Supplementary Figure 2 shows pathway analysis of the TCGA-COAD cohort and the ChangKang cohort.

BACKGROUND:Colon cancer is genetically heterogeneous, necessitating standardized genomic analyses for cross-cohort comparisons. Although The Cancer Genome Atlas-colon adenocarcinoma (TCGA-COAD) is a widely used dataset, its comparability to other ethnically different populations remains unclear. This study systematically compares the genomic characteristics of TCGA-COAD and ChangKang, a Chinese colon cancer cohort, using an identical data processing pipeline to minimize methodologic biases. METHODS:Whole-exome sequencing data from both cohorts were uniformly processed to analyze five key genomic features: tumor mutational burden (TMB), microsatellite instability (MSI), significantly mutated genes, mutational signatures, and copy-number variation (CNV). Samples were classified into hypermutated and non-hypermutated subgroups for further comparisons. RESULTS:The TCGA-COAD cohort exhibited a higher overall TMB, driven by a greater proportion of hypermutated samples. However, the hypermutated subgroup of the ChangKang cohort included more ultramutated cases with polymerase epsilon (POLE) exonuclease domain mutations, leading to a higher subgroup TMB. MSI was more prevalent in TCGA-COAD, whereas significantly mutated gene frequencies varied, with lower APC and ACVR2A mutation rates in the ChangKang cohort. CNV patterns were largely similar though CNV frequencies were higher in TCGA-COAD. CONCLUSIONS:Despite differences in subgroup distributions and mutation frequencies, the overall genomic characteristics of colon cancer remain consistent between these ethnically different cohorts. This suggests that cross-population analyses are feasible when standardized processing methods are applied. IMPACT:This study provides a systematic, unbiased comparison of TCGA-COAD and the Chinese ChangKang cohort, demonstrating that the genomic characteristics remain largely consistent across ethnically distinct populations.

Supplementary Figure 1 shows the comparison of clinical data of the two cohorts.

Supplementary Figure 4 shows the correlation of TMB between ChangKang and the subsampled cohort.

Supplementary Figure 5 shows genomic characteristics of the subsampled ChangKang cohort.

Supplementary Figure 3 shows the sequencing characteristics in the TCGA and ChangKang cohorts.

Supplementary Table 1 summarizes the clinical data for the two cohorts.

Supplementary Table 5 compares MSI sample proportions across clinical subgroups between the TCGA-COAD and ChangKang cohorts.

Supplementary Table 3 shows the comparison of mutational frequency across 10 canonical oncogenic pathways between the TCGA-COAD and ChangKang cohorts.

Supplementary Table 4 compares hypermutated sample proportions across clinical subgroups between the TCGA-COAD and ChangKang cohorts.

We curated multiomic data including whole exome sequencing (WES) and high-resolution chest CT from patients diagnosed with NSCLC, and evaluated the potential for a multitask deep learning algorithm to predict tumor mutational status and WES signatures from CT. Taipei Veterans General Hospital (TVGH) initiated a multi-omic databank integrating clinical, radiological, pathological, and genomic data from NSCLC patients. A predefined pipeline identified INDEL and SNP mutations with high or moderate impact using Ensembl Variant Effect Predictor (VEP). PCA reduced dimensionality of the WES binary data, and t-SNE was used for clustering and visualization. Chest CT was performed using scanners with ≥ 20 detector rows, at end-inspiration with 120 kV, automated exposure control, and a 50 cm field of view. Patients were split 80% for training and 20% for validation. 3D CT image data centered on lesions were rotated, resampled and cropped. A 3D ResNet-18 network, pre-trained on Kinetics-400, was used for predicting common NSCLC mutations (EGFR, TP53), and the first and second principal component coefficients. A total of 317 patients were included, with 74% non-smokers.(Table 1) Of 109 LDCT detected nodules, only 27 were from smokers. PCA of 12, 023 gene mutations identified five principal components. t-SNE visualized the clustered distribution of the top three genes with the highest weights: EGFR, TP53, and RBM10, along with tumor mutation burden (TMB) and EGFR co-mutation status. Performance of the deep learning algorithm in the validation set (n=63) showed: EGFR mutation prediction: AUROC 71%, PPV 71%; TP53 mutation prediction: AUROC 76%, NPV 94%, sensitivity 0.84, specificity 0.72.Pearson correlation for prediction of the coefficients of the genetic signature principal components from CT images were 0.37 and 0.47. Deep learning algorithms can infer genetic mutation signatures from CT radiologic features. Further optimization is needed. Hsu-Ching Huang, Samira Masoudi, Daniel Halmos, Chien-jung Huang, Yu-Chao Wang, Yi-Chen Yeh, Yung-Hung Luo, Han-Shui Hsu, Yuh-Min Chen, Chun-Ku Chen, Sandip Patel, Albert Hsiao. Multiomic analysis of Taiwanese NSCLC: insights from a predominantly non-smoking cohort [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 1090.

Tumor mutation burden (TMB), the total number of nonsynonymous mutations in the tumor genome, is a well-established biomarker for predicting responses to immune checkpoint blockade (ICB) therapy across various cancers. Patients with high TMB tend to exhibit better responses to ICB. Recently, targeted gene panels have been developed to estimate TMB before treatment. These panels are enriched for calcium ion-binding genes. However, a direct link between TMB and calcium ion-binding genes has not been reported in the literature to date. The association between TMB and calcium ion-binding genes was analyzed using mutation data from The Cancer Genome Atlas (TCGA) database. In addition, a pan-cancer model was constructed to estimate TMB based solely on calcium ion-binding genes. The model’s predictive power for ICB response was validated using independent datasets. Finally, enrichment analysis was performed to investigate the biological connections between calcium ion-binding genes and TMB. Calcium ion-binding genes were enriched among the TMB-predictive model genes in 27 out of 33 cancer types. Among these, 19 cancer types exhibited strong predictive performance, with R² values greater than 0.5 in our pan-cancer model based on calcium ion-binding genes. The model effectively estimated TMB and identified ICB responders in independent datasets, including lung adenocarcinoma and melanoma. Enrichment analysis further suggested that calcium ion-binding genes may influence TMB through signal transduction pathways. These findings establish a novel association between calcium ion-binding genes and TMB, demonstrating the feasibility of a pan-cancer TMB estimation model based on calcium ion-binding genes. This approach may enhance TMB estimation and improve ICB response prediction across multiple cancers.

Gastrointestinal (GI) cancers account for more than one-third of cancer-related mortality, and the prognosis for late-stage patients remains poor. Immunotherapy has been proven to extend the survival of patients at advanced stages; however, challenges persist in patient selection and overcoming drug resistance. Tumor-infiltrating lymphocytes (TILs) and tertiary lymphoid structures (TLS) in the tumor microenvironment (TME) have been found to be associated with anti-tumor immune responses. ‘Hot tumors’ with high levels of infiltration tend to respond better to immune checkpoint inhibitor (ICI) therapy, making them potential biomarkers for ICI treatment. To explore potential biomarkers for predicting immunotherapy response and prognosis in GI cancers, we downloaded the gene expression profiles of seven GI cancers from The Cancer Genome Atlas (TCGA) database and characterized their TME, classifying the samples into hot/cold tumor subgroups. Furthermore, we developed a computational framework to construct cancer-specific hot tumor classification models with only a few genes. External independent datasets and qPCR experiments were used to verify the performance of our few-gene models. We constructed cancer-specific few-gene models to identify hot tumors for GI cancers with only two to nine genes. The results showed that B cells are important for hot tumor determination, and the identified hot tumors are significantly associated with TLS. They not only overexpress TLS marker genes but are also associated with the presence of TLS in whole-slide images. Further, a two-gene qPCR model was developed to effectively distinguish between hot and cold tumor subgroups in cholangiocarcinoma, providing an opportunity for stratifying patients with hot tumors in clinical settings. In conclusion, our established few-gene models, which can be easily integrated into clinical practice, can distinguish hot and cold tumor subgroups, and may serve as potential biomarkers for predicting ICI response.

Background Cholangiocarcinoma is a challenging malignancy with limited responses to conventional therapies, particularly immune checkpoint inhibitor therapy. Tumor-infiltrating lymphocytes (TILs) and tertiary lymphoid structures (TLSs) are key components of the tumor microenvironment (TME) and have been implicated in the immune response to cancer. However, the role and difference of TLSs and TILs in patients with cholangiocarcinoma remains unclear. This study elucidates their contributions to the TME.Methods We examined 16 tumor samples from a single-arm, phase II trial of nivolumab plus modified gemcitabine and S-1 and various datasets. Immunohistochemistry and RNA sequencing were employed to assess TLSs and TILs presence and activity. Differential gene expression and signature of immune cell composition were examined by GeoMx Digital Spatial Profiler and Cancer Transcriptome Altas analysis.Results TLS-positive (N=7) patients demonstrated significantly better immunotherapy outcomes compared with TLS-negative (N=9) patients, including higher objective response rates (71% vs 0%) and disease control rates (100% vs 67%). The presence of TLSs correlated with improved progression-free and overall survival (p=0.03). TLSs were associated with “inflamed” tumors characterized by substantial immune infiltration, particularly involving T and B cells. Gene expression analyses identified significant upregulation of B cell-related genes in TLSs. Additionally, TLSs exhibited higher properties of memory B cells and myeloid dendritic cells but lower levels of innate immune cells compared with TILs. T cells within TLSs showed elevated expression of precursor-exhausted-related genes and lower cytotoxicity signature. Furthermore, TILs in TLS-positive tumors had higher levels of exhaustion signatures compared with TILs in TLS-negative tumors. Clinical data corroborated these findings, with higher PD-L1 and LAG-3 expression in TLS-positive tumors.Conclusion Our findings revealed that TILs in TLS-positive tumors have more exhausted T cell signature and PD-1 and LAG-3 protein expression in CCA which support our clinical finding. TLSs can predict favorable immunotherapy responses in patients with cholangiocarcinoma, highlighting their potential as a biomarker and therapeutic target to enhance treatment efficacy.

Background: Increasingly, more evidence has shown that inflammation stress and the tumor microenvironment pose a negative effect on targeted therapy. The neutrophil-to-lymphocyte ratio is considered to be a surrogate biomarker of inflammation and can predict pazopanib treatment effect in non-adipocytic soft-tissue sarcoma (STS). The role of the pan-immune-inflammation value (PIV) in STS is still yet to be determined. Objectives: We sought whether the pre-treatment PIV could be applied to predict the response of pazopanib in STS. Design: We conducted a retrospective analysis of 75 patients who had been treated with pazopanib for recurrent or metastatic non-adipocytic STS. Methods: Our cohort was stratified into either a pre-treatment high PIV group with PIV ⩾310 ( n = 45) or a low PIV group with PIV <310 ( n = 30). We compared their clinical features and outcomes. Cox regression analysis was employed to determine the risk factors of disease progression and mortality. Kaplan–Meier survival curves were utilized to assess both the progression-free survival (PFS) and overall survival (OS). Results: The results revealed that a pre-treatment high PIV (⩾310) is a risk factor for progression under pazopanib (hazard ratio: 1.91; 95% confidence interval: 1.08–3.36; p = 0.025). The median PFS and OS of the pre-treatment high PIV group were found to be significantly lower than the low PIV group (0.33 vs 0.75 years; p = 0.023, 0.46 vs 1.63 years; p = 0.025). Conclusion: High pre-treatment PIV in STS patients may indicate an elevated risk of disease progression and mortality. Pre-treatment PIV reflects inflammation stress and acts as a practical biomarker for STS patients treated with pazopanib.

The lncRNA tumor protein translationally controlled 1-antisense RNA 1 (TPT1-AS1) is known for its oncogenic role in various cancers, but its impact on the pathological progression of prostate cancer remains unclear. Our previous study demonstrated that the RE1-silencing transcription factor (REST) regulates neuroendocrine differentiation (NED) in prostate cancer (PCA) by derepressing specific long non-coding RNAs (lncRNAs), including TPT1-AS1. In this study, we revealed that TPT1-AS1 is overexpressed in LNCaP and C4-2B cells after IL-6 and enzalutamide treatment. By analyzing The Cancer Genome Atlas (TCGA) prostate adenocarcinoma dataset, we detected upregulated TPT1-AS1 expression in neuroendocrine-associated PCA but not in prostate adenocarcinoma. Single-cell RNA sequencing data further confirmed the increased TPT1-AS1 levels in neuroendocrine prostate cancer (NEPC) cells. Surprisingly, functional experiments indicated that TPT1-AS1 overexpression had no stimulatory effect on NED in LNCaP cells and that TPT1-AS1 knockdown did not inhibit IL-6-induced NED. Transcriptomic analysis revealed the essential role of TPT1-AS1 in synaptogenesis and autophagy activation in neuroendocrine differentiated PCA cells induced by IL-6 and enzalutamide treatment. TPT1-AS1 was found to regulate the expression of autophagy-related genes that maintain neuroendocrine cell survival through autophagy activation. In conclusion, our data expand the current knowledge of REST-repressed lncRNAs in NED in PCA and highlight the contribution of TPT1-AS1 to protect neuroendocrine cells from cell death rather than inducing NED. Our study suggested that TPT1-AS1 plays a cytoprotective role in NEPC cells; thus, targeting TPT1-AS1 is a potential therapeutic strategy.

Hepatocellular carcinoma (HCC) is a leading cause of death worldwide. Current therapies are effective for HCC patients with early disease, but many patients suffer recurrence after surgery and have a poor response to chemotherapy. Therefore, new therapeutic targets are needed. We analyzed gene expression profiles between HCC tissues and normal adjacent tissues from public databases and found that the expression of genes involved in lipid metabolism was significantly different. The analysis showed that AKR1C3 was upregulated in tumors, and high AKR1C3 expression was associated with a poorer prognosis in HCC patients. In vitro, assays demonstrated that the knockdown of AKR1C3 or the addition of the AKR1C3 inhibitor indomethacin suppressed the growth and colony formation of HCC cell lines. Knockdown of AKR1C3 in Huh7 cells reduced tumor growth in vivo. To explore the mechanism, we performed pathway enrichment analysis, and the results linked the expression of AKR1C3 with prostaglandin F2 alpha (PGF2α) downstream target genes. Suppression of AKR1C3 activity reduced the production of PGF2α, and supplementation with PGF2α restored the growth of indomethacin-treated Huh7 cells. Knockdown of the PGF receptor (PTGFR) and treatment with a PTGFR inhibitor significantly reduced HCC growth. We showed that indomethacin potentiated the sensitivity of Huh7 cells to sorafenib. In summary, our results indicate that AKR1C3 upregulation may promote HCC growth by promoting the production of PGF2α, and suppression of PTGFR limited HCC growth. Therefore, targeting the AKR1C3-PGF2α-PTGFR axis may be a new strategy for the treatment of HCC.

Immunotherapy in combination with chemotherapy is the current treatment of choice for frontline programmed cell death ligand 1 (PD-L1)-positive gastric cancer. However, the best treatment strategy remains an unmet medical need for elderly or fragile patients with gastric cancer. Previous studies have revealed that PD-L1 expression, Epstein-Barr virus association, and microsatellite instability-high (MSI-H) are the potential predictive biomarkers for immunotherapy use in gastric cancer. In this study, we showed that PD-L1 expression, tumor mutation burden, and the proportion of MSI-H were significantly elevated in elderly patients with gastric cancer who were older than 70 years compared with patients younger than 70 years from analysis of The Cancer Genome Atlas gastric adenocarcinoma cohort [≥70/<70: MSI-H: 26.8%/15.0%, P =0.003; tumor mutation burden: 6.7/5.1 Mut/Mb, P =0.0004; PD-L1 mRNA: 5.6/3.9 counts per million mapped reads, P =0.005]. In our real-world study, 416 gastric cancer patients were analyzed and showed similar results (≥70/<70: MSI-H: 12.5%/6.6%, P =0.041; combined positive score ≥1: 38.1%/21.5%, P <0.001). We also evaluated 16 elderly patients with gastric cancer treated with immunotherapy and revealed an objective response of 43.8%, a median overall survival of 14.8 months, and a median progression-free survival of 7.0 months. Our research showed that a durable clinical response could be expected when treating elderly patients with gastric cancer with immunotherapy, and this approach is worth further study.

Genomic biomarkers predicting immune checkpoint inhibitor (ICI) treatment outcomes for Asian metastatic melanoma have been rarely reported. This study presents data on next‐generation sequencing (NGS) and tumour microenvironment biomarkers in 33 cases.

Sclerosing pneumocytoma is a rare and distinct lung neoplasm whose histogenesis and molecular alterations are the subject of ongoing research. Our recent study revealed that AKT1 internal tandem duplications (ITD), point mutations, and short indels were present in almost all tested sclerosing pneumocytomas, suggesting that AKT1 mutations are a major driving oncogenic event in this tumor. Although the pathogenic role of AKT1 point mutations is well established, the significance of AKT1 ITD in oncogenesis remains largely unexplored. We conducted comprehensive genomic and transcriptomic analyses of sclerosing pneumocytoma to address this knowledge gap. RNA-sequencing data from 23 tumors and whole-exome sequencing data from 44 tumors were used to obtain insights into their genetic and transcriptomic profiles. Our analysis revealed a high degree of genetic and transcriptomic similarity between tumors carrying AKT1 ITD and those with AKT1 point mutations. Mutational signature analysis revealed COSMIC signatures 1 and 5 as the prevailing signatures of sclerosing pneumocytoma, associated with the spontaneous deamination of 5-methylcytosine and an unknown etiology, respectively. RNA-sequencing data analysis revealed that the sclerosing pneumocytoma gene expression profile is characterized by activation of the PI3K/AKT/mTOR pathway, which exhibits significant similarity between tumors harboring AKT1 ITD and those with AKT1 point mutations. Notably, an upregulation of SOX9, a transcription factor known for its involvement in fetal lung development, was observed in sclerosing pneumocytoma. Specifically, SOX9 expression was prominent in the round cell component, whereas it was relatively lower in the surface cell component of the tumor. To the best of our knowledge, this is the first comprehensive investigation of the genomic and transcriptomic characteristics of sclerosing pneumocytoma. Results of the present study provide insights into the molecular attributes of sclerosing pneumocytoma and a basis for future studies of this enigmatic tumor.