Predictive Model of Risk Factors of High Flow Nasal Cannula Using Machine Learning in COVID-19

BackgroundWith the rapid increase in the number of COVID-19 patients in Japan, the number of patients receiving oxygen at home has also increased rapidly, and some of these patients have died. An efficient approach to identify high-risk patients with slowly progressing and rapidly worsening COVID-19, and to avoid missing the timing of therapeutic intervention will improve patient prognosis and prevent medical complications. MethodsPatients admitted to medical institutions in Japan from November 14, 2020 to April 11, 2021 and registered in the COVID-19 Registry Japan were included. Risk factors for patients with High Flow Nasal Cannula invasive respiratory management or higher were comprehensively explored using machine learning. Age-specific cohorts were created, and severity prediction was performed for the patient surge period and normal times, respectively. ResultsWe were able to obtain a model that was able to predict severe disease with a sensitivity of 57% when the specificity was set at 90% for those aged 40-59 years, and with a specificity of 50% and 43% when the sensitivity was set at 90% for those aged 60-79 years and 80 years and older, respectively. We were able to identify lactate dehydrogenase level (LDH) as an important factor in predicting the severity of illness in all age groups. DiscussionUsing machine learning, we were able to identify risk factors with high accuracy, and predict the severity of the disease. Using machine learning, we were able to identify risk factors with high accuracy, and predict the severity of the disease. We plan to develop a tool that will be useful in determining the indications for hospitalisation for patients undergoing home care and early hospitalisation.

Preoperative Evaluation of the Texture of Pituitary Adenoma by MRI-Based Three-Dimensional Reconstruction / 中国医科大学学报

Objective To investigate the usefulness of three-dimensional reconstruction in the preoperative evaluation of the texture of pituitary tumors. Methods Seventy patients with pituitary tumors admitted to our hospital between January 2015 and July 2018 were enrolled in the study. All patients underwent enhanced MRI scanning before surgery. They were classified into the soft group, medium group, and tough group according to the tumor texture. The patient's clinical data, MRI images, and surgery conditions were collected.The Mimics software was used to reconstruct the three-dimensional models of pituitary adenomas. The volume and surface area of different tumor signal groups were calculated and analyzed. In addition, the relationships between tumor size, tumor resection, and postoperative complications were analyzed. Results The three-dimensionally reconstructed model of the pituitary adenoma had a clear outline and was consistent with the tumor area in the MRI images. The calculated average threshold accurately segmented the images. Grouped by the classification of texture, the differences of the proportions of each part were statistically significant (P ＜ 0.01). According to the ordinal polytomous logistics regression analysis, the proportion of the volume of the higher part positively correlated with the tumor texture (P ＜0.05), and the ratio of the surface area of the medium part to the overall surface area positively correlated with the tumor texture (P ＜ 0.05).Conclusion The use of Mimics software for 3 D reconstruction of preoperative MRI images can accurately predict the tumor texture in pituitary tumors and can provide a basis for the choice of surgical methods.