Retrospective Study
Copyright ©The Author(s) 2021.
World J Gastrointest Oncol. Sep 15, 2021; 13(9): 1184-1195
Published online Sep 15, 2021. doi: 10.4251/wjgo.v13.i9.1184
Table 3 Comparison of predictive performances between different pathological classifications and prognostic models
Pathological classification/prognostic modelAUC (95%CI)
AIC
3-yr overall survival
5-yr overall survival
Differentiation grade0.626 (0.608–0.644)0.621 (0.602–0.639)25971
Lauren classification0.666 (0.647–0.683)0.681 (0.663–0.699)25923
Modified Lauren classification0.679 (0.661–0.696)0.702 (0.685–0.719)25877
DeLong tests for AUCs
Differentiation grade vs LaurenP < 0.001P < 0.001–
Lauren vs modified LaurenP < 0.001P < 0.001–
Modified Lauren vs differentiation gradeP < 0.001P < 0.001–
Novel prognostic model 0.803 (0.786–0.819)0.804 (0.787–0.820)20010
Age, tumor size, retrieved lymph nodes, pT stage, pN stage, modified Lauren classification
Control model0.776 (0.759–0.793)0.776 (0.759–0.793)20144
AJCC 8th pTNM stage (pT stage, pN stage)
AIC: Akaike's Information Criterion; AJCC: American Joint Committee on Cancer; AUC: Area under curve; pN stage: Pathological N stage; pT stage: Pathological T stage. A higher area under the curve indicated better model discrimination and a lower Akaike's Information Criterion indicates superior model-fitting; Differentiation grade, well vs moderate vs poor vs undifferentiation; Lauren classification, intestinal type vs diffuse type vs mixed type; Modified Lauren classification, distal non-diffuse vs proximal non-diffuse vs diffuse type.