Multiple Level CT Radiomics Features Preoperatively Predict Lymph Node Metastasis in Esophageal Cancer: A Multicentre Retrospective Study

Abstract

Background: Lymph node (LN) metastasis is the most important prognostic factor in esophageal squamous cell carcinoma (ESCC). Traditional clinical factor and existing methods based on CT images are insufficiently effective in diagnosing LN metastasis. A more efficient method to predict LN status based on CT image is needed. Methods: In this multicenter retrospective study, 411 patients with pathologically confirmed ESCC were registered from two hospitals. Quantitative image features including handcrafted-, computer vision-(CV-), and deep-features were extracted from preoperative arterial phase CT images for each patient. A handcrafted-, CV-, and deep-radiomics signature were built, respectively. Then, multiple radiomics models were constructed by merging independent clinical risk factor into radiomics signatures. The performance of models were evaluated with respect to the discrimination, calibration, and clinical usefulness. Finally, an independent external validation cohort was used to validate the model's predictive performance. Results: Five, seven, and nine features were selected for building handcrafted-, CV-, and deep-radiomics signatures from extracted features, respectively. Those signatures were statistically significant different between LN-positive and LN-negative patients in all cohorts (p < 0.001). The developed multiple level CT radiomics model that integrates multiple radiomics signatures with clinical risk factor, was superior to traditional clinical factors and the results reported by existing methods, and achieved satisfactory discrimination performance with C-statistic of 0.875 in development cohort, 0.874 in internal validation cohort and 0.840 in independent external validation cohort. Nomogram and decision curve analysis (DCA) further confirmed our method may serve as an effective tool for clinicians to evaluate the risk of LN metastasis in patients with ESCC and further choose treatment strategy. Wu et al. Radiomics Approach Predicting ESCC LNM Conclusions: The proposed multiple level CT radiomics model which integrate multiple level radiomics features into clinical risk factor can be used for preoperative predicting LN metastasis of patients with ESCC.

WITHDRAWN: Prognostic Significance of Pyruvate Kinase M2 Expression in Esophageal Squamous Cell Carcinoma and Its Meta-Analysis

“…Similarly, some studies have suggested that PKM2 can be used as a prognostic marker for pancreatic ductal cell carcinoma (PDCA), breast cancer, hepatocellular carcinoma (HCC), and gallbladder carcinoma [23][24][25]. In addition, lymph node metastasis is the most important prognostic factor in ESCC [26], accurate nodal staging is crucial for ESCC [27]. Our study showed that PKM2 over-expression promoted lymph node metastasis of ESCC, suggesting that PKM2 may be a molecular target for lymph node metastasis of ESCC.…”

Machine Learning and Radiomics Applications in Esophageal Cancers Using Non-Invasive Imaging Methods—A Critical Review of Literature

“…The major advantage of deep learning-based feature extraction is that no specific domain knowledge is required for feature engineering, and the representative and high-level features can be learned in a completely automatic manner. Recent works [ 88 , 89 ] have shown that the automatically learned deep features of neural networks can outperform the hand-crafted ones in some applications. One key challenge of applying deep networks in clinical decision making is that deep networks are black box models with multilayer nonlinear operations, thus the reasoning behind the results from deep networks are very difficult to interpret clinically.…”

“…There were four studies [ 88 , 118 , 119 , 120 ] focusing on the lymph node metastasis status of EC, with three studies using CT and one study using MRI. Shen et al [ 120 ] built a nomogram incorporating radiomics features, CT-reported suspicious lymph node number and tumor location, which showed good discrimination of lymph node status with a C-index of 0.75 in the validation set.…”

A Nomogram Based on CT Deep Learning Signature: A Potential Tool for the Prediction of Overall Survival in Resected Non-Small Cell Lung Cancer Patients

“…15,[17][18][19][20][21] However, hand-crafted features are low-order and susceptible to noise, and may not be adequate for unveiling the characteristics of tumors. 22 Currently, deep learning based on a neural network structure has shown great potential in medical images. 23,24 It can automatically extract high-level features from pixel images for tumor classification, segmentation and detection.…”

“… 15 , 17–21 However, hand-crafted features are low-order and susceptible to noise, and may not be adequate for unveiling the characteristics of tumors. 22 …”