simulation for resident education. Residents from 6 Chicago area urology programs participated in a MRI/US fusion biopsy workshop, which incorporated a 30 minute didactic session followed by hands-on simulation. The workshop was facilitated by fellowship-trained university faculty members and company representatives of MRI/US fusion technologies who provided teaching assistance and verbal feedback. Participants completed pre- and post-test nonvalidated 4-item questionnaires graded on a Likert scale. Information on resident prior experience with TRUS and MRI/US fusion biopsies was also collected. Pre-and postquestionnaires were compared with paired t tests for each survey domain (P <.05 were considered significant). Thirty-three residents (PGY 1-6, median PGY 3) participated in the workshop. 13 (40.6%) residents reported performing between 51 and 100 TRUS biopsies previously. Twenty-one (65.6%) reported being familiar with PIRADS v2 interpretation of prostate MR imaging, however 17 (53.1%) had never previously performed a MRI/US fusion biopsy. Analysis of pre- and post-test questionnaires showed significant increases in all 4 survey domains (P <.05). Residents demonstrated increased familiarity with indications for fusion biopsy (mean difference = +0.59), preparation for fusion biopsy (mean difference = +1.16), methods of MRI to TRUS image registration (mean difference = +1.38), and the advantages/disadvantages of perineal vs TRUS fusion biopsy (mean difference = +1.25). This workshop model which combines didactics followed by hands-on simulation training is an effective method for increasing the knowledge and familiarity with MRI/US fusion biopsy of trainees.

Differential Clinical Presentation, Intraoperative Management Strategies, and Surgical Outcomes After Endoscopic Endonasal Treatment of Cystic Sellar Masses

Cystic sellar masses (CSMs) pose diagnostic and therapeutic challenges associated with subtotal cyst wall resection, cerebrospinal fluid (CSF) leak repair, and disease recurrence. Current magnetic resonance imaging (MRI) interpretation often cannot reliably differentiate CSMs, mandating adaptable intraoperative strategies. We reviewed our diagnostic and therapeutic experience after endoscopic endonasal approaches (EEAs) for CSMs. A retrospective record review of patients with CSM managed via EEA at the University of Southern California from 2011 to 2018 was conducted. Patient demographics, preoperative characteristics, surgical details, pathologic findings, and postoperative outcomes were assessed. Analysis included 47 patients (mean age, 43.2 years); of these, 78.7% were women. Preoperative symptoms included headache (76.6%) and vision loss (42.6%). Histologically verified sellar pathology included 27 Rathke cleft cysts (RCCs) (57.4%), 17 cystic pituitary adenomas (CPAs) (36.2%), 2 arachnoid cysts (4.3%), and 1 xanthogranuloma (2.1%). Twelve patients (70.6%) with CPAs underwent complete resection and 5 (29.4%) underwent subtotal resection. All 27 patients with RCC and 2 patients with arachnoid cyst underwent complete fenestration and drainage. One xanthogranuloma was completely resected. There were 14 intraoperative (29.8%) and 4 postoperative CSF leaks (8.5%). Headaches, vision, and endocrinopathy improved in 69.2%, 80.0%, and 33.3% of patients with CPA and 73.9%, 71.4%, and 40.9% of patients with RCC, respectively. There were 2 RCC recurrences and 1 CPA recurrence over the follow-up period. Surgeons must prepare for versatile management strategies of CSMs based on pretest probability associated with MRI and intraoperative findings. Outcomes after EEA for CSMs show low complication profiles and excellent rates of headache and visual improvement, albeit lower rates of endocrine normalization.

BI-RADS, C-RADS, CAD-RADS, LI-RADS, Lung-RADS, NI-RADS, O-RADS, PI-RADS, TI-RADS: Reporting and Data Systems.

In this online presentation, nine American College of Radiology Reporting and Data Systems, from BI-RADS to TI-RADS, are individually reviewed, and the publication trends for each and expert resour...

Improving diagnosis accuracy of brain volume abnormalities during childhood with an automated MP2RAGE-based MRI brain segmentation

Background and purposeIt can be challenging to depict brain volume abnormalities in the pediatric population on magnetic resonance imaging (MRI). The aim of the study was to evaluate the inter-radiologist reliability in brain MRI interpretation, including brain volume assessment and the efficiency of an automated brain segmentation. Materials and methodsWe performed a single-center prospective study including 44 patients aged six months to five years recruited from the University Hospital, having a 1.5T brain MRI using a MP2RAGE sequence. All MRI were randomly and blindly reviewed by one junior and two senior pediatric radiologists. Inter-observer agreements were assessed using Fleiss’ kappa coefficient. Brain volumetry (total intracranial volume (TIV), brain parenchyma, and cerebrospinal fluid volumes) was estimated using the MorphoBox prototype. Clinical head circumference (HC) and z scores were reported. A Pearson correlation coefficient was calculated between brain volumes with HC. ResultsTwenty-four brain MRI examinations were normal and twenty were pathological. Brain volume abnormalities were poorly detected by junior and senior radiologists: sensitivities 16.67% [confidence interval 4.7–44.8], 33.33% [13–60] and 30.7% [12–58] and specificities 93.75% [79–98], 84.38% [68–93] and 77% [60–88], respectively. Brain volume apart, interobserver kappa coefficients were 0.93 between junior and seniors as well as between seniors. Brain volumes were significantly correlated with HC (P<0.0001). In patients with normal MRI, brain parenchyma volumes increased regularly with age. Low brain volume was easier to identify with automated quantification. ConclusionBrain volume was poorly appreciated by radiologists. The fully automated brain segmentation used can provide quantitative data to better diagnose, describe, and follow-up brain volume abnormalities.

Detection of Local Recurrence in Patients with Head and Neck Squamous Cell Carcinoma Using Voxel-Based Color Maps of Initial and Final Area under the Curve Values Derived from DCE-MRI.

Early detection of local recurrence is important to increase the chance of cure because local recurrence is the main cause of treatment failure in head and neck squamous cell carcinoma. We evaluated the added value of voxel-based color maps of dynamic contrast-enhanced MR imaging compared with conventional MR imaging alone for detecting local recurrence of head and neck squamous cell carcinoma. We retrospectively enrolled 63 consecutive patients with head and neck squamous cell carcinoma after definitive treatment and posttreatment surveillance MR imaging studies that demonstrated focal enhancement at the primary site. Three independent readers assessed conventional MR imaging and a pair of color maps of initial and final 90-second time-signal intensity areas under the curve from dynamic contrast-enhanced MR imaging. The sensitivities, specificities, and accuracies of both conventional MR imaging alone and combined interpretation of conventional and dynamic contrast-enhanced MR imaging were assessed using the clinicopathologic diagnosis as the criterion standard. κ statistics were used to evaluate interreader agreement. There were 28 patients with subsequently documented local recurrence and 35 with posttreatment change. Adding dynamic contrast-enhanced MR imaging to conventional MR imaging significantly increased the diagnostic accuracies for detecting local recurrence (48%-54% versus 87%-91%; P < .05), with excellent interreader agreement (κ = 0.8; 95% CI, 0.67-0.92 to κ = 0.81; 95% CI, 0.69-0.93). By all 3 readers, the specificities were also significantly improved by adding dynamic contrast-enhanced MR imaging to conventional MR imaging (22%-43% versus 87%-91%; P < .001) without sacrificing the sensitivities (68%-82% versus 86%-89%; P > .05). Adding voxel-based color maps of initial and final 90-second time-signal intensity areas under the curve from dynamic contrast-enhanced MR imaging to conventional MR imaging increases the diagnostic accuracy to detect local recurrence in head and neck squamous cell carcinoma by improving the specificity without sacrificing the sensitivity.

A new method for diagnosis of anterior cruciate ligament tear: MRI with maximum flexion of knee in the prone position: A case control study.

The diagnosis of anterior cruciate ligament tear can be made by physical examination and magnetic resonance imaging (MRI) in the supine position. In cases where the tear is partially evaluated on MRI, the choice of treatment may vary. The purpose of the study was to investigate the efficiency of MRI at maximum knee flexion in the prone position and to compare the images with findings of the ACL detected during surgery. Sixty-one patients with partial ACL tears with meniscal and cartilage lesions requiring arthroscopic knee surgery were included in the study between 2017 and 2019. MRI of these patients was prescribed at maximum knee flexion in the prone position. Then, an arthroscopic operation was performed on 61 patients and the findings (intact, partial or total tear of ACL) were recorded. The ACL was evaluated as being intact and partial or total tear. The statistical significance of the efficacy of MRI in the supine position with the knee at maximum flexion in the prone position was compared. It was found that, of 61 patients with suspected partial ACL tears, 25 patients had intact ACLs, 22 patients had partial tears and 14 patients had total ACL tears, through the interpretation of MRIs of the prone position by the radiologist. In the arthroscopic surgery of 61 patients, 20 patients had intact ACLs, 27 patients had a partial tear and 14 patients had a total tear. The MRI results with maximum knee flexion in the prone position were more compatible with the findings of the arthroscopic surgery. It could be considered that MRI with maximum knee flexion in the prone position may also be guiding in the diagnosis and treatment of patients with partial anterior cruciate ligament rupture.

Is There a Role for MRI in the Management of Placenta Accreta Spectrum?

The diagnosis of placenta accreta spectrum (PAS) is based on evaluation of risk factors and imaging findings. Ultrasound (US) is widely available and relatively cheap; magnetic resonance imaging (MRI) is expensive and not universally available. We review performance, interpretation, accuracy, and pitfalls of MRI, and whether it has a role in the management of PAS. Recent meta-analyses showed excellent diagnostic accuracy in depth of placental invasion. Specific MRI features (intraplacental T2 dark bands, myometrial disruption, uterine bulge, and uteroplacental or parametrial hypervascularity) were shown to have significant association with clinical outcomes. The presence of at least three MRI signs was associated with complicated delivery; the presence of six or more MRI signs was associated with massive bleeding, hysterectomy, or extensive bladder repair. MRI is performed on patients with high suspicion for PAS based on risk factors and abnormal US findings. Thus, the pretest probability is very high, artificially inflating sensitivity and specificity. The FIGO consensus recommendation states that MRI is not essential for diagnosis of PAS. When selectively performed, and interpreted by experts, it may be useful in evaluation of areas difficult to see on US and in determining the full extent of placenta percreta. Imaging findings should be reviewed together with clinical information, an attempt made to reconcile any differences in interpretation, and if findings remain discordant, management should err on the side of the most severe assessment of this highly morbid condition which can lead to maternal mortality if unrecognized.

Prediction of Dyslexia from Eye Movements Using Machine Learning

Dyslexia is a reading disability and a language disorder where the individual exhibits difficulty in reading, writing, speaking, and trouble in spelling words. Early prediction of dyslexia can help dyslexics to get early support or intervention through remedial teaching. There is no remarkable computational model for the prediction of dyslexia in the literature. Existing methods to diagnose dyslexia include oral and written assessments, analysis and interpretation of Magnetic Resonance Imaging (MRI), functional MRI (fMRI), and Electroencephalogram (EEG). These methods require every instance to be interpreted by the domain expert in all stages whereas rigorously trained and tested computational models need subject expert intervention only at the end. In this paper, a prediction model has been proposed that uses statistical methods to differentiate dyslexics from non-dyslexics using their eye movement. The eye movements are tracked with an eye tracker. Eye movement has many features like fixations, saccades, transients, and distortions. From the raw data of eye tracker, high-level features are extracted using Principal Component Analysis. This paper proposes a Particle Swarm Optimization (PSO)-based Hybrid Kernel SVM-PSO for the prediction of dyslexia in individuals. The proposed model gives better predictive accuracy of 95% compared to a Linear SVM model. The proposed model is validated on 187 subjects by tracking their eye movements while reading. It is observed that eye movement data along with machine learning can be used for building models of high predictive accuracy. The proposed model can be used as a screening tool for the diagnosis of dyslexia in schools.

Magnetic Resonance Assessment of Peritoneal Carcinomatosis: Is There a True Benefit From Diffusion-Weighted Imaging?

PurposeTo assess the added value of diffusion weighted imaging (DWI) with intermediate (500 s/mm2) and high (1000 s/mm2) b values when combined to conventional contrast-enhanced magnetic resonance imaging (MRI) in identifying peritoneal neoplastic involvement. MethodsTwenty-four patients with peritoneal carcinomatosis from gastrointestinal or gynecological tumors were retrospectively evaluated. All patients underwent peritonectomy with hyperthermic intraoperative chemotherapy and 1.5 T MRI including DWI with 500 s/mm2 and 1000 s/mm2 b values within 1 month from surgery. Images were independently reviewed by 2 radiologists with different experience in abdominal MRI in 3 separate reading sessions, the first including conventional MR images alone (T2-weighted, T1-weighted pre- and post gadolinium injection), the second conventional MRI and DWI with a b value of 500 s/mm2 (b 500-DWI), and the third conventional MRI and DWI with a b value of 1000 s/mm2 (b 1000-DWI). Apparent diffusion coefficient maps were included in the DWI analyses. Peritoneal dissemination was assessed in 9 anatomical sites, including right and left subphrenic space, paracolic gutters, small bowel mesentery, greater omentum, gastric-bowel serosa, free peritoneal surfaces, rectosigmoid-colon mesentery, and pelvis. The presence or absence of peritoneal dissemination for each patient and for each site was scored using a 5-point confidence scale. Sensitivity, specificity, and area under the curve (AUC) for identifying per-site peritoneal implants were calculated for each reader at each reading session. Interobserver agreement was evaluated using kappa statistics. ResultsFor both readers, the sensitivity and AUC values resulting from combined interpretation of conventional MRI and DWI (both b500-DWI and b1000-DWI) were significantly higher than those of conventional MRI alone (P < 0.001). The added value of DWI was greater for the less experienced reader (sensitivity 0.55, specificity 0.73, AUC 0.64 on conventional MRI; sensitivity 0.75, specificity 0.72, AUC 0.74 on b500-DWI; sensitivity 0.87, specificity 0.72, AUC 0.80 on b1000-DWI) than for the more experienced reader (sensitivity 0.63, specificity 0.75, AUC 0.70 on conventional MRI; sensitivity 0.76, specificity 0.77, AUC 0.77 on b500-DWI; sensitivity 0.85, specificity 0.72, AUC 0.79 on b1000-DWI), although the differences between the 2 observers were not statistically significant. Interobserver agreement resulted to be fair (κ = 0.30) when dealing with conventional MRI alone. The addition of b500-DWI and b1000-DWI to conventional MRI allowed to reach a substantial agreement (κ = 0.75). ConclusionsThe combined interpretation of high b value DWI and conventional MRI provides increased sensitivity and diagnostic performance in detection of peritoneal carcinomatosis in oncologic patients.

Artificial Intelligence-Based Classification of Breast Lesions Imaged With a Multiparametric Breast MRI Protocol With Ultrafast DCE-MRI, T2, and DWI.

We investigated artificial intelligence (AI)-based classification of benign and malignant breast lesions imaged with a multiparametric breast magnetic resonance imaging (MRI) protocol with ultrafast dynamic contrast-enhanced MRI, T2-weighted, and diffusion-weighted imaging with apparent diffusion coefficient mapping. We analyzed 576 lesions imaged with MRI, including a consecutive set of biopsied malignant (368) and benign (149) lesions, and an additional set of 59 benign lesions proven by follow-up. We used deep learning methods to interpret ultrafast dynamic contrast-enhanced MRI and T2-weighted information. A random forests classifier combined the output with patient information (PI; age and BRCA status) and apparent diffusion coefficient values obtained from diffusion-weighted imaging to perform the final lesion classification. We used receiver operating characteristic (ROC) analysis to evaluate our results. Sensitivity and specificity were compared with the results of the prospective clinical evaluation by radiologists. The area under the ROC curve was 0.811 when only ultrafast dynamics was used. The final AI system that combined all imaging information with PI resulted in an area under the ROC curve of 0.852, significantly higher than the ultrafast dynamics alone (P = 0.002). When operating at the same sensitivity level of radiologists in this dataset, this system produced 19 less false-positives than the number of biopsied benign lesions in our dataset. Use of adjunct imaging and PI has a significant contribution in diagnostic performance of ultrafast breast MRI. The developed AI system for interpretation of multiparametric ultrafast breast MRI may improve specificity.