Kenji Kawada

The benefits of fibroblast growth factor receptor (FGFR) inhibitor therapy in several malignancies are often limited by the emergence of drug resistance. Recent studies have uncovered its molecular mechanisms, particularly in urothelial cancer and cholangiocarcinoma, but not yet in colorectal cancer (CRC). Here, we investigated FGFR inhibitor resistance mechanisms using patient-derived CRC stem-cell (SC) spheroids. To obtain FGFR inhibitor-resistant CRC-SC spheroid lines, we performed long-term in vitro treatment with FDA-approved pan-FGFR inhibitors, erdafitinib and futibatinib. The paired drug-resistant and parental spheroid stem cells were subjected to RNA-sequencing for mutational and transcriptional profiling. We examined the involvement of differentially expressed genes in resistance mechanisms by functional tests and CRISPR-mediated gene disruption. We established 7 FGFR inhibitor-resistant CRC-SC lines from 4 parental lines that responded well to FGFR inhibitors. All resistant lines showed EGFR mRNA upregulation and PTPRO mRNA downregulation, correlating with reduced sensitivities to the pan-FGFR inhibitors, erdafitinib and futibatinib. EGF stimulation of these lines induced MAPK and PI3K-AKT pathway activation more potently than that of their parental lines. Additionally, PTPRO gene disruption in the parental lines conferred resistance to erdafitinib and futibatinib by upregulating EGFR. Importantly, the combination treatment of resistant lines with an EGFR inhibitor erlotinib effectively suppressed the MAPK and PI3K-AKT signaling. In conclusion, Enhanced EGFR signaling, partly driven by PTPRO downregulation, was a key mechanism of acquired resistance to FGFR inhibitors in RAS/RAF wild-type CRC. The combination treatment with FGFR and EGFR inhibitors can be a promising strategy to overcome this resistance.

Background: Owing to significant clinical heterogeneity, the achievement of accurate survival forecasting for individuals with colorectal cancer and peritoneal metastasis continues to be a complex undertaking. We aimed to transcend traditional prognostic limitations by evaluating machine learning boosting models against standard regression-based methods in terms of estimating overall survival (OS). Methods: Utilizing a multi-institutional registry of 150 patients diagnosed with synchronous peritoneal metastasis of colorectal cancer, we integrated 124 clinicopathological variables to refine our predictive models. Beyond standard preprocessing—including standardization and median imputation—we rigorously compared XGBoost and LightGBM against Ridge, Lasso, and linear regression via five-fold cross-validation. To specifically address right-censoring, an XGBoost Cox model was implemented and validated using Harrell’s C-index, with SHAP and LIME providing essential model interpretability. Results: Boosting models consistently outperformed linear alternatives, which struggled with high error rates and negative R2 values. Specifically, XGBoost achieved an MAE of 475 ± 60 and an RMSE of 585 ± 88. The XGBoost Cox model reached a C-index of 0.64 ± 0.06. SHAP analysis highlighted inflammatory markers and peritoneal disease extent as the most influential prognostic drivers. Conclusions: While boosting models offer a clear accuracy advantage over linear methods, their prognostic power remains moderate. These findings underscore the potential of ensemble learning in oncology, yet mandate external validation before these tools can be integrated into clinical decision-making.

Aim: The adoption of intracorporeal anastomosis (IA) in minimally invasive colectomy for colon cancer has increased; however, its mid‐term oncological outcomes compared with those of extracorporeal anastomosis (EA) remain unclear. Methods: This multicenter retrospective cohort study analyzed patients who underwent minimally invasive curative colectomy for colon cancer between 2018 and 2022 from the Kyoto Colorectal Surgery Group database. Propensity score matching (1:4) was applied to balance patient, tumor, and surgical characteristics between IA and EA groups. Primary outcome was 2‐year recurrence‐free survival (RFS). Secondary outcomes included 2‐year cumulative incidence of peritoneal, anastomotic, and distant recurrence, and postoperative complications. Results: Among 1904 eligible patients (IA, n = 253; EA, n = 1651), matching yielded 224 IA and 862 EA cases with well‐balanced baseline covariates. After matching, the 2‐year RFS tended to be better in the IA group (hazard ratio [HR], 0.66; 95% confidence interval [CI], 0.41–1.06), though not statistically significant. The incidences of peritoneal and anastomotic recurrence were similar between the groups (risk difference [RD], 0.13%; 95% CI, −0.76% to 1.02%; and RD, −0.25%; 95% CI, −0.61% to 0.10%, respectively), whereas distant recurrence was lower in the IA group (RD, −4.79%; 95% CI, −6.75% to −2.83%). Subgroup analyses suggested a greater reduction in distant recurrence in patients with advanced tumors. Conclusion: IA demonstrated a comparable risk of peritoneal and anastomotic recurrence and suggested a potential reduction in distant recurrence without an associated deterioration in 2‐year RFS. These findings support IA as an oncologically reasonable option, even for advanced‐stage tumors.

Background: Prognostic prediction for colorectal cancer patients with peritoneal metastasis remains challenging due to clinical and biological heterogeneity. We aimed to evaluate the utility of machine learning, comparing high-performance boosting models with interpretable regression approaches for overall survival (OS) prediction. Methods: We analyzed a multi-institutional registry cohort of 150 colorectal cancer patients with synchronous peritoneal metastases. A total of 124 variables were included; continuous variables were standardized, categorical variables were one-hot encoded, and missing values were imputed using the median. Models included XGBoost, LightGBM, Ridge regression, Lasso regression, and linear regression. Training was performed with 3-fold cross-validation, and hyperparameters were optimized using Optuna. Evaluation metrics included mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R²). Model interpretability was assessed using SHAP values, LIME local explanations, and regression coefficients. Results: Boosting models consistently outperformed linear models. XGBoost achieved the best performance (MAE 424, RMSE 526, R² = 0.04), while LightGBM showed comparable accuracy. In contrast, Ridge, Lasso, and Linear regression yielded high errors (MAE > 900, RMSE > 1200) with negative R² values, indicating poor predictive ability. SHAP analysis highlighted systemic inflammation markers (CRP, BUN), surgical assessment of tumor depth, operative factors (time, bleeding), and peritoneal metastasis characteristics as major determinants of OS. LIME analyses further provided case-specific interpretability, identifying feature contributions in long-, intermediate-, and short-term survivors. Conclusion: Boosting models, particularly XGBoost, demonstrated superior performance compared with traditional regression models in predicting OS for colorectal cancer patients with peritoneal metastasis, although absolute predictive accuracy remains modest. Integration of SHAP and LIME linked model outputs with clinically plausible prognostic factors, enhancing interpretability. Ensemble learning may provide a promising adjunct for prognostic assessment and should be validated in larger, genomically enriched cohorts.

Suppl. Fig. S2 with its legend on page 2

Suppl. Fig. S4 with its legend on page 2

Suppl. Fig. S3 with its legend on page 2