Search

Objective To establish a fungal image-assisted classification model using deep learning technology. Methods The microscope images of people infected with Aspergillus, Saccharomyces and Cryptococcus neoformans were retrospectively collected from the Eighth Medical Center of PLA General Hospital from September 2020 to April 2021. The images were randomly divided into training set, validation set and test set according to the ratio of 7∶1.5∶1.5. The improved MobileNetV2 network structure was trained using the training set, a convolutional neural network (CNN) fungal image 11 classification model based on multi-scale attention mechanism was constructed and the parameters were debugged based on the validation set. Machine identification results were taken as the gold standard, the performance of the model on 11 fungal image classification tasks was evaluated, and the results were shown by precision, recall and F1 value. In addition, the performance of the proposed model with 5 classic CNN models were compared, and the results were measured in terms of model parameters, memory usage, frames per second (FPS), accuracy, and area under the curve (AUC) of receiver operating characteristic curve. Results A total of 7666 fungal microscope images were collected, including 2781, 4115, and 770 images of Aspergillus, Saccharomyces, and Cryptococcus neoformans, respectively. Among them, there were 5366 training images, 1150 validation images, and 1150 test images. The improved MobileNetV2 model had high performance for the classification of 11 fungal images in the test set. The precision rate was distributed between 96.36% and 100%, the recall rate was distributed between 96.53% and 100%, and the F1 value was distributed between 97.01% and 100%. The parameters, memory usage, FPS, accuracy, and AUC of the improved MobileNetV2 model were 4.22 M, 356.89 M, 573, (99.09±0.18)%, and 0.9944±0.0018, respectively, and the comprehensive performance was better than 5 kinds of classical networks. Conclusion The proposed fungal image classification model based on the improved MobileNetV2 can obtain higher fungal image recognition ability while maintaining low computational cost, with an overall performance better than classical CNN model.

Purpose: To determine the clinical features of traumatic ciliochoroidal detachment (CCD), and to evaluate the surgical outcomes., Methods: We retrospectively reviewed the records of 37 consecutive patients with traumatic CCD who underwent surgical procedures, including ciliary body suturing, transscleral cyclophotocoagulation, and cyclocryopexy. A complete ocular examination was performed at pre-surgery and at periodical post-surgery follow-ups. We compared visual acuity (VA), intraocular pressure (IOP), and morphologic changes with UBM among the different surgical procedures at the pre-surgery and periodical follow-ups., Results: The mean IOP was 6.62 mmHg, and the median VA was 20/200 at baseline. The mean final IOP was 11.03 mmHg, and the final median VA improved to 20/50. IOPs were significantly different in post-surgery compared with those at baseline (P = 0.000) among the ciliary body suturing, cyclophotocoagulation, and cyclocryopexy groups. However, no significant differences were noted at each follow-up among the 3 groups (P > 0.05). The post-surgical morphological figures consisted of complete reattachment, partial reattachment, and the complete detachment. Cyclocryopexy (71.4%), suturing (68.4%), and cyclophotocoagulation (63.6%) produced similar surgical outcomes of the complete reattachment based on UBM images., Conclusion: Prompt treatment and periodic follow-ups are necessary after traumatic CCD, based on accurate dimensions and configuration by UBM. The appropriate choice of surgical procedures is pivotal for an optimal outcome.

Purpose: To investigate postoperative inhibition of cell proliferation in animals treated with exogenous p27Kip1 during an experimental glaucoma filtration surgery (GFS) by histological and ultrastructural examinations., Methods: The primer was designed according to p27Kip1 gene sequence of GenBank. The p27Kip1-expressing adenovirus-mediated Ad-p27 was constructed according to standard techniques. Gene therapy was performed by subconjunctival delivery of Ad-p27, mitomycin C (MMC) and PBS served as controls. Histological and ultastructural changes at surgical sites were observed for 28 days postoperatively., Results: Histologically, evident cellular proliferation induced by Ad-p27 was observed on day 7, with appreciable filtering cavity. However, thin conjunctival layers were noted, and the number of fibroblasts decreased in the Tenon's capsule on day 14. On day 28, the filtering cavity partially disappeared. The histological features of the animals with MMC delivery were similar to those treated by Ad-p27. In contrast, in the PBS treatment group, thick conjunctival layers were observed, and the number of goblet cells and fibroblasts, increased markedly. The filtering cavity disappeared at postoperative day 21. Twenty eight day postoperatively, ultrastructural findings showed that most conjunctival epithelial cells in the Ad-p27 treatment group were under static state, and that organelles were inactive. The number of goblet cells, and the secretion of mucin, were decreased. The amount of fibroblasts also decreased with partial apoptotic cells. The ultrastructural features presented by rabbits with the MMC delivery resembled those treated with Ad-p27. In the PBS treatment group, the number of conjunctival epithelial cells increased; the amount of goblet cells increased as the secretion of mucin was strengthened, and a substantial amount of fibroblasts were observed with active and productive organelles., Conclusion: Ad-p27 inhibits the proliferation of Tenon's capsule fibroblasts at surgical sites, and suppresses the formation of scars, thereby promoting surgical efficacy.