Objective To retrospectively analyze the clinical features of invasive pulmonary aspergillosis (IPA) in intensive care unit (ICU), so as to improve the level of clinical diagnosis and treatment. Methods A total of 81 patients diagnosed as IPA from March, 2017 to March, 2022 in the ICU of The First Affiliated Hospital of USTC, Division of Life Sciences and Medicine, University of Science and Technology of China were selected as infection group. A total of 81 non-IPA patients with pulmonary infection and Aspergillus negative sputum culture were selected as the control group. The host factors, Acute Physiology and Chronic Health Assessment Ⅱ score at admission, underlying diseases, clinical symptoms and signs, relevant laboratory test results, and lung CT findings were compared between the two groups. Univariate analysis and multivariate conditional logistic regression analysis were used to identify the risk factors for the occurrence of pulmonary aspergillosis in IPA patients in ICU. At the same time, the types of aspergillus in the IPA group and the outcomes of the two groups at 28 days after ICU admission were analyzed. Results Of the 81 IPA patients, 4 were proven diagnosed and 77 were putative diagnosed. IPA patients were mainly infected with Aspergillus fumigatus and Aspergillus flavus. Symptoms and signs such as fever, cough and expectoration, dyspnea and pulmonary rales occurred in both groups. The level of procalcitonin in IPA group was higher than that in non-IPA group, and the difference was statistically significant (P=0.016). The positive rate of serum galactomannan antigen test (GM test) in the IPA group was higher than that in the non-IPA group, and the differences was statistically significant (P=0.000). The incidence of pulmonary imaging cavities in IPA group was higher than that in non-IPA group, and the difference was statistically significant (P=0.022). Univariate analysis showed that central venous catheterization, septic shock, complete parenteral nutrition, chronic obstructive pulmonary disease, and immunosuppression were risk factors for IPA (P<0.05); Multivariate conditional logistic regression analysis showed that complete parenteral nutrition, chronic obstructive pulmonary disease, and immunosuppression were independent risk factors for IPA (P<0.05). The 28-day fatality rate in IPA group was higher than that in non-IPA group (55.6% vs. 34.6%, P=0.007). Conclusions IPA patients have no specific clinical symptoms and signs, and are mainly infected with Aspergillus fumigatus and Aspergillus flavus; GM test has guiding significance for the diagnosis of IPA. Serum GM test and pulmonary imaging have cavity findings that are helpful for the diagnosis of IPA. Patients with a history of chronic obstructive pulmonary disease, immunosuppression, or complete parenteral nutrition need to be on high alert for the possibility of IPA during ICU stay.
Objective To evaluate medical students’ perceptions and attitudes toward artificial intelligence (AI)-assisted diagnosis of renal cell carcinoma (RCC), and to analyze their educational needs regarding AI in pathological diagnosis. Methods A questionnaire survey (including closed and open-ended questions) was conducted to assess medical students’ perceptions, attitudes, and educational needs concerning AI-assisted RCC diagnosis. Participants included medical students from different specialties and standardized training residents. The questionnaire covered demographic information, perceptions and attitudes toward AI, and AI-related educational needs. Results A total of 249 respondents completed the survey. The majority were standardized training residents, mostly aged 23-26 years, and 40.96% had practical experience in pathological diagnosis of RCC. The median scores for most closed-ended questions were 4. Respondents generally considered “efficiency” and “improved accuracy” as the most prominent advantages of AI, with timeliness, automated diagnosis, reduction of human error, and precise diagnosis being the most emphasized aspects. Analysis of AI-related educational needs revealed high-frequency keywords such as “expanding sample size” “balanced responsibility allocation” and “enhancing collaboration skills.” Conclusion Medical students hold a positive attitude toward AI and its application in RCC diagnosis, but there remains a lack of formal AI-related education.