A Predictive Nomogram for Lymph Node Metastasis in Supraglottic Laryngeal Squamous Cell Carcinoma

Evaluating the predictive value of log odds of positive lymph nodes on postoperative survival in patients with laryngeal cancer: a SEER population-based study

Traditionally, the AJCC TNM staging system has been the primary tool for assessing the severity and prognosis of laryngeal cancer. Although several studies have demonstrated that the log odds of positive lymph nodes (LODDS) offers superior predictive accuracy compared to the TNM staging for other cancers, there is limited research for laryngeal cancer. This study analyzed data from SEER database (2000-2019). Independent risk factors for survival were identified using univariate and multivariate Cox regression analyses, and different prognostic models were constructed based on the multivariate analysis results. The predictive performance of these models was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) values. The results indicated that LODDS subgroup, age, marital status, histologic grade, T-stage, and N-stage were consistent independent prognostic factors for overall survival (OS) and cancer-specific survival (CSS). Assessment metrics showed that the multivariate model, which incorporated both LODDS and N staging, outperformed the individual N staging and LODDS models in predicting postoperative prognosis in laryngeal cancer patients. Overall, the multivariate model constructed in this study is a superior tool for predicting the postoperative status of laryngeal cancer.

Development and validation of a machine learning model to predict the risk of lymph node metastasis in early-stage supraglottic laryngeal cancer

Background

Cervical lymph node metastasis (LNM) is a significant factor that leads to a poor prognosis in laryngeal cancer. Early-stage supraglottic laryngeal cancer (SGLC) is prone to LNM. However, research on risk factors for predicting cervical LNM in early-stage SGLC is limited. This study seeks to create and validate a predictive model through the application of machine learning (ML) algorithms.

Methods

The training set and internal validation set data were extracted from the Surveillance, Epidemiology, and End Results (SEER) database. Data from 78 early-stage SGLC patients were collected from Fujian Provincial Hospital for independent external validation. We identified four variables associated with cervical LNM and developed six ML models based on these variables to predict LNM in early-stage SGLC patients.

Results

In the two cohorts, 167 (47.44%) and 26 (33.33%) patients experienced LNM, respectively. Age, T stage, grade, and tumor size were identified as independent predictors of LNM. All six ML models performed well, and in both internal and independent external validations, the eXtreme Gradient Boosting (XGB) model outperformed the other models, with AUC values of 0.87 and 0.80, respectively. The decision curve analysis demonstrated that the ML models have excellent clinical applicability.

Conclusions

Our study indicates that combining ML algorithms with clinical data can effectively predict LNM in patients diagnosed with early-stage SGLC. This is the first study to apply ML models in predicting LNM in early-stage SGLC patients.

New nomogram for predicting lymph node positivity in pancreatic head cancer

BACKGROUND

Pancreatic cancer is one of the most malignant cancers worldwide, and it mostly occurs in the head of the pancreas. Existing laparoscopic pancreaticoduodenectomy (LPD) surgical techniques have has undergone a learning curve, a wide variety of approaches for the treatment of pancreatic cancer have been proposed, and the operation has matured. At present, pancreatic head cancer has been gradually changing from "surgeons' evaluation of anatomical resection" to "biologically inappropriate resection". In this study, the risk of lymph node metastasis in pancreatic head cancer was predicted using common preoperative clinical indicators.

METHODS

The preoperative clinical data of 191 patients with pancreatic head cancer who received LPD in the First Affiliated Hospital of Jilin University from May 2016 to December 2021 were obtained. A univariate regression analysis study was conducted, and the indicators with a significance level of P<0.05 were included in the univariate logistic regression analysis into multivariate. Lastly, a nomogram was built based on age, tumor size, leucocyte,albumin(ALB), and lymphocytes/monocytes(LMR). The model with the highest resolution was selected by obtaining the area under a curve. The clinical net benefit of the prediction model was examined using decision curve analyses.Risk stratification was performed by combining preoperative CT scan with existing models.

RESULTS

Multivariate logistic regression analysis found age, tumor size, WBC, ALB, and LMR as five independent factors. A nomogram model was constructed based on the above indicators. The model was calibrated by validating the calibration curve within 1000 bootstrap resamples. The ROC curve achieved an AUC of 0.745(confidence interval of 95%: 0.673-0.816), thus indicating that the model had excellent discriminative skills. DCA suggested that the predictive model achieved a high net benefit in the nearly entire threshold probability range.

CONCLUSIONS

This study has been the first to investigate a nomogram for preoperative prediction of lymphatic metastasis in pancreatic head cancer. The result suggests that age, ALB, tumor size, WBC, and LMR are independent risk factors for lymph node metastasis in pancreatic head cancer. This study may provide a novel perspective for the selection of appropriate continuous treatment regimens, the increase of the survival rate of patients with pancreatic head cancer, and the selection of appropriate neoadjuvant therapy patients.

Application of an Interpretable Machine Learning Model to Predict Lymph Node Metastasis in Patients with Laryngeal Carcinoma

Objectives

A more accurate preoperative prediction of lymph node metastasis (LNM) plays a decisive role in the selection of treatment in patients with laryngeal carcinoma (LC). This study aimed to develop a machine learning (ML) prediction model for predicting LNM in patients with LC.

Methods

We collected and retrospectively analysed 4887 LC patients with detailed demographical characteristics including age at diagnosis, race, sex, primary site, histology, number of tumours, T-stage, grade, and tumour size in the National Institutes of Health (NIH) Surveillance, Epidemiology, and End Results (SEER) database from 2005 to 2015. A correlation analysis of all variables was evaluated by the Pearson correlation. Independent risk factors for LC patients with LNM were identified by univariate and multivariate logistic regression analyses. Afterward, patients were randomly divided into training and test sets in a ratio of 8 to 2. On this basis, we established logistic regression (LR), k-nearest neighbor (KNN), support vector machine (SVM), extreme gradient boosting (XGBoost), random forest (RF), and light gradient boosting machine (LightGBM) algorithm models based on ML. The area under the receiver operating characteristic curve (AUC) value, accuracy, precision, recall rate, F1-score, specificity, and Brier score was adopted to evaluate and compare the prediction performance of the models. Finally, the Shapley additive explanation (SHAP) method was used to interpret the association between each feature variable and target variables based on the best model.

Results

Of the 4887 total LC patients, 3409 were without LNM (69.76%), and 1478 had LNM (30.24%). The result of the Pearson correlation showed that variables were weakly correlated with each other. The independent risk factors for LC patients with LNM were age at diagnosis, race, primary site, number of tumours, tumour size, grade, and T-stage. Among six models, XGBoost displayed a better performance for predicting LNM, with five performance metrics outperforming other models in the training set (AUC: 0.791 (95% CI: 0.776–0.806), accuracy: 0.739, recall rate: 0.638, F1-score: 0.663, and Brier score: 0.165), and similar results were observed in the test set. Moreover, the SHAP value of XGBoost was calculated, and the result showed that the three features, T-stage, primary site, and grade, had the greatest impact on predicting the outcomes.

Conclusions

The XGBoost model performed better and can be applied to forecast the LNM of LC, offering a valuable and significant reference for clinicians in advanced decision-making.