Drug resistance is one of the key factors affecting the effectiveness of cancer treatment methods, including chemotherapy, radiotherapy, and immunotherapy. Its occurrence is related to factors such as mRNA expression and methylation within cancer cells. If drug resistance in patients can be accurately identified early, doctors can devise more effective treatment plans, which is of great significance for improving patients' survival rates and quality of life. Cancer drug resistance prediction based on artificial intelligence (AI) technology has emerged as a current research hotspot, demonstrating promising application prospects in guiding clinical individualized and precise medication for cancer patients. This review aims to comprehensively summarize the research progress in utilizing AI algorithms to analyze multi-omics data including genomics, transcriptomics, epigenomics, proteomics, metabolomics, radiomics, and histopathology, for predicting cancer drug resistance. It provides a detailed exposition of the processes involved in data processing and model construction, examines the current challenges faced in this field and future development directions, with the aim of better advancing the progress of precision medicine.

Objective:To investigate the risk factors for aneurysm rupture and post-intervention complications in intracranial arterial stenosis patients with intracranial aneurysms.Methods:A retrospective analysis was performed; 238 intracranial arterial stenosis patients with intracranial aneurysms (306 intracranial aneurysms) admitted to Cerebrovascular Disease Department, Neurosurgery Center, Zhujiang Hospital, Southern Medical University from January 2018 to August 2022 were chosen. Ruptured group and unruptured group were divided according to the rupture of intracranial aneurysms. Additionally, 139 patients who underwent interventional therapy and had complete follow-up data were divided into 2 groups according to occurrence of post-intervention complications. Univariate and multivariate Logistic regression analyses were used to identify the risk factors for aneurysm rupture and post-intervention complications.Results:(1) Of 238 patients, 269 unruptured aneurysms and 37 ruptured aneurysms were noted. Univariate regression analysis showed that significant difference was noted between the ruptured group and unruptured group in female ratio, aneurysm distribution, proportion of irregular shaped aneurysms, percentages of patients with increased white blood cell count, neutrophil count, total cholesterol and D-2 polymer, and percentage of patients with decreased blood lymphocyte count ( P<0.05). Multivariate Logistic regression analysis showed that irregular shaped aneurysms ( OR=12.393, 95% CI: 4.114-37.332, P<0.001), elevated neutrophil count ( OR=18.753, 95% CI: 6.555-53.648, P<0.001), and increased D-2 polymer ( OR=4.410, 95% CI: 1.758-11.065, P=0.002) were independent risk factors for aneurysm rupture in intracranial arterial stenosis patients with intracranial aneurysms. (2) Of the 139 patients, 57 had complications and 82 had no complications. Univariate regression analysis showed that the proportion of patients with hypertension history, distribution of arterial stenosis, and proportion of patients with elevated blood D-2 polymer were significantly different between patients with and without complications ( P<0.05); while multivariate Logistic regression analysis did not identify these 3 indexes as independent risk factors for post-intervention complications ( P>0.05). Conclusion:Patients with irregular shaped aneurysms, elevated blood neutrophil count and D-2 polymer trend to have aneurysm rupture; hypertension history, arterial stenosis, and elevated D-2 polymer have impact on postoperative complications in intracranial arterial stenosis patients with intracranial aneurysms.

Objective To investigate the expression and biological value of heme oxygenase I (HO-1) in gliomas. Methods Fifty-six patients with gliomas, admitted to and accepted surgery in our hospital from January 2010 to January 2015, were chosen in our study; WHO grade I was noted in 4 patients, grade Ⅱ in 16, grade Ⅲ in 10, and grade IV in 26 patients; patients with grade I and Ⅱ were as low-grade glioma group and patients with grade Ⅲ and IV were as high-grade glioma group. The HO-1 expression in the two groups was detected by immunohistochemistry. R-langrage survival tool was used to analyze the relation between HO-1 expression and prognosis of 1107 patients with gliomas selected from The Cancer Genome Atlas (TCGA) database. Results Significant differences of HO-1 expression were observed in different grades of gliomas (P<0.05), and HO-1 high-expression rate in the high-grade gliomas (75%) was significantly higher than that in the low-grade group (30%, P<0.05). HO-1 protein expression level in the high-grade gliomas was significantly higher than that in the low-grade group (P<0.05). Moreover, area under the curve (AUC) of the receiver operating characteristic curve was suggested that the HO-1 could be an ideal determine factor (AUC=0.747, P=0.002). Log rank analysis indicated that the accumulate survival rate in patients with low HO-1 expression was significantly higher than that in patients with high HO-1 expression (P<0.05). TCGA database analysis showed that simultaneous survival rate in patients with low HO-1 expression was significantly higher than that in patients with high HO-1 expression (P<0.05). Conclusion Expression of HO-1 is correlated with the malignant degrees and prognoses of gliomas, and it has potential to be a novel biological marker for the diagnosis and treatment of gliomas; furthermore, HO-1 could also be a target for the study and treatment of gliomas.

Objective To investigate the sorting nexin 10 (SNX10) expression in glioma tissues and its relationship with prognosis of the patients.Methods Thirty glioma specimens,collected from surgery and conformed by pathology in our hospital from January 2007 to December 2012,were used in our study,and in them,9 were WHO grade Ⅰ and Ⅱ and 21 were WHO grade Ⅲ and Ⅳ;and 30 nonneoplastic tissue specimens collected during decompression were used as control group.Immunohistochemical staining using polyclonal SNX10 antibody was performed on paraffin embedded specimens.The staining intensity was stratified as absent (-),weak (+),moderate (++) and strong (4+++).The relationships of SNX10 expression with several clinic pathologic indicators and prognosis were analyzed.Results High SNX10 expression was noted in 12 specimens and low SNX10 expression in 18 specimens of the glioma group.High SNX10 expression was noted in 25 specimens and low SNX10 expression in 5 specimens of the control group;the high SNX10 expression rate in glioma tissues was significantly lower than that in non-neoplastic brain tissues (P＜0.05);the high SNX10 expression rate in high-grade glioma tissues was significantly lower than that in low-grade glioma tissues (P＜0.05).The median survival time ofglioma patients with high SNX10 expression was 22.50±8.27 months,and that of glioma patients with low SNX10 expression was 15.50±0.99 months.The survival rate of glioma patients with low SNX10 expression was significantly lower than that of glioma patients with high SNX10 expression (34％ vs.65％,P＜0.05).By Cox multi-factor risk scale model,the expression level of SNX10 and grading of tumors were identified as the independent risk factors of patient's post-operative death;following the decreased SNX10 expression,the risk of postoperative death increased 1.983 times (95％ confidence interval=1.602-2.314,P＜0.05).Conclusions Decreased SNX10 expression is associated with occurrence and development of gliomas,and has a significant effect on patients' post-operative survival time.Decreased SNX10 expression level may be an important index of poor prognosis in glioma patients.