On the reliability of single-camera markerless systems for overground gait monitoring.

BACKGROUND AND OBJECTIVE: Motion analysis is crucial for effective and timely rehabilitative interventions on people with motor disorders. Conventional marker-based (MB) gait analysis is highly time-consuming and calls for expensive equipment, dedicated facilities and personnel. Markerless (ML) systems may pave the way to less demanding gait monitoring, also in unsupervised environments (i.e., in telemedicine). However,scepticism on clinical usability of relevant outcome measures has hampered its use. ML is normally used to analyse treadmill walking, which is significantly different from the more physiological overground walking. This study aims to provide end-users with instructions on using a single-camera markerless system to obtain reliable motion data from overground walking, while clinicians will be instructed on the reliability of obtained quantities. METHODS: The study compares kinematics obtained from ML systems to those concurrently obtained from marker-based systems, considering different stride counts and subject positioning within the capture volume. RESULTS: The findings suggest that five straight walking trials are sufficient for collecting reliable kinematics with ML systems. Precision on joint kinematics decreased at the boundary of the capture volume. Excellent correlation was found between ML and MB systems for hip and knee angles (0.92

Preserving Data Privacy and Accuracy of Human Pose Estimation Software Based on CNN s for Remote Gait Analysis.

In the last years there have been significant improvements in the accuracy of real-time 3D skeletal data estimation software. These applications based on convolutional neural networks (CNNs) can playa key role in a variety of clinical scenarios, from gait analysis to medical diagnosis. One of the main challenges is to apply such intelligent video analytic at a distance, which requires the system to satisfy, beside accuracy, also data privacy. To satisfy privacy by default and by design, the software has to run on "edge" computing devices, by which the sensitive information (i.e., the video stream) is elaborated close to the camera while only the process results can be stored or sent over the communication network. In this paper we address such a challenge by evaluating the accuracy of the state-of-the-art software for human pose estimation when run "at the edge". We show how the most accurate platforms for pose estimation based on complex and deep neural networks can become inaccurate due to subs amp ling of the input video frames when run on the resource constrained edge devices. In contrast, we show that, starting from less accurate and "lighter" CNNs and enhancing the pose estimation software with filters and interpolation primitives, the platform achieves better real-time performance and higher accuracy with a deviation below the error tolerance of a marker-based motion capture system.

Enabling Gait Analysis in the Telemedicine Practice through Portable and Accurate 3D Human Pose Estimation.

Human pose estimation (HPE) through deep learning-based software applications is a trend topic for markerless motion analysis. Thanks to the accuracy of the state-of-the-art technology, HPE could enable gait analysis in the telemedicine practice. On the other hand, delivering such a service at a distance requires the system to satisfy multiple and different constraints like accuracy, portability, real-time, and privacy compliance at the same time. Existing solutions either guarantee accuracy and real-time (e.g., the widespread OpenPose software on well-equipped computing platforms) or portability and data privacy (e.g., light convolutional neural networks on mobile phones). We propose a portable and low-cost platform that implements real-time and accurate 3D HPE through an embedded software on a low-power off-the-shelf computing device that guarantees privacy by default and by design. We present an extended evaluation of both accuracy and performance of the proposed solution conducted with a marker-based motion capture system (i.e., Vicon) as ground truth. The results show that the platform achieves real-time performance and high-accuracy with a deviation below the error tolerance when compared to the marker-based motion capture system (e.g., less than an error of 5∘ on the estimated knee flexion difference on the entire gait cycle and correlation 0.91<ρ<0.99). We provide a proof-of-concept study, showing that such portable technology, considering the limited discrepancies with respect to the marker-based motion capture system and its working tolerance, could be used for gait analysis at a distance without leading to different clinical interpretation.

Effectiveness of MicroShunt in Patients with Primary Open-Angle and Pseudoexfoliative Glaucoma: A Retrospective European Multicenter Study.

PURPOSE: To evaluate the effectiveness and safety of the Preserflo MicroShunt implant (Santen) in patients with primary open-angle glaucoma (POAG) and pseudoexfoliation glaucoma (PXG). DESIGN: Retrospective, open-label, multicenter study. PARTICIPANTS: Patients with insufficiently controlled primary POAG or PXG who underwent a standalone MicroShunt implantation procedure. METHODS: Consecutive patients with POAG and PXG who underwent surgery with the ab externo minimally invasive glaucoma surgery device Preserflo MicroShunt with mitomycin C. MAIN OUTCOME MEASURES: Primary end points were mean change in intraocular pressure (IOP) and number of hypotensive medications from baseline through month 12. Success was defined as an IOP of 18 mmHg or less and an IOP reduction of 20% or more, with (qualified) or without (complete) any hypotensive medication. RESULTS: Among the 130 patients who underwent MicroShunt implantation, 104 fulfilled the inclusion and exclusion criteria and were included in the analysis. Eighty-one eyes (77.9%) were diagnosed with POAG and 23 eyes (22.1%) were diagnosed with PXG. The mean age was 71.4 ± 12.6 years, and 45 patients (43.3%) were women. Mean IOP was lowered significantly from 25.1 ± 6.5 mmHg at baseline to 14.1 ± 3.4 mmHg at month 12 (P < 0.0001). At month 12, 27 eyes (26.0%) were categorized as complete successes and 61 eyes (58.7%) were categorized as qualified successes. The mean number of hypotensive medications was reduced significantly from 3.0 ± 1.0 medications at the preoperative visit to 0.77 ± 0.95 medication at month 12 (P < 0.001). Throughout the study, 19 eyes (18.3%) required needling and 14 eyes (13.5%) underwent surgical revision. Eight eyes (7.7%) showed hyphema and 5 eyes (4.8%) showed choroidal detachment. These were resolved with medical therapy without sequelae. Four patients underwent subsequent surgeries, and 2 patients underwent trabeculectomy (at months 3 and 6): One patient underwent transscleral cyclophotocoagulation at month 3 and 1 patient underwent MicroPulse cyclophotocoagulation at month 4. CONCLUSIONS: In this retrospective study, the MicroShunt effectively lowered IOP and the need for IOP-lowering medications.

European study of the efficacy of a cross-linked gel stent for the treatment of glaucoma.

PURPOSE: To investigate the efficacy and safety of a cross-linked gel stent (XEN45) with or without cataract surgery in the treatment of glaucoma patients. SETTING: Five university hospitals. DESIGN: Prospective multicenter clinical trial. METHODS: Patients with glaucoma inadequately controlled by treatment or poor compliance or intolerance to topical therapy were included. Patients were divided into those who had an implant only (Solo Group; phakic and pseudophakic patients) and those who had an implant combined with cataract surgery (Combo Group). Differences in mean intraocular pressure (IOP) and number of medications between the baseline preoperative visit and study end (12 months), and the rate of qualified and complete success, were evaluated. RESULTS: The Solo Group comprised 115 glaucoma patients (43 phakic and 72 pseudophakic) and the Combo Group comprised 56 patients. Compared with baseline, mean IOP (23.9 ± 7.6 to 15.5 ± 3.9) and number of medications (3.0 ± 1.1 to 0.5 ± 1.0) decreased significantly at 12-month follow-up (P < .001 for both). A >20% and >30% reduction in IOP from baseline was achieved by 72.3% and 52.6% of patients, respectively. IOP at 1 week postoperatively was a predictor of success, and the needling rate was inversely correlated with early postoperative IOP. The number of preoperative medications and patient age were not significantly associated with failure. CONCLUSIONS: Insertion of a cross-linked gel stent alone or combined with phacoemulsification might be effective and safe in the treatment of open-angle glaucoma, with a substantial reduction in IOP and number of medications.

Exploring the gap between diagnostic research outputs and clinical use of OCT for diagnosing glaucoma.

AIMS: To explore the gap between diagnostic research outputs and clinical use of optical coherence tomography (OCT) in glaucoma and assess the reliability of a specific reference database when applied to a morphological imaging parameter for diagnostic purposes. METHODS: Consecutive subjects enrolled in the Multicenter Italian Glaucoma Imaging Study (MIGIS) have been included in this cross-sectional, comparative evaluation of diagnostic tests study. Patients underwent measurement of global and sectorial peripapillary retinal nerve fibre thickness (pRNFL) and minimum rim width (MRW) by OCT. The sensitivity and specificity of reference-database categorical classifications were calculated by means of 2×2 tables and sensitivity was compared with that of the corresponding continuous parameter extracted from the receiver operating characteristic (ROC) curves by matching the specificity. RESULTS: 280 Caucasian subjects have been included. At matched specificities, the sensitivity of pRNFL categorical classifications was statistically similar to that of the corresponding continuous parameters, whereas the sensitivity of the MRW categorical classifications was significantly lower than that of the corresponding continuous parameters. CONCLUSIONS: The diagnostic accuracy of reference database classifications might be lower than that extrapolated from the ROC curves of continuous parameters used in diagnostic research. The gap between the accuracy of these two approaches may be used to estimate the reliability of a specific reference database when applied to a continuous parameter for diagnostic purposes.

Macular versus nerve fibre layer versus optic nerve head imaging for diagnosing glaucoma at different stages of the disease: Multicenter Italian Glaucoma Imaging Study.

PURPOSE: To compare the diagnostic accuracy of minimum rim width (MRW), peripapillary retinal nerve fibre layer (pRNFL) and multilayered macular analysis by Spectralis SD-OCT (Heidelberg Engineering, Germany) in discriminating perimetric glaucoma at different stages of the disease from healthy eyes. METHODS: In this multicentre, prospective, evaluation of diagnostic tests study, multilayered macular analysis and MRW and pRNFL were obtained from one eye of 197 glaucoma (76 early, 68 moderate and 53 advanced) and of 83 healthy controls from the Multicenter Italian Glaucoma Imaging Study (MIGIS). The reference standard for classifying eyes as glaucomatous and for staging the disease was the visual field. The main outcome measures were area under the ROC curve (AUC) and sensitivity at fixed specificity (95%). RESULTS: Average MRW and average pRNFL showed the highest and similar diagnostic accuracy in both the whole study population (AUC 0.968 and 0.939) and early glaucoma (AUC 0.956 and 0.929). Among the macular parameters, the three innermost retinal layers combined as the Ganglion Cell Complex provided the highest diagnostic accuracy (AUC 0.931) in the whole population, which was statistically similar to that of average pRNFL but inferior to that of average MRW. Compared to both average MRW and pRNFL, all macular parameters showed statistically significant lower accuracy in early glaucoma, but accuracy in moderate and advanced glaucoma showed no statistically significant differences among three protocols. CONCLUSION: The diagnostic accuracy of MRW, pRNFL and macular analysis by Spectralis SD-OCT is overall good. MRW and pRNFL analysis performs statistically and clinically better than macular analysis to discriminate early glaucoma from healthy eyes.

Central corneal thickness and glaucoma treatment: an Italian multicenter cross-sectional study.

PURPOSE: Supposedly, prostaglandin analogs (PGA) could reduce the central corneal thickness (CCT), while topical carbonic anhydrase inhibitors (TCAI) could increase CCT. The aim of the study was to evaluate clinically significant CCT effects in patients treated with PGA or TCAI. METHODS: At least 50 glaucomatous patients were saved on the Italian Glaucoma Register from 16 different glaucoma centers. About 816 glaucomatous patients were found in the register; of these, 316 were recruited in this study because they were treated with PGA or TCAI. The diagnosis of glaucoma was based on visual field examination, optic nerve head analysis, intraocular pressure (IOP) measurements, and gonioscopy. Two age-matched subgroups were created: one treated with PGA and the other with TCAI. CCT, ophthalmoscopic cup/disc ratio (CDR), mean deviation (MD), pattern standard deviation (PSD), and IOP were considered for both eyes of each patient. Student t-test was used to compare the 2 subgroups. RESULTS: The mean age of the PGA group was 66.35±12.17 years, while 65.17±12.52 years was for the TCAI group. No significant difference was found for CCT (543.75±35 µm and 544±35 µm, respectively), CDR (0.55±0.2 and 0.53±0.2, respectively), MD (-4.5±4.9 dB and -5.4±6.4 dB, respectively), PSD (4.6±3.4 and 4.6±4.9, respectively), and IOP (15.9±3.3 mmHg and 15.7±2.9 mmHg, respectively) between the 2 subgroups. A significant (P<0.001) correlation was found between CCT and CDR and between CCT and IOP. CONCLUSION: No significant difference in CCT was found between patients treated with PGA and TCAI, suggesting that these topical medications did not statistically and clinically change the CCT.

Lidocaine versus ropivacaine for topical anesthesia in cataract surgery(1).

PURPOSE: To assess the anesthetic efficacy and safety of topical ropivacaine versus topical lidocaine in cataract surgery. SETTING: Institute of Ophthalmology, University of Modena and Reggio Emilia, Modena, Italy. METHODS: This prospective controlled randomized double-blind study comprised 64 patients scheduled for planned routine cataract extraction. Patients were randomized into 2 groups; 1 received topical ropivacaine 1% and the other, topical lidocaine 4%. The duration of surgery, intraoperative and early postoperative complications, and the need for supplemental intracameral anesthesia were recorded. Intraoperative and postoperative subjective pain was quantified by patients using a scale from 1 to 10. An endothelial cell count was performed preoperatively and 2 months after surgery. RESULTS: The mean endothelial cell density decreased from 2334 cells/mm(2) +/- 496 (SD) to 2016 +/- 674 cells/mm(2) in the ropivacaine group and from 2519 +/- 404 cells/mm(2) to 1847 +/- 607 cells/mm(2) in the lidocaine group. The difference in cell density between groups was not significant before (P =.154) or after surgery (P =.329); however, the difference in mean cell loss between groups was statistically significant (P =.031). The duration of surgery and intraoperative complications were the same in both groups. Four patients in the ropivacaine group and 5 in the lidocaine group required supplemental anesthesia (P >.05). The mean subjective analog pain score was slightly higher in the lidocaine group (P >.05). The day after surgery, 12 eyes in the ropivacaine group and 6 in the lidocaine group had transient corneal edema (P =.150). CONCLUSIONS: Topical ropivacaine performed at least as well as topical lidocaine in efficacy and safety in cataract surgery. It provided sufficient and long-lasting analgesia without the need for supplemental intracameral anesthesia in most cases.