Computed Tomography-Based Predictive Model for the Probability of Lymph Node Metastasis in Gastric Cancer: A Meta-analysis

Machine learning models for prediction of lymph node metastasis in patients with gastric cancer: a Chinese single-centre study with external validation in an Asian American population

OBJECTIVE

To develop and validate machine learning (ML)-based models to predict lymph node metastasis (LNM) in patients with gastric cancer (GC).

DESIGN

Retrospective cohort study.

SETTING

Second Affiliated Hospital of Soochow University.

PARTICIPANTS

A total of 500 inpatients from the Second Affiliated Hospital of Soochow University, collected retrospectively between 1 April 2018 and 31 March 2023, were used as the training set, while 824 Asian patients from the Surveillance, Epidemiology and End Results database comprised the external validation set.

MAIN OUTCOME MEASURES

Prediction models were developed using multiple ML algorithms, including logistic regression, support vector machine, k-nearest neighbours, naive Bayes, decision tree (DT), gradient boosting DT, random forest and artificial neural network (ANN). The predictive value of these models was validated and evaluated through receiver operating characteristic curves, precision-recall (PR) curves, calibration curves, decision curve analysis and accuracy metrics.

RESULTS

Among the ML algorithms, the ANN outperformed others, achieving the highest accuracy (0.722; 95% CI: 0.692 to 0.751), precision (0.732; 95% CI: 0.694 to 0.776), F1 score (0.733; 95% CI: 0.695 to 0.773), specificity (0.728; 95% CI: 0.684 to 0.770) and area under the PR curve (0.781; 95% CI: 0.740 to 0.821) in the external validation results. Moreover, it demonstrated superior calibration and clinical utility. Shapley Additive Explanations analysis identified the depth of invasion, tumour size and Lauren classification as the most influential predictors of LNM in patients with GC. Furthermore, a user-friendly web application was developed to provide individual prediction results.

CONCLUSIONS

This study introduces an accurate, reliable and clinically applicable approach for predicting the risk of LNM in patients with GC. The model demonstrates its potential to enhance the personalised management of GC in diverse populations, supported by external validation and an accessible web application for practical use.

Preoperative predictive model for the probability of lymph node metastasis in gastric cancer: a retrospective study

Background

Effectively diagnosing lymph node (LN) metastasis (LNM) is crucial in determining the condition of patients with gastric cancer (GC). The present study was devised to develop and validate a preoperative predictive model (PPM) capable of assessing the LNM status of individuals with GC.

Methods

A retrospective analysis of consecutive GC patients from two centers was conducted over the period from January 2021 to December 2023. These patients were utilized to construct a 289-patient training cohort for identifying LNM-related risk factors and developing a PPM, as well as a 90-patient testing cohort used for PPM validation.

Results

Of the GC patients included in the training cohort, 67 (23.2%) and 222 (76.8%) were respectively LNM negative and positive. Risk factors independently related to LNM status included cT3 invasion (P = 0.001), CT-reported LN (+) (P = 0.044), and CA199 value (P = 0.030). LNM risk scores were established with the following formula: score = -2.382 + 0.694×CT-reported LN status (+: 1; -: 0)+2.497×invasion depth (cT1: 0; cT2: 1; cT3: 2)+0.032×CA199 value. The area under the curve (AUC) values for PPM and CT-reported LN status were 0.753 and 0.609, respectively, with a significant difference between them (P < 0.001). When clinical data from the testing cohort was included in the PPM, the AUC values for the PPM and CT-reported LN status were 0.756 and 0.568 (P < 0.001).

Conclusions

The established PPM may be an effective technique for predicting the LNM status of patients preoperatively. This model can better diagnose LNM than CT-reported LN status alone, this model is better able to diagnose LNM.

A MRI radiomics-based model for prediction of pelvic lymph node metastasis in cervical cancer

BACKGROUND

Cervical cancer (CC) is a common malignancy of the female reproductive tract, and preoperative prediction of lymph node metastasis (LNM) is essential. This study aims to design and validate a magnetic resonance imaging (MRI) radiomics-based predictive model capable of detecting LNM in patients diagnosed with CC.

METHODS

This retrospective analysis incorporated 86 and 38 CC patients into the training and testing groups, respectively. Radiomics features were extracted from MRI T2WI, T2WI-SPAIR, and axial apparent diffusion coefficient (ADC) sequences. Selected features identified in the training group were then used to construct a radiomics scoring model, with relevant LNM-related risk factors having been identified through univariate and multivariate logistic regression analyses. The resultant predictive model was then validated in the testing cohort.

RESULTS

In total, 16 features were selected for the construction of a radiomics scoring model. LNM-related risk factors included worse differentiation (P < 0.001), more advanced International Federation of Gynecology and Obstetrics (FIGO) stages (P = 0.03), and a higher radiomics score from the combined MRI sequences (P = 0.01). The equation for the predictive model was as follows: -0.0493-2.1410 × differentiation level + 7.7203 × radiomics score of combined sequences + 1.6752 × FIGO stage. The respective area under the curve (AUC) values for the T2WI radiomics score, T2WI-SPAIR radiomics score, ADC radiomics score, combined sequence radiomics score, and predictive model were 0.656, 0.664, 0.658, 0.835, and 0.923 in the training cohort, while these corresponding AUC values were 0.643, 0.525, 0.513, 0.826, and 0.82 in the testing cohort.

CONCLUSIONS

This MRI radiomics-based model exhibited favorable accuracy when used to predict LNM in patients with CC. Relative to the use of any individual MRI sequence-based radiomics score, this predictive model yielded superior diagnostic accuracy.