BACKGROUND:C-reactive protein (CRP) is a widely used inflammatory biomarker that has been related to colorectal cancer risk. However, observational studies are prone to confounding and reverse causality. METHODS:We conducted a two-sample Mendelian randomization using CRP genome-wide association study (GWAS) data from 59,605 Korean individuals and colorectal cancer GWAS data from 23,572 cases and 48,700 controls from an East Asia population. The analysis had 80% power to detect an OR of 1.12 for colorectal cancer risk per twofold increase in CRP levels. RESULTS:Genetically predicted serum CRP levels (per twofold increase) were not significantly associated with colorectal cancer risk (inverse-variance weighted OR, 0.995; 95% confidence interval, 0.893-1.109; P value, 0.929). Null findings remained consistent in sensitivity analyses excluding the horizontally pleiotropic effect. CONCLUSIONS:Despite sufficient statistical power, little evidence supported a causal association between CRP and colorectal cancer risk in East Asians. IMPACT:Our findings suggest that CRP is unlikely to be a key determinant of colorectal carcinogenesis, aligning with prior studies in European populations.

Supplementary Table 1: Study population and exclusion criteria by cohorts as established by the ACC reproductive factor working group

BACKGROUND:Both genetic and epigenetic changes, particularly deoxyribonucleic acid (DNA) methylation, play crucial roles in gastric cancer development. Nuts, highlighted for their anti-inflammatory effects, may help maintain proper DNA methylation patterns. Thus, we hypothesized that high intake of nuts may interact with the DNA methylation-related genetic factor, DNA methyltransferase 1 (DNMT1) rs16999593 polymorphism in gastric cancer risk. OBJECTIVES:We aimed to investigate the interaction between dietary nut consumption and DNMT1 rs16999593 polymorphism, exploring their combined effects on gastric cancer risk. METHODS:This case-control study at the National Cancer Center in Korea investigated the impact of nut consumption and DNMT1 rs16999593 polymorphism on gastric cancer risk, involving 1245 participants. Genetic analysis and dietary intake assessment were performed. Logistic regression models estimated odds ratios (ORs) and 95% confidence intervals (CIs). RESULTS:High nut consumption was strongly associated with reduced risk of gastric cancer among women (adjusted OR: 0.42; 95% CI: 0.20, 0.87; P for trend = 0.026 for ≥3 servings/wk group vs. none). The dominant model of DNMT1 rs16999593 exhibited an inverse trend in gastric cancer risk (total: OR: 0.75; 95% CI: 0.54, 1.05), although not statistically significant. However, those with high nut consumption and homozygous for the major allele (T) showed lower gastric cancer risk, with significant interaction in overall (OR: 0.52; 95% CI: 0.27, 1.01; P-interaction = 0.045) and in women (OR: 0.17; 95% CI: 0.05, 0.57; P-interaction = 0.001). CONCLUSIONS:Our study demonstrates a significant inverse association between nut consumption and gastric cancer risk, particularly in women. On a multiplicative scale, our findings highlight the potential negative interactions between nut consumption and DNMT1 rs16999593 genetic polymorphism in gastric cancer risk in the total population and in women.

BACKGROUND:Gastric cancer (GC) is a global health concern, with incidence and mortality varying significantly across populations. The spatial heterogeneity of GC may reflect differences in prevention strategies and their effectiveness, highlighting the need to identify high-risk areas and their possible contributing factors. OBJECTIVE:Leveraging the decline of GC incidence since 2012 affected by nationwide cancer prevention programs in South Korea, this study aimed to identify high- and low-risk clusters and explore the changes in spatial clusters of GC incidence between 2009-2013 and 2014-2018 before and after the decline, respectively. In addition, we identified geographic characteristics associated with cluster. METHOD:Using national cancer registry and geographic data across 243 districts, we performed spatial clustering analyses in each period. We identified the high- and low-risk areas as the overlapping clusters by two common clustering approaches of local Moran's I and Getis Ord Gi*. Then, we compared 23 geographic characteristics between high- and low-risk areas and two periods. RESULTS:The average GC incidence rate declined in both high- (96-79 per 100,000 people) and low-risk districts (73-61) between 2009-2013 and 2014-2018. While cluster locations remained stable, low-risk districts expanded notably (26-37) and high-risk districts slightly reduced (31-28). As opposed to little geographic characteristics that showed the significant difference between high- and low-risk areas in the early period, the later period gave large green space, frequent regular walking, and reduced self-related obesity in the low-risk area, in addition to sociodemographic advantages, compared to those in the high-risk area (p-value <0.002). CONCLUSION:Our findings suggest the potential effectiveness of lifestyle- and/or environment-focused prevention for GC incidence and the need of locally-tailored strategies to reduce cancer burden.

Background: Enzymes encoded by Helicobacter pylori are involved in various metabolic processes. Using a Mendelian randomization (MR) framework, we investigated the metabolites associated with six antibodies against H. pylori and explored the potential underlying biological pathways. Methods: A meta-analysis of 65 genome-wide association studies was conducted to assess the genetic predisposition to approximately 3,550 metabolites. Summary-level data for > 10 million genetic variants associated with H. pylori antibodies were extracted from the UK Biobank. MR analysis was performed using inverse-variance weighting, weighted median, and Egger regression to ensure robust findings. Significant metabolites were further analyzed using enrichment analysis to identify the relevant biological pathways. Results: A total of 100 metabolites were positively associated with H. pylori antibodies (cytotoxin-associated gene A [CagA], 16; catalase, 25; GroEL, 16; outer membrane protein [OMP], 22; urease [UREA], 15; and vacuolating cytotoxin [VacA], 6). These metabolites were linked to pathways involving the metabolism of alanine, aspartate, glutamate, glycine, serine, threonine, purine, and tryptophan in four antibodies. Additionally, 67 metabolites were negatively associated with H. pylori antibodies (CagA, 11; catalase, 6; GroEL, 13; OMP, 11; UREA, 11; and VacA, 15). These metabolites were primarily involved in pyrimidine metabolism in the three antibodies. Conclusion: Our study identified numerous metabolites linked to H. pylori antibody levels, indicating that metabolic alterations are associated with infection. These changes were particularly enriched in pathways involved in amino acid, nucleotide, and coenzyme metabolism and biosynthesis. These findings highlight the systemic metabolic impact of H. pylori infection and offer insights into the biological mechanisms underlying host-pathogen interactions.

Background:Gastrointestinal (GI) cancers are a significant health concern in South Korea. Recently, machine learning (ML) models have emerged as powerful tools to support early screening efforts and identify people at risk before disease onset. However, the low incidence of GI malignancies in prospective cohorts leads to severe class imbalance, often causing ML models to favor the majority "healthy" class at the expense of clinical sensitivity. Objective:This study aimed to evaluate class imbalance mitigation strategies and develop ML-based GI cancer risk prediction models using noninvasive and minimally invasive predictors linked to modifiable behavioral and metabolic risk factors. Methods:We analyzed a prospective cohort (n=7652) with 156 incident GI cancer cases (2%) identified over a 14-year follow-up period. The data were randomly split into training (5356/7652, 70%) and testing (2296/7652, 30%) sets. To address class imbalance while preserving observed population structure, we developed a patient-centered undersampling technique (PCUSTe) based on the logic of frequency-matched case-control studies. PCUSTe was compared with commonly used resampling approaches, including synthetic minority oversampling (SMOTE), adaptive synthetic sampling (ADASYN), and SMOTE with edited nearest neighbors (ENN). Six classifiers were implemented, including both batch and incremental training variants. To account for the prior shift introduced by resampling, probability correction was applied. Model performance was evaluated on the independent test set using a classification threshold equal to the observed event proportion (cumulative incidence) in the training data and then across thresholds reflecting incidence values between 1% and 5%. Primary performance metrics included sensitivity, specificity, Matthews correlation coefficient, and area under the receiver operating characteristic curve (AUC). Results:Models trained using PCUSTe demonstrated improved sensitivity compared with standard resampling techniques, particularly for more complex classifiers. The incrementally trained stochastic gradient descent model achieved the highest overall performance trained on PCUSTe data with a sensitivity of 0.77 (95% CI 0.64-0.89), specificity of 0.65 (95% CI 0.63-0.67), AUC of 0.77 (95% CI 0.70-0.84), and Matthews correlation coefficient of 0.12 (95% CI 0.08-0.16). In contrast, logistic regression achieved balanced performance without resampling (sensitivity 0.70, 95% CI 0.57-0.83; specificity 0.71, 95% CI 0.69-0.72; AUC 0.75, 95% CI 0.68-0.82). Our results showed that PCUSTe primarily enhanced sensitivity in more complex models at the expense of specificity. Conclusions:Integrating epidemiological principles, including covariate frequency matching and threshold selection based on the observed cumulative incidence in the training data, improved minority class detection in GI cancer risk prediction. However, model performance varied by algorithm, and in some cases, decision threshold adjustment alone achieved comparable or superior results to data resampling. These findings highlight the importance of carefully selecting imbalance mitigation strategies based on modeling objectives. The resulting models achieved sensitivity levels that may be suitable for early risk identification in cohort settings and could contribute to personalized risk stratification and targeted prevention or screening strategies.

Abstract Combination chemotherapy has long been central to the management of colorectal cancer, and treatment patterns have evolved in response to emerging clinical evidence, changing guidelines, and the availability or withdrawal of specific agents. This study aims to characterize changes in medication utilization over time by providing a descriptive analysis of temporal changes in anticancer drug use among Korean colorectal cancer patients treated at the National Cancer Center (NCC) of Korea. Two patient cohorts were examined: individuals treated between 2000-2004 and those treated between 2010-2020. Sixteen anticancer agents administered to ≥10 patients in both cohorts were included. A total of 1,188 patients from the 2000-2004 cohort and 1,043 patients from the 2010-2020 cohort were analyzed using Stata/BE 18.0. In the 2000-2004 cohort, monotherapy was the most predominant treatment approach (55.89%), followed by two-drug combination therapy (17.68%). Among monotherapies, 5-fluorouracil (5-FU) was most frequently used (45.96%). The most common two-drug regimen during this period was capecitabine plus 5-FU (4.88%). By 2010-2020, treatment patterns shifted as two-drug combination chemotherapy became more common than monotherapy (38.73% vs. 32.02%). The 5-FU plus oxaliplatin combination emerged as the most frequently used regimen (25.98%). Capecitabine was the most common regimen for monotherapy (21%). Although the use of capecitabine monotherapy increased in 2010-2020 compared to 2000-2004 (21% vs. 5.56%), the difference in overall capecitabine usage was not statistically significant (p = 0.872), likely because capecitabine had frequently been used in combination regimens in earlier years. In contrast, use of 5-FU significantly declined (p < 0.001), while use of oxaliplatin and bevacizumab markedly increased (both p < 0.001). Meanwhile, tegafur/uracil and megestrol usage decreased significantly (both p < 0.001). This descriptive analysis highlights shifts in chemotherapy use among colorectal cancer patients treated at the NCC of Korea across two distinct periods. The earlier cohort (2000-2004) was characterized by predominant use of 5-FU-based monotherapy, whereas the later cohort (2010-2020) showed a marked transition toward two-drug combinations. These patterns mirror global changes in anticancer therapy drug usage. Further analyses should link treatment patterns to clinical outcomes, including evaluations of overall survival (OS) and disease-free survival (DFS), clarification of treatment lines, and resolution of overlapping-prescription or regimen-coding issues. This study was approved by the Institutional Review Board of the NCC (NCC2025-0115). Citation Format: Yu Jin Lim, Jeongseon Kim. Anti-cancer drug usage analysis among Korean patients with colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 293.

Supplementary Table 3: Pooled relative risks for recategorized age at menarche and age at menopause & incident thyroid cancer risk, Overall and papillary type

Susceptibility transcription factors (TF) whose DNA bindings are altered by genetic variants regulating colorectal cancer (CRC) risk genes remain poorly defined. Using generalized linear mixed models, we analyze 218 TF ChIP-Seq datasets alongside GWAS data from 100,204 CRC cases and 154,587 controls of East Asian and European ancestries. We identify 51 TFs and TF-cofactor interactions, including VDR-cofactors, as key regulators of CRC risk. Integrating these TF insights with transcriptome-wide association studies (TWAS), we further evaluate associations between genetically predicted gene expression, alternative splicing, and alternative polyadenylation with CRC risk, using RNA-seq data from 364 Asian-ancestry and 707 European-ancestry individuals. Multi-ancestry TWAS identify 222 risk genes, including 95 novel genes and 48 potentially druggable targets. Single-cell analysis provides additional functional evidence supporting ~45% of these genes, and experimental validation confirms oncogenic roles for RHPN2, IRS2, and TXN. Our findings elucidate key TF-gene regulatory networks and uncover novel CRC risk genes.

Supplementary Table 2: Distribution of total cases according to histology according to participating cohorts

Supplementary Figure 3: Forest plots of the pooled hazard ratios (HRs) and 95% confidence intervals (CIs) generated by combining cohort-specific HRs for the association between reproductive factors and the overall risk of thyroid cancer in the Asia Cohort Consortium. A - Forest plot for the pooled HRs and CIs for breastfeeding status and thyroid cancer risk, overall B - Forest plot for the pooled HRs and CIs for postmenopausal status and thyroid cancer risk, overall C - Forest plot for the pooled HRs and CIs for age at menopause and thyroid cancer risk, overall

This study aimed to examine differences in the association between reproductive factors and breast cancer (BC) risk across ethnic groups, particularly Asians and non-Asians, and to explore temporal trends through meta-analysis. The study focused on epidemiologic research published up to August 31, 2022, examining reproductive factors related to BC risk and family history. All effect sizes were calculated using a random-effect model. The protective effect of the higher number of childbirths against BC was stronger in Asians than in Europeans or Americans (childbirths ≥ 2 vs. 1; Asians, relative risk [RR]: 0.66, 95% CI: 0.59-0.74; Europeans, RR: 0.89, 95% CI: 0.86-0.92; Americans, RR: 0.91, 95% CI: 0.87-0.96). Similarly, the effect of high parity was more pronounced in Asians than in Americans and Europeans (Asians, RR: 0.72, 95% CI: 0.58-0.89; Europeans, RR: 0.81, 95% CI: 0.74-0.88; Americans, RR: 0.84, 95% CI: 0.76-0.92). In contrast, no significant differences among populations were found in BC risks associated with combined hormone replacement therapy use. While the association between family history and BC risk appeared to differ by ethnicity, no temporal change was observed (< 2010, RR: 1.58, 95% CI: 1.40-1.78; ≥ 2010, RR: 1.57, 95% CI: 1.46-1.67). These results suggest that some reproductive factors associated with BC differ across ethnicities and time trends, perhaps due to the prevalence of reproductive factors and the baseline hazard of BC.

Supplementary Methods 1: Details on the development of the Asia Cohort Consortium reproductive factor working group protocol

Supplementary Figure 4: Forest plots of the pooled hazard ratios (HRs) and 95% confidence intervals (CIs) generated by combining cohort-specific HRs for the association between reproductive factors and the overall risk of thyroid cancer in the Asia Cohort Consortium. A - Forest plot for the pooled HRs and CIs for oral contraceptive use and thyroid cancer risk, overall B - Forest plot for the pooled HRs and CIs for hormone replace therapy use and thyroid cancer risk, overall

Supplementary Figure 5: Forest plot of stratified analysis between number of children/deliveries and thyroid cancer risk by birth years in the Asia Cohort Consortium. The Pooled Hazard Ratios (HRs) with 95% Confidence intervals (CIs) were generated by combining cohort-specific HRs. Models were adjusted for smoking status, alcohol drinking status and Body mass index. a Significant (p-value <0.05) trend across categories of the reproductive factor. b Significant (p-value <0.05) for interaction indicating a modifying effect on the association between the reproductive factor and thyroid cancer risk.

Supplementary Figure 6: Forest plot of pooled hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between reproductive factors and thyroid cancer risk, by age of diagnosis in the Asia Cohort Consortium. The Pooled Hazard Ratios (HRs) with 95% Confidence intervals (CIs) were generated by combining cohort-specific HRs. Models were adjusted for smoking status, alcohol drinking status and Body mass index. a Significant (p-value <0.05). b The model included all 9 cohorts. c The model for Breastfeeding included 6 cohorts, that for Oral contraceptive use included 5 cohorts and that for hormone replacement therapy included 6 cohorts.

Supplementary Figure 1: Forest plots of the pooled hazard ratios (HRs) and 95% confidence intervals (CIs) generated by combining cohort-specific HRs for the association between reproductive factors and the overall risk of thyroid cancer in the Asia Cohort Consortium. A - Forest plot for the pooled HRs and CIs for age at menarche and thyroid cancer risk, overall B - Forest plot for the pooled HRs and CIs for age at first delivery and thyroid cancer risk, overall

Supplementary Figure 2: Forest plots of the pooled hazard ratios (HRs) and 95% confidence intervals (CIs) generated by combining cohort-specific HRs for the association between reproductive factors and the overall risk of thyroid cancer in the Asia Cohort Consortium. A - Forest plot for the pooled HRs and CIs for parity status and thyroid cancer risk, overall B - Forest plot for the pooled HRs and CIs for number of children/deliveries and thyroid cancer risk, overall C - Forest plot for the pooled HRs and CIs for recategorized number of children/deliveries and thyroid cancer risk, overall

OBJECTIVE:The role of overall macronutrient balance in gastric cancer (GC) risk remains unclear. This study investigated the associations of the Ketogenic Ratio (KR) and Simplified Ketogenic Ratio (sKR), as composite indicators of macronutrient composition, with GC risk, and examined potential modification by the ADIPOQ rs1501299 genotype. METHODS:A hospital-based case-control study was conducted among 376 GC cases and 752 controls. Dietary intake was assessed using a validated semi-quantitative food frequency questionnaire. Logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). Interactions between KR, sKR, the ADIPOQ rs1501299 polymorphism, and GC risk were evaluated using multiplicative terms in regression models. RESULTS:The ADIPOQ rs1501299 polymorphism alone was not associated with GC risk. Higher KR and sKR were significantly associated with reduced GC risk (KR: OR = 0.64; 95% CI = 0.44-0.93; sKR: OR = 0.59; 95% CI = 0.40-0.86). Notably, a significant gene-diet interaction was observed strictly for sKR, where this association was more pronounced among overweight/obese individuals carrying the ADIPOQ rs1501299 G/T or T/T genotypes (OR = 0.33, 95% CI: 0.16-0.66, P interaction = 0.038). In contrast, no significant interaction was found between KR and the ADIPOQ rs1501299 genotype. CONCLUSION:Higher KR and sKR are inversely associated with GC risk. The ADIPOQ rs1501299 genotype significantly modifies the association between sKR and GC risk, specifically among overweight/obese individuals. Future studies should focus on developing and validating standard measurements in nutritional epidemiology to elucidate biological pathways from gene-diet interactions.