Development and validation of a prognostic nomogram for early stage non-small cell lung cancer: a study based on the SEER database and a Chinese cohort

Survival analysis and clinicopathological features of patients with stage IA lung adenocarcinoma

Background

With the development of medical technology and change of life habits, early-stage lung adenocarcinoma (LUAD) has become more common. This study aimed to systematically analyzed clinicopathological factors associated to the overall survival (OS) of patients with Stage IA LUAD.

Methods

A total of 5942 Stage IA LUAD patients were obtained from the Surveillance, Epidemiology, and End Results (SEER) database. Kaplan-Meier methods and log-rank tests were used to compare the differences in OS. A nomogram constructed based on the Cox regression was evaluated by Concordance index (C index), calibration curve, decision curve analysis (DCA) and area under curve (AUC). And 136 patients were recruited from Shandong Province Hospital for external validation.

Results

Cox analysis regression indicated that 12 factors, such as Diagnosis to Treatment Interval (DTI) and Income Level, were independent prognostic factors and were included to establish the nomogram. The C-index of our novel model was 0.702, 0.724 and 0.872 in the training, internal and external validation cohorts, respectively. The 3-year and 5-year survival AUCs and calibration curves showed excellent agreement in each cohort. Some new factors in the SEER database, including DTI and Income Level, were firstly confirmed as independent prognostic factors of Stage IA LUAD patients. The distribution of these factors in the T1a, T1b, and T1c subgroups differed and had different effects on survival.

Conclusion

We summarized 12 factors that affect prognosis and constructed a nomogram to predict OS of Stage IA LUAD patients who underwent operation. For the first time, new SEER database parameters, including DTI and Income Level, were proved to be survival-related.

Prognostic factors and survival prediction of resected non-small cell lung cancer with ipsilateral pulmonary metastases: a study based on the Surveillance, Epidemiology, and End Results (SEER) database

BACKGROUND

Prognostic factors and survival outcomes of non-small cell lung cancer (NSCLC) with Ipsilateral pulmonary metastasis (IPM) are not well-defined. Thus, this study intended to identify the prognostic factors for these patients and construct a predictive nomogram model.

METHODS

One thousand, seven hundred thirty-two patients with IPM identified between 2000 to 2019 were from the Surveillance, Epidemiology, and End Results (SEER) database. Independent prognostic factors were identified using multivariate Cox regression analyses. Nomograms were constructed to predict the overall survival (OS), C-index, the area under the curve (AUC), and the calibration curve to determine the predictive accuracy and discrimination; the decision curve analysis was used to confirm the clinical utility.

RESULTS

Patients were randomly divided into training (n = 1213) and validation (n = 519) cohorts. In the training cohort, the multivariable analysis demonstrated that age, sex, primary tumor size, N status, number of regional lymph nodes removed, tumor grade, and chemotherapy were independent prognostic factors for IPM. We constructed a 1-year, 3-year, and 5-year OS prediction nomogram model using independent prognostic factors. The C-index of this model for OS prediction was 0.714 (95% confidence interval [CI], 0.692 to 0.773) in the training cohort and 0.695 (95% CI, 0.660 to 0.730) in the validation cohort. Based on the AUC of the receiver operating characteristic analysis, calibration plots, and decision curve analysis, we concluded that the prognosis model of IPM exhibited excellent performance. Patients with total nomogram points greater than 96 were considered high-risk.

CONCLUSION

We constructed and internally validated a nomogram to predict 1-year, 3-year, and 5-year OS for NSCLC patients with IPM according to independent prognostic factors. This nomogram demonstrated good calibration, discrimination, clinical utility, and practical decision-making effects for the prognosis of NSCLC patients with IPM.

Two machine learning-based nomogram to predict risk and prognostic factors for liver metastasis from pancreatic neuroendocrine tumors: a multicenter study

BACKGROUND

Pancreatic neuroendocrine tumors (PNETs) are one of the most common endocrine tumors, and liver metastasis (LMs) are the most common location of metastasis from PNETS; However, there is no valid nomogram to predict the diagnosis and prognosis of liver metastasis (LMs) from PNETs. Therefore, we aimed to develop a valid predictive model to aid physicians in making better clinical decisions.

METHODS

We screened patients in the Surveillance, Epidemiology, and End Results (SEER) database from 2010-2016. Feature selection was performed by machine learning algorithms and then models were constructed. Two nomograms were constructed based on the feature selection algorithm to predict the prognosis and risk of LMs from PNETs. We then used the area under the curve (AUC), receiver operating characteristic (ROC) curve, calibration plot and consistency index (C-index) to evaluate the discrimination and accuracy of the nomograms. Kaplan-Meier (K-M) survival curves and decision curve analysis (DCA) were also used further to validate the clinical efficacy of the nomograms. In the external validation set, the same validation is performed.

RESULTS

Of the 1998 patients screened from the SEER database with a pathological diagnosis of PNET, 343 (17.2%) had LMs at the time of diagnosis. The independent risk factors for the occurrence of LMs in PNET patients included histological grade, N stage, surgery, chemotherapy, tumor size and bone metastasis. According to Cox regression analysis, we found that histological subtype, histological grade, surgery, age, and brain metastasis were independent prognostic factors for PNET patients with LMs. Based on these factors, the two nomograms demonstrated good performance in model evaluation.

CONCLUSION

We developed two clinically significant predictive models to aid physicians in personalized clinical decision-makings.