Development and Validation of Prognostic Nomograms for Patients with Primary Gastrointestinal Non-Hodgkin Lymphomas

Deep learning model for predicting the survival of patients with primary gastrointestinal lymphoma based on the SEER database and a multicentre external validation cohort

PURPOSE

Due to the rarity of primary gastrointestinal lymphoma (PGIL), the prognostic factors and optimal management of PGIL have not been clearly defined. We aimed to establish prognostic models using a deep learning algorithm for survival prediction.

METHODS

We collected 11,168 PGIL patients from the Surveillance, Epidemiology, and End Results (SEER) database to form the training and test cohorts. At the same time, we collected 82 PGIL patients from three medical centres to form the external validation cohort. We constructed a Cox proportional hazards (CoxPH) model, random survival forest (RSF) model, and neural multitask logistic regression (DeepSurv) model to predict PGIL patients' overall survival (OS).

RESULTS

The 1-, 3-, 5-, and 10-year OS rates of PGIL patients in the SEER database were 77.1%, 69.4%, 63.7%, and 50.3%, respectively. The RSF model based on all variables showed that the top three most important variables for predicting OS were age, histological type, and chemotherapy. The independent risk factors for PGIL patient prognosis included sex, age, race, primary site, Ann Arbor stage, histological type, symptom, radiotherapy, and chemotherapy, according to the Lasso regression analysis. Using these factors, we built the CoxPH and DeepSurv models. The DeepSurv model's C-index values were 0.760 in the training cohort, 0.742 in the test cohort, and 0.707 in the external validation cohort, which demonstrated that the DeepSurv model performed better compared to the RSF model (0.728) and the CoxPH model (0.724). The DeepSurv model accurately predicted 1-, 3-, 5- and 10-year OS. Both calibration curves and decision curve analysis curves demonstrated the superior performance of the DeepSurv model. We developed the DeepSurv model as an online web calculator for survival prediction, which can be accessed at http://124.222.228.112:8501/ .

CONCLUSIONS

This DeepSurv model with external validation is superior to previous studies in predicting short-term and long-term survival and can help us make better-individualized decisions for PGIL patients.

Development and External Validation of a Nomogram to Predict Cancer-Specific Survival in Patients with Primary Intestinal Non-Hodgkin Lymphomas

Purpose

Primary intestinal non-Hodgkin lymphoma (PINHL) is a biologically and clinically heterogeneous disease. Few individual prediction models are available to establish prognoses for PINHL patients. Herein, a novel nomogram was developed and verified to predict long-term cancer-specific survival (CSS) rates in PINHL patients, and a convenient online risk calculator was created using the nomogram.

Materials and Methods

Data on PINHL patients from January 1, 2004, to December 31, 2015, obtained from the Surveillance, Epidemiology, and End Results (SEER) database (n = 2372; training cohort), were analyzed by Cox regression to identify independent prognostic parameters for CSS. The nomogram was internally and externally validated in a SEER cohort (n = 1014) and a First Affiliated Hospital of Guangzhou University of Chinese Medicine (FAHGUCM) cohort (n = 37), respectively. Area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA) were used to evaluate nomogram performance.

Results

Five independent predictors were identified, namely, age, marital status, Ann Arbor Stage, B symptoms, and histologic type. The nomogram showed good performance in discrimination and calibration, with C-indices of 0.772 (95% CI: 0.754–0.790), 0.763 (95% CI: 0.734–0.792), and 0.851 (95% CI: 0.755–0.947) in the training, internal validation, and external validation cohorts, respectively. The calibration curve indicated that the nomogram was accurate, and DCA showed that the nomogram had a high clinical application value. AUC values indicated that the prediction accuracy of the nomogram was higher than that of Ann Arbor Stage (training cohort: 0.804 vs 0.630; internal validation cohort: 0.800 vs 0.637; external validation cohort: 0.811 vs 0.598), and Kaplan–Meier curves indicated the same.

Conclusion

A nomogram was developed to assist clinicians in predicting the survival of PINHL patients and in making optimal treatment decisions. An online calculator based on the nomogram was made available at https://cuifenzhang.shinyapps.io/DynNomapp/.

Second Primary Malignancies in Patients with Pancreatic Neuroendocrine Neoplasms: A Population-Based Study on Occurrence, Risk Factors, and Prognosis

Background

This study aimed to evaluate the risk factors of developing second primary malignancies (SPMs) among patients with pancreatic neuroendocrine neoplasms (pNENs) and the prognosis of pNENs patients with SPMs (pSPMs) using data from the Surveillance, Epidemiology, and End Results (SEER) database.

Methods

Data from patients diagnosed with pNENs between 1988 and 2016 were extracted. A case-control study was conducted to investigate the risk factors of developing SPMs among patients with pNENs. Meanwhile, cox regression analysis was also conducted to obtain the independent prognostic factors in pSPMs.

Results

Of 7,630 patients with pNENs, 326 developed SPMs. Patients with pNENs who had not undergone surgery and had been diagnosed in recent periods had a higher risk of developing SPMs. The following independent prognostic predictors for pSPMs were identified: age, latency period, SEER stage, radiotherapy, and surgery.

Conclusions

These findings may improve the surveillance of risk factors for developing SPMs in patients with pNENs and the prognostic risk factors in pSPMs.