Fumikazu Koyama
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Fumikazu Koyama has a total of 351 co-authors. Here are the top co-authors by number of shared publications:
Hisao Fujii
Department of Gastrointestinal Endoscopy and IBD Center, Heiwakai Yoshida Hospital, Nara, Japan
20 shared publicationsYoshiyuki Nakajima
Department of Surgery, Nara Medical University, Kashihara, Nara, Japan
18 shared publicationsTakayuki Nakamoto
Department of Surgery, Nara Medical University, Nara, Japan
13 shared publicationsTadashi Nakagawa
Department of Surgery, Nara Medical University, Kashihara, Nara, Japan
12 shared publicationsMasayuki Sho
Nara Medical University, Nara, Japan
10 shared publications- Takeshi Ueda
Department of Surgery, Minami-Nara General Medical Center, Nara, Japan
10 shared publications
Latest Publications
- Open Access
- 19 November 2026
- Published by Wiley in IJU Case Reports
- Vol. 9 (6)
Robot‐Assisted Transabdominal Repair of Complex Rectourethral Fistula Using Pedicled Gracilis Muscle Interposition With Concomitant Salvage Prostatectomy: A Case Report
Introduction: Rectourethral fistula is a challenging condition for which no standard surgical approach has yet been established. We report a case successfully treated with robot‐assisted radical prostatectomy, transabdominal fistula repair, and pedicled gracilis muscle interposition. Case Presentation: A 77‐year‐old man developed a complex rectourethral fistula after open resection of a tumor located between the prostate and rectum without preoperative biopsy. Postoperative pathological review revealed prostatic adenocarcinoma, raising concerns that future salvage surgery would be technically difficult in the event of recurrence. Therefore, we performed robot‐assisted radical prostatectomy with transabdominal fistula repair and pedicled gracilis muscle interposition. Postoperative imaging demonstrated viable interposed muscle without fistula recurrence. Conclusion: Robot‐assisted surgery enabled precise placement of a pedicled gracilis muscle flap in the deep pelvis. This case, together with the accompanying surgical video, demonstrates the feasibility of this approach for complex rectourethral fistula repair.
- 19 November 2026
- Published by Wiley in IJU Case Reports
- Vol. 9 (6)
Open AccessRobot‐Assisted Transabdominal Repair of Complex Rectourethral Fistula Using Pedicled Gracilis Muscle Interposition With Concomitant Salvage Prostatectomy: A Case Report
Introduction: Rectourethral fistula is a challenging condition for which no standard surgical approach has yet been established. We report a case successfully treated with robot‐assisted radical prostatectomy, transabdominal fistula repair, and pedicled gracilis muscle interposition. Case Presentation: A 77‐year‐old man developed a complex rectourethral fistula after open resection of a tumor located between the prostate and rectum without preoperative biopsy. Postoperative pathological review revealed prostatic adenocarcinoma, raising concerns that future salvage surgery would be technically difficult in the event of recurrence. Therefore, we performed robot‐assisted radical prostatectomy with transabdominal fistula repair and pedicled gracilis muscle interposition. Postoperative imaging demonstrated viable interposed muscle without fistula recurrence. Conclusion: Robot‐assisted surgery enabled precise placement of a pedicled gracilis muscle flap in the deep pelvis. This case, together with the accompanying surgical video, demonstrates the feasibility of this approach for complex rectourethral fistula repair.
- Open Access
- 31 October 2026
- Published by Korean Association for the Study of Intestinal Diseases in Intestinal Research
- Vol. 24 (4) , 776-784
Colitis-associated colorectal neoplasia in ulcerative colitis with primary sclerosing cholangitis: a nationwide study
Background/Aims: Ulcerative colitis (UC)-associated colorectal neoplasia (UCAN) in patients with UC and primary sclerosing cholangitis (PSC) has not been studied well in Japan. This retrospective study examined the clinicopathological features and prognosis of UCAN in patients with PSC-UC. Methods: A total of 808 patients with UCAN were enrolled from 1983 to 2020 and categorized into PSC (PSC-UCAN, n = 26) and no PSC (UCAN-alone, n = 782) groups. Clinicopathological features were compared between the 2 groups, and the 10-year overall survival (OS) and cancer-specific survival (CSS) were analyzed. Results: The PSC-UCAN group had a shorter UC duration before UCAN diagnosis (12.8 years vs. 16.9 years, P= 0.044), were younger at UCAN diagnosis (47.8 years vs. 53.3 years, P= 0.046), and developed UCAN more frequently in the right-sided colon (34.6% vs. 15.9%, P= 0.028) than the UCAN-alone group. The PSC-UCAN group showed a trend toward a lower proportion of high-grade dysplasia (19.2% vs. 30.7%) and a higher proportion of early-stage cancers (53.9% vs. 31.2%). The 10-year OS (64.6% vs. 79.3%, P=0.080) and CSS (80.8% vs. 83.9%, P=0.60) were comparable. Conclusions: Patients with PSC-UCAN showed earlier and younger development of UCAN than patients with only UCAN, with a high prevalence in the right-sided colon. Early-stage cancer was more frequently observed in the PSC-UCAN group, despite the shorter duration of UC. Patients with PSC-UC probably benefit from early initiation of surveillance colonoscopy.
- 31 October 2026
- Published by Korean Association for the Study of Intestinal Diseases in Intestinal Research
- Vol. 24 (4) , 776-784
Open AccessColitis-associated colorectal neoplasia in ulcerative colitis with primary sclerosing cholangitis: a nationwide study
Background/Aims: Ulcerative colitis (UC)-associated colorectal neoplasia (UCAN) in patients with UC and primary sclerosing cholangitis (PSC) has not been studied well in Japan. This retrospective study examined the clinicopathological features and prognosis of UCAN in patients with PSC-UC. Methods: A total of 808 patients with UCAN were enrolled from 1983 to 2020 and categorized into PSC (PSC-UCAN, n = 26) and no PSC (UCAN-alone, n = 782) groups. Clinicopathological features were compared between the 2 groups, and the 10-year overall survival (OS) and cancer-specific survival (CSS) were analyzed. Results: The PSC-UCAN group had a shorter UC duration before UCAN diagnosis (12.8 years vs. 16.9 years, P= 0.044), were younger at UCAN diagnosis (47.8 years vs. 53.3 years, P= 0.046), and developed UCAN more frequently in the right-sided colon (34.6% vs. 15.9%, P= 0.028) than the UCAN-alone group. The PSC-UCAN group showed a trend toward a lower proportion of high-grade dysplasia (19.2% vs. 30.7%) and a higher proportion of early-stage cancers (53.9% vs. 31.2%). The 10-year OS (64.6% vs. 79.3%, P=0.080) and CSS (80.8% vs. 83.9%, P=0.60) were comparable. Conclusions: Patients with PSC-UCAN showed earlier and younger development of UCAN than patients with only UCAN, with a high prevalence in the right-sided colon. Early-stage cancer was more frequently observed in the PSC-UCAN group, despite the shorter duration of UC. Patients with PSC-UC probably benefit from early initiation of surveillance colonoscopy.
- Open Access
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Treatment Patterns and Survival Outcomes in Locally Recurrent Rectal Cancer
BackgroundLocally recurrent rectal cancer (LRRC) remains a major clinical challenge, with limited high-quality evidence because of marked clinical and treatment heterogeneity. Although complete resection is the most important prognostic factor, the role of particle therapy remains unclear. This nationwide multicenter cohort study aimed to evaluate real-world treatment patterns and survival outcomes in patients with LRRC, with particular focus on particle therapy.MethodsPatients diagnosed with LRRC at 77 institutions in Japan between 2011 and 2023 were included and categorized according to treatment for the current recurrence: surgery, particle therapy, conventional radiotherapy with or without chemotherapy, chemotherapy alone, or palliative care. Overall survival (OS) was compared across treatment modalities, with a predefined subgroup analysis of patients with initial LRRC without distant metastases.ResultsAmong 2346 patients, 1210 underwent surgery, 392 received particle therapy, 256 conventional radiotherapy, 372 chemotherapy alone, and 116 palliative care. Median follow-up was 5.28 years. Three-year OS was 82.2% after surgery and 83.7% after particle therapy. In the subgroup analysis, 3- and 5-year OS was 88.1% and 72.5% after R0 resection, 73.0% and 48.2% after R1 resection, and 84.6% and 61.3% after particle therapy. R1 resection was associated with longer survival than conventional radiotherapy or chemotherapy alone, whereas survival after R2 resection was similar to these nonsurgical treatments.ConclusionThe findings reaffirm complete resection as the optimal treatment goal. Particle therapy was also associated with favorable survival in selected patients, supporting its potential role as a local treatment option when complete surgical resection is not feasible.
- 22 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessTreatment Patterns and Survival Outcomes in Locally Recurrent Rectal Cancer
BackgroundLocally recurrent rectal cancer (LRRC) remains a major clinical challenge, with limited high-quality evidence because of marked clinical and treatment heterogeneity. Although complete resection is the most important prognostic factor, the role of particle therapy remains unclear. This nationwide multicenter cohort study aimed to evaluate real-world treatment patterns and survival outcomes in patients with LRRC, with particular focus on particle therapy.MethodsPatients diagnosed with LRRC at 77 institutions in Japan between 2011 and 2023 were included and categorized according to treatment for the current recurrence: surgery, particle therapy, conventional radiotherapy with or without chemotherapy, chemotherapy alone, or palliative care. Overall survival (OS) was compared across treatment modalities, with a predefined subgroup analysis of patients with initial LRRC without distant metastases.ResultsAmong 2346 patients, 1210 underwent surgery, 392 received particle therapy, 256 conventional radiotherapy, 372 chemotherapy alone, and 116 palliative care. Median follow-up was 5.28 years. Three-year OS was 82.2% after surgery and 83.7% after particle therapy. In the subgroup analysis, 3- and 5-year OS was 88.1% and 72.5% after R0 resection, 73.0% and 48.2% after R1 resection, and 84.6% and 61.3% after particle therapy. R1 resection was associated with longer survival than conventional radiotherapy or chemotherapy alone, whereas survival after R2 resection was similar to these nonsurgical treatments.ConclusionThe findings reaffirm complete resection as the optimal treatment goal. Particle therapy was also associated with favorable survival in selected patients, supporting its potential role as a local treatment option when complete surgical resection is not feasible.
- Open Access
- 16 June 2026
- Published by Georg Thieme Verlag KG in Endoscopy
- Vol. 58 (S 04) , E771-E772
3-D pathological analysis of varicose microvascular vessels in sessile serrated lesions seen on endoscopy
We have clarified the vascular structure of colorectal lesions by contrasting three-dimensional (3D) reconstructed images of pathological slides using SYNAPSE VINCENT with in vivo endoscopic images.[1] We showed that the blood vessels visible in magnified blue light imaging (BLI) images are capillaries located within 80 μm of the surface.[2] Sessile serrated lesions (SSLs) exhibit varying macroscopic and vascular morphologies, necessitating differentiation from hyperplastic polyps due to their malignant potential.[3] Characteristic findings of SSLs include varicose microvascular vessels (VMVs), which are considered useful in differentiating them from hyperplastic polyps, but their histopathological basis are unclear.[4] This study aims to clarify the anatomy of VMVs by comparing 3D reconstructed images of a part of SSLs containing VMVs with BLI. Endoscopic and pathological vascular images of two SSL patients are shown. An outline of this process is presented in [Fig. 1]. The video shows the process of comparing 3D reconstruction of histopathology with actual BLI and analyzing the corresponding vascular depth from the mucosal surface ([Video 1]). First, a 3D vascular model of the entire SSL was constructed. The vascular structure was maintained in the SSL, from the deep mucosa, running up and down along the sides of the glands, and forming capillary plexuses at the surface. One-to-one correspondence between the BLI and the vascular network of the 3D model revealed the anatomical location of the VMVs. All three VMVs in the two cases were dilated capillaries within the mucosa and were located within approximately 80 μm from the surface. Video 1 The identification of varicose microvascular vessels in sessile serrated lesions by comparing endoscopic images and 3D reconstructed pathological images using the Synapse VINCENT software program (Fujifilm Co., Tokyo, Japan). It was confirmed that the capillary network and VMVs observed by magnified BLI is limited to within 80 μm from the surface. Comparing the endoscopic images of observed lesions with the pathological images of the corresponding areas improves the accuracy of endoscopic diagnosis. Endoscopy_UCTN_Code_TTT_1AQ_2AB Received: 03 April 2026 Accepted after revision: 29 May 2026 Article published online: 16 June 2026 © 2026. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. (https://creativecommons.org/licenses/by/4.0/). Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany
- 16 June 2026
- Published by Georg Thieme Verlag KG in Endoscopy
- Vol. 58 (S 04) , E771-E772
Open Access3-D pathological analysis of varicose microvascular vessels in sessile serrated lesions seen on endoscopy
We have clarified the vascular structure of colorectal lesions by contrasting three-dimensional (3D) reconstructed images of pathological slides using SYNAPSE VINCENT with in vivo endoscopic images.[1] We showed that the blood vessels visible in magnified blue light imaging (BLI) images are capillaries located within 80 μm of the surface.[2] Sessile serrated lesions (SSLs) exhibit varying macroscopic and vascular morphologies, necessitating differentiation from hyperplastic polyps due to their malignant potential.[3] Characteristic findings of SSLs include varicose microvascular vessels (VMVs), which are considered useful in differentiating them from hyperplastic polyps, but their histopathological basis are unclear.[4] This study aims to clarify the anatomy of VMVs by comparing 3D reconstructed images of a part of SSLs containing VMVs with BLI. Endoscopic and pathological vascular images of two SSL patients are shown. An outline of this process is presented in [Fig. 1]. The video shows the process of comparing 3D reconstruction of histopathology with actual BLI and analyzing the corresponding vascular depth from the mucosal surface ([Video 1]). First, a 3D vascular model of the entire SSL was constructed. The vascular structure was maintained in the SSL, from the deep mucosa, running up and down along the sides of the glands, and forming capillary plexuses at the surface. One-to-one correspondence between the BLI and the vascular network of the 3D model revealed the anatomical location of the VMVs. All three VMVs in the two cases were dilated capillaries within the mucosa and were located within approximately 80 μm from the surface. Video 1 The identification of varicose microvascular vessels in sessile serrated lesions by comparing endoscopic images and 3D reconstructed pathological images using the Synapse VINCENT software program (Fujifilm Co., Tokyo, Japan). It was confirmed that the capillary network and VMVs observed by magnified BLI is limited to within 80 μm from the surface. Comparing the endoscopic images of observed lesions with the pathological images of the corresponding areas improves the accuracy of endoscopic diagnosis. Endoscopy_UCTN_Code_TTT_1AQ_2AB Received: 03 April 2026 Accepted after revision: 29 May 2026 Article published online: 16 June 2026 © 2026. The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. (https://creativecommons.org/licenses/by/4.0/). Georg Thieme Verlag KG Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany
- Open Access
- 19 May 2026
- Published by Oxford University Press (OUP) in Inflammatory Bowel Diseases
- Vol. 32 (9) , 1750-1759
Clinicopathological features and cancer-specific survival in young patients with ulcerative colitis–associated colorectal neoplasia: a nationwide cohort study in Japan
Background: The incidence of ulcerative colitis–associated colorectal neoplasia (UCAN) is increasing, but clinicopathological features and outcomes in young patients remain unclear. We aimed to clarify the characteristics and survival of UCAN diagnosed before the age of 40 using years in a nationwide Japanese cohort study conducted by the Japanese Society for Cancer of the Colon and Rectum. Methods: We retrospectively analyzed 1212 patients with UCAN treated at 43 institutions in Japan during the period from 1983 to 2020. Patients were categorized as young (<40 years) or non-young (≥40 years). Clinicopathological features, cancer-specific survival (CSS), relapse-free survival (RFS), and survival after recurrence (SAR) were compared. Results: Young-onset UCAN accounted for 252 patients (20.8%) and showed a shorter interval from UC onset to carcinogenesis (14.0 vs 17.0 years; P < .001), more frequent pancolitis (81.3% vs 74.8%; P = .012), and a higher proportion of poorly differentiated, mucinous, or signet-ring cell carcinoma (20.2% vs 13.3%; P < .001). Five-year CSS was significantly lower in young patients (85.1% vs 89.8%; P = .039), whereas RFS was comparable (86.2% vs 86.3%; P = .750). Minimally invasive surgery (MIS) was independently associated with improved CSS (hazard ratio [HR], 0.593; 95% CI, 0.356-0.987; P = .045). The SAR remained poor in both age groups (23.6% vs 28.5%; P = .468). Conclusions: When diagnosed in patients aged <40 years, UCAN is characterized by earlier carcinogenesis, extensive disease, and aggressive histology, with worse CSS. The potential role of MIS in improving prognosis warrants further prospective investigation, while the outcome of SAR remains poor irrespective of patient age.
- 19 May 2026
- Published by Oxford University Press (OUP) in Inflammatory Bowel Diseases
- Vol. 32 (9) , 1750-1759
Open AccessClinicopathological features and cancer-specific survival in young patients with ulcerative colitis–associated colorectal neoplasia: a nationwide cohort study in Japan
Background: The incidence of ulcerative colitis–associated colorectal neoplasia (UCAN) is increasing, but clinicopathological features and outcomes in young patients remain unclear. We aimed to clarify the characteristics and survival of UCAN diagnosed before the age of 40 using years in a nationwide Japanese cohort study conducted by the Japanese Society for Cancer of the Colon and Rectum. Methods: We retrospectively analyzed 1212 patients with UCAN treated at 43 institutions in Japan during the period from 1983 to 2020. Patients were categorized as young (<40 years) or non-young (≥40 years). Clinicopathological features, cancer-specific survival (CSS), relapse-free survival (RFS), and survival after recurrence (SAR) were compared. Results: Young-onset UCAN accounted for 252 patients (20.8%) and showed a shorter interval from UC onset to carcinogenesis (14.0 vs 17.0 years; P < .001), more frequent pancolitis (81.3% vs 74.8%; P = .012), and a higher proportion of poorly differentiated, mucinous, or signet-ring cell carcinoma (20.2% vs 13.3%; P < .001). Five-year CSS was significantly lower in young patients (85.1% vs 89.8%; P = .039), whereas RFS was comparable (86.2% vs 86.3%; P = .750). Minimally invasive surgery (MIS) was independently associated with improved CSS (hazard ratio [HR], 0.593; 95% CI, 0.356-0.987; P = .045). The SAR remained poor in both age groups (23.6% vs 28.5%; P = .468). Conclusions: When diagnosed in patients aged <40 years, UCAN is characterized by earlier carcinogenesis, extensive disease, and aggressive histology, with worse CSS. The potential role of MIS in improving prognosis warrants further prospective investigation, while the outcome of SAR remains poor irrespective of patient age.
- Open Access
- 8 April 2026
- Published by MDPI AG in Bioengineering
- Vol. 13 (4) , 434
Prognostic Power of Ensemble Learning in Colorectal Cancer with Peritoneal Metastasis: A Multi-Institutional Analysis
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.
- 8 April 2026
- Published by MDPI AG in Bioengineering
- Vol. 13 (4) , 434
Open AccessPrognostic Power of Ensemble Learning in Colorectal Cancer with Peritoneal Metastasis: A Multi-Institutional Analysis
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.
- Open Access
- 21 January 2026
- Published by MDPI AG in Preprints.org
Machine Learning-Based Survival Time Prediction in Colorectal Cancer with Peritoneal Metastasis: A Multi-Institutional Registry-Based Study
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.
- 21 January 2026
- Published by MDPI AG in Preprints.org
Open AccessMachine Learning-Based Survival Time Prediction in Colorectal Cancer with Peritoneal Metastasis: A Multi-Institutional Registry-Based Study
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.
- Open Access
- 3 December 2025
- Published by Oxford University Press (OUP) in Journal of Crohn's and Colitis
- Vol. 20 (2)
Background mucosal inflammation affects colorectal cancer prognosis in ulcerative colitis: a nationwide, multicenter study
Background and aims: Chronic background mucosal inflammation contributes to colorectal cancer (CRC) development in ulcerative colitis (UC), but its prognostic impact is unclear. We evaluated whether background mucosal inflammation documented at cancer diagnosis is associated with oncologic outcomes. Methods: This retrospective study analyzed 1189 UC patients diagnosed with CRC using a nationwide, multicenter database in Japan. Patients were classified as CRC within the UC-involved area (within-area) or outside the UC-involved area (outside-area), based on tumor location relative to the UC disease extent documented endoscopically at cancer diagnosis. The primary end point was 5-year recurrence-free survival (RFS), and the secondary end point was 5-year cancer-specific survival (CSS). In within-area cases, inflammation severity was assessed using the Mayo Endoscopic Score (MES), stratified as Inactive, Mild–Moderate, and Severe. Results: Of the 723 eligible patients, 683 had within-area and 40 outside-area CRC. Five-year RFS was significantly lower in within-area than outside-area CRCs (75.1% vs 87.6%, P = 0.022). Multivariable Cox regression analysis of RFS revealed this classification as an independent prognostic factor (hazard ratio = 2.99, 95% confidence interval: 1.09-8.18, P = 0.030). A significant difference was also observed in 5-year CSS (P = 0.038). Among within-area cases, higher MES was associated with stepwise declines in RFS (P = 0.150), and a similar, statistically significant gradient in CSS (P = 0.048). Conclusions: Background mucosal inflammation at cancer diagnosis is associated with significantly worse prognosis of CRC in UC patients. Systematic endoscopic assessment at cancer diagnosis may aid prognostic stratification and inform management.
- 3 December 2025
- Published by Oxford University Press (OUP) in Journal of Crohn's and Colitis
- Vol. 20 (2)
Open AccessBackground mucosal inflammation affects colorectal cancer prognosis in ulcerative colitis: a nationwide, multicenter study
Background and aims: Chronic background mucosal inflammation contributes to colorectal cancer (CRC) development in ulcerative colitis (UC), but its prognostic impact is unclear. We evaluated whether background mucosal inflammation documented at cancer diagnosis is associated with oncologic outcomes. Methods: This retrospective study analyzed 1189 UC patients diagnosed with CRC using a nationwide, multicenter database in Japan. Patients were classified as CRC within the UC-involved area (within-area) or outside the UC-involved area (outside-area), based on tumor location relative to the UC disease extent documented endoscopically at cancer diagnosis. The primary end point was 5-year recurrence-free survival (RFS), and the secondary end point was 5-year cancer-specific survival (CSS). In within-area cases, inflammation severity was assessed using the Mayo Endoscopic Score (MES), stratified as Inactive, Mild–Moderate, and Severe. Results: Of the 723 eligible patients, 683 had within-area and 40 outside-area CRC. Five-year RFS was significantly lower in within-area than outside-area CRCs (75.1% vs 87.6%, P = 0.022). Multivariable Cox regression analysis of RFS revealed this classification as an independent prognostic factor (hazard ratio = 2.99, 95% confidence interval: 1.09-8.18, P = 0.030). A significant difference was also observed in 5-year CSS (P = 0.038). Among within-area cases, higher MES was associated with stepwise declines in RFS (P = 0.150), and a similar, statistically significant gradient in CSS (P = 0.048). Conclusions: Background mucosal inflammation at cancer diagnosis is associated with significantly worse prognosis of CRC in UC patients. Systematic endoscopic assessment at cancer diagnosis may aid prognostic stratification and inform management.
- Open Access
- 24 January 2025
- Published by Wiley in Annals of Gastroenterological Surgery
- Vol. 9 (4) , 750-760
Clinical significance of preoperative Glasgow prognostic score in patients with colorectal cancer and synchronous peritoneal metastases
Background: Chemotherapy is the typical choice for treating colorectal cancer with synchronous peritoneal metastases. Nonetheless, surgical resection may be chosen if the metastases are resectable. Unfortunately, there is no reliable preoperative or intraoperative prognostic indicator. This study aimed to determine the prognostic significance of the preoperative Glasgow prognostic score (GPS) in colorectal cancer patients with synchronous peritoneal metastases.Methods: We conducted a prospective study on 143 patients with colorectal cancer and concurrent peritoneal metastases. Our analysis included prognostic factors, such as the GPS, using data from the institutional observational study by the Japanese Society for Cancer of the Colon and Rectum.Results: The 3‐year survival rates for the GPS0 or 1 and GPS2 groups were 32.7% and 14.3%, respectively, with a significantly worse prognosis in the GPS2 group (p = 0.003). Multivariate analysis identified GPS2 (p = 0.006) and the peritoneal cancer index (PCI) (p = 0.029) or the Japanese surgical peritoneal metastasis grade (p = 0.009) as independent poor prognostic factors. Additionally, the GPS0 or 1 group with total resection of peritoneal metastases had a significantly better prognosis than the non‐resection group (p < 0.001); however, there was no difference between the GPS2 group with total peritoneal resection and the non‐resection group (p = 0.713).Conclusions: Preoperative GPS2 is an independent poor prognostic factor in patients with colorectal cancer and synchronous peritoneal metastases, and surgical resection does not improve prognosis in patients with GPS2. Preoperative GPSs may be used as indicators for surgical resection of synchronous peritoneal metastases.
- 24 January 2025
- Published by Wiley in Annals of Gastroenterological Surgery
- Vol. 9 (4) , 750-760
Open AccessClinical significance of preoperative Glasgow prognostic score in patients with colorectal cancer and synchronous peritoneal metastases
Background: Chemotherapy is the typical choice for treating colorectal cancer with synchronous peritoneal metastases. Nonetheless, surgical resection may be chosen if the metastases are resectable. Unfortunately, there is no reliable preoperative or intraoperative prognostic indicator. This study aimed to determine the prognostic significance of the preoperative Glasgow prognostic score (GPS) in colorectal cancer patients with synchronous peritoneal metastases.Methods: We conducted a prospective study on 143 patients with colorectal cancer and concurrent peritoneal metastases. Our analysis included prognostic factors, such as the GPS, using data from the institutional observational study by the Japanese Society for Cancer of the Colon and Rectum.Results: The 3‐year survival rates for the GPS0 or 1 and GPS2 groups were 32.7% and 14.3%, respectively, with a significantly worse prognosis in the GPS2 group (p = 0.003). Multivariate analysis identified GPS2 (p = 0.006) and the peritoneal cancer index (PCI) (p = 0.029) or the Japanese surgical peritoneal metastasis grade (p = 0.009) as independent poor prognostic factors. Additionally, the GPS0 or 1 group with total resection of peritoneal metastases had a significantly better prognosis than the non‐resection group (p < 0.001); however, there was no difference between the GPS2 group with total peritoneal resection and the non‐resection group (p = 0.713).Conclusions: Preoperative GPS2 is an independent poor prognostic factor in patients with colorectal cancer and synchronous peritoneal metastases, and surgical resection does not improve prognosis in patients with GPS2. Preoperative GPSs may be used as indicators for surgical resection of synchronous peritoneal metastases.
- Open Access
- 1 January 2025
- Published by S. Karger AG in Inflammatory Intestinal Diseases
- Vol. 10 (1) , 151-154
A Case of Postoperative Pouch-Vaginal Fistula of Ulcerative Colitis Successfully Treated with Estriol Vaginal Tablet
Introduction: Ileal pouch-vaginal fistula (PVF) is a severe complication that can occur following surgery for ulcerative colitis (UC). Most cases of PVF are managed surgically, and reports on successful closure through conservative treatment alone are limited. We report the first documented case of PVF closure without stoma formation, successfully treated with antibiotics and estriol vaginal tablets. Case Presentation: A 65-year-old woman was diagnosed in her 50s with total colitis-type UC. She developed steroid-dependent, refractory disease, prompting the indication of infliximab therapy. However, infliximab failed to maintain remission, necessitating restorative proctocolectomy with ileal pouch-anal anastomosis followed by loop ileostomy. The postoperative course was uneventful, and the ileostomy was closed 6 months later. Two years after surgery, she developed fever, diarrhea, and vaginal discharge containing fecal fluid. The endoscopic evaluation identified a PVF secondary to pouchitis. She was admitted to the hospital, placed on fasting, and treated with antibiotics and estriol vaginal tablets. Endoscopy 18 days after initiating of estriol therapy revealed vaginal wall thickening and PVF closure, and she was subsequently discharged. Estriol vaginal tablet administration continued for 3 months, and no recurrence has been observed 9 years following surgery. Conclusion: Estriol vaginal tablets may be serve as an effective conservative treatment option for PVF following surgery for UC.
- 1 January 2025
- Published by S. Karger AG in Inflammatory Intestinal Diseases
- Vol. 10 (1) , 151-154
Open AccessA Case of Postoperative Pouch-Vaginal Fistula of Ulcerative Colitis Successfully Treated with Estriol Vaginal Tablet
Introduction: Ileal pouch-vaginal fistula (PVF) is a severe complication that can occur following surgery for ulcerative colitis (UC). Most cases of PVF are managed surgically, and reports on successful closure through conservative treatment alone are limited. We report the first documented case of PVF closure without stoma formation, successfully treated with antibiotics and estriol vaginal tablets. Case Presentation: A 65-year-old woman was diagnosed in her 50s with total colitis-type UC. She developed steroid-dependent, refractory disease, prompting the indication of infliximab therapy. However, infliximab failed to maintain remission, necessitating restorative proctocolectomy with ileal pouch-anal anastomosis followed by loop ileostomy. The postoperative course was uneventful, and the ileostomy was closed 6 months later. Two years after surgery, she developed fever, diarrhea, and vaginal discharge containing fecal fluid. The endoscopic evaluation identified a PVF secondary to pouchitis. She was admitted to the hospital, placed on fasting, and treated with antibiotics and estriol vaginal tablets. Endoscopy 18 days after initiating of estriol therapy revealed vaginal wall thickening and PVF closure, and she was subsequently discharged. Estriol vaginal tablet administration continued for 3 months, and no recurrence has been observed 9 years following surgery. Conclusion: Estriol vaginal tablets may be serve as an effective conservative treatment option for PVF following surgery for UC.