Objective To develop and assess a machine learning model using the Shapley additive explanations(SHAP)method to predict fracture risk in multiple myeloma(MM)patients.Methods A retrospective study analyzed 181 MM patients in Zhejiang University Medical School Affiliated First Hospital from June 2021 to June 2024.Data included patient information,lab tests,medical history,and disease staging.Univariate analysis and recursive feature elimination(RFE)were employed for the purpose of variable selection.Predictive models were developed utilizing extreme gradient boosting(XGBoost),random forest(RF),light gradient boosting machine(LightGBM),and Logistic regression(LR).The performance of these models was evaluated through 5-fold cross-validation,and SHAP values were utilized to assess variable contributions in the optimal model.Results A total of 181 MM patients were included,with 50 in fracture group and 131 in non fracture group.RFE identified five key variables,notably including ferritin and B-type natriuretic peptide.The area under receiver operating characteristic curve values for the XGBoost,RF,LightGBM,and LR models were 0.861,0.846,0.755,and 0.780,respectively,with XGBoost demonstrating superior performance.SHAP analysis revealed that B-type natriuretic peptide was the most influential variable in the XGBoost model.Conclusion The XGBoost model demonstrates efficacy in predicting fracture risk among MM patients,with SHAP values enhancing its interpretability.

Objective To develop and assess a machine learning model using the Shapley additive explanations(SHAP)method to predict fracture risk in multiple myeloma(MM)patients.Methods A retrospective study analyzed 181 MM patients in Zhejiang University Medical School Affiliated First Hospital from June 2021 to June 2024.Data included patient information,lab tests,medical history,and disease staging.Univariate analysis and recursive feature elimination(RFE)were employed for the purpose of variable selection.Predictive models were developed utilizing extreme gradient boosting(XGBoost),random forest(RF),light gradient boosting machine(LightGBM),and Logistic regression(LR).The performance of these models was evaluated through 5-fold cross-validation,and SHAP values were utilized to assess variable contributions in the optimal model.Results A total of 181 MM patients were included,with 50 in fracture group and 131 in non fracture group.RFE identified five key variables,notably including ferritin and B-type natriuretic peptide.The area under receiver operating characteristic curve values for the XGBoost,RF,LightGBM,and LR models were 0.861,0.846,0.755,and 0.780,respectively,with XGBoost demonstrating superior performance.SHAP analysis revealed that B-type natriuretic peptide was the most influential variable in the XGBoost model.Conclusion The XGBoost model demonstrates efficacy in predicting fracture risk among MM patients,with SHAP values enhancing its interpretability.

Objective: To describe the procedure for early corneal epithelial basement membrane(EBM) repair and regeneration in rabbits with corneal penetrating injury. Methods: Forty-two New Zealand white rabbits were divided into modeling 1-, 3-, 5-, 7-, 14-, 21-, and 30-day groups using a random number table method, with 6 rabbits in each group; the right eyes were selected as the experimental eyes.Another 6 New Zealand white rabbits without any treatment were taken as the normal control group.A 2.0-mm trephine was used to ablate a full-thickness button of the central corneal tissue of each rabbit.The corneas were observed by slit lamp biomicroscopy at the respective time points after the trephined injury.Corneal epithelial fluorescein staining was used to evaluate re-epithelialization with Image J software and haze grading was evaluated with the Fantes classification.Hematoxylin-eosin staining was used to observe the healing process of the cornea.Transmission electron microscopy was conducted to assess the regeneration of the EBM and the reconstruction of the cornea.The study protocol was approved by the Ethics Committee of Guangxi Medical University (No.201811031). The use and care of the experimental animals complied with the Statement for the Use of Animals in Ophthalmic and Vision Research. Results: The corneal epithelial fluorescein areas in modeling 1-, 3-, 5-, 7-, and 14-day group were (4.00±0.10), (3.11±0.10), (2.00±0.06), (0.90±0.04) and (0.67±0.03)mm², respectively, with a significant difference among them (F=3 398.88, P<0.01). With the increasing of time after modeling, the corneal epithelium fluorescein area was gradually reduced, showing significant differences between any two groups (all at P<0.05), and the staining was disappeared at 21 and 30 days after modeling.The corneal haze grades were 3.44±0.53, 0.67±0.25, 1.33±0.50, 2.11±0.60, 2.44±0.53, 3.22±0.44 and 3.78±0.44 in modeling 1-, 3-, 5-, 7-, 14-, 21-, and 30-day group, respectively. The corneal opacity score gradually decreased during 1-5 days after modeling and gradually increased during 5-30 days after modeling, with a significant difference among them (F=51.182, P<0.01). Hematoxylin-eosin staining revealed that a fibrin clot formed in the wound area, and a single layer of epithelium covered the area at initial 3 days after modeling, and a large number of fibroblasts and some extracellular matrix were found at 5 days after modeling.At 21 and 30 days after modeling, the collagen fibers were tightly arranged in the anterior stroma.Transmission electron microscopy showed that the wound was filled with irregular collagen fibers and myofibroblasts.The stroma was remodeled at 21 days after modeling, and defective regeneration of the EBM was detected at 21 and 30 days after modeling. Conclusions Corneal fibrosis initiates after corneal penetrating injury in rabbits and gradually aggravates, the EBM regenerates defectively.

Glioma is the most frequently observed primary tumor of the central nervous system in adults. Among the glioma cases, more than three quarters of patients suffer from high-grade gliomas. High-grade glioma is not only a high-degree malignant tumor but is also an easily recurring disease after surgery with a very poor prognosis. Radiotherapy plus concomitant chemotherapy after operation is the standard treatment strategy for high-grade gliomas, which could increase the survival rate of patients. However, the curative effect is really not satisfactory because it could only guarantee a limited survival time. Over the recent years, molecular-targeted treatment has increasingly drawn the attention of scholars with the continuous development in glioma treatment, thereby becoming the hotspot among researchers. Vascular endothelial growth factor (VEGF) is highly expressed in glioma and in the tissues surrounding the cancer cells. VEGF could regulate tumor growth by inducing endothelial cell proliferation, growth, migration, and by increasing the vascular permeability. Hence, VEGF becomes an effective target for the treatment of glioma. Bevacizumab is a monoclonal antibody that can specifically prevent the combination of VEGF and its receptor, thereby inhibiting the formation of tumor blood vessels. At the same time, bevacizumab can normalize the tumor blood vessels, improve the permeability of blood vessels, and increase the effectiveness of drug concentration in the tumor tissues, thereby achieving anticancer efficacy. In this paper, the mechanism of bevacizumab is introduced. The research progress in the application of bevacizumab alone, as well as in combination with chemotherapy or other drugs, for the high-grade glioma treatment will be summarized.