Radiomics, which involves the extraction of large numbers of quantitative features from medical images, has attracted attention in cancer research. In radiomics analysis, tumor segmentation is a crucial step. In this study, we evaluated the potential application of radiomics for predicting the histology of early stage non-small cell lung cancer (NSCLC) by analyzing interobserver variability in tumor delineation. Forty patient datasets were included in this study, 21 involving adenocarcinomas and 19 involving squamous cell carcinomas. All patients underwent stereotactic body radiotherapy treatment. In total, 476 features were extracted from each dataset, representing treatment planning, computed tomography images, and gross tumor volume (GTV). The definition of GTV can significantly affect the histology prediction. Therefore, in the present study, the effect of interobserver tumor delineation variability on radiomic features was evaluated by preparing 4 volumes of interest (VOIs) for each patient, as follows: the original GTV (which was delineated at treatment planning); two GTVs delineated retrospectively by radiation oncologists; and a semi-automatic GTV contoured by a medical physicist. Radiomic features extracted from each VOI were then analyzed using a naïve Bayesian model. Area-under-the-curve (AUC) analysis showed that interobserver variability in delineation is a significant factor in radiomics performance. Nevertheless, with 8 selected features, AUC values averaged over the VOIs were high (0.725 ± 0.070). The present study indicated that radiomics has potential for predicting early stage NSCLC histology despite variability in delineation. The high prediction accuracy implies that noninvasive histology evaluation by radiomics is a promising clinical application.
Radiomics Histology Non-small-cell lung cancer (NSCLC) Prediction Machine learning
This is a preview of subscription content, log in to check access.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no competing interests.
This work was partially supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (JSPS) KAKENHI JP Scientific Research (C), Grant number 15K08691.
The present study is ethically approved by institutional review board in the University of Tokyo Hospital. The reference number is 3372. This article does not involve any studies performed with animals.
Written informed consent was obtained from all patients whose data were used in this study.
Aerts HJ, Velazquez ER, Leijenaar RT, Parmar C, Grossmann P, Carvalho S, et al. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach. Nat Commun. 2014;5:4006.PubMedPubMedCentralGoogle Scholar
Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, Van Stiphout RGPM, Granton P, et al. Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer. 2012;48(4):441–6.CrossRefPubMedPubMedCentralGoogle Scholar
Ellingson BM, Lai A, Harris RJ, Selfridge JM, Yong WH, Das K, et al. Probabilistic radiographic atlas of glioblastoma phenotypes. Am J Neuroradiol. 2013;34(3):533–40.CrossRefPubMedGoogle Scholar
Boellaard R. Need for standardization of 18F-FDG PET/CT for treatment response assessments. J Nucl Med. 2011;52(Suppl 2):93S–100S.CrossRefPubMedGoogle Scholar
King AD, Chow K-K, Yu K-H, Mo FKF, Yeung DKW, Yuan J, et al. Head and neck squamous cell carcinoma: diagnostic performance of diffusion-weighted MR imaging for the prediction of treatment response. Radiology. 2013;266(2):531–8.CrossRefPubMedGoogle Scholar
Baek HJ, Kim HS, Kim N, Choi YJ, Kim YJ. Percent change of perfusion skewness and kurtosis: a potential imaging biomarker for early treatment response in patients with newly diagnosed glioblastomas (histographic pattern). Radiology. 2012;264(3):834–43.CrossRefPubMedGoogle Scholar
Xu R, Kido S, Suga K, Hirano Y, Tachibana R, Muramatsu K, et al. Texture analysis on 18F-FDG PET/CT images to differentiate malignant and benign bone and soft-tissue lesions. Ann Nucl Med. 2014;28(9):926–35.CrossRefPubMedGoogle Scholar
Nie K, Chen JH, Yu HJ, Chu Y, Nalcioglu O, Su MY. Quantitative analysis of lesion morphology and texture features for diagnostic prediction in breast MRI. Acad Radiol AUR. 2008;15(12):1513–25.CrossRefGoogle Scholar
Wu W, Parmar C, Grossmann P, Quackenbush J, Lambin P, Bussink J, et al. Exploratory study to identify radiomics classifiers for lung cancer histology. Front Oncol. 2016;6:71.PubMedPubMedCentralGoogle Scholar
Parmar C, Velazquez ER, Leijenaar R, Jermoumi M, Carvalho S, Mak RH, et al. Robust radiomics feature quantification using semiautomatic volumetric segmentation. PLoS One. 2014;9(7):1–8.CrossRefGoogle Scholar
Gaede S, Olsthoorn J, Louie AV, Palma D, Yu E, Yaremko B, et al. An evaluation of an automated 4D-CT contour propagation tool to define an internal gross tumour volume for lung cancer radiotherapy. Radiother Oncol. 2011;101(2):322–8.CrossRefPubMedGoogle Scholar
Haralick R, Shanmugan K, Dinstein I. Textural features for image classification. IEEE Trans Syst Man Cybern. 1973;1973:610–21.CrossRefGoogle Scholar
Galloway MM. Texture analysis using gray level run lengths. Comput Graph Image Process. 1975;4(2):172–9.CrossRefGoogle Scholar
Chu A, Sehgal CM, Greenleaf JF. Use of gray value distribution of run lengths for texture analysis. Pattern Recognit Lett. 1990;11(6):415–9.CrossRefGoogle Scholar
Dasarathy BV, Holder EB. Image characterizations based on joint gray level-run length distributions. Pattern Recognit Lett. 1991;12(8):497–502.CrossRefGoogle Scholar
Thibault G, Fertil B, Navarro C, Pereira S, Cau P, Levy N, Sequeira J and Mari J-L. Texture indexes and gray level size zone matrix: application to cell nuclei classification. In: 10th International conference on pattern recognition and information processing. Minsk; 2009. pp. 140–145.Google Scholar
Amadasun M, King R. Textural features corresponding to textural properties. IEEE Trans Syst Man Cybern. 1989;19(5):1264–73.CrossRefGoogle Scholar
Vallières M, Freeman CR, Skamene SR, El Naqa I. A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Phys Med Biol. 2015;60(14):5471–96.CrossRefPubMedGoogle Scholar
Ratanamahatana C, Gunopulos D. Feature selection for the naive bayesian classifier using decision trees. Appl Artif Intell. 2003;17(5–6):475–87.CrossRefGoogle Scholar
Bishop CM. Pattern recognition and machine learning, pattern recognition. Berlin: Springer; 2006. p. 791–9.Google Scholar
Paez JG, Janne PA, Lee JC, Tracy S, Greulich H, Gabriel S, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science. 2004;304(5676):1497–500.CrossRefPubMedGoogle Scholar