Classification of fungal genera from microscopic images using artificial intelligence.

Microscopic image examination is fundamental to clinical microbiology and often used as the first step to diagnose fungal infections. In this study, we present classification of pathogenic fungi from microscopic images using deep convolutional neural networks (CNN). We trained well-known CNN architectures such as DenseNet, Inception ResNet, InceptionV3, Xception, ResNet50, VGG16, and VGG19 to identify fungal species, and compared their performances. We collected 1079 images of 89 fungi genera and split our data into training, validation, and test datasets by 7:1:2 ratio. The DenseNet CNN model provided the best performance among other CNN architectures with overall accuracy of 65.35% for top 1 prediction and 75.19% accuracy for top 3 predictions for classification of 89 genera. The performance is further improved (>80%) after excluding rare genera with low sample occurrence and applying data augmentation techniques. For some particular fungal genera, we obtained 100% prediction accuracy. In summary, we present a deep learning approach that shows promising results in prediction of filamentous fungi identification from culture, which could be used to enhance diagnostic accuracy and decrease turnaround time to identification.

Invasive Nocardiosis Versus Colonization at a Tertiary Care Center: Clinical and Radiological Characteristics.

Objective: To describe the clinical and radiographic findings in a large cohort of patients with positive cultures for Nocardia emphasizing the differences between invasive disease and colonization. Patients and Methods: We conducted a single-center, retrospective cohort study of 133 patients with a positive Nocardia isolate between August 1, 1998, and November 30, 2018, and a computed tomography (CT) of the chest within 30 days before or after the bacteria isolation date. Results: Patients with colonization were older (71 vs 65 years; P=.004), frequently with chronic obstructive pulmonary disease (56.8% vs 16.9%; P<.001) and coronary artery disease (47.7% vs 27%, P=.021), and had Nocardia isolated exclusively from lung specimens (100% vs 83.1%; P=.003). On CT of the chest, they had frequent airway disease (84.1% vs 51.7%; P<.001). Patients with invasive nocardiosis had significantly (P<.05) more diabetes, chronic kidney disease, solid organ transplant, use of corticosteroids, antirejection drugs, and prophylactic sulfa. They had more fever (25.8% vs 2.3%; P<.001), cutaneous lesions (14.6% vs 0%; P=.005), fatigue (18% vs 0%; P=.001), pulmonary nodules (52.8% vs 27.3%; P=.006), and free-flowing pleural fluid (63.6% vs 29.4%; P=.024). The patterns of nodule distribution were different-diffuse for invasive nocardiosis and peribronchiolar for Nocardia colonization. Conclusion: The isolation of Nocardia in sputum from a patient with respiratory symptoms does not equal active infection. Only by combining clinical and chest CT findings, one could better differentiate between invasive nocardiosis and Nocardia colonization.

Cladophialophora bantiana and Nocardia farcinica infection simultaneously occurring in a kidney transplant recipient: Case report and literature review.

Solid organ transplant recipients are at increased risk of acquiring devastating infections with unusual pathogens. Nocardia are aerobic actinomycetes that affect the lungs, brain, skin and soft tissue. Cladophialophora species are dematiaceous fungi that overwhelmingly cause infections in the brain. Both organisms carry a high mortality rate. We present the first reported renal transplant case with Cladophialophora bantiana involving the renal allograft with concurrent invasive nocardiosis involving the lungs and brain.

Invasive Nocardiosis in Transplant and Nontransplant Patients: 20-Year Experience in a Tertiary Care Center.

OBJECTIVE: To present the clinical characteristics and outcome of transplant and nontransplant patients with invasive nocardiosis. PATIENTS AND METHODS: We conducted a retrospective chart review of 110 patients 18 years and older diagnosed with culture-proven invasive nocardiosis (defined as the presence of clinical signs and/or radiographic abnormalities) between August 1, 1998, and November 30, 2018. Information on demographic, clinical, radiographic, and microbiological characteristics as well as mortality was collected. RESULTS: One hundred ten individuals with invasive nocardiosis were identified, of whom 54 (49%) were transplant and 56 nontransplant (51%) patients. Most transplant patients were kidney and lung recipients. The overall mean age was 64.9 years, and transplant patients had a higher prevalence of diabetes and chronic kidney disease. A substantial proportion of nontransplant patients were receiving corticosteroids (39%), immunosuppressive medications (16%), and chemotherapy (9%) and had chronic obstructive pulmonary disease (20%), rheumatologic conditions (18%), and malignant neoplasia (18%). A higher proportion of transplant patients (28%) than nontransplant patients (4%) received trimethoprim-sulfamethoxazole prophylaxis. In both groups, the lung was the most common site of infection. Seventy percent of all Nocardia species isolated were present in almost equal proportion: N brasiliensis (16%), N farcinica (16%), N nova (15%), N cyriacigeorgia (13%), and N asteroides (11%). More than 90% of isolates were susceptible to trimethoprim-sulfamethoxazole, linezolid, and amikacin. There was no significant difference in mortality between the 2 groups at 1, 6, and 12 months after the initial diagnosis. CONCLUSION: The frequency of invasive Nocardia infection was similar in transplant and nontransplant patients and mortality at 1, 6, and 12 months was similar in both groups. Trimethoprim-sulfamethoxazole prophylaxis failed to prevent Nocardia infection.