Methods: The accuracy of Webgazer.js for software-based gaze tracking is tested under different lighting conditions. Predefined time delays of a prototype diagnosis task automation script are contrasted against with manual delays based on human time estimation to understand how automation influences diagnosis accuracy. SLI diagnosis binary classifier was built and tested based on randomised parameters. The obtained results were cross-compared to Singlims_ES.exe for equality.

Results: Webgazer.js achieved an average accuracy of 88.755% under global lighting conditions, 61.379% under low lighting conditions and 52.7% under face-focused lighting conditions. The diagnosis task automation script found to execute with actual time delays with a deviation percentage no more than 0.04%, while manually executing time delays based on human time estimation resulted in a deviation percentage of not more than 3.37%. One-tailed test probability value produced by both the newly built classifier and Singlims_ES were observed to be similar up to three decimal places.

METHODS: The pre- and post-operative CT images of 55 patients undergoing DC surgery were analyzed. The ICV was measured by segmenting every slice of the CT images, and compared with estimated ICV calculated using the 1-in-10 sampling strategy and processed using the SBI method. An independent t test was conducted to compare the ICV measurements between the two different methods. The calculation using this method was repeated three times for reliability analysis using the intraclass correlations coefficient (ICC). The Bland-Altman plot was used to measure agreement between the methods for both pre- and post-operative ICV measurements.

RESULTS: The mean ICV (±SD) were 1341.1±122.1ml (manual) and 1344.11±122.6ml (SBI) for the preoperative CT data. The mean ICV (±SD) were 1396.4±132.4ml (manual) and 1400.53±132.1ml (SBI) for the post-operative CT data. No significant difference was found in ICV measurements using the manual and the SBI methods (p=.983 for pre-op, and p=.960 for post-op). The intrarater ICC showed a significant correlation; ICC=1.00. The Bland-Altman plot showed good agreement between the manual and the SBI method.

METHODS: A double-blind quasi-experiment was carried out on NC (n = 43) and NCI (n = 33) groups. Participants in each group were randomly assigned into treatment and control programs groups. The treatment group underwent auditory-cognitive training, whereas the control group was assigned to watch documentary videos, three times per week, for 8 consecutive weeks. Study outcomes that included Montreal Cognitive Assessment, Malay Hearing in Noise Test, Dichotic Digit Test, Gaps in Noise Test and Pitch Pattern Sequence Test were measured at 4-week intervals at baseline, and weeks 4, 8 and 12.

MATERIALS AND METHODS: This was a retrospective study using computed tomography (CT) scans from 3 hospitals. Inclusion criteria were scans with 1-5 nodules of diameter ≥5 mm; exclusion criteria were poor-quality scans or those with nodules measuring <5mm in diameter. In the lesion detection phase, 2,147 nodules from 219 scans were used to develop and train the deep learning 3D-CNN to detect lesions. The 3D-CNN was validated with 235 scans (354 lesions) for sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) analysis. In the path planning phase, Bayesian optimization was used to propose possible needle trajectories for lesion biopsy while avoiding vital structures. Software-proposed needle trajectories were compared with actual biopsy path trajectories from intraprocedural CT scans in 150 patients, with a match defined as an angular deviation of <5° between the 2 trajectories.

RESULTS: The model achieved an overall AUC of 97.4% (95% CI, 96.3%-98.2%) for lesion detection, with mean sensitivity of 93.5% and mean specificity of 93.2%. Among the software-proposed needle trajectories, 85.3% were feasible, with 82% matching actual paths and similar performance between supine and prone/oblique patient orientations (P = .311). The mean angular deviation between matching trajectories was 2.30° (SD ± 1.22); the mean path deviation was 2.94 mm (SD ± 1.60).