Intravoxel incoherent motion and diffusion kurtosis imaging and their machine-learning-based texture analysis for detection and assessment of prostate cancer severity at 3 T.

OBJECTIVES: To evaluate the role of combined intravoxel incoherent motion and diffusion kurtosis imaging (IVIM-DKI) and their machine-learning-based texture analysis for the detection and assessment of severity in prostate cancer (PCa). MATERIALS AND METHODS: Eighty-eight patients underwent MRI on a 3 T scanner after giving informed consent. IVIM-DKI data were acquired using 13 b values (0-2000 s/mm2) and analyzed using the IVIM-DKI model with the total variation (TV) method. PCa patients were categorized into two groups: clinically insignificant prostate cancer (CISPCa) (Gleason grade ≤ 6) and clinically significant prostate cancer (CSPCa) (Gleason grade ≥ 7). One-way analysis-of-variance, t test, and receiver operating characteristic analysis was performed to measure the discriminative ability to detect PCa using IVIM-DKI parameters. A chi-square test was used to select important texture features of apparent diffusion coefficient (ADC) and IVIM-DKI parameters. These selected texture features were used in an artificial neural network for PCa detection. RESULTS: ADC and diffusion coefficient (D) were significantly lower (p < 0.001), and kurtosis (k) was significantly higher (p < 0.001), in PCa as compared with benign prostatic hyperplasia (BPH) and normal peripheral zone (PZ). ADC, D, and k showed high areas under the curves (AUCs) of 0.92, 0.89, and 0.88, respectively, in PCa detection. ADC and D were significantly lower (p < 0.05) as compared with CISPCa versus CSPCa. D for detecting CSPCa was high, with an AUC of 0.63. A negative correlation of ADC and D with GS (ADC, ρ = -0.33; D, ρ = -0.35, p < 0.05) and a positive correlation of k with GS (ρ = 0.22, p < 0.05) were observed. Combined IVIM-DKI texture showed high AUC of 0.83 for classification of PCa, BPH, and normal PZ. CONCLUSION: D, f, and k computed using the IVIM-DKI model with the TV method were able to differentiate PCa from BPH and normal PZ. Texture features of combined IVIM-DKI parameters showed high accuracy and AUC in PCa detection.

Fully automatic deep learning-based lung parenchyma segmentation and boundary correction in thoracic CT scans.

PURPOSE: The proposed work aims to develop an algorithm to precisely segment the lung parenchyma in thoracic CT scans. To achieve this goal, the proposed technique utilized a combination of deep learning and traditional image processing algorithms. The initial step utilized a trained convolutional neural network (CNN) to generate preliminary lung masks, followed by the proposed post-processing algorithm for lung boundary correction. METHODS: First, the proposed method trained an improved 2D U-Net CNN model with Inception-ResNet-v2 as its backbone. The model was trained on 32 CT scans from two different sources: one from the VESSEL12 grand challenge and the other from AIIMS Delhi. Further, the model's performance was evaluated on a test dataset of 16 CT scans with juxta-pleural nodules obtained from AIIMS Delhi and the LUNA16 challenge. The model's performance was assessed using evaluation metrics such as average volumetric dice coefficient (DSCavg), average IoU score (IoUavg), and average F1 score (F1avg). Finally, the proposed post-processing algorithm was implemented to eliminate false positives from the model's prediction and to include juxta-pleural nodules in the final lung masks. RESULTS: The trained model reported a DSCavg of 0.9791 ± 0.008, IoUavg of 0.9624 ± 0.007, and F1avg of 0.9792 ± 0.004 on the test dataset. Applying the post-processing algorithm to the predicted lung masks obtained a DSCavg of 0.9713 ± 0.007, IoUavg of 0.9486 ± 0.007, and F1avg of 0.9701 ± 0.008. The post-processing algorithm successfully included juxta-pleural nodules in the final lung mask. CONCLUSIONS: Using a CNN model, the proposed method for lung parenchyma segmentation produced precise segmentation results. Furthermore, the post-processing algorithm addressed false positives and negatives in the model's predictions. Overall, the proposed approach demonstrated promising results for lung parenchyma segmentation. The method has the potential to be valuable in the advancement of computer-aided diagnosis (CAD) systems for automatic nodule detection.

Radiologists' Rating for Comparative Qualitative Assessment of Intravoxel Incoherent Motion Using Novel Analysis Methods.

OBJECTIVE: The objective was to assess qualitative interpretability and quantitative precision and reproducibility of intravoxel incoherent motion ( IVIM) parametric images evaluated using novel IVIM analysis methods for diagnostic accuracy. METHODS: Intravoxel incoherent motion datasets of 55 patients (male/female = 41:14; age = 17.8 ± 5.5 years) with histopathology-proven osteosarcoma were analyzed. Intravoxel incoherent motion parameters-diffusion coefficient ( D ), perfusion fraction ( f ), and perfusion coefficient ( D* )-were estimated using 5 IVIM analysis methods-(i) biexponential (BE) model, (ii) BE-segmented fitting 2-parameter (BESeg-2), (iii) BE-segmented fitting 1-parameter (BESeg-1), (iv) BE model with total variation penalty function (BE + TV), and (v) BE model with Huber penalty function (BE + HPF). Qualitative scoring in a 5-point Likert scale (uninterpretable: 1; poor: 2; fair: 3; good: 4; excellent: 5) was performed by 2 radiologists for 4 criteria: (a) tumor shape and margin, (b) morphologic correlation, (c) noise suppression, and (d) overall interpretability. Interobserver agreement was evaluated using Spearman rank-order correlation ( rs ). Precision and reproducibility were evaluated using within-subject coefficient of variation (wCV) and between-subject coefficient of variation (bCV). RESULTS: BE + TV and BE + HPF produced significantly ( P < 10 -3 ) higher qualitative scores for D (fair-good [3.3-3.8]) than BE (poor [2.3]) and for D* (poor-fair [2.2-2.7]) and f (fair-good [3.2-3.8]) than BE, BESeg-2, and BESeg-1 ( D* : uninterpretable-poor [1.3-1.9] and f : poor-fair [1.5-3]). Interobserver agreement for qualitative scoring was rs = 0.48-0.59, P < 0.009. BE + TV and BE + HPF showed significantly ( P < 0.05) improved reproducibility in estimating D (wCV: 24%-31%, bCV: 21%-31% improvement) than the BE method and D* (wCV: 4%-19%, bCV: 5%-19% improvement) and f (wCV: 25%-49%, bCV: 25%-47% improvement) than BE, BESeg-2, and BESeg-1 methods. CONCLUSIONS: BE + TV and BE + HPF demonstrated qualitatively and quantitatively improved IVIM parameter estimation and may be considered for clinical use further.

A proposed methodology for detecting the malignant potential of pulmonary nodules in sarcoma using computed tomographic imaging and artificial intelligence-based models.

The presence of lung metastases in patients with primary malignancies is an important criterion for treatment management and prognostication. Computed tomography (CT) of the chest is the preferred method to detect lung metastasis. However, CT has limited efficacy in differentiating metastatic nodules from benign nodules (e.g., granulomas due to tuberculosis) especially at early stages (<5 mm). There is also a significant subjectivity associated in making this distinction, leading to frequent CT follow-ups and additional radiation exposure along with financial and emotional burden to the patients and family. Even 18F-fluoro-deoxyglucose positron emission technology-computed tomography (18F-FDG PET-CT) is not always confirmatory for this clinical problem. While pathological biopsy is the gold standard to demonstrate malignancy, invasive sampling of small lung nodules is often not clinically feasible. Currently, there is no non-invasive imaging technique that can reliably characterize lung metastases. The lung is one of the favored sites of metastasis in sarcomas. Hence, patients with sarcomas, especially from tuberculosis prevalent developing countries, can provide an ideal platform to develop a model to differentiate lung metastases from benign nodules. To overcome the lack of optimal specificity of CT scan in detecting pulmonary metastasis, a novel artificial intelligence (AI)-based protocol is proposed utilizing a combination of radiological and clinical biomarkers to identify lung nodules and characterize it as benign or metastasis. This protocol includes a retrospective cohort of nearly 2,000-2,250 sample nodules (from at least 450 patients) for training and testing and an ambispective cohort of nearly 500 nodules (from 100 patients; 50 patients each from the retrospective and prospective cohort) for validation. Ground-truth annotation of lung nodules will be performed using an in-house-built segmentation tool. Ground-truth labeling of lung nodules (metastatic/benign) will be performed based on histopathological results or baseline and/or follow-up radiological findings along with clinical outcome of the patient. Optimal methods for data handling and statistical analysis are included to develop a robust protocol for early detection and classification of pulmonary metastasis at baseline and at follow-up and identification of associated potential clinical and radiological markers.

Individualized closed-loop TMS synchronized with exoskeleton for modulation of cortical-excitability in patients with stroke: a proof-of-concept study.

Background: Repetitive TMS is used in stroke rehabilitation with predefined passive low and high-frequency stimulation. Brain State-Dependent Stimulation (BSDS)/Activity-Dependent Stimulation (ADS) using bio-signal has been observed to strengthen synaptic connections. Without the personalization of brain-stimulation protocols, we risk a one-size-fits-all approach. Methods: We attempted to close the ADS loop via intrinsic-proprioceptive (via exoskeleton-movement) and extrinsic-visual-feedback to the brain. We developed a patient-specific brain stimulation platform with a two-way feedback system, to synchronize single-pulse TMS with exoskeleton along with adaptive performance visual feedback, in real-time, for a focused neurorehabilitation strategy to voluntarily engage the patient in the brain stimulation process. Results: The novel TMS Synchronized Exoskeleton Feedback (TSEF) platform, controlled by the patient's residual Electromyogram, simultaneously triggered exoskeleton movement and single-pulse TMS, once in 10 s, implying 0.1 Hz frequency. The TSEF platform was tested for a demonstration on three patients (n = 3) with different spasticity on the Modified Ashworth Scale (MAS = 1, 1+, 2) for one session each. Three patients completed their session in their own timing; patients with (more) spasticity tend to take (more) inter-trial intervals. A proof-of-concept study on two groups-TSEF-group and a physiotherapy control-group was performed for 45 min/day for 20-sessions. Dose-matched Physiotherapy was given to control-group. Post 20 sessions, an increase in ipsilesional cortical-excitability was observed; Motor Evoked Potential increased by ~48.5 µV at a decreased Resting Motor Threshold by ~15.6%, with improvement in clinical scales relevant to the Fugl-Mayer Wrist/Hand joint (involved in training) by 2.6 units, an effect not found in control-group. This strategy could voluntarily engage the patient. Conclusion: A brain stimulation platform with a real-time two-way feedback system was developed to voluntarily engage the patients during the brain stimulation process and a proof-of-concept study on three patients indicates clinical gains with increased cortical excitability, an effect not observed in the control-group; and the encouraging results nudge for further investigations on a larger cohort.

Clinical potential and neuroplastic effect of targeted virtual reality based intervention for distal upper limb in post-stroke rehabilitation: a pilot observational study.

PURPOSE: A library of Virtual Reality (VR) tasks has been developed for targeted post-stroke rehabilitation of distal upper extremities. The objective of this pilot study was to evaluate the clinical potential of the targeted VR-based therapeutic intervention in a small cohort of patients specifically with chronic stroke. Furthermore, our aim was to explore the possible neuronal reorganizations in corticospinal pathways in response to the distal upper limb targeted VR-intervention. METHODOLOGY: Five patients with chronic stroke were enrolled in this study and were given VR-intervention of 20 sessions of 45 min each. Clinical Scales, cortical-excitability measures (using Transcranial Magnetic Stimulation): Resting Motor Threshold (RMT), and Motor Evoked Potential (MEP) amplitude, task-specific performance metrics i.e., Time taken to complete the task (TCT), smoothness of trajectory, relative % error were evaluated pre- and post-intervention to evaluate the intervention-induced improvements. RESULTS: Pre-to post-intervention improvements were observed in Fugl-Meyer Assessment (both total and wrist/hand component), Modified Barthel Index, Stroke Impact Scale, Motor Assessment Scale, active range of motion at wrist, and task-specific outcome metrics. Pre-to post-intervention ipsilesional RMT reduced (mean ∼9%) and MEP amplitude increased (mean ∼29µV), indicating increased cortical excitability at post-intervention. CONCLUSION: VR-training exhibited improved motor outcomes and cortical-excitability in patients with stroke. Neurophysiological changes observed in terms of improved cortical-excitability might be a consequence of plastic reorganization induced by VR-intervention.IMPLICATIONS FOR REHABILITATIONPost-stroke rehabilitation of distal upper extremities is crucial and needs targeted intervention to rehabilitate in the chronic phase of recovery.Virtual reality (VR) has emerged as a supplemental approach in post-stroke rehabilitation. However, its customization as per clinical need is still under research.This pilot study provides preliminary evidence of the clinical utility of the developed VR tasks targeted for distal upper extremities.

Post-Stroke Rehabilitation of Distal Upper Limb with New Perspective Technologies: Virtual Reality and Repetitive Transcranial Magnetic Stimulation-A Mini Review.

Upper extremity motor impairment is the most common sequelae in patients with stroke. Moreover, its continual nature limits the optimal functioning of patients in the activities of daily living. Because of the intrinsic limitations in the conventional form of rehabilitation, the rehabilitation applications have been expanded to technology-driven solutions, such as Virtual Reality and Repetitive Transcranial Magnetic Stimulation (rTMS). The motor relearning processes are influenced by variables, such as task specificity, motivation, and feedback provision, and a VR environment in the form of interactive games could provide novel and motivating customized training solutions for better post-stroke upper limb motor improvement. rTMS being a precise non-invasive brain stimulation method with good control of stimulation parameters, has the potential to facilitate neuroplasticity and hence a good recovery. Although several studies have discussed these forms of approaches and their underlying mechanisms, only a few of them have specifically summarized the synergistic applications of these paradigms. To bridge the gaps, this mini review presents recent research and focuses precisely on the applications of VR and rTMS in distal upper limb rehabilitation. It is anticipated that this article will provide a better representation of the role of VR and rTMS in distal joint upper limb rehabilitation in patients with stroke.

Pancreatic Mass Characterization Using IVIM-DKI MRI and Machine Learning-Based Multi-Parametric Texture Analysis.

Non-invasive characterization of pancreatic masses aids in the management of pancreatic lesions. Intravoxel incoherent motion-diffusion kurtosis imaging (IVIM-DKI) and machine learning-based texture analysis was used to differentiate pancreatic masses such as pancreatic ductal adenocarcinoma (PDAC), pancreatic neuroendocrine tumor (pNET), solid pseudopapillary epithelial neoplasm (SPEN), and mass-forming chronic pancreatitis (MFCP). A total of forty-eight biopsy-proven patients with pancreatic masses were recruited and classified into pNET (n = 13), MFCP (n = 6), SPEN (n = 4), and PDAC (n = 25) groups. All patients were scanned for IVIM-DKI sequences acquired with 14 b-values (0 to 2500 s/mm2) on a 1.5T MRI. An IVIM-DKI model with a 3D total variation (TV) penalty function was implemented to estimate the precise IVIM-DKI parametric maps. Texture analysis (TA) of the apparent diffusion coefficient (ADC) and IVIM-DKI parametric map was performed and reduced using the chi-square test. These features were fed to an artificial neural network (ANN) for characterization of pancreatic mass subtypes and validated by 5-fold cross-validation. Receiver operator characteristics (ROC) analyses were used to compute the area under curve (AUC). Perfusion fraction (f) was significantly higher (p < 0.05) in pNET than PDAC. The f showed better diagnostic performance for PDAC vs. MFCP with AUC:0.77. Both pseudo-diffusion coefficient (D*) and f for PDAC vs. pNET showed an AUC of 0.73. ADC and diffusion coefficient (D) showed good diagnostic performance for pNET vs. MFCP with AUC: 0.79 and 0.76, respectively. In the TA of PDAC vs. non-PDAC, f and combined IVIM-DKI parameters showed high accuracy ≥ 84.3% and AUC ≥ 0.84. Mean f and combined IVIM-DKI parameters estimated that the IVIM-DKI model with TV texture features has the potential to be helpful in characterizing pancreatic masses.

Chemotherapy response evaluation using diffusion weighted MRI in Ewing Sarcoma: A single center experience.

BACKGROUND: Non-invasive biomarkers for early chemotherapeutic response in Ewing sarcoma family of tumors (ESFT) are useful for optimizing existing treatment protocol. PURPOSE: To assess the role of diffusion-weighted magnetic resonance imaging (MRI) in the early evaluation of chemotherapeutic response in ESFT. MATERIAL AND METHODS: A total of 28 patients (mean age = 17.2 ± 5.6 years) with biopsy proven ESFT were analyzed prospectively. Patients underwent MRI acquisition on a 1.5-T scanner at three time points: before starting neoadjuvant chemotherapy (baseline), after first cycle chemotherapy (early time point), and after completion of chemotherapy (last time point). RECIST 1.1 criteria was used to evaluate the response to chemotherapy and patients were categorized as responders (complete and partial response) and non-responders (stable and progressive disease). Tumor diameter, absolute apparent diffusion coefficient (ADC), and normalized ADC (nADC) values in the tumor were measured. Baseline parameters and relative percentage change of parameters after first cycle chemotherapy were assessed for early detection of chemotherapy response. RESULTS: The responder:non-responder ratio was 21:7. At baseline, ADC ([0.864 ± 0.266 vs. 0.977 ± 0.246]) × 10-3mm2/s; P = 0.205) and nADC ([0.740 ± 0.254 vs. 0.925 ± 0.262] × 10-3mm2/s; P = 0.033) among responders was lower than the non-responders and predicted response to chemotherapy with AUCs of 0.6 and 0.735, respectively. At the early time point, tumor diameter (27% ± 14% vs. 4.6% ± 10%; P = 0.002) showed a higher reduction and ADC (75% ± 44% vs. 52% ± 72%; P = 0.039) and nADC (81% ± 44% vs. 48% ± 67%; P = 0.008) showed a higher increase in mean values among responders than the non-responders and identified chemotherapy response with AUC of 0.890, 0.723, and 0.756, respectively. CONCLUSION: Baseline nADC and its change after the first cycle of chemotherapy can be used as non-invasive surrogate markers of early chemotherapeutic response in patients with ESFT.

Non-invasive intravoxel incoherent motion MRI in prediction of histopathological response to neoadjuvant chemotherapy and survival outcome in osteosarcoma at the time of diagnosis.

BACKGROUND: Early prediction of response to neoadjuvant chemotherapy (NACT) is important to aid personalized treatment in osteosarcoma. Diffusion-weighted Intravoxel Incoherent Motion (IVIM) MRI was used to evaluate the predictive value for response to NACT and survival outcome in osteosarcoma. METHODS: Total fifty-five patients with biopsy-proven osteosarcoma were recruited prospectively, among them 35 patients were further analysed. Patients underwent 3 cycles of NACT (Cisplatin + Doxorubicin) followed by surgery and response adapted adjuvant chemotherapy. Treatment outcomes were histopathological response to NACT (good-response ≥ 50% necrosis and poor-response < 50% necrosis) and survival outcome (event-free survival (EFS) and overall survival (OS)). IVIM MRI was acquired at 1.5T at baseline (t0), after 1-cycle (t1) and after 3-cycles (t2) of NACT. Quantitative IVIM parameters (D, D*, f & D*.f) were estimated using advanced state-of-the-art spatial penalty based IVIM analysis method bi-exponential model with total-variation penalty function (BETV) at 3 time-points and histogram analysis was performed. RESULTS: Good-responders: Poor-responders ratio was 13 (37%):22 (63%). EFS and OS were 31% and 69% with 16.27 and 25.9 months of median duration respectively. For predicting poor-response to NACT, IVIM parameters showed AUC = 0.87, Sensitivity = 86%, Specificity = 77% at t0, and AUC = 0.96, Sensitivity = 86%, Specificity = 100% at t1. Multivariate Cox regression analysis showed smaller tumour volume (HR = 1.002, p = 0.001) higher ADC-25th-percentile (HR = 0.047, p = 0.005) & D-Mean (HR = 0.1, p = 0.023) and lower D*-Mean (HR = 1.052, p = 0.039) were independent predictors of longer EFS (log-rank p-values: 0.054, 0.0034, 0.0017, 0.0019 respectively) and non-metastatic disease (HR = 4.33, p < 10-3), smaller tumour-volume (HR = 1.001, p = 0.042), lower D*-Mean (HR = 1.045, p = 0.056) and higher D*.f-skewness (HR = 0.544, p = 0.048) were independent predictors of longer OS (log-rank p-values: < 10-3, 0.07, < 10-3, 0.019 respectively). CONCLUSION: IVIM parameters obtained with a 1.5T scanner along with novel BETV method and their histogram analysis indicating tumour heterogeneity were informative in characterizing NACT response and survival outcome in osteosarcoma.

Machine learning-based analysis of a semi-automated PI-RADS v2.1 scoring for prostate cancer.

Background: Prostate Imaging-Reporting and Data System version 2.1 (PI-RADS v2.1) was developed to standardize the interpretation of multiparametric MRI (mpMRI) for prostate cancer (PCa) detection. However, a significant inter-reader variability among radiologists has been found in the PI-RADS assessment. The purpose of this study was to evaluate the diagnostic performance of an in-house developed semi-automated model for PI-RADS v2.1 scoring using machine learning methods. Methods: The study cohort included an MRI dataset of 59 patients (PI-RADS v2.1 score 2 = 18, score 3 = 10, score 4 = 16, and score 5 = 15). The proposed semi-automated model involved prostate gland and zonal segmentation, 3D co-registration, lesion region of interest marking, and lesion measurement. PI-RADS v2.1 scores were assessed based on lesion measurements and compared with the radiologist PI-RADS assessment. Machine learning methods were used to evaluate the diagnostic accuracy of the proposed model by classification of PI-RADS v2.1 scores. Results: The semi-automated PI-RADS assessment based on the proposed model correctly classified 50 out of 59 patients and showed a significant correlation (r = 0.94, p < 0.05) with the radiologist assessment. The proposed model achieved an accuracy of 88.00% ± 0.98% and an area under the receiver-operating characteristic curve (AUC) of 0.94 for score 2 vs. score 3 vs. score 4 vs. score 5 classification and accuracy of 93.20 ± 2.10% and AUC of 0.99 for low score vs. high score classification using fivefold cross-validation. Conclusion: The proposed semi-automated PI-RADS v2.1 assessment system could minimize the inter-reader variability among radiologists and improve the objectivity of scoring.

Quantification of prostate tumour diameter and volume from MR images using 3D ellipsoid model and its impact on PI-RADS v2.1 assessment.

Maximum diameter and volume of the tumour provide important clinical information and are decision-making parameters for patients suspected with prostate cancer (PCa). The objectives of this study were to develop an automated method for 3D tumour measurement and compare it with the radiologist's manual assessment, as well as to investigate the impact of 3D tumour measurement on Prostate Imaging-Reporting and Data System version-2.1 (PI-RADS v2.1) scoring of prostate cancer. Tumour maximum diameter and volume were calculated using automated ellipsoid-fit method. For all PI-RADS scores, mean ± standard deviation range of tumour maximum diameter and volume measured using ellipsoid-fit method were 1.36 ± 0.28 to 1.97 ± 0.67 cm and 0.49 ± 0.31 to 1.05 ± 0.78 cc and manual assessment were in range of 0.73 ± 0.12 to 1.14 ± 0.25 cm and 0.36 ± 0.21 to 0.93 ± 0.39 cc, respectively. Ellipsoid-fit method showed significantly (p < 0.05) higher values for maximum diameter and volume than manual assessment. 3D measurement of tumour using ellipsoid-fit method was found to have higher maximum diameter and volume values (in 40-61% patients) compared to conventional assessment by radiologist, which may have an impact on PI-RADS v2.1 scoring system.

A novel perspective of associativity of upper limb motor impairment and cortical excitability in sub-acute and chronic stroke.

Background: The global inclination of stroke onset in earlier years of life and increased lifespan have resulted in an increased chronic post-stroke-related disability. The precise and simplistic approach such as the correlation of Fugl-Meyer Assessment (FMA) with Transcranial Magnetic Stimulation (TMS) parameters, Resting Motor Threshold (RMT) and Motor Evoked Potential (MEP), in patients with stroke might play a critical role, given the prognostic value of MEP, a measure of cortical excitability, and might be the key point in prescribing appropriate therapeutic strategies. Objective: The study aimed to determine the correlation of FMA-based impairment in the upper extremity function specifically of the wrist and hand with respect to the neurophysiological parameters of corticospinal tract integrity. Materials and methods: The Institutional Review Board approved the study and 67 (n) patients with stroke were enrolled in the Department of Neurology, AIIMS, New Delhi, India. The motor assessment was performed on patients by the upper extremity subset of Fugl-Meyer Assessment (FMA) and the clinical history was obtained. RMT and MEP of Extensor Digitorum Communis (EDC) muscle were measured via TMS. Results: A significant positive correlation was observed between Fugl-Meyer Assessment Wrist/Hand (FMA W/H) and MEP scores (r = 0.560, <0.001). Also, Fugl-Meyer Assessment Upper Extremity (FMA UE) scores demonstrated a moderate positive association with MEP responsiveness (r = 0.421, <0.001). Conclusion: MEP of the EDC muscle was found to be associated with sensorimotor control as measured by FMA. Moreover, FMA W/H score values might be a better prognostic indicator of EDC MEP responsiveness. Interestingly, a novel element comprising the range of FMA UE and FMA W/H components was observed to be a potential indicator of MEP responsiveness and could also indicate establishing FMA as a surrogate for TMS in resource-limited settings for prognostification.

Machine learning based reanalysis of clinical scores for distinguishing between ischemic and hemorrhagic stroke in low resource setting.

BACKGROUND: Identifying ischemic or hemorrhagic strokes clinically may help in situations where neuroimaging is unavailable to provide primary-care prior to referring to stroke-ready facility. Stroke classification-based solely on clinical scores faces two unresolved issues. One pertains to overestimation of score performance, while other is biased performance due to class-imbalance inherent in stroke datasets. After correcting the issues using Machine Learning theory, we quantitatively compared existing scores to study the capabilities of clinical attributes for stroke classification. METHODS: We systematically searched PubMed, ERIC, ScienceDirect, and IEEE-Xplore from 2001 to 2021 for studies that validated the Siriraj, Guys Hospital/Allen, Greek, and Besson scores for stroke classification. From included studies we extracted the reported cross-tabulation to identify and correct the above listed issues for an accurate comparative analysis of the performance of clinical scores. RESULTS: A total of 21 studies were included. Comparative analysis demonstrates Siriraj Score outperforms others. For Siriraj Score the reported sensitivity range (Ischemic Stroke-diagnosis) 43-97% (Median = 78% [IQR 65-88%]) is significantly higher than our calculated range 40-90% (Median = 70% [IQR 57-73%]), also the reported sensitivity range (Hemorrhagic Stroke-diagnosis) 50-95% (Median = 71% [IQR 64-82%]) is higher than our calculated range 34-86% (Median = 59% [IQR 50-79%]) which indicates overestimation of performance by the included studies. Guys Hospital/Allen and Greek Scores show similar trends. Recommended weighted-accuracy metric provides better estimate of the performance. CONCLUSION: We demonstrate that clinical attributes have a potential for stroke classification, however the performance of all scores varies across demographics, indicating the need to fine-tune scores for different demographics. To improve this variability, we suggest creating global data pool with statistically significant attributes. Machine Learning classifiers trained over such dataset may perform better and generalise at scale.

Design and Validation of Virtual Reality Task for Neuro-Rehabilitation of Distal Upper Extremities.

Stroke, affecting approximately 15 million people worldwide, has long been a global cause of death and disability. Virtual Reality (VR) has shown its potential as an assistive tool for post-stroke rehabilitation. The objective of this pilot study was to define the task-specific performance metrics of VR tasks to assess the performance level of healthy subjects and patients quantitatively and to obtain their feedback for improving the developed framework. A pilot prospective study was designed. We tested the designed VR tasks on forty healthy right-handed subjects to evaluate its potential. Qualitative trajectory plots and three quantitative performance metrics-time taken to complete the task, percentage relative error, and trajectory smoothness-were computed from the recorded data of forty healthy subjects. Two patients with stroke were also enrolled to compare their performance with healthy subjects. Each participant received one VR session of 90 min. No adverse effects were noticed throughout the study. Performance metrics obtained from healthy subjects were used as a reference for patients. Relatively higher values of task completion time and trajectory smoothness and lower values of relative % error was observed for the affected hands w.r.t the unaffected hands of both the patients. For the unaffected hands of both the patients, the performance levels were found objectively closer to that of healthy subjects. A library of VR tasks for wrist and fingers were designed, and task-specific performance metrics were defined in this study. The evaluation of the VR exercises using these performance metrics will help the clinicians to assess the patient's progress quantitatively and to design the rehabilitation framework for a future clinical study.

T1 mapping as a surrogate marker of chemotherapy response evaluation in patients with osteosarcoma.

OBJECTIVES: To characterize baseline T1 values of tumors, measure changes after the course of chemotherapy, and evaluate its potential as a marker of response assessment in patients with osteosarcoma. MATERIALS AND METHODS: A total of 35 patients (male:female = 27:8; age = 17.9 ± 6 years; metastatic:localized = 11:24) with biopsy-proven osteosarcoma were analyzed prospectively. All patients underwent magnetic resonance imaging before neoadjuvant chemotherapy (NACT) (baseline) and after NACT completion (follow-up), followed by surgery and histopathological evaluation. Histopathological necrosis served as the gold standard for assessing chemotherapy response (responder: ≥50% necrosis and non-responder: <50% necrosis). Three-dimensional spoiled gradient recalled echo images were acquired with varying flip angles (50, 100, 200, and 300) using a 1.5 Tesla scanner. T1 values were estimated in healthy muscle tissue and tumors at baseline and follow-up, and the relative percentage changes after NACT (ΔT1) were evaluated, and histogram analysis was performed to characterize the T1 value of the tumor and predict the NACT response using the Pearson correlation coefficient (r) and receiver operating characteristic curve analysis. RESULTS: At baseline, a significantly higher T1-mean (830.96 ± 193.88 msec versus 683.29 ± 210.00 msec; p = 0.003) and lower T1-skewness (0.86 ± 1.66 versus 1.60 ± 1.55; p = 0.02) were observed in osteosarcoma than healthy tissue. Responder:non-responder ratio was 13:22. At baseline, a significantly higher T1-mean (936.14 ± 193.17 msec versus 768.82 ± 169.25 msec; p = 0.018) and lower T1-skewness (-0.17 ± 0.85 versus 1.47 ± 1.73; p = 0.003) were observed among responders, than non-responders. After NACT, ΔT1 in tumor was significantly higher among responders than non-responders (-27.22 ± 12.17% versus -15.70 ± 18.99%; p = 0.034). ΔT1-mean and ΔT1-skewness after NACT were moderately correlated (r =  - 0.4, p = 0.030; r = 0.4, p = 0.011) with histopathological necrosis. For predicting NACT response, T1-mean and T1-skewness jointly produced an area under the curve (AUC) = 0.80 at baseline, and ΔT1-mean and ΔT1-skewness jointly produced AUC = 0.76 after NACT. CONCLUSION: The mean and skewness of T1 values in osteosarcoma are potential non-invasive imaging markers for chemotherapy response assessment.

Device for Assessing Knee Joint Dynamics During Magnetic Resonance Imaging.

BACKGROUND: Knee assessment with and without load using magnetic resonance imaging (MRI) can provide information on knee joint dynamics and improve the diagnosis of knee joint diseases. Performing such studies on a routine MRI-scanner require a load-exerting device during scanning. There is a need for more studies on developing loading devices and evaluating their clinical potential. PURPOSE: Design and develop a portable and easy-to-use axial loading device to evaluate the knee joint dynamics during the MRI study. STUDY TYPE: Prospective study. SUBJECTS: Nine healthy subjects. FIELD STRENGTH/SEQUENCE: A 0.25 T standing-open MRI and 3.0 T MRI. PD-T2 -weighted FSE, 3D-fast-spoiled-gradient-echo, FS-PD, and CartiGram sequences. ASSESSMENT: Design and development of loading device, calibration of loads, MR safety assessment (using projectile angular displacement, torque, and temperature tests). Scoring system for ease of doing. Qualitative (by radiologist) and quantitative (using structural similarity index measure [SSIM]) image-artifact assessment. Evaluation of repeatability, comparison with various standing stances load, and loading effect on knee MR parameters (tibiofemoral bone gap [TFBG], femoral cartilage thickness [FCT], tibial cartilage thickness [TCT], femoral cartilage T2 -value [FCT2], and tibia cartilage T2 -value [TCT2]). The relative percentage change (RPC) in parameters due to the device load was computed. STATISTICAL TEST: Pearson's correlation coefficient (r). RESULTS: The developed device is conditional-MR safe (details in the manuscript and supplementary materials), 15 × 15 × 45 cm3 dimension, and <3 kg. The ease of using the device was 4.9/5. The device introduced no visible image artifacts, and SSIM of 0.9889 ± 0.0153 was observed. The TFBG intraobserver variability (absolute difference) was <0.1 mm. Interobserver variability of all regions of interest was <0.1 mm. The load exerted by the device was close to the load during standing on both legs in 0.25 T scanner with r > 0.9. Loading resulted in RPC of 1.5%-11.0%, 7.9%-8.5%, and -1.5% to 13.0% in the TFBG, FCT, and TCT, respectively. FCT2 and TCT2 were reduced in range of 1.5-2.7 msec and 0.5-2.3 msec due to load. DATA CONCLUSION: The proposed device is conditionally MR safe, low cost (material cost < INR 6000), portable, and effective in loading the knee joint with up to 50% of body weight. EVIDENCE LEVEL: 1 TECHNICAL EFFICACY: Stage 1.

IVIM-DKI for differentiation between prostate cancer and benign prostatic hyperplasia: comparison of 1.5 T vs. 3 T MRI.

OBJECTIVE: To implement an advanced spatial penalty-based reconstruction to constrain the intravoxel incoherent motion (IVIM)-diffusion kurtosis imaging (DKI) model and investigate whether it provides a suitable alternative at 1.5 T to the traditional IVIM-DKI model at 3 T for clinical characterization of prostate cancer (PCa) and benign prostatic hyperplasia (BPH). MATERIALS AND METHODS: Thirty-two patients with biopsy-proven PCa were recruited for MRI examination (n = 16 scanned at 1.5 T, n = 16 scanned at 3 T). Diffusion-weighted imaging (DWI) with 13 b values (b = 0 to 2000 s/mm2 up to 3 averages, 1.5 T: TR = 5.774 s, TE = 81 ms and 3 T: TR = 4.899 s, TE = 100 ms), T2-weighted, and T1-weighted imaging were used on the 1.5 T and 3 T MRI scanner, respectively. The IVIM-DKI signal was modeled using the traditional IVIM-DKI model and a novel model in which the total variation (TV) penalty function was combined with the traditional model to optimize non-physiological variations. Paired and unpaired t-tests were used to compare intra-scanner and scanner group differences in IVIM-DKI parameters obtained using the novel and the traditional models. Analysis of variance with post hoc test and receiver operating characteristic (ROC) curve analysis were used to assess the ability of parameters obtained using the novel model (at 1.5 T) and the traditional model (at 3 T) to characterize prostate lesions. RESULTS: IVIM-DKI modeled using novel model with TV spatial penalty function at 1.5 T, produced parameter maps with 50-78% lower coefficient of variation (CV) than traditional model at 3 T. Novel model estimated higher D with lower D*, f and k values at both field strengths compared to traditional model. For scanner differences, the novel model at 1.5 T estimated lower D* and f values as compared to traditional model at 3 T. At 1.5 T, D and f values were significantly lower with k values significantly higher in tumor than BPH and healthy tissue. D (AUC: 0.98), f (AUC: 0.82), and k (AUC: 0.91) parameters estimated using novel model showed high diagnostic performance in cancer lesion detection at 1.5 T. DISCUSSION: In comparison with the IVIM-DKI model at 3 T, IVIM-DKI signal modeled with the TV penalty function at 1.5 T showed lower estimation errors. The proposed novel model can be utilized for improved detection of prostate lesions.

A semi-automatic framework based upon quantitative analysis of MR-images for classification of femur cartilage into asymptomatic, early OA, and advanced-OA groups.

To develop a semi-automatic framework for quantitative analysis of biochemical properties and thickness of femur cartilage using magnetic resonance (MR) images and evaluate its potential for femur cartilage classification into asymptomatic (AS), early osteoarthritis (OA), and advanced OA groups. In this study, knee joint MRI data (fat suppressed-proton density-weighted and multi-echo T2-weighted images) of eight AS-volunteers (data acquired twice) and 34 OA patients including 20 early OA (16 Grade-I and 4 Grade-II), 14 advanced-OA (Grade-III) were acquired at 3.0T MR scanner. Modified Outerbridge classification criteria was performed for the clinical evaluation of data by an experienced radiologist. Cartilage segmentation, T2-mapping, 2D-WearMap generation, and subregion analysis were performed semi-automatically using in-house developed algorithms. The intraclass correlation coefficient (ICC) and coefficient of variation (CV) were computed for testing the reproducibility of T2 values. One-way analysis of variance with Tukey-Kramer post hoc test was performed for evaluating the differences among the groups. The performance of individual T2 and thickness, as well as their combination using logistic regression, were evaluated with receiver operating characteristics (ROC) curve analysis. The interscan agreement based on the ICC index was 0.95 and the CV was 2.45 ± 1.33%. T2 mean of values greater than 75th percentile showed sensitivity and specificity of 94.1% and 81.3% (AUC = 0.93, cut-off value = 47.9 ms) in differentiating AS volunteers versus OA group, while sensitivity and specificity of 90.0% and 81.3% (AUC = 0.90, cut-off value = 47.9 ms) in differentiating AS volunteers versus early OA groups, respectively. In the differentiation of early OA versus advanced-OA group, ROC results of combination (T2 and thickness) showed the highest sensitivity and specificity of 85.7%, and 70.0% (AUC = 0.79, cut-off value = 0.39) compared with individual T2 and thickness features, respectively. A computer-aided quantitative evaluation of femur cartilage degeneration showed promising results and can be used to assist clinicians in diagnosing OA.

Clinical Effectiveness of Non-Immersive Virtual Reality Tasks for Post-Stroke Neuro-Rehabilitation of Distal Upper-Extremities: A Case Report.

A library of non-immersive Virtual Reality (VR) tasks were developed for post-stroke rehabilitation of distal upper extremities. The objective was to evaluate the rehabilitation impact of the developed VR-tasks on a patient with chronic stroke. The study involved a 50-year-old male patient with chronic (13 month) stroke. Twenty VR therapy sessions of 45 min each were given. Clinical scales, cortical-excitability measures, functional MRI (fMRI), and diffusion tensor imaging (DTI) data were acquired pre-and post-therapy to evaluate the motor recovery. Increase in Fugl-Meyer Assessment (wrist/hand) by 2 units, Barthel Index by 5 units, Brunnstrom Stage by 1 unit, Addenbrooke's Cognitive Examination by 3 units, Wrist Active Range of Motion by 5° and decrease in Modified Ashworth Scale by 1 unit were observed. Ipsilesional Motor Evoked Potential (MEP) amplitude (obtained using Transcranial Magnetic Stimulation) was increased by 60.9µV with a decrease in Resting Motor Threshold (RMT) by 7%, and contralesional MEP amplitude was increased by 56.2µV with a decrease in RMT by 7%. The fMRI-derived Laterality Index of Sensorimotor Cortex increased in precentral-gyrus (from 0.28 to 0.33) and in postcentral-gyrus (from 0.07 to 0.3). The DTI-derived FA-asymmetry decreased in precentral-gyrus (from 0.029 to 0.024) and in postcentral-gyrus (from 0.027 to 0.017). Relative reduction in task-specific performance metrics, i.e., time taken to complete the task (31.6%), smoothness of trajectory (76.7%), and relative percentage error (80.7%), were observed from day 1 to day 20 of the VR therapy. VR therapy resulted in improvement in clinical outcomes in a patient with chronic stroke. The research also gives insights to further improve the overall system of rehabilitation.