Artificial intelligence is an advanced method to identify novel anticancer targets and discover novel drugs from biology networks because the networks can effectively preserve and quantify the interaction between components of cell systems underlying human diseases such as cancer. Here, we review and discuss how to employ artificial intelligence approaches to identify novel anticancer targets and discover drugs. First, we describe the scope of artificial intelligence biology analysis for novel anticancer target investigations. Second, we review and discuss the basic principles and theory of commonly used network-based and machine learning-based artificial intelligence algorithms. Finally, we showcase the applications of artificial intelligence approaches in cancer target identification and drug discovery. Taken together, the artificial intelligence models have provided us with a quantitative framework to study the relationship between network characteristics and cancer, thereby leading to the identification of potential anticancer targets and the discovery of novel drug candidates.

Background: Tyrosine kinase inhibitor (TKI) treatment has significantly improved the prognosis of oncogenic-driven lung adenocarcinoma (LUAD). However, drug resistance limits the long-term benefits of patients. Therefore, there is a pressing need to explore the mechanism of TKI resistance and identify new therapeutic targets. It is possible to overcome TKI resistance by inducing tumor cell death through a new process called ferroptosis. Aberrations in ferroptosis, which is a kind of regulated cell death (RCD), has been confirmed to be involved in the development and progression of multiple tumors, and is closely related to patient survival. At present, the role of ferroptosis in TKI resistance remains unclear. Methods: Ferroptosis-related factors were isolated by expression characteristics analysis based on the multi-omics data of LUADs and normal lung tissues from The Cancer Genome Atlas (TCGA) database. Next, expression of selected ferroptosis-related factors and prognosis were analyzed. Subsequently, the differences in the expression of selected ferroptosis-related factors before and after TKI resistance on a variety of LUAD cell lines were analyzed to identify the factors that were involved in TKI resistance. Finally, the therapeutic effects were confirmed in vitro by targeting the selected ferroptosis-related factors with small molecule compounds. Results: Glutathione Peroxidase 4 (GPX4), a ferroptosis-related factor, was up-regulated in tumor tissue of LUADs, and correlated with the prognosis of patients. By detecting the expression change of GPX4 before and after TKI resistance in a variety of LUAD cell lines, we confirmed that the inhibition of GPX4 could overcome epidermal growth factor receptor (EGFR)-TKI resistance by inducing ferroptosis. Conclusions: GPX4 could serve as a novel therapeutic target for EGFR-TKI resistance in LUAD.

Inflammatory processes are essential for innate immunity and contribute to carcinogenesis in various malignancies, such as colorectal cancer, esophageal cancer and lung cancer. Pharmacotherapies targeting inflammation have the potential to reduce the risk of carcinogenesis and improve therapeutic efficacy of existing anti-cancer treatment. Non-steroidal anti-inflammatory drugs (NSAIDs), comprising a variety of structurally different chemicals that can inhibit cyclooxygenase (COX) enzymes and other COX-independent pathways, are originally used to treat inflammatory diseases, but their preventive and therapeutic potential for cancers have also attracted researchers’ attention. Pharmacogenomic variability, including distinct genetic characteristics among different patients, can significantly affect pharmacokinetics and effectiveness of NSAIDs, which might determine the preventive or therapeutic success for cancer patients. Hence, a more comprehensive understanding in pharmacogenomic characteristics of NSAIDs and cancer-related inflammation would provide new insights into this appealing strategy. In this review, the up-to-date advances in clinical and experimental researches targeting cancer-related inflammation with NSAIDs are presented, and the potential of pharmacogenomics are discussed as well.

Not all patients can benefit from chemotherapy due to the various of tumor type, stage, location, and the different distribution of immune cells in tumor immune microenvironment (TIME). Immune cells are widely involved in every step of cancer progression, including immune escape, metastasis, drug response, and prognosis. In this study, we explored the transcriptome data of 10 solid tumors treated with chemotherapy to identify the role of immune cells. We downloaded the transcriptome and mutation data of 10 cancers from TCGA databases, and used ESTIMATE and CIBERSORT algorithms to assess the proportion of immune cells in the TIME. According to the proportion of specific immune cell infiltration (SICI) of CD8⁺ T cells and M1 macrophages to group the patients, we found that compared with the SICI low and medium groups, the SICI high group had a larger tumor mutation burden, more gene mutations with targeted drugs, more activation of immune checkpoints (PD‐1, PD‐L1, CTLA‐4, LAG‐3, TIM‐3, and TIGIT), and immune molecules (CD8a, CD80, CD86, TLR2, HLA‐A, HLA‐B, and CD11a) (p < 0.05). Therefore, we can select an appropriate treatment for patients by clarifying the proportion of immune infiltration of CD8⁺ T cells and M1 macrophages in the TIME.

The drug therapy for non-small cell lung cancer (NSCLC) have always been issues of poisonous side effect, acquired drug resistance and narrow applicable population. In this study, we built a novel network analysis method (difference- correlation- enrichment- causality- node), which was based on the difference analysis, Spearman correlation network analysis, biological function analysis and Bayesian causality network analysis to discover new therapeutic target of NSCLC in the sequencing data of BEAS-2B and 7 NSCLC cell lines. Our results showed that, as a proteasome subunit coding gene in the central of cell cycle network, PSMD2 was associated with prognosis and was an independent prognostic factor for NSCLC patients. Knockout of PSMD2 inhibited the proliferation of NSCLC cells by inducing cell cycle arrest, and exhibited marked increase of cell cycle blocking protein p21, p27 and decrease of cell cycle driven protein CDK4, CDK6, CCND1 and CCNE1. IPA and molecular docking suggested bortezomib has stronger affinity to PSMD2 compared with reported targets PSMB1 and PSMB5. In vitro and In vivo experiments demonstrated the inhibitory effect of bortezomib in NSCLC with different driven mutations or with tyrosine kinase inhibitors resistance. Taken together, bortezomib could target PSMD2, PSMB1 and PSMB5 to inhibit the proteasome degradation of cell cycle check points, to block cell proliferation of NSCLC, which was potential optional drug for NSCLC patients.

Background: Circulating tumor DNA (ctDNA) is valuable for detecting minimal residual disease (MRD). However, studies involving long-term blood sampling are required to comprehensively interpret the clinical use of ctDNA analyses. Methods: We conducted a prospective multicenter cohort study (LUNGCA) for dynamic ctDNA monitoring in lung cancer patients receiving curative-intent surgery. ctDNA analysis was conducted on preoperative plasma samples, at postoperative 3 days and 1 month, and then every 3-6 months for up to 3 years. Results: In total, 233 non-small cell lung cancer (NSCLC) patients and 2336 longitudinal plasma samples were included; the median follow-up was 51.4 months. Post-comprehensive treatment (after radical surgery + necessary adjuvant therapy) MRD status was better at predicting relapse than postoperative MRD status (positive predictive value: 100% vs 90.0%; negative predictive value: 90.3% vs 90.1%). Patients with positive pre-adjuvant ctDNA and targetable mutations in tumor tissues had improved recurrence-free survival (RFS) with corresponding adjuvant tyrosine kinase inhibitor (TKI) treatment (hazard ratio [HR] = 0.01, P = .005), but adjuvant chemotherapy failed to improve RFS (HR = 0.6, P = .491). Of patients receiving adjuvant therapies, patients with a negative- or positive-negative ctDNA change pattern had favorable and similar RFS (P = .419), whereas patients with a positive- or negative-positive pattern had worse RFS (P < .001). TKI therapy was more effective than chemotherapy in clearing ctDNA. Post-relapse ctDNA negativity was associated with favorable OS (HR = 0.4; P = .029). Conclusions: Comprehensive interpretation of dynamic ctDNA monitoring data can inform precise whole-course postsurgical management of NSCLC patients.