Prognostic importance of lymph node yield after curative resection of gastroenteropancreatic neuroendocrine tumours

Application of machine learning algorithms to predict lymph node metastasis in gastric neuroendocrine neoplasms

Background

Neuroendocrine neoplasms (NENs) are tumors that originate from secretory cells of the diffuse endocrine system and typically produce bioactive amines or peptide hormones. This paper describes the development and validation of a predictive model of the risk of lymph node metastasis among gastric NEN patients based on machine learning platform.

Methods

In this investigation, data from 1256 patients were used, of whom 119 patients from the First Affiliated Hospital of Soochow University in China and 1137 cases from the surveillance epidemiology and end results (SEER) database were combined. Six machine learning algorithms, including the logistic regression model (LR), random forest (RF), decision tree (DT), Naive Bayes (NB), support vector machine (SVM), and k-nearest neighbor algorithm (KNN), were used to build the predictive model. The performance of the models was evaluated using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity.

Results

Among the 1256 patients with gastric NENs, 276 patients (21.97 %) developed lymph node metastasis. T stage, tumor size, degree of differentiation, and sex were predictive factors of lymph node metastasis. The RF model achieved the best predictive performance among the six machine learning models, with an AUC, accuracy, sensitivity, and specificity of 0.81, 0.78, 0.76, and 0.82, respectively.

Conclusion

The RF model provided the best prediction and can help physicians determine the lymph node metastasis risk of gastric NEN patients to formulate individualized medical strategies.

Prognostic factors for relapse in resected gastroenteropancreatic neuroendocrine neoplasms: A systematic review and meta-analysis

BACKGROUND

Gastroenteropancreatic neoplasms (GEP-NENs)can potentially be cured through surgical resection, but only 42-57% achieve 5-year disease-free survival.There is a lack of consensus regarding the factorsassociated withrelapse followingresection ofGEP-NENs.

METHODS

Asystematic review identified studies reporting factors associated with relapse in patients with GEP-NENs following resection of a primary tumour. Meta-analysis was performed to identify the factors prognostic for relapse-free survival (RFS)oroverall survival (OS).

RESULTS

63 studies comprising 13,715 patients were included; 56 studies reported on pancreatic NENs (12,418 patients), 24 reported on patients with grade 1-2 tumours (4,735 patients). Median follow-up was 44.2 months, median RFS was 32 months. Pooling of multivariable analyses of GEP-NENs (all sites and grades) found the following factors predicted worse RFS (all p values < 0.05): vascular resection performed, metastatic disease resected, grade 2 disease, grade 3 disease, tumour size > 20 mm, R1 resection, microvascular invasion, perineural invasion, Ki-67 > 5% and any lymph node positivity. In a subgroup of studies comprising exclusively of grade 1-2 GEP-NENs, R1 resection, perineural invasion, grade 2 disease, any lymph node positivity and tumour size > 20 mm predicted worse RFS (all p values < 0.05). Few OSdata were available for pooling; in univariableanalysis(entire cohort), grade 2 predicted worse OS (p = 0.007), whileR1 resectiondid not (p = 0.14).

CONCLUSIONS

The factors prognostic for worse RFS following resection of a GEP-NEN identified in this meta-analysis could be included in post-curative treatment surveillance clinical guidelines and inform the stratification and inclusion criteria of future adjuvant trials.

Impact of Examined Lymph Node Count and Lymph Node Density on Overall Survival of Penile Cancer

BACKGROUND

Few studies have explored the optimal examined lymph node count and lymph node density cutoff values that could be used to predict the survival of patients with penile cancer. We further clarify the prognostic value of lymph node density and examined lymph node count in penile cancer.

METHODS

The Surveillance, Epidemiology, and End Results (SEER) database was explored to recruit penile cancer patients from 2010 to 2015. A retrospective analysis of penile cancer patients' data from the First Affiliated Hospital of Anhui Medical University was performed for verification (2006-2016). The cutoff values of examined lymph node count and lymph node density were performed according to the ROC curve. Kaplan-Meier survival analysis was used to compare survival differences among different groups. Univariate and multivariate Cox proportional hazard regression analyses were used to determine the significant variables. On the basis of Cox proportional hazards regression model, a nomogram was established and validated by calibration plot diagrams and concordance index (C-index).

RESULTS

A total of 528 patients in the Surveillance, Epidemiology, and End Results cohort and 156 patients in the Chinese cohort were included in this study. Using the ROC curve, we found that the recommended cutoff values of ELN and LND were 13 and 9.3%, respectively (P <0.001). Kaplan-Meier curves suggested the significant differences of overall survival among different examined lymph nodes and lymph node density. Multivariate analysis indicated ELN and LND were independent prognostic factor for OS of penile cancer patients. Nomogram showed the contribution of ELN and LND to predicting OS was large. The C-index at 3-, and 5-year were 0.744 for overall survival (95% CI 0.711-0.777).

CONCLUSIONS

The more lymph nodes examined, the lower the density of lymph nodes, and the higher the long-term survival rate of penile cancer. We recommended 13 examined lymph nodes and lymph node density >9.3% as the cutoff value for evaluating the prognosis of penile cancer patients.

Risk factors for lymph node metastasis in gastric neuroendocrine tumor: a retrospective study

Background

Lymph node metastasis (LNM) plays a vital role in the determination of clinical outcomes in patients with gastric neuroendocrine tumor (G-NET). Preoperative identification of LNM is helpful for intraoperative lymphadenectomy. This study aims to investigate risk factors for LNM in patients with G-NET.

Methods

We performed a retrospective study involving 37 patients in non-LNM group and 82 patients in LNM group. Data of demographics, preoperative lab results, clinical–pathological results, surgical management, and postoperative situation were compared between groups. Significant parameters were subsequently entered into logistic regression for further analysis.

Results

Patients in LNM group exhibited older age (p = 0.011), lower preoperative albumin (ALB) (p = 0.003), higher carcinoembryonic antigen (CEA) (p = 0.020), higher International normalized ratio (p = 0.034), longer thrombin time (p = 0.018), different tumor location (p = 0.005), higher chromogranin A positive rate (p = 0.045), and higher Ki-67 expression level (p = 0.002). Logistic regression revealed ALB (p = 0.043), CEA (p = 0.032), tumor location (p = 0.013) and Ki-67 (p = 0.041) were independent risk factors for LNM in G-NET patients.

Conclusions

ALB, CEA, tumor location, and Ki-67 expression level correlate with the risk of LNM in patients with G-NET.