Laryngeal disease classification is a relatively hard task in medical image processing resulting from its complex structures and varying viewpoints in data collection. Some existing methods try to tackle this task via the convolutional neural network, but they more or less ignore the intrinsic difficulty differences among different input samples and suffer from high training complexity. In order to better resolve these problems, an end-to-end Hierarchical Dynamic Convolutional Network (HDCNet) is proposed, which can dynamically process the input samples based on their difficulty. For the easy-classified samples, the HDCNet processes them with a smaller resolution and a relatively small network, while the difficult samples are passed to a large network with a larger resolution for more accurate classification results. Furthermore, a Feature Reuse Module (FRM) is designed to transfer the features learned by the small network to the corresponding block in the deep network to enhance the overall performance of some rather complicated samples. To validate the effectiveness of the proposed HDCNet, comprehensive experiments are conducted on the public available laryngeal disease classification dataset and HDCNet provides superior performances compared with other current state-of-the-art methods.

Objective: To compare the recurrence rate of retinopathy of prematurity (ROP) after treatment with 0.3 mg vs. 0.25 mg ranibizumab. Subjects: All patients with ROP who underwent intravitreal injection of ranibizumab in Hainan General Hospital between January 2014 and May 2020 were included in this retrospective study. Methods: Eighty-two cases (146 eyes) who received intravitreal injection of 0.25 mg ranibizumab were included in the conventional-dose group, and 59 cases (108 eyes) who received intravitreal injection of 0.3 mg ranibizumab were included in the high-dose group. The two groups were further divided into the 25-28-week, 29-31-week, 32-34-week, and 35-36-week GA subgroups. The differences between the conventional-dose group and the high-dose group in gestational age (GA), birth weight (BW), age at initial injection (weeks), incidence of systemic diseases, the recurrence rate of ROP, and age at retinal vascularization completed (weeks) were analyzed. Results: GA, BW, age at initial injection, and the incidence of systemic diseases were not significantly different between the conventional-dose group and the high-dose group (p > 0.05). The recurrence rates of ROP were significantly lower in the 25-28-week, 29-31-week, and 32-34-week subgroups of the high-dose group than in the same subgroups of the conventional-dose group (p < 0.05). Within the conventional-dose group, the recurrence rate of ROP was significantly lower in the 32-34-week and 35-36-week subgroups than in the 25-28-week and 29-31-week subgroups (p < 0.05). Within the high-dose group, the recurrence rate of ROP was not significantly different between the four subgroups (p > 0.05). Retinal vascularization was completed at a later age in the 32-34-week subgroup of the high-dose group than in the 32-34-week subgroup of the conventional-dose group (p < 0.05) but was not significantly different between the two groups at any other GA range (p > 0.05). No severe ocular or systemic complications occurred in any patient. Conclusion: Treatment with 0.3 mg ranibizumab can reduce the recurrence rate of ROP without prolonging retinal vascularization or causing serious systemic complications. Therefore, this dose may be an appropriate therapeutic dose for ROP.