Immune-related adverse events (irAEs) are a range of complications associated with the use of immune-checkpoint inhibitors (ICIs). Two major classes of ICIs widely used are Cytotoxic T-Lymphocyte Antigen 4 (CTLA4) and Programmed Cell death-1 (PD-1)/Programmed death-ligand 1 (PD-L1) inhibitors. High-grade irAEs are life-threatening and often cause a severe decline in performance status in such that patients do not qualify for any further anticancer treatments. It is difficult to generalize the evidence in the current literature on risk factors or biomarkers for the entire class of ICIs as the studies so far are either disease-specific (e.g., lung cancer or melanoma) or ICI agent-specific (e.g., pembrolizumab, ipilimumab) or irAE-specific (e.g., pneumonitis or gastritis). In this review, risk factors and biomarkers to consider before initiating or monitoring ICI are listed with a practical purpose in day-to-day practice. Risk factors are grouped into demographics and social history, medical history, and medication history, tumor-specific and agent-specific risk factors. A higher risk of irAE is associated with age <60 years, high body mass index, women on CTLA4 and men on PD-1/PD-L1 agents, and chronic smokers. Patients with significant kidney (Stage IV-V), cardiac (heart failure, coronary artery disease, myocardial infarction, hypertension), and lung (asthma, pulmonary fibrosis, and chronic obstructive pulmonary disease) are at a higher risk of respective organ-specific irAEs. Pre-existing autoimmune disease and chronic use of certain drugs (proton pump inhibitors, diuretics, anti-inflammatory drugs) also increase the irAE-risk. Biomarkers are categorized into circulating blood counts, cytokines, autoantibodies, HLA genotypes, microRNA, gene expression profiling, and serum proteins. The blood counts and certain protein markers (albumin and thyroid-stimulating hormone) are readily accessible in current practice. High neutrophil-lymphocyte ratio, eosinophil/monocyte/lymphocyte counts; TSH and troponins at diagnosis and drop in the white count and lymphocyte count can predict irAE. Other biomarkers with limited evidence are cytokines, autoantibodies, HLA genotypes, microRNA, and gene expression profiling. With fast-expanding approvals for ICIs in various cancer types, knowledge on risk factors and biomarkers can help providers assess the irAE-risk of their patients. Prospective disease and agent-specific studies are needed to provide further insight on this essential aspect of ICI therapy.

Biliary tract cancers (BTC) are often diagnosed at advanced stages and have a grave outcome due to limited systemic options. Gemcitabine and cisplatin combination (GC) has been the first-line standard for more than a decade. Second-line chemotherapy (CT) options are limited. Targeted therapy or TT (fibroblast growth factor 2 inhibitors or FGFR2, isocitrate dehydrogenase 1 or IDH-1, and neurotrophic tyrosine receptor kinase or NTRK gene fusions inhibitors) have had reasonable success, but <5% of total BTC patients are eligible for them. The use of immune checkpoint inhibitors (ICI) such as pembrolizumab is restricted to microsatellite instability high (MSI-H) patients in the first line. The success of the TOPAZ-1 trial (GC plus durvalumab) is promising, with numerous trials underway that might soon bring targeted therapy (pemigatinib and infrigatinib) and ICI combinations (with CT or TT in microsatellite stable cancers) in the first line. Newer targets and newer agents for established targets are being investigated, and this may change the BTC management landscape in the coming years from traditional CT to individualized therapy (TT) or ICI-centered combinations. The latter group may occupy major space in BTC management due to the paucity of targetable mutations and a greater toxicity profile.

The effective management of biliary tract cancers (BTCs) has been hampered by limited options for systemic therapy. In recent years, the focus on precision medicine has made technologies such as next-generation sequencing (NGS) accessible to clinicians to identify targetable mutations in BTCs in tumor tissue (primarily) as well as blood, and to treat them with targeted therapies when possible. It has also expanded our understanding of functional pathways associated with genetic alterations and opened doors for identifying novel targets for treatment. Recent advances in the precision medicine approach allowed us to identify new molecular markers in BTCs, such as epigenetic changes (methylation and histone modification) and non-DNA markers such as messenger RNA, microRNA, and long non-coding RNA. It also made detecting these markers from non-traditional sources such as blood, urine, bile, and cytology (from fine-needle aspiration and biliary brushings) possible. As these tests become more accessible, we can see the integration of different molecular markers from all available sources to aid physicians in diagnosing, assessing prognosis, predicting tumor response, and screening BTCs. Currently, there are a handful of approved targeted therapies and only one class of immunotherapy agents (immune checkpoint inhibitors or ICIs) to treat BTCs. Early success with new targets, vascular endothelial growth factor receptor (VEGFR), HER2, protein kinase receptor, and Dickkopf-1 (DKK1); new drugs for known targets, fibroblast growth factor receptors (FGFRs) such as futabatinib, derazantinib, and erdafitinib; and ICIs such as durvalumab and tremelimumab is encouraging. Novel immunotherapy agents such as bispecific antibodies (bintrafusp alfa), arginase inhibitors, vaccines, and cellular therapy (chimeric antigen receptor—T cell or CAR-T, natural killer cells, tumor-infiltrating lymphocytes) have the potential to improve outcomes of BTCs in the coming years.

Adverse drug events (ADEs) significantly impact patient safety and health outcomes. Manual ADE detection from clinical narratives is time-consuming, labor-intensive, and costly. Recent advancements in large language models (LLMs), including transformer-based architectures such as Bidirectional Encoder Representations from Transformers (BERT) and Generative Pretrained Transformer (GPT) series, offer promising methods for automating ADE extraction from clinical data. These models have been applied to various aspects of pharmacovigilance and clinical decision support, demonstrating potential in extracting ADE-related information from real-world clinical data. Additionally, chatbot-assisted systems have been explored as tools in clinical management, aiding in medication adherence, patient engagement, and symptom monitoring. This narrative review synthesizes the current state of LLMs in ADE detection from a clinical perspective, organizing studies into categories such as human-facing decision support tools, immune-related ADE detection, cancer-related and non-cancer-related ADE surveillance, and personalized decision support systems. In total, 39 articles were included in this review. Across domains, LLM-driven methods have demonstrated promising performances, often outperforming traditional approaches. However, critical limitations persist, such as domain-specific variability in model performance, interpretability challenges, data quality and privacy concerns, and infrastructure requirements. By addressing these challenges, LLM-based ADE detection could enhance pharmacovigilance practices, improve patient safety outcomes, and optimize clinical workflows.

Mucin-5AC (MUC5AC) is a heavily glycosylated gel-forming secreted mucin with a reliable prognostic value when detected in multiple malignancies. It is highly prevalent (70%) in PDA and is nonexistent in normal pancreatic tissues. Retrospective studies on PDA tumor tissue (detected by immunohistochemistry or IHC)) have investigated the prognostic value of MUC5AC expression but were equivocal. Some studies associated it with poor outcomes (survival or pathological features such as lymph node disease, vascular/neural invasion in resected tumors), while others have concluded that it is a good prognostic marker. The examination of expression level threshold (5%, 10%, or 25%) and the detected region (apical vs. cytoplasmic) were variable among the studies. The maturation stage and glycoform of MUC5AC detected also differed with the Monoclonal antibody (Mab) employed for IHC. CLH2 detects less mature/less glycosylated versions while 45M1 or 21-1 detect mature/more glycosylated forms. Interestingly, aberrantly glycosylated variants of MUC5AC were detected using lectin assays (Wheat Germ Agglutinin-MUC5AC), and Mabs such as NPC-1C and PAM4 have are more specific to malignant pancreatic tissues. NPC-1C and PAM4 antibody reactive epitopes on MUC5AC are immunogenic and could represent specific changes on the native MUC5AC glycoprotein linked to carcinogenesis. It was never studied to predict treatment response.

The combination of stereotactic body radiation therapy (SBRT) plus immune checkpoint inhibitors (ICI) must be explored to treat advanced primary liver tumors such as hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA). Limited retrospective reviews and case reports/series suggest this combination can be effective and safe in both cancer types. With ICIs moving into the first line (IMbrave 150, HIMALAYA, and TOPAZ-1) to manage these cancers, identifying a suitable population for this approach is challenging. Patients with macrovascular invasion (MVI)-positive HCC (especially if larger veins are involved) or recurrent HCCs post-locoregional therapies (such as transarterial radioembolization (TARE), transarterial chemoembolization (TACE), or ablation), as well as those ineligible for bevacizumab or tyrosine kinase inhibitors (TKIs), should be the focus of exploring this combination in HCC. Unresectable or oligometastatic CCA patients who cannot tolerate gemcitabine/cisplatin (GC) or those who progressed on GC without durvalumab and do not have targetable mutations could also be considered for this approach. In both HCC and CCA disease groups, SBRT plus ICI can be examined post-ICI as these two modalities act synergistically to enhance anti-tumor activity (based on pre-clinical studies). Large-scale randomized trials are needed to identify the subsets of primary liver cancers suitable for this approach and to clearly define its clinical benefit.