Head and neck squamous cell carcinoma: prediction of cervical lymph node metastasis by dual-energy CT texture analysis with machine learning
This study was conducted in order to evaluate a novel risk stratification model using dual-energy CT (DECT) texture analysis of head and neck squamous cell carcinoma (HNSCC) with machine learning to (1) predict associated cervical lymphadenopathy and (2) compare the accuracy of spectral versus single-energy (65 keV) texture evaluation for endpoint prediction.
Eighty-seven patients with HNSCC were evaluated. Texture feature extraction was performed on virtual monochromatic images (VMIs) at 65 keV alone or different sets of multi-energy VMIs ranging from 40 to 140 keV, in addition to iodine material decomposition maps and other clinical information. Random forests (RF) models were constructed for outcome prediction with internal cross-validation in addition to the use of separate randomly selected training (70%) and testing (30%) sets. Accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were determined for predicting positive versus negative nodal status in the neck.
Depending on the model used and subset of patients evaluated, an accuracy, sensitivity, specificity, PPV, and NPV of up to 88, 100, 67, 83, and 100%, respectively, could be achieved using multi-energy texture analysis. Texture evaluation of VMIs at 65 keV alone or in combination with only iodine maps had a much lower accuracy.
Multi-energy DECT texture analysis of HNSCC is superior to texture analysis of 65 keV VMIs and iodine maps alone and can be used to predict cervical nodal metastases with relatively high accuracy, providing information not currently available by expert evaluation of the primary tumor alone.
• Texture features of HNSCC tumor are predictive of nodal status.
• Multi-energy texture analysis is superior to analysis of datasets at a single energy.
• Dual-energy CT texture analysis with machine learning can enhance noninvasive diagnostic tumor evaluation.
KeywordsMultidetector computed tomography Machine learning Artificial intelligence Head and neck neoplasms Computer-assisted diagnosis
Area under the receiver operating curve
Head and neck squamous cell carcinoma
Negative predictive value
Positive predictive value
Region of interest
Virtual monochromatic image
This work was partly supported by a grant from the Rossy Cancer Network. R.F. is a clinical research scholar supported by the FRQS (Fonds de recherche en santé du Québec).
Compliance with ethical standards
The scientific guarantor of this publication is R. Forghani.
Conflict of interest
The authors of this manuscript declare relationships with the following companies: R.F. has acted as consultant and speaker for GE Healthcare and is a founding partner and stockholder of 4Intel Inc. B.F. is a founding partner and stockholder of 4Intel Inc.
Statistics and biometry
B.F. has significant statistical and informatics expertise and performed the mathematical and statistical analyses.
Ethics approval was obtained by the Institutional Review Board of the Jewish General Hospital (CIUSSS West-Central Montreal).
Written informed consent was waived by the Institutional Review Board.
• performed at one institution
- 1.Som PM, Brandwein-Gensler MS (2011) Lymph nodes of the neck. In: Som PM, Curtin HD (eds) Head and neck imaging. Mosby, St. LouisGoogle Scholar
- 2.Forghani R, Yu E, Levental M, Som PM, Curtin HD (2014) Imaging evaluation of lymphadenopathy and patterns of lymph node spread in head and neck cancer. Expert Rev Anticancer Ther. https://doi.org/10.1586/14737140.2015.978862:1-18
- 3.Kostakoglu L (2011) PET/CT Imaging. In: Som PM, Curtin HD (eds) Head and neck imaging. Mosby, St. LouisGoogle Scholar
- 5.Abu-Ghanem S, Yehuda M, Carmel NN et al (2016) Elective neck dissection vs observation in early-stage squamous cell carcinoma of the oral tongue with no clinically apparent lymph node metastasis in the neck: a systematic review and meta-analysis. JAMA Otolaryngol Head Neck Surg 142:857–865CrossRefGoogle Scholar
- 8.Medina JE (2017) Cancer of the neck. In: Myers J, Hanna E, Myers EN (eds) Cancer of the head and neck. Wolters Kluwer, Philadelphia, pp 427–453Google Scholar
- 19.Foncubierta-Rodriguez A, Jimenez del Toro OA, Platon A, Poletti PA, Muller H, Depeursinge A (2013) Benefits of texture analysis of dual energy CT for computer-aided pulmonary embolism detection. Conf Proc IEEE Eng Med Biol Soc 2013:3973–3976Google Scholar
- 21.Depeursinge A, Foncubierta-Rodriguez A, Vargas A et al (2013) Rotation-covariant texture analysis of 4D dual-energy CT as an indicator of local pulmonary perfusion. 2013 IEEE 10th international symposium on biomedical imaging (ISBI), San Francisco, CA, USA, pp 145–148Google Scholar
- 31.Yamauchi H, Buehler M, Goodsitt MM, Keshavarzi N, Srinivasan A (2016) Dual-energy CT-based differentiation of benign posttreatment changes from primary or recurrent malignancy of the head and neck: comparison of spectral Hounsfield units at 40 and 70 keV and iodine concentration. AJR Am J Roentgenol 206:580–587CrossRefGoogle Scholar
- 33.Forghani R, Kelly H, Yu E et al (2017) Low-energy virtual monochromatic dual-energy computed tomography images for the evaluation of head and neck squamous cell carcinoma: a study of tumor visibility compared with single-energy computed tomography and user acceptance. J Comput Assist Tomogr 41:565–571CrossRefGoogle Scholar
- 34.Forghani R (2015) Advanced dual-energy CT for head and neck cancer imaging. Expert Rev Anticancer Ther. https://doi.org/10.1586/14737140.2015.1108193:1-13
- 39.Ueno Y, Forghani B, Forghani R et al (2017) Endometrial carcinoma: MR imaging-based texture model for preoperative risk stratification—a preliminary analysis. Radiology. https://doi.org/10.1148/radiol.2017161950:161950
- 43.Caruana R, Karampatziakis N, Yessenalina A (2008) An empirical evaluation of supervised learning in high dimensions. ICML '08, Proceedings of the 25th International conference on Machine learning. ACM, Helsinki, pp 96–103Google Scholar
- 44.Hastie T, Tibshirani R, Friedman J (2009) The elements of statistical learning: Data mining, inference, and prediction, Second edition. Springer Series in Statistics, Springer-VerlagGoogle Scholar
- 45.Liaw A, Wiener M (2002) Classification and regression by randomForest. R News 2:18–22Google Scholar
- 46.Yang X, Wu K, Li S et al (2017) MFAP5 and TNNC1: potential markers for predicting occult cervical lymphatic metastasis and prognosis in early stage tongue cancer. Oncotarget 8:2525–2535Google Scholar