The rapid development of artificial intelligence (AI), especially generative AI (GenAI), has already brought, and will continue to bring, revolutionary changes to our daily production and life, as well as create new opportunities and challenges for diagnostic and therapeutic practices in the medical field. Haihe Hospital of Tianjin University collaborates with the National Supercomputer Center in Tianjin, Tianjin University, and other institutions to carry out research in areas such as smart healthcare, smart services, and smart management. We have conducted research and development of a full-process disease diagnosis and treatment assistance system based on GenAI in the field of smart healthcare. The development of this project is of great significance. The first goal is to upgrade and transform the hospital's information center, organically integrate it with existing information systems, and provide the necessary computing power storage support for intelligent services within the hospital. We have implemented the localized deployment of three models: Tianhe "Tianyuan", WiNGPT, and DeepSeek. The second is to create a digital avatar of the chief physician/chief physician's voice and image by integrating multimodal intelligent interaction technology. With generative intelligence as the core, this solution provides patients with a visual medical interaction solution. The third is to achieve deep adaptation between generative intelligence and the entire process of patient medical treatment. In this project, we have developed assistant tools such as intelligent inquiry, intelligent diagnosis and recognition, intelligent treatment plan generation, and intelligent assisted medical record generation to improve the safety, quality, and efficiency of the diagnosis and treatment process. This study introduces the content of a full-process disease diagnosis and treatment assistance system, aiming to provide references and insights for the digital transformation of the healthcare industry.
Artificial Intelligence ; Humans ; Delivery of Health Care ; Generative Artificial Intelligence

Objective:To explore the mechanism of follicular helper T (Tfh) cells, i.e. CD4 +CXCR5 +T cells, and the secreted cytokine programmed death factor 1 (PD-1) in the pathogenesis of tuberculosis, and to explore the significance of Tfh cells and PD-1 in the treatment of tuberculosis. Methods:Flow cytometry was used to detect the changes of Tfh cells and PD-1 in mononuclear cells during the treatment cycle of tuberculosis.Results:Before treatments, the ratio of Tfh cells/CD4 +T cells in peripheral blood mononuclear cells in the pulmonary tuberculosis group was 3.37%±0.45%, which was significantly higher than 2.21%±0.47% of the healthy control group ( P<0.01), and significantly higher than 2.39%±0.38% after treatments ( P<0.01). Before treatments, the ratio of CD4 +CXCR5 +PD-1 +T cells/Tfh cells in the peripheral blood of the tuberculosis group was 25.33%±10.08%, which was significantly higher than 8.42%±2.31% of the healthy control group ( P<0.01), and significantly higher than 11.35%±2.65% after treatments ( P<0.01). After treatments, the levels of Tfh cells and PD-1 in the sputum smear-negative group and the sputum smear-negative group were lower than that before treatments, and the difference between the groups was statistically significant (all P<0.05). Conclusions:The levels of Tfh cells and PD-1 in patients with tuberculosis are significantly higher than those in healthy people, and after drug treatment, the levels of both can be reduced. With the prolongation of the treatment cycle, the sputum smear-transforming group and the non-negative group began to show significant differences. In the course of pulmonary tuberculosis, monitoring changes in Tfh cells and PD-1 levels is helpful for the diagnosis of tuberculosis, and has certain guiding significance for its treatment and outcome.

Objective To explore a rapid and accurate method for the diagnosis of smear negative tuberculosis meningitis (TBM). Methods Sixty-seven patients with TBM were selected from Tianjin Haihe Hospital from June 2014 to June 2016, and 118 patients with non-tuberculous meningitis (NTBM) in the same period were chosen as control group, including bacterial meningitis (BM) group (n=61) and viral meningitis (VM) group (n=57). The laboratory routine, biochemical and immune indicators were tested with the specimens of both the blood and cerebrospinal fluid of all the patients. The Logistic regression equation was established for the diagnosis of TBM, and the diagnostic efficacy of which was evaluated by the receiver operating characteristic curve (ROC). Results The predictive regression equations of the TBM with BM, VM and NTBM (BM + VM) were obtained when BM group was used as a control: PRE_BM=1/1 +e-(-5.298+0.196 × ESAT-6+ 0.119 × CFP-10-2.968 × PCT+2.206 × ADA_CSF+0.705 × GLU_CSF+0.093 × LDH_CSF), PRE_VM=1/1+e-(-6.907+0.394 × ESAT-6-0.120 × Na+2.633 × ADA_CSF-0.088 × Cl_CSF) and PRE_NTBM=1/1+e-(0.683+0.099×ESAT-6+0.063×CFP-10-2.645×PCT+1.393×ADA_CSF+1.342×TbAb_CSF)respectively. When BM group was served as a control, the sensitivity, specificity, positive and negative predictive values of the regression for the diagnosis of TBM were 97.01%(89.63%-99.64%), 98.36%(91.20%-99.96%), 98.48%(91.84%-99.96%) and 96.77%(88.83%-99.61%), respectively.When VM group was served as a control, which were 94.03%(85.41%-98.35%), 94.74%(85.38%-98.90%), 95.45%(87.29%-99.05%) and 93.10%(83.27%-98.09%), respectively. When NTBM group was served as control, which were 94.03%(85.41%~98.35%), 90.68%(83.93%-95.25%), 85.14%(74.96%-92.34%) and 96.40%(91.03%-99.01%), respectively. Conclusion The predictive regression equation could be used as early diagnostic TBM with high sensitivity and specificity, which should be popularized in clinical practice, while, according to the higher negative predictive value, the negative results of which could be used to rule out of the TBM and non-empirical medication.

Objective To establish a diagnostic model of multiple cytokines for differential diagnosis of tuberculous pleural effusion , and compare its diagnostic accuracy with tuberculosis infected T cells detection ( T-SPOT.TB ) in order to evaluate its diagnostic performance.Methods Case-control study.Totally 147 patients with pleural fluid in Tianjin Haihe Hospital were enrolled and categorized as tuberculous pleural effusion group ( n=95 ) and malignant pleural effusion group ( n=52 ) from December 2011 to June 2013.Pleural effusion cytokines including interferon-γ( IFN-γ) , C-X-C motif chemokine 10 (CXCL-10), tumor necrosis factor-α(TNF-α), vascular endothelial growth factor (VEGF), IL-2, IL-16, IL-17, IL-27 and IL-33 were tested by liquid chip technology and analyzed by Binary Logistic regression and receiver operating characteristic curve (ROC), and the pleural effusion was also detected by tuberculosis infected T cells detection ( T-SPOT.TB) as a control.Results The comparison of the AUC of cytokines is:CXCL-10>IL-27>IFN-γ>IL-33 >IL-17>IL-16>TNF-α>VEGF>IL-2; After that, CXCL-10, IFN-γ, IL-27 and IL-33 were included the Binary Logistic regression model.The regression equation is P=1/1+e-( -16.851+0.390 ×IFN-γ+0.006 ×IL-27+0.020 ×IL-33).The AUC, sensitivity and specificity of the diagnostic model were 99.5%, 96.84%, and 98.08%, respectively.Both AUC and sensitivity of the diagnostic model were superior to those of any single index.Compared with T-SPOT.TB (0.995 ±0.003), the AUC of the diagnostic model (0.921 ±0.023) was significantly greater ( Z=3.235, P <0.01), but no significant difference was found when it comes to diagnostic results consistency (χ2 =0.062 5, P>0.05).The Kappa of the two methods was 0.795, which meant fine agreement of the evaluations of the two raters.Conclusion The application of liquid array technology of high sensitivity and repeatability with high throughput provided a novel insight and method in the clinical diagnosis , treatment and prevention for tuberculous pleural effusion scientifically and accurately.

Objective To explore the diagnostic value of combined detection of the liquid array technology, interfer-on (IFN)-γand IFN-γ-inducible protein (IP)-10 in the rapid, accurate diagnosis and differential diagnosis of tuberculous pleural effusions. Methods Patients with transudative pleural effusions were divided into tuberculous pleural effusion group (n=52) and malignant pleural effusion group (n=38). The method of T-SPOT.TB was used to detect the number of effec-tor T cells sensitized to Mycobacterium tuberculosis and spot forming cells (SFCs). The liquid array technology was used to detect the level of IFN-γand IP-10. Logistic regression was used to analyze and compare the diagnostic value of the two-method combination. Results The diagnostic sensitivity, specificity and the area under the ROC curve (AUC) of T-SPOT. TB were 90.38%, 84.21%, and 0.938 (95%CI:0.867-0.978), respectively. The diagnostic sensitivity, specificity and AUC of combined detection of IFN-γand IP-10 were 98.08%, 97.37%, and 0.995 (95%CI:0.951-1.000), respectively. There was no significant difference in the diagnostic sensitivity and specificity between the two methods, and the diagnostic agreement for the two diagnostic methods was fine (Kappa=0.703). The difference of AUC between the methods was significantly differ-ent (Z=1.996, P<0.05). The method of combined detection of IFN-γand IP-10 showed the larger AUC (AUC=0.995). Con-clusion The combined diagnosis meets the clinical needs of rapid, accurate diagnosis and differential diagnosis for tuber-culous pleural effusion by simultaneously assaying the level of IFN-γand IP-10 using the liquid array technology.