Development and Validation of a 18F-FDG PET-Based Radiomic Model for Evaluating Hypermetabolic Mediastinal–Hilar Lymph Nodes in Non-Small-Cell Lung Cancer

Abstract

BackgroundAccurate evaluation of lymph node (LN) status is critical for determining the treatment options in patients with non-small cell lung cancer (NSCLC). This study aimed to develop and validate a 18F-FDG PET-based radiomic model for the identification of metastatic LNs from the hypermetabolic mediastinal–hilar LNs in NSCLC.MethodsWe retrospectively reviewed 259 patients with hypermetabolic LNs who underwent pretreatment 18F-FDG PET/CT and were pathologically confirmed as NSCLC from two centers. Two hundred twenty-eight LNs were allocated to a training cohort (LN = 159) and an internal validation cohort (LN = 69) from one center (7:3 ratio), and 60 LNs were enrolled to an external validation cohort from the other. Radiomic features were extracted from LNs of PET images. A PET radiomics signature was constructed by multivariable logistic regression after using the least absolute shrinkage and selection operator (LASSO) method with 10-fold cross-validation. The PET radiomics signature (model 1) and independent predictors from CT image features and clinical data (model 2) were incorporated into a combined model (model 3). A nomogram was plotted for the complex model, and the performance of the nomogram was assessed by its discrimination, calibration, and clinical usefulness.ResultsThe area under the curve (AUC) values of model 1 were 0.820, 0.785, and 0.808 in the training, internal, and external validation cohorts, respectively, showing good diagnostic efficacy for lymph node metastasis (LNM). Furthermore, model 2 was able to discriminate metastatic LNs in the training (AUC 0.780), internal (AUC 0.794), and external validation cohorts (AUC 0.802), respectively. Model 3 showed optimal diagnostic performance among the three cohorts, with an AUC of 0.874, 0.845, and 0.841, respectively. The nomogram based on the model 3 showed good discrimination and calibration.ConclusionsOur study revealed that PET radiomics signature, especially when integrated with CT imaging features, showed the ability to identify true and false positives of mediastinal–hilar LNM detected by PET/CT in patients with NSCLC, which would help clinicians to make individual treatment decisions.

Developing a primary tumor and lymph node 18F-FDG PET/CT-clinical (TLPC) model to predict lymph node metastasis of resectable T2-4 NSCLC

“…Herein we demonstrated that the combination of radiomic features of primary tumor and lymph node based on PET/CT improved the LNM diagnostic sensitivity, specificity and accuracy from 0.59, 0.38, and 0.50 of the TPC model and 0.92, 0.41, and 0.70 of the LPC model to 0.93, 0.75, 0.85 of the TLPC model in the testing set. The combination yielded better diagnostic efficacy than most of the existing forecasts, with values much higher than those of previous studies (for which sensitivity, specificity, and accuracy in the testing set ranged from 0.64, 0.94, 0.77 to 0.82, 0.89, 0.85, respectively) (Wang et al 2017 ; Cong et al 2020a , 2020b ; Ouyang et al 2021 ; Zheng et al 2021 ). These findings supported the necessity of comprehensive analysis of each lesion when evaluating the clinical stage of lung cancer patients; this approach may also be necessary for evaluating therapeutic efficacy or prognosis of lung cancer patients.…”

“…High-throughput radiomics has recently emerged as a powerful approach for identification of imaging biomarkers that can be used to build decision-support systems for cancer treatment (Hyun et al 2019 ) ( Lee et al 2015 ). Machine learning can be significantly effective for object detection and classification, and it is being increasingly used to help clinicians predict LNM based on radiomic features of primary tumors or lymph nodes (Goldstraw et al 2016 ; Cong et al 2020a ; Ouyang et al 2021 ; Zheng et al 2021 ; Scrivener et al 2016 ; Li et al 2015 ) with area under the curve (AUC) ranging from 0.77 to 0.86 . However, few studies have combined radiomic features of primary tumor and lymph node extracted from PET/CT images to build a LNM prediction model by applying radiomics and machine learning (ML).…”

Radiomics in Oncological PET Imaging: A Systematic Review—Part 1, Supradiaphragmatic Cancers

“…PET radiomics in pulmonary oncology gathered 107 articles [ 83 , 84 , 85 , 86 , 87 , 88 , 89 , 90 , 91 , 92 , 93 , 94 , 95 , 96 , 97 , 98 , 99 , 100 , 101 , 102 , 103 , 104 , 105 , 106 , 107 , 108 , 109 , 110 , 111 , 112 , 113 , 114 , 115 , 116 , 117 , 118 , 119 , 120 , 121 , 122 , 123 , 124 , 125 , 126 , 127 , 128 , 129 , 130 , 131 , 132 , 133 ,…”

Transfer-Learning Deep Radiomics and Hand-Crafted Radiomics for Classifying Lymph Nodes from Contrast-Enhanced Computed Tomography in Lung Cancer

“…Similar studies have tried to predict the nodal status of lung cancer and other disease entities [23][24][25]. The authors of [23] applied a LASSO [26] model to directly estimate the N-stage on a per-patient basis using clinical information and features derived from PET/CT. They evaluated the model discrimination, not the calibration, in a single train/test split and a small external validation cohort.…”

“…Predicted outcomes were composed of the severity of pulmonary lesions, epidermal growth factor receptor status, or survival outcomes [28][29][30][31][32][33]. In contrast, we focus on the per-node dignity of thoracic lymph nodes, where rather little research has been conducted [23,24]. However, the potential synergies of applying radiomics and machine learning to the primary tumor and the thoracic lymph nodes remain subject to future research.…”