Dongyang Du1, Hui Feng1, Wenbing Lv1, Saeed Ashrafinia2,3, Qingyu Yuan4, Quanshi Wang4, Wei Yang1, Qianjin Feng1, Wufan Chen1, Arman Rahmim3,5,6, Lijun Lu7. 1. School of Biomedical Engineering and Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, Guangdong, China. 2. Department of Electrical & Computer Engineering, Johns Hopkins University, Baltimore, MD, 21218, USA. 3. Department of Radiology, Johns Hopkins University, Baltimore, MD, 21287, USA. 4. Nanfang PET Center, Nanfang Hospital, Southern Medical University, Guangzhou, 510515, Guangdong, China. 5. Department of Radiology, University of British Columbia, Vancouver, BC, V6T 1Z1, Canada. 6. Department of Physics & Astronomy, University of British Columbia, Vancouver, BC, V6T 1Z1, Canada. 7. School of Biomedical Engineering and Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou, 510515, Guangdong, China. ljlubme@gmail.com.
Abstract
PURPOSE: To identify optimal machine learning methods for radiomics-based differentiation of local recurrence versus inflammation from post-treatment nasopharyngeal positron emission tomography/X-ray computed tomography (PET/CT) images. PROCEDURES: Seventy-six nasopharyngeal carcinoma (NPC) patients were enrolled (41/35 local recurrence/inflammation as confirmed by pathology). Four hundred eighty-seven radiomics features were extracted from PET images for each patient. The diagnostic performance was investigated for 42 cross-combinations derived from 6 feature selection methods and 7 classifiers. Of the original cohort, 70 % was applied for feature selection and classifier development, and the remaining 30 % used as an independent validation set. The diagnostic performance was evaluated using area under the ROC curve (AUC), test error, sensitivity, and specificity. Furthermore, the performance of the radiomics signatures against routine features was statistically compared using DeLong's method. RESULTS: The cross-combination fisher score (FSCR) + k-nearest neighborhood (KNN), FSCR + support vector machines with radial basis function kernel (RBF-SVM), FSCR + random forest (RF), and minimum redundancy maximum relevance (MRMR) + RBF-SVM outperformed others in terms of accuracy (AUC 0.883, 0.867, 0.892, 0.883; sensitivity 0.833, 0.864, 0.831, 0.750; specificity 1, 1, 0.873, 1) and reliability (test error 0.091, 0.136, 0.150, 0.136). Compared with conventional metrics, the radiomics signatures showed higher AUC values (0.867-0.892 vs. 0.817), though the differences were not statistically significant (p = 0.462-0.560). CONCLUSION: This study identified the most accurate and reliable machine learning methods, which could enhance the application of radiomics methods in the precision of diagnosis of NPC.
PURPOSE: To identify optimal machine learning methods for radiomics-based differentiation of local recurrence versus inflammation from post-treatment nasopharyngeal positron emission tomography/X-ray computed tomography (PET/CT) images. PROCEDURES: Seventy-six nasopharyngeal carcinoma (NPC) patients were enrolled (41/35 local recurrence/inflammation as confirmed by pathology). Four hundred eighty-seven radiomics features were extracted from PET images for each patient. The diagnostic performance was investigated for 42 cross-combinations derived from 6 feature selection methods and 7 classifiers. Of the original cohort, 70 % was applied for feature selection and classifier development, and the remaining 30 % used as an independent validation set. The diagnostic performance was evaluated using area under the ROC curve (AUC), test error, sensitivity, and specificity. Furthermore, the performance of the radiomics signatures against routine features was statistically compared using DeLong's method. RESULTS: The cross-combination fisher score (FSCR) + k-nearest neighborhood (KNN), FSCR + support vector machines with radial basis function kernel (RBF-SVM), FSCR + random forest (RF), and minimum redundancy maximum relevance (MRMR) + RBF-SVM outperformed others in terms of accuracy (AUC 0.883, 0.867, 0.892, 0.883; sensitivity 0.833, 0.864, 0.831, 0.750; specificity 1, 1, 0.873, 1) and reliability (test error 0.091, 0.136, 0.150, 0.136). Compared with conventional metrics, the radiomics signatures showed higher AUC values (0.867-0.892 vs. 0.817), though the differences were not statistically significant (p = 0.462-0.560). CONCLUSION: This study identified the most accurate and reliable machine learning methods, which could enhance the application of radiomics methods in the precision of diagnosis of NPC.
Authors: M Al-Sarraf; M LeBlanc; P G Giri; K K Fu; J Cooper; T Vuong; A A Forastiere; G Adams; W A Sakr; D E Schuller; J F Ensley Journal: J Clin Oncol Date: 1998-04 Impact factor: 44.544
Authors: A T C Chan; P M L Teo; R K Ngan; T W Leung; W H Lau; B Zee; S F Leung; F Y Cheung; W Yeo; H H Yiu; K H Yu; K W Chiu; D T Chan; T Mok; K T Yuen; F Mo; M Lai; W H Kwan; P Choi; P J Johnson Journal: J Clin Oncol Date: 2002-04-15 Impact factor: 44.544
Authors: Philippe Lambin; Emmanuel Rios-Velazquez; Ralph Leijenaar; Sara Carvalho; Ruud G P M van Stiphout; Patrick Granton; Catharina M L Zegers; Robert Gillies; Ronald Boellard; André Dekker; Hugo J W L Aerts Journal: Eur J Cancer Date: 2012-01-16 Impact factor: 9.162
Authors: Marco Gerlinger; Andrew J Rowan; Stuart Horswell; James Larkin; David Endesfelder; Eva Gronroos; Pierre Martinez; Nicholas Matthews; Aengus Stewart; Charles Swanton; M Math; Patrick Tarpey; Ignacio Varela; Benjamin Phillimore; Sharmin Begum; Neil Q McDonald; Adam Butler; David Jones; Keiran Raine; Calli Latimer; Claudio R Santos; Mahrokh Nohadani; Aron C Eklund; Bradley Spencer-Dene; Graham Clark; Lisa Pickering; Gordon Stamp; Martin Gore; Zoltan Szallasi; Julian Downward; P Andrew Futreal Journal: N Engl J Med Date: 2012-03-08 Impact factor: 91.245
Authors: Ober Van Gómez; Joaquin L Herraiz; José Manuel Udías; Alexander Haug; Laszlo Papp; Dania Cioni; Emanuele Neri Journal: Cancers (Basel) Date: 2022-06-14 Impact factor: 6.575