AI aiding in diagnosing, tracking recovery of COVID-19 using deep learning on Chest CT scans.

Coronavirus (COVID-19) has spread throughout the world, causing mayhem from January 2020 to this day. Owing to its rapidly spreading existence and high death count, the WHO has classified it as a pandemic. Biomedical engineers, virologists, epidemiologists, and people from other medical fields are working to help contain this epidemic as soon as possible. The virus incubates for five days in the human body and then begins displaying symptoms, in some cases, as late as 27 days. In some instances, CT scan based diagnosis has been found to have better sensitivity than RT-PCR, which is currently the gold standard for COVID-19 diagnosis. Lung conditions relevant to COVID-19 in CT scans are ground-glass opacity (GGO), consolidation, and pleural effusion. In this paper, two segmentation tasks are performed to predict lung spaces (segregated from ribcage and flesh in Chest CT) and COVID-19 anomalies from chest CT scans. A 2D deep learning architecture with U-Net as its backbone is proposed to solve both the segmentation tasks. It is observed that change in hyperparameters such as number of filters in down and up sampling layers, addition of attention gates, addition of spatial pyramid pooling as basic block and maintaining the homogeneity of 32 filters after each down-sampling block resulted in a good performance. The proposed approach is assessed using publically available datasets from GitHub and Kaggle. Model performance is evaluated in terms of F1-Score, Mean intersection over union (Mean IoU). It is noted that the proposed approach results in 97.31% of F1-Score and 84.6% of Mean IoU. The experimental results illustrate that the proposed approach using U-Net architecture as backbone with the changes in hyperparameters shows better results in comparison to existing U-Net architecture and attention U-net architecture. The study also recommends how this methodology can be integrated into the workflow of healthcare systems to help control the spread of COVID-19.

Primary 23-gauge sutureless vitrectomy for rhegmatogenous retinal detachment.

AIMS: To report a prospective non-comparative consecutive interventional study on the safety and efficacy of 23-Gauge transconjunctival sutureless pars plana vitrectomy for primary rhegmatogenous retinal detachment (RRD). MATERIALS AND METHODS: Fifty eyes of 50 consecutive patients were recruited between June 2007 and January 2008. All surgeries were performed using the one-step 23-Gauge system with angled incisions. The surgical protocol consisted of a minimum of eight clinical visits: baseline, 1 day, 1 week, 1-, 3- and 6- months after the initial surgery. The endpoints were anatomical, functional results and complications arising from the surgery. RESULTS: Anatomical success was achieved in 82% of cases (41 out of 50) with single surgery and rose to 98% (49 out of 50) with additional surgery. Mean visual acuity improved from logMAR 0.48(SD 0.36) to 0.26(SD 0.31), P < 0.001. Two cases with ocular hypotony, defined as an intraocular pressure ≤ 6mmHg, that were associated with a choroidal detachment were seen. CONCLUSIONS: Acceptable anatomical and functional success rates can be achieved with primary 23-Gauge transconjunctival sutureless vitrectomy for RRD. We found that the approach technique is different from conventional vitrectomy and the complications arising from post surgical hypotony and leakage from sclerotomies are potentially higher compared to 20-Gauge vitrectomy.

Does the presence of an epiretinal membrane alter the cleavage plane during internal limiting membrane peeling?

PURPOSE: To determine whether the presence of a clinically and/or microscopically detectable epiretinal membrane (ERM) alters the cleavage plane during internal limiting membrane (ILM) peeling. DESIGN: Retrospective, observational, immunohistochemical study of ILM specimens using archival formalin-fixed, paraffin-embedded tissue. PARTICIPANTS: Fifty-one patients who had had ILM excision. METHODS: Fifty-one ILM specimens peeled during vitrectomy for various etiologies were examined by light microscopy. The removal of ILM was assisted using Trypan blue (n = 30), indocyanine green (n = 7), or brilliant blue G (n = 14). Monoclonal antibodies to glial fibrillary acidic protein and to neurofilament protein were used to detect glial or neuronal cells respectively on the vitreous or retinal surfaces of the ILM. Specimens were divided into 2 groups: ILM peeled for full-thickness macular hole (MH; n = 31) and ILM peeled after removal of clinically detectable ERM (n = 20). MAIN OUTCOME MEASURES: Primary outcome measure was the localization of immunohistochemical markers to neuronal or glial cells on the vitreous or retinal surfaces of ILM. The secondary outcome measure was the correlation of the results of the primary measure with the dyes used to facilitate ILM peeling. RESULTS: Glial and/or neuronal cells were detected on the retinal surface of the ILM in 10 of 31 (32%) of the MH ILM specimens and in 13 of 20 (65%) of the ILM peeled after ERM excision; the difference was significant (P = 0.02). There was no association between the presence of neuronal and glial cells with the type of dye used (P = 0.2). Of the 23 ILM specimens with cells attached to the retinal surface, 21 (91%) were associated with clinical and/or histologic evidence of ERM and 2 (9%) were not. The correlation between the presence of cells on the vitreous and the retinal surfaces of ILM was high (P<0.0001). CONCLUSIONS: The findings suggest that ERM may be associated with sub-ILM changes that alter the plane of separation during ILM peeling. This study does not confirm any influence of dyes on the cleavage plane during surgery.