Development and validation of lymph node ratio-based nomograms for primary duodenal adenocarcinoma after surgery

Abstract

Background: The prediction models for primary duodenal adenocarcinoma (PDA) are deficient. This study aimed to determine the predictive value of the lymph node ratio (LNR) in PDA patients and to establish and validate nomograms for predicting overall survival (OS) and cancer-specific survival (CSS) for PDAs after surgical resection.

Methods: We extracted the demographics and clinicopathological information of PDA patients between 2004 and 2018 from the Surveillance, Epidemiology and End Results database. After screening cases, we randomly divided the enrolled patients into training and validation groups. X-tile software was used to obtain the best cut-off value for the LNR. Univariate and multivariate Cox analyses were used in the training group to screen out significant variables to develop nomograms. The predictive accuracy of the nomograms was evaluated by the concordance index (C-index), calibration curves, area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA). Finally, four risk groups were created based on quartiles of the model scores.

Results: A total of 978 patients were included in this study. The best cut-off value for the LNR was 0.47. LNR was a negative predictive factor for both OS and CSS. Age, sex, grade, chemotherapy and LNR were used to construct the OS nomogram, while age, grade, chemotherapy, the number of lymph nodes removed and LNR were incorporated into the CSS nomogram. The C-index, calibration curves and AUC of the training and validation sets revealed their good predictability. DCA showed that the predictive value of the nomograms was superior to that of the American Joint Committee on Cancer (AJCC) TNM staging system (8th edition). In addition, risk stratification demonstrated that patients with higher risk correlated with poor survival.

Conclusions: The LNR was an adverse prognostic determinant for PDAs. The nomograms provided an accurate and applicable tool to evaluate the prognosis of PDA patients after surgery.

Keywords: cancer-specific survival; lymph node ratio; nomogram; overall survival; primary duodenal adenocarcinoma.

Figure 1

The flow chart of included patients.

Figure 2

Kaplan–Meier curves of OS (A) and CSS (B) for all included patients stratified by LNR. OS, overall survival; CSS, cancer-specific survival; LNR, lymph node ratio.

Figure 3

(A) Nomogram predicting the 1-, 3- and 5-year OS rates for PDA patients after surgery. (B) Nomogram predicting the 1-, 3- and 5-year CSS rates for PDA patients after surgery. OS, overall survival; PDA, primary duodenal adenocarcinoma; CSS, cancer-specific survival.

Figure 4

Calibration curves for OS prediction in the training cohort (A) and the validation cohort (E). Decision curve analysis of the AJCC 8th edition staging system, nomogram and LNR for the 1- (B), 3- (C) and 5-year (D) OS rates of PDA patients from the training cohort. Decision curve analysis of the AJCC 8th edition staging system, nomogram and LNR for the 1- (F), 3- (G) and 5-year (H) OS rates of PDA patients from the validation cohort. OS, overall survival; LNR, lymph node ratio; PDA, primary duodenal adenocarcinoma. For calibration curves, red, blue and green lines represent 1, 3, and 5 years, respectively.

Figure 5

Calibration curves for CSS prediction in the training cohort (A) and the validation cohort (E). Decision curve analysis of the AJCC 8th edition staging system, nomogram and LNR for the 1- (B), 3- (C) and 5-year (D) CSS rates of PDA patients from the training cohort. Decision curve analysis of the AJCC 8th edition staging system, nomogram and LNR for the 1- (F), 3- (G) and 5-year (H) CSS rates of PDA patients from the validation cohort. CSS, cancer-specific survival; LNR, lymph node ratio; PDA, primary duodenal adenocarcinoma. For calibration curves, red, blue and green lines represent 1, 3, and 5 years, respectively.

Figure 6

Kaplan–Meier curves of OS for the training group (A) and validation group (B) stratified by different risk scores. Kaplan–Meier curves of CSS for the training group (C) and validation group (D) stratified by different risk scores. OS, overall survival; CSS, cancer-specific survival.