Differentiating the pathological subtypes of primary lung cancer for patients with brain metastases based on radiomics features from brain CT images

Abstract

Objectives It is of high clinical importance to identify the primary lesion and its pathological types for patients with brain metastases (BM). The purpose of this study is to investigate the feasibility and accuracy of differentiating the primary adenocarcinoma (AD) and squamous cell carcinoma (SCC) of non-small-cell lung cancer (NSCLC) for patients with BM based on radiomics from brain contrast-enhanced computer tomography (CECT) images. Methods A total of 144 BM patients (94 male, 50 female) were enrolled in this study with 102 with primary lung AD and 42 with SCC, respectively. Radiomics features from manually contoured tumors were extracted using python. Mann-Whitney U test and the least absolute shrinkage and selection operator (LASSO) logistic regression were applied to select relative radiomics features. Binary logistic regression and support vector machines (SVM) were applied to build models with radiomics features alone and with radiomics features plus age and sex. Results Fourteen features were selected from a total of 105 radiomics features for the final model building. The area under the curves (AUCs) and accuracy of SVM and binary logistic regression models were 0.765 vs. 0.769, 0.795 vs.0.828, and 0.716 vs. 0.726, 0.768 vs. 0.758, respectively, for models with radiomics features alone and models with radiomics features plus sex and age. Conclusions Brain CECT radiomics are promising in differentiating primary AD and SCC to achieve optimal therapeutic management in patients with BM from NSCLC. Key Points• It is of high clinical importance to identify the primary lesion and its pathological types for patients with brain metastases (BM) to define the prognosis and treatment. • Few studies had investigated the feasibility and accuracy of differentiating the pathological subtypes of primary non-small-cell lung cancer between adenocarcinoma (AD) and squamous cell carcinoma (SCC) for patients with BM based on radiomics from brain contrast-enhanced CT (CECT) images, although CECT images are often the initial imaging modality to screen for metastases and are recommended on equal footing with MRI for the detection of cerebral metastases. • Brain CECT radiomics are promising in differentiating primary AD and SCC to achieve optimal therapeutic management in patients with BM from NSCLC with a highest area under the curve (AUC) of 0.828 and an accuracy of 0.758, respectively. Keywords Brain tumors . Metastasis . Non-small-cell lung cancer . Adenocarcinoma . Squamous cell carcinoma Abbreviations AD Adenocarcinoma AUC Area under the curve BM Brain metastases CECT Contrast-enhanced computer tomography ECCR Ethics Committee in Clinical Research GLCM Gray-level co-occurrence matrix GLRLM Gray-level run-length matrix GLZLM Gray-level zone length matrix LASSO Least absolute shrinkage and selection operator Ji Zhang and Juebin Jin are equal contributors.

[Retracted] Value of CT Radiomics and Clinical Features in Predicting Bone Metastases in Patients with NSCLC

“…Since Lambin et al [7] proposed the concept of radiomics in 2012, it has developed rapidly by extracting quantitative features of lesions through high-throughput images to describe tumor phenotype and heterogeneity from a macroscopic perspective. A large number of studies have also focused on the relationship between radiomics and lymph node metastasis and brain metastasis [8][9][10], including metastasis of NSCLC [11]. In view of this, this study aimed to apply the method of radiomics to explore the CT radiomic characteristics and clinical characteristics of NSCLC patients and their predictive value for the occurrence of bone metastases.…”

Development and validation a radiomics nomogram for diagnosing occult brain metastases in patients with stage IV lung adenocarcinoma

“…Recently, many studies aimed to construct CT radiomic models to assess the relationship between lung cancer and BM ( 19 , 20 ). Zhang et al extracted radiomics features from contrast-enhanced brain CT to differentiate the pathological subtypes of primary lung cancer ( 21 ).…”

Beyond Glioma: The Utility of Radiomic Analysis for Non-Glial Intracranial Tumors

“…However, the radiomics analysis outperformed the radiologists' readings in all cases. Another interesting study was intended to investigate the differentiability of lung cancer, breast cancer, and melanoma metastases using 2D and 3D features with different gray-level quantizations (8,16,32,64,128) and a random forest classifier in CE T1 images [30]. A mean AUC of 0.87 was reported for the top four features out of 43 in an optimal dataset (3D, 32 gray levels).…”

“…Moreover, the use of CE CT imaging-based radiomics to differentiate adenocarcinoma from squamous cell cancer in lung cancer patients with BMs was also reported. Finally, a promising AUC of 0.83 was reported using binary logistic regression paired with clinical data as a classifier [32].…”