KFWC: A Knowledge-Driven Deep Learning Model for Fine-grained Classification of Wet-AMD.

BACKGROUND AND OBJECTIVES: Automated diagnosis using deep neural networks can help ophthalmologists detect the blinding eye disease wet Age-related Macular Degeneration (AMD). Wet-AMD has two similar subtypes, Neovascular AMD and Polypoidal Choroidal Vasculopathy (PCV). However, due to the difficulty in data collection and the similarity between images, most studies have only achieved the coarse-grained classification of wet-AMD rather than a fine-grained one of wet-AMD subtypes. Therefore, designing and building a deep learning model to diagnose neovascular AMD and PCV is a great challenge. METHODS: To solve this issue, in this paper, we propose a Knowledge-driven Fine-grained Wet-AMD Classification Model (KFWC) to enhance the model's accuracy in the fine-grained disease classification with insufficient data. We innovatively introduced a two-stage method. In the first stage, we present prior knowledge of 10 lesion signs through pre-training; in the second stage, the model implements the classification task with the help of human knowledge. With the pre-training of priori knowledge of 10 lesion signs from input images, KFWC locates the powerful image features in the fine-grained disease classification task and therefore achieves better classification. RESULTS: To demonstrate the effectiveness of KFWC, we conduct a series of experiments on a clinical dataset collected in cooperation with a Grade III Level A ophthalmology hospital in China. The AUC score of KFWC reaches 99.71%, with 6.69% over the best baseline and 4.14% over ophthalmologists. KFWC can also provide good interpretability and effectively alleviate the pressure of data collection and annotation in the field of fine-grained disease classification for wet-AMD. CONCLUSIONS: The model proposed in this paper effectively solves the difficulties of small data volume and high image similarity in the wet-AMD fine-grained classification task through a knowledge-driven approach. Besides, this method effectively relieves the pressure of data collection and annotation in the field of fine-grained classification. In the diagnosis of wet-AMD, KFWC is superior to previous work and human ophthalmologists.

The Relationship between VEGF-C, TAM, and Lymph Node Metastasis in Oral Cancer.

Objective: To investigate the relationship between vascular endothelial growth factor-C (VEGF-C) and tumor-associated macrophages in oral cancer (TAMs) with lymph node metastasis. Method: From January 2018 to January 2022, 155 cases of oral cancer tissues and 165 cases of normal mucosal tissues were collected from oral surgical resection tissues or biopsy specimens in Hebei Eye Hospital. Oral cancer tissues were observed. The control group had normal mucosal tissues. The clinical and immune parameters were observed and the treatment of oral cancer is also briefly discussed. Results: The number of TAMs and the expression of VEGF-C in oral cancer tissues were significantly higher than those in normal tissues (P < 0.05). The lymphatic vessel density, the number of TAMs, and the expression of VEGF-C in the metastatic group were higher than in nonmetastatic group, and the lymphatic vessel density, the number of TAMs, and the expression of VEGF-C in the paracancerous tissues were higher than central tumor tissue in the metastatic group (P < 0.05). Univariate analysis showed that the number of TAMs was related to the histological stage and the pathological type of oral cancer (P > 0.05). The expression of VEGF-C was associated with the histological stage of oral cancer (P < 0.05). Compared with the immune function after different treatments, the contents of CD4+ in both groups was higher than before, and the combined treatment group was increased more than single treatment group (P < 0.05). The contents of CD3+ and CD8+ in the two groups were lower than before, and the combined treatment group was decreased higher than combined treatment group (P < 0.05). Conclusions: The number of TAMs and the expression of VEGF-C in oral cancer tissues are higher than normal tissues. The number of TAMs and the expression of VEGF-C are higher in patients with lymph node metastasis. TAMs and VEGF-C may play an important role in lymph node metastasis of oral cancer. Integrated traditional Chinese and Western medicine can improve the immune function of patients with oral cancer and may improve the therapeutic efficacy.

Comparison of effectiveness between complete internal limiting membrane peeling and internal limiting membrane peeling with preservation of the central fovea in combination with 25G vitrectomy for the treatment of high myopic foveoschisis.

Vitrectomy combined with internal limiting membrane peeling (ILMP) treats high myopic foveoschisis with good results, but there is a risk of iatrogenic macular holes, which may be reduced by preserving the internal limiting membrane of the central fovea. This study compared complete ILMP with partial ILMP, both combined with 25G vitrectomy, for the treatment of high myopic foveoschisis.Thirty-three cases (33 eyes) of high myopic foveoschisis were retrospectively analyzed. They underwent 25G vitrectomy combined with C3F8 gas filling; Group A (20 patients, 20 eyes) underwent complete ILMP, while Group B (13 patients, 13 eyes) underwent partial ILMP to preserve the internal limiting membrane of the central fovea. Patients were followed up for 6 months and underwent best corrected visual acuity (BCVA), intraocular pressure, diopter, slit lamp microscopy, indirect ophthalmoscope, A-mode and/or B-mode ultrasound, and optical coherence tomography. Surgical complications were recorded.Up to the last follow-up, the BCVA improved and central fovea thickness (CFT) was lower compared with before surgery in both groups (All P < .05). There were no significant differences in BCVA and CFT at each time point between the groups (P > .05). Most of the postoperative retinas of the 2 groups were completely reattached, with disappearance of the macular retinoschisis cavity. Two patients in the Group A and none in the Group B developed a macular hole during follow-up (P = .508).The results did not support the superiority of partial ILMP over complete ILMP in reduced incidence of macular hole. Both methods had a similar curative effect.