A nomogram for cancer-specific survival of lung adenocarcinoma patients: A SEER based analysis

Hong Guo; Guole Nie; Xin Zhao; Jialu Liu; Kaihua Yu; Yulan Li

doi:10.1016/j.sopen.2024.10.003

A nomogram for cancer-specific survival of lung adenocarcinoma patients: A SEER based analysis

Surg Open Sci. 2024 Oct 18:22:13-23. doi: 10.1016/j.sopen.2024.10.003. eCollection 2024 Dec.

Authors

Hong Guo^{1

2}, Guole Nie³, Xin Zhao², Jialu Liu¹, Kaihua Yu¹, Yulan Li^{1

4}

Affiliations

¹ First Clinical Medical College, Lanzhou University, Lanzhou 730000, China.
² Department of Anesthesiology, Inner Mongolia Hospital of Peking University Cancer Hospital, The Affiliated People's Hospital of Inner Mongolia Medical University, Hohhot 10020, China.
³ Department of Colorectal Hernia Surgery, Binzhou Medical University Hospital, Binzhou, 256600, China.
⁴ Department of Anesthesiology, First Hospital of Lanzhou University, Lanzhou University, Lanzhou 730000, China.

Abstract

Background: Non-small cell lung cancer (NSCLC) accounts for 85 % of lung cancer cases. Among NSCLC subtypes, lung adenocarcinoma (LUAD) stands as the most prevalent. Regrettably, LUAD continues to exhibit a notably unfavorable overall prognosis. This study's primary aim was to develop and validate prognostic tools capable of predicting the likelihood of cancer-specific survival (CSS) in patients with LUAD.

Methods: We retrospectively collected 21,099 patients diagnosed with LUAD between 2010 and 2015, and 8290 patients diagnosed between 2004 and 2009 from SEER database. The cohort of 21,099 patients served as the prognostic group for the exploration of LUAD-related prognostic risk factors. The cohort of 8290 patients was designated for external validation. We created a training set and an internal validation set in the prognostic group for the development and internal validation of CSS nomograms. CSS predictors were identified through the least absolute shrinkage and selection operator (Lasso) regression analysis. Prognostic model was constructed via Cox hazard regression analysis, presented in the form of both static and dynamic network-based nomograms.

Results: Several independent prognostic factors were incorporated into the construction of nomogram. The nomogram accurately predicted CSS at 1, 3, and 5 years, with respective AUC values of 0.769, 0.761, and 0.748 for the training group, and 0.741, 0.752, and 0.740 for the testing group. The study demonstrated a strong agreement between anticipated and actual CSS values, supported by decision curve analysis (DCA) and time-dependent calibrated curves. High-risk patients based on the nomogram exhibiting significantly lower survival rates compared to their low-risk counterparts according to Kaplan-Meier (K-M) curves. The nomogram demonstrates excellent predictive power in the external validation cohort.

Conclusions: A dependable and user-friendly nomogram has been developed, available in both static and online dynamic calculator formats, to facilitate healthcare professionals in accurately estimating the likelihood of CSS for patients diagnosed LUAD.

Keywords: American Joint Committee on Cancer Staging (AJCC); Cancer-specific survival (CSS); Lung adenocarcinoma (LUAD); Nomogram; SEER.