Aim of the study: Nomograms have been proposed to assess prognosis following curative surgery for gastric cancer. The objective of the current study was to evaluate the performance of the Gastric Cancer Collaborative Group nomograms developed in 2014 by Kim et al., using a cohort of patients from a 10-year single institution experience in gastric cancer management. Patients and methods: We retrospectively reviewed patients who underwent curative-intent surgery for histologically confirmed gastric cancer at First Surgical Clinic of Padua University Hospital (Italy) from January 2010 to May 2020. Univariable and multivariable Cox proportional hazard models were employed to assess the effect of the variables of interest on mortality and recurrence. Multivariable analysis was performed by considering the variables included in the Gastric Cancer Collaborative Group nomograms in order to validate them. The performance of the nomograms was evaluated using Harrell's C-index and calibration plots. Results: Overall, 168 patients were included, with a median follow-up of 20.1 months. On multivariable analysis, tumor location, lymph node ratio, and pathological T stage were associated with recurrence; age, tumor location, lymph node ratio, and pT stage were associated with OS (overall survival). The nomograms had good discriminatory capability to classify both OS (C-index: 0.75) and DFS (disease-free survival) (C-index 0.72). The corrected C-Index for DFS based on the AJCC staging system revealed better prediction (C-Index 0.75), while the corrected C-Index for OS had worse discrimination ability compared with the current nomogram (C-Index 0.72). Conclusions: The Gastric Cancer Collaborative Group nomograms demonstrated good performances in terms of prediction of both OS and DFS on external validation. The two nomograms are easy to apply, and variables included are widely available to most facilities.
Validation of a Nomogram to Predict Long Term Outcomes After Curative Surgery for Gastric Cancer in an Italian Cohort of Patients
Capelli, G;Tonello, A S;Chiminazzo, V;Lorenzoni, G;Bao, Q R;Gregori, D;Pucciarelli, S;Spolverato, G
2021
Abstract
Aim of the study: Nomograms have been proposed to assess prognosis following curative surgery for gastric cancer. The objective of the current study was to evaluate the performance of the Gastric Cancer Collaborative Group nomograms developed in 2014 by Kim et al., using a cohort of patients from a 10-year single institution experience in gastric cancer management. Patients and methods: We retrospectively reviewed patients who underwent curative-intent surgery for histologically confirmed gastric cancer at First Surgical Clinic of Padua University Hospital (Italy) from January 2010 to May 2020. Univariable and multivariable Cox proportional hazard models were employed to assess the effect of the variables of interest on mortality and recurrence. Multivariable analysis was performed by considering the variables included in the Gastric Cancer Collaborative Group nomograms in order to validate them. The performance of the nomograms was evaluated using Harrell's C-index and calibration plots. Results: Overall, 168 patients were included, with a median follow-up of 20.1 months. On multivariable analysis, tumor location, lymph node ratio, and pathological T stage were associated with recurrence; age, tumor location, lymph node ratio, and pT stage were associated with OS (overall survival). The nomograms had good discriminatory capability to classify both OS (C-index: 0.75) and DFS (disease-free survival) (C-index 0.72). The corrected C-Index for DFS based on the AJCC staging system revealed better prediction (C-Index 0.75), while the corrected C-Index for OS had worse discrimination ability compared with the current nomogram (C-Index 0.72). Conclusions: The Gastric Cancer Collaborative Group nomograms demonstrated good performances in terms of prediction of both OS and DFS on external validation. The two nomograms are easy to apply, and variables included are widely available to most facilities.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
