Copyright
©The Author(s) 2025.
World J Gastrointest Surg. Sep 27, 2025; 17(9): 107326
Published online Sep 27, 2025. doi: 10.4240/wjgs.v17.i9.107326
Published online Sep 27, 2025. doi: 10.4240/wjgs.v17.i9.107326
Table 1 Univariate analysis of risk factors for delayed postoperative bleeding in patients with early colorectal cancer and precancerous lesions, n (%)
Factors | n | DPOB group (n = 30) | Non-DPOB group (n = 272) | χ2 | P value |
Gender | 0.822 | 0.365 | |||
Male | 178 | 20 (66.67) | 158 (58.09) | ||
Female | 124 | 10 (33.33) | 114 (41.91) | ||
Age (years) | 1.276 | 0.259 | |||
< 60 | 122 | 15 (50.00) | 107 (39.34) | ||
≥ 60 | 180 | 15 (50.00) | 165 (60.66) | ||
Comorbidities | |||||
Hypertension | 30 | 4 (13.33) | 26 (9.56) | 0.430 | 0.512 |
Diabetes | 25 | 2 (6.67) | 23 (8.46) | 0.114 | 0.736 |
COPD | 12 | 2 (6.67) | 10 (3.68) | 0.633 | 0.426 |
Smoking | 73 | 5 (16.67) | 68 (25.00) | 1.024 | 0.312 |
Alcohol consumption | 68 | 9 (30.00) | 59 (21.69) | 1.069 | 0.301 |
Number of lesions | 7.294 | 0.007 | |||
Single | 215 | 15 (50.00) | 200 (73.53) | ||
Multiple | 87 | 15 (50.00) | 72 (26.47) | ||
Lesion size (cm) | 5.787 | 0.016 | |||
< 3 | 173 | 11 (36.67) | 162 (59.56) | ||
≥ 3 | 129 | 19 (63.33) | 110 (40.44) | ||
Lesion location | 7.706 | 0.006 | |||
Colon | 163 | 9 (30.00) | 154 (56.62) | ||
Rectum | 139 | 21 (70.00) | 118 (43.38) | ||
Degree of submucosal fibrosis | 5.434 | 0.020 | |||
None or mild | 282 | 25 (83.33) | 257 (94.49) | ||
Severe | 20 | 5 (16.67) | 15 (5.51) | ||
Intraoperative bleeding | 107 | 16 (53.33) | 91 (33.46) | 4.667 | 0.031 |
Table 2 Variable assignment
Factors | Variables | Assignments |
DPOB | Y | No: 0, yes: 1 |
Number of lesions | X1 | Single: 0, multiple: 1 |
Lesion size (cm) | X2 | < 3: 0, ≥ 3: 1 |
Lesion location | X3 | Colon: 0, rectum: 1 |
Degree of submucosal fibrosis | X4 | None or mild: 0, severe: 1 |
Intraoperative bleeding | X5 | No: 0, yes: 1 |
Table 3 Multivariate analysis of risk factors for delayed postoperative bleeding in patients with early colorectal cancer and precancerous lesions
Factors | β | SE | Wald | P value | Exp (β) | 95%CI |
Number of lesions | 0.904 | 0.414 | 4.756 | 0.029 | 2.469 | 1.096-5.563 |
Lesion size (cm) | 0.856 | 0.417 | 4.210 | 0.040 | 2.353 | 1.039-5.328 |
Lesion location | 1.145 | 0.437 | 6.861 | 0.009 | 3.144 | 1.334-7.407 |
Degree of submucosal fibrosis | 1.490 | 0.628 | 5.639 | 0.018 | 4.439 | 1.297-15.188 |
Intraoperative bleeding | 0.651 | 0.417 | 2.438 | 0.118 | 1.917 | 0.847-4.338 |
Table 4 Receiver operating characteristic analysis data for the predictive performance of the risk model for delayed postoperative bleeding
Factor | AUC | SE | 95%CI | Cutoff | Specificity (%) | Sensitivity (%) |
Modeling group | 0.831 | 0.058 | 0.717-0.944 | 0.26 | 96.13 | 60.00 |
Validation group | 0.821 | 0.080 | 0.664-0.979 | 0.13 | 75.82 | 90.00 |
Table 5 10-fold cross-validation data
Sequence number | AUC of the validation set |
1 | 0.877 |
2 | 0.786 |
3 | 0.835 |
4 | 0.733 |
5 | 0.722 |
6 | 0.774 |
7 | 0.748 |
8 | 0.810 |
9 | 0.916 |
10 | 0.746 |
Mean | 0.795 (95%CI: 0.702-0.887) |
- Citation: Qian J, Tao YL, Zheng SS. Risk modeling of delayed postoperative bleeding after endoscopic submucosal dissection for early colorectal cancer and precancerous lesions. World J Gastrointest Surg 2025; 17(9): 107326
- URL: https://www.wjgnet.com/1948-9366/full/v17/i9/107326.htm
- DOI: https://dx.doi.org/10.4240/wjgs.v17.i9.107326