Background: FOLFIRI (5-FU + leucovorin + irinotecan) plus ramucirumab is one of the standards in second-line metastatic colorectal cancer (CRC) patients progressing after treatment with oxaliplatin/fluoropyrimidine with bevacizumab, but there is no evidence on its efficacy without prior bevacizumab. Moreover, VEGF-D has not been confirmed as a predictive biomarker for ramucirumab’s efficacy, either. Methods: The RAINCLOUD study was a multicenter, single-arm, phase II trial conducted in Japan. Patients with recurrent CRC pretreated with fluoropyrimidine and oxaliplatin without bevacizumab were analyzed. The primary endpoint was progression-free survival (PFS). The secondary endpoints measured were overall survival (OS), overall response rate (ORR), and safety. Results: A total of 48 patients were enrolled from 15 sites between September 2017 and September 2020. Their median age was 63.5 years (25~77), 20.1% had a right-sided tumor, and 68.8% had RAS-mutant cancer. The median PFS was 8.9 months (90% CI: 6.3–11.8), so the primary endpoint was met. Their median OS and ORR were 22.3 months (95% CI: 17.4-NA) and 41.7% (95% CI: 4.9–7.6), respectively. An incidence of grade 3/4 adverse events that reached over 5% applied to neutropenia (44%), leucopenia (10%), and hypertension (8%). In the biomarker analysis, the serum VEGF-D levels post-treatment were higher than those pre-treatment, but the PFS in those with high VEGF-D levels trended towards being worse than that in those with low VEGF-D (7.6M/5.6M (p = 0.095; HR: 0.56)). Instead, those with low TSP-2 had a better PFS than those with high TSP-2 (7.5M/4.3M (p = 0.022; HR: 0.45)). Conclusions: Our data suggested that FOLFIRI plus ramucirumab was effective and tolerable for CRC refractory to fluoropyrimidine and oxaliplatin without anti-angiogenesis. Serum VEGF-D levels may not be predictive but TSP-2 may be a potential prognostic biomarker for ramucirumab’s efficacy.

Background: The short-term outcomes of chemoradiotherapy (CRT) with S-1 (a combination of tegafur, gimeracil, and oteracil) following limited surgery for patients with T1 or T2 lower rectal cancer have shown encouraging results. Objectives: This study was designed to delve deeper into the long-term outcomes of CRT with S-1 after limited surgery, with the goal of evaluating both the long-term efficacy and potential risks associated with this treatment approach in patients diagnosed with T1 or T2 lower rectal cancer. Methods: This was conducted as a multicenter, single-arm, prospective phase II trial. The patient population consisted of individuals clinically diagnosed with either T1 or T2 lower rectal or anal canal cancer, with a maximum tumor diameter of 30 mm and classified as N0 or M0. Patients underwent local excision or endoscopic resection. After surgery, CRT with S-1 was administered to patients meeting several criteria, including the confirmation of well-differentiated or moderately differentiated adenocarcinoma, negative surgical margins, submucosal invasion depth of ≥1000 µm, and high tumor-budding grade (2/3). The primary endpoint of this study was relapse-free survival, while secondary endpoints included local recurrence-free survival, overall survival, anal sphincter preservation rate, and safety. Results: A total of 52 patients were included, with pathological diagnoses revealing T1 in 36 patients and T2 in 16 patients. The 3-year and 5-year relapse-free survival rates were 90.17% and 85.87%, respectively. The 3-year and 5-year local recurrence-free survival rates were 90.17% and 88.07%, respectively, while the 3-year and 5-year overall survival rates were 94.03% and 91.94%, respectively. Conclusions: CRT with S-1 after limited surgery for T1 lower rectal cancer demonstrated favorable outcomes in terms of recurrence, survival, and local control rates while effectively maintaining anal function in patients. However, further treatment approaches may be necessary to improve outcomes for patients diagnosed with stage T2 lower rectal cancer

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.

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.