Abstract | BACKGROUND: METHODS: By the highest Youden index (sensitivity + 1-specificity), the optimal cut-off values of testing markers for postoperative major complications and lymph node metastasis were identified. Multivariate regression analysis was used to reveal independent predictors for primary lymph node metastasis, postoperative major complications and progression-free survival (PFS). Nomograms based on independent predictors were constructed, and the discrimination and calibration were evaluated. RESULTS: A nomogram predicting primary lymph node metastasis was based on four risky independent predictors: American Society of Anesthesiologists (ASA) score 3-4, preoperative albumin (ALB) <41.15 g/L, poor differentiation and multiple liver metastases. The performance of the model was acceptable in predicting lymph node metastasis, with an area under the receiver operating characteristic curve (AUROC) of 0.655 (95% CI: 0.591-0.739). Calibration curves and the Hosmer-Lemeshow test revealed desirable model calibration (chi-square: 13.26, P=0.815). In the multivariate analysis, preoperative lactate dehydrogenase (LDH) ≥202.5 U/L [odds ratio (OR) =2.084, 95% confidence interval (CI): 1.039-4.181, P=0.039] and operation time ≥350.5 min (OR =2.848, 95% CI: 1.418-5.723, P=0.003) were independently associated with the presence of postoperative major complications. A nomogram predicting PFS was constructed based on poor differentiation, positive lymph node metastasis, bilobar liver distribution and R0 resection with good discrimination (C-index: 0.656±0.021) and calibration. CONCLUSIONS: This study established predictive nomograms specifically incorporating preoperative ALB and LDH levels for the prediction of primary lymph node metastasis and prognosis in synchronous CRLM patients with simultaneous resection, which have favourable discrimination and calibration to make individualized predictions.
|
Authors | Jinjie Yao, Qichen Chen, Yiqiao Deng, Jianjun Zhao, Xinyu Bi, Zhiyu Li, Zhen Huang, Yefan Zhang, Jianguo Zhou, Hong Zhao, Jianqiang Cai |
Journal | Annals of palliative medicine
(Ann Palliat Med)
Vol. 10
Issue 4
Pg. 4220-4231
(Apr 2021)
ISSN: 2224-5839 [Electronic] China |
PMID | 33832313
(Publication Type: Journal Article)
|
Topics |
- Colorectal Neoplasms
- Humans
- Liver Neoplasms
- Lymphatic Metastasis
- Nomograms
- Prognosis
- Retrospective Studies
|