Hit or miss: The Dilemma of Specimen Selection for Microbiological Diagnosis of Rhino-Orbital-Cerebral- COVID-Associated Mucormycosis (ROCM-CAM).

BACKGROUND: We diagnosed various cases of rhino-orbital-cerebral- COVID-associated Mucormycosis (ROCM-CAM) during India's second wave of COVID-19. This helped formulate novel suggestions for improving laboratory output, applicable anywhere in the world. METHOD: To diagnose ROCM-CAM by microbiological methods, we used direct microscopy and conventional culture on various clinical samples within the shortest turn-around time. DESIGN: Prospective single-center observational study Participants: patients with ROCM-CAM Results: Of 113 suspected cases of ROCM-CAM during May 2021, direct microscopy and culture could confirm the disease in 87.61% and 44.25% of patients, respectively. The highest pathogen isolation was seen from maxillary bone fragments, FESS-guided biopsy from pterygopalatine fossae, nasal turbinates and nasal mucosal biopsy. Direct microscopy could diagnose the disease in almost 40% of patients within 24 hours and 60% within two days. Conventional cultures yielded Rhizopus spp. (86%) as the commonest fungal pathogen followed by Mucor spp. (12%) within 7 days. Deep tissue biopsies are more useful for rapid diagnosis than superficial specimens. Routine fungal cultures can supplement case detection and help prognosticate survivors. CONCLUSION: The management of ROCM is a surgical emergency. The diagnosis of the condition must therefore be prompt and precise. Despite ongoing antifungal therapy, nasal mucosal tissue, FESSguided, and intra-operative tissue biopsies showed the pathogen's highest diagnostic yield. The diagnostic index improved further when multiple (4-5) high-quality specimens were collected. Nasal swabs and crusts, among the most commonly requested specimens worldwide, were found to have an overall low diagnostic potential.

Deep Learning Assisted COVID-19 Detection using full CT-scans

The ongoing pandemic of COVID-19 has shown the limitations of our current medical institutions. There is a need for research in automated diagnosis for speeding up the process while maintaining accuracy and reducing the computational requirements. This work proposes an automated diagnosis of COVID-19 infection from CT scans of the patients using deep learning technique. The proposed model, ReCOV-101, uses full chest CT scans to detect varying degrees of COVID-19 infection. To improve the detection accuracy, the CT-scans were preprocessed by employing segmentation and interpolation. The proposed scheme is based on the residual network that takes advantage of skip connection, allowing the model to go deeper. The model was trained on a single enterprise-level GPU. It can easily be provided on a network’s edge, reducing communication with the cloud, often required for larger neural networks. This work aims to demonstrate a less hardware-intensive approach for COVID-19 detection with excellent performance that can be combined with medical equipment and help ease the examination procedure. With the proposed model, an accuracy of 94.9% was achieved.