Radiol Oncol 2021; 55(1): 1-6. doi: 10.2478/raon-2020-0068 1 review Artificial intelligence in musculoskeletal oncological radiology Matjaz Vogrin1,2, Teodor Trojner1, Robi Kelc1,2 1 Department of Orthopaedic Surgery, University Medical Center Maribor, Slovenia 2 Faculty of Medicine, University of Maribor, Slovenia Radiol Oncol 2021; 55(1): 1-6. Received 1 June 2020 Accepted 29 September 2020 Correspondence to: Assist. Prof. Robi Kelc, M.D., Ph.D., Department of Orthopaedic Surgery, University Medical Centre Maribor, Ljubljanska ulica 5; SI-2000 Maribor. E-mail: robi.kelc @gmail.com Disclosure: No potential conflicts of interest were disclosed. Background. Due to the rarity of primary bone tumors, precise radiologic diagnosis often requires an experienced musculoskeletal radiologist. In order to make the diagnosis more precise and to prevent the overlooking of potentially dangerous conditions, artificial intelligence has been continuously incorporated into medical practice in recent dec- ades. This paper reviews some of the most promising systems developed, including those for diagnosis of primary and secondary bone tumors, breast, lung and colon neoplasms. Conclusions. Although there is still a shortage of long-term studies confirming its benefits, there is probably a consid- erable potential for further development of computer-based expert systems aiming at a more efficient diagnosis of bone and soft tissue tumors. Key words: artificial intelligence; deep learning; tumor recognition; cancer imaging; image segmentation Introduction Primary bone and soft tissue neoplasms present a minority among neoplastic lesions. Due to their rarity, precise radiologic diagnosis often requires an experienced radiologist with special interests in musculoskeletal oncology. To surmount the chal- lenge of making precise diagnoses, and more im- portantly, to prevent overlooking potentially fatal conditions, attempts to incorporate artificial intelli- gence and its related techniques into medical prac- tice have occurred in the last decades (Figure 1). Being first introduced by McCarthy in the 1950s, artificial intelligence (AI) is a general term that describes computer machines that imitate hu- man intelligence.1 Machine learning, a subset of AI, uses computational algorithms, which learn with experience and therefore improve the per- formance of tasks.2 The rapid progress of compu- tational power and big data availability allowed the emergence of an even more specialized sub- field of machine learning, called deep learning. It is a promising method capable of processing raw data to perform detection or classification tasks.3 Deep learning algorithms, implemented as arti- ficial neural networks, mimic biological nervous systems.4 The network is organized in layers com- posed of interconnected nodes imitating archi- tecture in a biologic brain.2,5 Nodes are weighted individually for the purpose of increasing data extraction. In order to classify data, weights are automatically and dynamically optimized during the training phase.5,6 Regarding the layers, three different kinds are present in each neural network. It begins with an input layer, which receives input data, followed by numerous hidden layers extract- ing the pattern within the data. It terminates with the output layer, which produces results or output data (Figure 2).5 Among the different types of artificial neural networks, convolutional neural networks, in par- ticular, gained attention in radiology due to their high performance in recognizing images.7 By cal- culating the intensity of each pixel or voxel, to- gether with evaluating complex patterns in each image, they provide reliable quantitative image Radiol Oncol 2021; 55(1): 1-6. Kelc R et al. / Artificial intelligence in musculoskeletal oncological radiology2 interpretations, eventually surpassing human per- formance.8 However, to build an intelligent machine a train- ing phase is required, which requires sufficient computational power and large datasets. The latter is obtained from radiological images, which is in the domain of radiomics. Radiomics is a process of quantitative extraction of a high number of seman- tic and agnostic features from diagnostic images.9 Its approaches, like feature extraction and feature engineering techniques, are essential in the forma- tion of AI applications.10 Artificial intelligence in cancer imaging (oncologic radiology) Until recently, radiologists’ decisions were based predominantly on his or her experience of recog- nizing patterns and appreciating various features of each tumor including size, location, intensity, and surface characteristics, all combined with pa- tients’ demographic data. However, after many consecutive image interpretations, they were con- sciously or subconsciously faced with fatigue11, which could lead to errors and potentially jeopard- ize patients’ health and own credibility.12 AI’s potentials are developing exponentially. Instead of qualitative and subjective image inter- pretations, it allows quantifiable and objective data extraction with the ability to reproduce the same results.13 Furthermore, by quantifying information otherwise not detectable to humans, AI may com- plement clinical decision-making.5,13 Computer-aided detection (CADe) and comput- er-aided diagnosis (CADx) algorithms have been used for the last two decades14 predominantly in mammography15, detection of lung16, and colon17 malignancies. In contrast to CAD algorithms, which only highlight the features they have been exactly trained for, actual AI systems continue to learn and improve in time. By focusing on the spe- cific diagnosis, systems learn to discover new typi- cal patterns that have not been linked with the dis- ease before.18 For example, Beck et al. developed a machine learning-based computer program named “Computational Pathologist (C-Path)” to automat- ically analyze breast cancer and predict its prog- nosis.19 Importantly, regardless of all well-known histological characteristics that were implemented into the program, C-Path recognized surrounding stroma as another important prognostic factor.19 Convolutional neural networks have also given optimistic results in automated detection.6 It has been used for automated detection of liver tumors on CT scans with high detection accuracy and pre- cision of 93% and 67%, respectively.20 Similarly, a deep learning-based CADe for detection of brain metastasis on magnetic resonance imaging (MRI) has been developed and achieved a sensitivity of 96% and reasonable specificity.21 In general, characterization is composed of segmentation, diagnosis, and staging of tumors.13 Image segmentation is the process used in cancer imaging to outline pathological area and distin- guish it from non-pathological adjacent tissue. It can range from planar measurements to advanced 3-dimensional assessment of tumor volume.13 Tumors have traditionally been manually labelled by radiologists, which is indeed time-consuming, as well as a subject of interobserver variability.13,22 Thus, implementation of AI into automated image segmentation could potentially take over, increase the quality and reproducibility of measurements, and also save time.5,13 For example, machine learn- ing has been used for breast density segmentation on mammography, which turns out to be as ac- curate as manual ones.23 Ye et al. successfully pro- posed and verified a fully automatic nasopharyn- geal carcinoma segmentation method based on dual-sequence MRI and convolutional neural net- work.24 The mean dice similarity coefficient (DSC) of the models with only T1 sequence, only T2 se- quence, and dual sequence were 0.620 ± 0.0642, FIGURE 1. Schematic illustration of the hierarchy of artificial intelligence and its machine learning and deep learning subfields. Radiol Oncol 2021; 55(1): 1-6. Kelc R et al. / Artificial intelligence in musculoskeletal oncological radiology 3 0.642 ± 0.118, and 0.721 ± 0.036, respectively. The combination of different features acquired from T1 and T2 sequences significantly improved the seg- mentation accuracy.24 Ability to quantitatively extract tumor features has great potential in the process of making diag- nosis. With machine learning, Liu et al. quantita- tively represented radiological traits characteris- tics of lung nodules and showed improved accu- racy of cancer diagnosis in pulmonary nodules.25 Convolutional neural network has also shown to be an effective and objective method that provides an accurate diagnosis of pancreatic cancer.26 Another important aspect of tumor characteri- zation includes staging. Mainly, TNM classifica- tion is used to assess the extent of primary tumor, lymph nodes, and metastases and therefore classi- fy the lesion in a specific stage. However, attempts to extend the TNM cancer staging system have been made. For instance, CAD has shown to be a promising method of evaluation of tumor extent and multifocality in invasive breast cancer patients and therefore expanding the staging algorithms.27 In tumor response monitoring, many AI ap- proaches have shown some potential. For example, AI and machine learning have been successfully implemented into the pre-procedural prediction of trans-arterial chemoembolization treatment outcomes in patients with hepatocellular carci- noma using clinical and imaging features.28 Ha et al. demonstrated promising results in the utiliza- tion of convolutional neural network to predict neoadjuvant chemotherapy response prior to the first cycle of therapy in breast carcinoma using baseline MRI tumor datasets.29 Positive response to chemotherapy led to decreased tumor metabolism, opening a potential opportunity to detect lower ac- cumulation rates of a radiotracer.30 Differentiating between responders and non-responders based on low-dose 18F-FDG PET/MRI scans might, therefore, be another opportunity of the implementation of convolutional neural networks.6 Complementary to radiologic diagnosis, ad- ditional advanced methods have been proposed, promising even further advances in cancer man- agement. Liquid biopsy based on circular tumor DNA (ctDNA) analysis may importantly improve early tumor detection, diagnosis, monitoring ther- apy, and progression in time.31 Contrary to stand- ard tissue biopsies, liquid biopsies taken from blood may provide us with detailed biochemical characteristics of the neoplastic lesion and detect potential metastasis.32 What is more, a combina- tion of liquid biopsies and radiomics, supervised by deep learning may significantly improve cancer management in the future. FIGURE 2. Schematic presentation of a neural network. Regions of interests (ROI) are defined, either by user or by an automated computer process. These present the input cells (in pink) a neural network. For each ROI the neural network extracts and compute features within the hidden layers (in grey) by using pre-trained data sets. Finally, the output cell offers the final results in different possible forms (yes/no, final diagnosis, probability of malignancy etc.). Radiol Oncol 2021; 55(1): 1-6. Kelc R et al. / Artificial intelligence in musculoskeletal oncological radiology4 Artificial intelligence in skeletal tumors The first attempts to introduce computational power into diagnostic procedures of primary bone tumors date back to the 1960s.33 Based on Bayes’ formula, a computer program accurately predicted a bone tumor diagnosis in 77.9% of cases. Later in 1980, the same author set a milestone by publish- ing an article about computed-based radiographic grading of bone tumor destruction.34 This was a cornerstone for further research and implementa- tion of neural networks into the diagnosis of focal bone lesions.35,36 A scarce number of articles regarding AI and primary bone tumors have been published so far, while considerably more has been done on the de- tection and segmentation of bone metastasis. For example, Burns et al. successfully identified and segmented sclerotic lesions in the thoracolumbar spine using CADe techniques. The sensitivity for lesions detection was 79%.37 What is more, Wang et al. developed a Siamese convolutional neural network to research the potential of automated spinal metastasis detection in MRI. The proposed approach accurately detected all spinal metastatic lesions with a false-positive rate of 0.4 per case.38 Another research proposed a machine learning- based whole-body automatic disease classification tool to distinguish benign processes and malignant bone lesions in 18F-NaF PET/CT images.39 Healthy and tumorous bone differs in numer- ous characteristics. Unlike healthy osseous tissue, which consists of cortical and trabecular part, pri- mary bone malignancies may penetrate cortex and spread into adjacent soft tissue, as well as cause swelling around the bone or even weaken the bone architecture and lead to pathological fracture.40 Radiologically, they differ in absorption rate, which can be quantitatively evaluated. For exam- ple, CADx has been used to detect and classify pri- mary bone tumors into benign and malignant le- sions using x-ray images. In their study, Ping et al. an overall greater intensity of pixels for malignant bone tumors compared to benign bone tumors.41 Another study by Bandyopadhyay et al. proposed a CADx method to automatically analyze bone x- ray images. By integrating several classifiers, the method achieved accurate decisions regarding a bone-destruction pattern, stage, and grade of can- cer in 85% of cases.42 When describing sarcomas, diagnosed on MRI, features like tumor size, shape, and enhancement pattern are estimated and taken into consideration along with patient’s demographic data.2 Machine learning and artificial neural network excel in quantifying and extracting supplementary fea- tures, which can correlate with clinical character- istics, diagnosis, and outcomes. Most of these are out of human visual perception and include inter- voxel relationships, image intensity analysis, and filtered images analysis.43 Deep learning-based algorithm has also been developed to predict sur- vival rates in patients with synovial sarcoma.44 Its prediction was more precise compared to the Cox proportional hazard model, which is a commonly used regression model in medical research. In primary bone tumors, bone tumor matrix, its density, and zone of transition represent suit- able characteristics than may be classified through deep learning techniques. In fact, recurrent con- volutional neural network outshined experienced musculoskeletal radiologists in bone tumor matrix classification with 86% vs. 72%, respectively.45 Li et al. proposed a super label guided convolutional neural network to classify CT images of bone tu- mors.46 In comparison, results exceeded the classic convolutional neural network. However, the classi- fication included only nine types of the most com- mon skeletal tumors. Limitations and future directions There are indeed some limitations of AI. First, it could potentially still be a subject of interobserv- er variability, due to different algorithms used in a neural networks of different AI systems or un- equal learning stages in which the systems pro- cess a specific task. Standardization is mandatory to establish a large database. Data also needs to be accessible in order to integrate them into large sets. Prior to that quality check, labelling, classifi- cation, and segmentation need to be done manu- ally by experts, making the process expensive and time-consuming.13 However, introducing deep learning-based techniques to the extensive quality ground-truth training datasets is essential for the development of accurate algorithms.47,48 Also, ethi- cal dilemmas should be taken into consideration. When dealing with systems that operate with enor- mous amounts of data, patients’ privacy as well as human dignity may be jeopardized, unless meticu- lous safety mechanisms are implemented. There are also no long-term follow-up studies available thus far. On the contrary, the appreciating results Radiol Oncol 2021; 55(1): 1-6. Kelc R et al. / Artificial intelligence in musculoskeletal oncological radiology 5 of the already published studies and the presenta- tion of a commercially available application is only a matter of time. Undoubtedly, all kinds of artificial intelligence are persistently being integrated into the com- plex management of musculoskeletal tumors and tumors of other sites. Deep learning-based tech- niques are expected to minimalize false positive rates as well as assure accurate decisions and di- agnoses.49 Further automatization of radiological tasks is expected to take place in the future. Among physicians, radiologists in particular are required to perform many time-consuming tasks, like im- age segmentation, delineation of regions of inter- est, and image annotation. AI techniques have an enormous potential to transform their workflow, which will allow them to focus on more meaning- ful tasks.11 On the other hand, “imaging is not an isolated measure of disease.”13 Neoplastic lesions are com- plex conditions, following DNA mutations that cause abnormal cellular proliferation.50 Despite many mutations being discovered and related to specific malignancies, intertumoural and in- tratumoural heterogeneity exist.51 Undoubtedly, molecular approaches, like genetic biomarkers and molecular imaging have already significantly contributed to a better understanding of cancer management. Finally, combining radiomics with other aspects of a broad family of “-omics”, includ- ing genomics, proteomics, and metabolomics, and therefore drastically expanding datasets available for advanced AI modalities, might move us closer to the precision medicine.52 References 1. Han X-G, Tian W. Artificial intelligence in orthopedic surgery: current state and future perspective. Chin Med J (Engl) 2019; 132: 2521-3. doi: 10.1097/ CM9.0000000000000479 2. Gyftopoulos S, Lin D, Knoll F, Doshi AM, Rodrigues TC, Recht MP. Artificial intelligence in musculoskeletal imaging: current status and future direc- tions. AJR Am J Roentgenol 2019; 213: 506-13. doi: 10.2214/AJR.19.21117 3. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015; 521: 436-44. doi: 10.1038/nature14539 4. King BF. Guest editorial: discovery and artificial intelligence. ARJ Am J Roentgenol 2017; 209: 1189-90. doi: 10.2214/AJR.17.19178 5. Pesapane F, Codari M, Sardanelli F. Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine. Eur Radiol Exp 2018; 2: 35. doi: 10.1186/s41747-018-0061-6 6. Daldrup-Link H. Artificial intelligence applications for pediatric oncology im- aging. Pediatr Radiol 2019; 49: 1384-90. doi: 10.1007/s00247-019-04360-1 7. Yasaka K, Abe O. Deep learning and artificial intelligence in radiology: cur- rent applications and future directions. PLoS Med 2018; 15: e1002707. doi: 10.1371/journal.pmed.1002707 8. Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. Artificial intelligence in radiology. Nat Rev Cancer 2018; 18: 500-10. doi: 10.1038/ s41568-018-0016-5 9. Gillies RJ, Kinahan PE, Hricak H. Radiomics: images are more than pic- tures, they are data. Radiology 2016; 278: 563-77. doi: 10.1148/ra- diol.2015151169 10. Koçak B, Durmaz EŞ, Ateş E, Kılıçkesmez Ö. Radiomics with artificial intelli- gence: a practical guide for beginners. Diagn Interv Radiol 2019; 25: 485-95. doi: 10.5152/dir.2019.19321 11. Hirschmann A, Cyriac J, Stieltjes B, Kober T, Richiardi J, Omoumi P. Artificial intelligence in musculoskeletal imaging: review of current literature, chal- lenges, and trends. Semin Musculoskelet Radiol 2019; 23: 304-11. doi: 10.1055/s-0039-1684024 12. Bruno MA, Walker EA, Abujudeh HH. Understanding and confronting our mistakes: the epidemiology of error in radiology and strategies for error reduction. RadioGraphics 2015; 35: 1668-76. doi: 10.1148/rg.2015150023 13. Bi WL, Hosny A, Schabath MB, Giger ML, Birkbak NJ, Mehrtash A, et al. Artificial intelligence in cancer imaging: clinical challenges and applications. CA Cancer J Clin 2019; 69: 127-57. doi: 10.3322/caac.21552 14. Takahashi R, Kajikawa Y. Computer-aided diagnosis: a survey with biblio- metric analysis. Int J Med Inform 2017; 101: 58-67. doi: 10.1016/j.ijmed- inf.2017.02.004 15. Katzen J, Dodelzon K. A review of computer aided detection in mammog- raphy. Clin Imaging 2018; 52: 305-9. doi: 10.1016/j.clinimag.2018.08.014 16. Al Mohammad B, Brennan PC, Mello-Thoms C. A review of lung cancer screening and the role of computer-aided detection. Clin Radiol 2017; 72: 433-42. doi: 10.1016/j.crad.2017.01.002 17. Regge D, Halligan S. CAD: How it works, how to use it, performance. Eur J Radiol 2013; 82: 1171-6. doi: 10.1016/j.ejrad.2012.04.022 18. Kohli M, Prevedello LM, Filice RW, Geis JR. Implementing machine learning in radiology practice and research. AJR Am J Roentgenol 2017; 208: 754-60. doi: 10.2214/AJR.16.17224 19. Beck AH, Sangoi AR, Leung S, Marinelli RJ, Nielsen TO, van de Vijver MJ, et al. Systematic analysis of breast cancer morphology uncovers stromal features associated with survival. Sci Transl Med 2011; 3: 108ra113. doi: 10.1126/ scitranslmed.3002564 20. Afifi A, Nakaguchi T. Unsupervised detection of liver lesions in CT im- ages. Annu Conf IEEE Eng Med Biol Soc 2015; 2015: 2411-4. doi: 10.1109/ EMBC.2015.7318880 21. Zhang M, Young GS, Chen H, Li J, Qin L, McFaline-Figueroa JR, et al. Deep- learning detection of cancer metastases to the brain on MRI. J Magn Reson Imaging 2020: 52: 1227-36. doi: 10.1002/jmri.27129 22. Li XA, Tai A, Arthur DW, Buchholz TA, MacDonald S, Marks LB, et al. Variability of target and normal structure delineation for breast-cancer radiotherapy: a RTOG multi-institutional and multi-observer study. Int J Radiat Oncol Biol Phys 2009; 73: 944-51. doi: 10.1016/j.ijrobp.2008.10.034 23. Kallenberg M, Petersen K, Nielsen M, Ng AY, Diao P, Igel C, et al. Unsupervised deep learning applied to breast density segmentation and mammographic risk scoring. IEEE Trans Med Imaging 2016; 35: 1322-31. doi: 10.1109/ TMI.2016.2532122 24. Ye Y, Cai Z, Huang B, He Y, Zeng P, Zou G, et al. Fully-automated segmentation of nasopharyngeal carcinoma on dual-sequence MRI using convolutional neural networks. Front Oncol 2020; in press. doi: 10.3389/fonc.2020.00166 25. Liu Y, Balagurunathan Y, Atwater T, Antic S, Li Q, Walker RC, et al. Radiological image traits predictive of cancer status in pulmonary nodules. Clin Cancer Res 2017; 23: 1442-9. doi: 10.1158/1078-0432.CCR-15-3102 26. Liu SL, Li S, Guo YT, Zhou YP, Zhang ZD, Li S, et al. Establishment and applica- tion of an artificial intelligence diagnosis system for pancreatic cancer with a faster region-based convolutional neural network. Chin Med J (Engl) 2019; 132: 2795-2803. doi: 10.1097/CM9.0000000000000544 27. Song SE, Seo BK, Cho KR, Woo OH, Son GS, Kim C, et al. Computer-aided detection (CAD) system for breast MRI in assessment of local tumor extent, nodal status, and multifocality of invasive breast cancers: preliminary study. Cancer Imaging 2015; 15: 1. doi: 10.1186/s40644-015-0036-2 Radiol Oncol 2021; 55(1): 1-6. Kelc R et al. / Artificial intelligence in musculoskeletal oncological radiology6 28. Abajian A, Murali N, Savic LJ, Laage-Gaupp FM, Nezami N, Duncan JS, et al. Predicting treatment response to intra-arterial therapies of hepatocel- lular carcinoma using supervised machine learning - an artificial intel- ligence concept. J Vasc Interv Radiol 2018; 29: 850-57.e1. doi: 10.1016/j. jvir.2018.01.769 29. Ha R, Chin C, Karcich J, Liu MZ, Chang P, Mutasa S, et al. Prior to initiation of chemotherapy, can we predict breast tumor response? Deep learning convolutional neural networks approach using a breast MRI tumor dataset. J Digit Imaging 2019; 32: 693-701. doi: 10.1007/s10278-018-0144-1 30. Kleis M, Daldrup-Link H, Matthay K, Goldsby R, Lu Y, Schuster T, et al. Diagnostic value of PET/CT for the staging and restaging of pediatric tumors. Eur J Nucl Med Mol Imaging 2009; 36: 23-36. doi: 10.1007/s00259-008- 0911-1 31. Cheng F, Su L, Qian C. Circulating tumor DNA: a promising biomarker in the liquid biopsy of cancer. Oncotarget 2016; 7: 48832-41. doi: 10.18632/ oncotarget.9453 32. Mader S, Pantel K. Liquid biopsy: current status and future perspectives. Oncol Res Treat 2017; 40: 404-8. doi: 10.1159/000478018 33. Lodwick GS, Haun CL, Smith WE, Keller RF, Robertson ED. Computer diagnosis of primary bone tumors. Radiology 1963; 80: 273-5. doi: 10.1148/80.2.273 34. Lodwick GS, Wilson AJ, Farrell C, Virtama P, Dittrich F. Determining growth rates of focal lesions of bone from radiographs. Radiology 1980; 134: 577- 83. doi: 10.1148/radiology.134.3.6928321 35. Reinus WR, Wilson AJ, Kalman B, Kwasny S. Diagnosis of focal bone lesions using neural networks. Invest Radiol 1994; 29: 606-11. doi: 10.1097/00004424-199406000-00002 36. Piraino DW, Amartur SC, Richmond BJ, Schils JP, Thome JM, Belhobek GH, et al. Application of an artificial neural network in radiographic diagnosis. J Digit Imaging 1991; 4: 226. doi: 10.1007/BF03173904 37. Burns JE, Yao J, Wiese TS, Muñoz HE, Jones EC, Summers RM. Automated detection of sclerotic metastases in the thoracolumbar spine at CT. Radiology 2013; 268: 69-78. doi: 10.1148/radiol.13121351 38. Wang J, Fang Z, Lang N, Yuan H, Su M-Y, Baldi P. A multi-resolution approach for spinal metastasis detection using deep Siamese neural networks. Comput Biol Med 2017; 84: 137-46. doi: 10.1016/j.compbi- omed.2017.03.024 39. Perk T, Bradshaw T, Chen S, Im H, Cho S, Perlman S, et al. Automated clas- sification of benign and malignant lesions in 18F-NaF PET/CT images using machine learning. Phys Med Biol 2018; 63: 225019. doi: 10.1088/1361- 6560/aaebd0 40. Costelloe CM, Madewell JE. Radiography in the initial diagnosis of pri- mary bone tumors. AJR Am J Roentgenol 2013; 200: 3-7. doi: 10.2214/ AJR.12.8488 41. Ping YY, Yin CW, Kok LP. Computer aided bone tumor detection and clas- sification using x-ray images. In: Abu Osman NA, Ibrahim F, Wan Abas WAB, Abdul Rahman HS, Ting H-N, editors. IFMBE proceedings. 4th Kuala Lumpur International Conference on Biomedical Engineering 2008. Berlin, Heidelberg: Springer; 2008. p. 544-7. doi: 10.1007/978-3-540-69139-6_136 42. Bandyopadhyay O, Biswas A, Bhattacharya BB. Bone-cancer assessment and destruction pattern analysis in long-bone x-ray image. J Digit Imaging 2019; 32: 300-13. doi: 10.1007/s10278-018-0145-0 43. McBee MP, Awan OA, Colucci AT, Ghobadi CW, Kadom N, Kansagra AP, et al. Deep learning in radiology. Acad Radiol 2018; 25: 1472-80. doi: 10.1016/j. acra.2018.02.018 44. Han I, Kim JH, Park H, Kim H-S, Seo SW. Deep learning approach for sur- vival prediction for patients with synovial sarcoma. Tumour Biol 2018; 40: 1010428318799264. doi: 10.1177/1010428318799264 45. Reicher JJ, Palo Alto VA, Do BH, Nguyen M, Beaulieu CF. Single-input bone tumor diagnosis based on convolutional neural network classification of bone tumor matrix. SIIM 2018 Annual Meeting, May 31-June 2, National Harbor, MD, 2018. 46. Li Y, Zhou W, Lv G, Luo G, Zhu Y, Liu J. Classification of bone tumor on CT im- ages using deep convolutional neural network. In: Kůrková V, Manolopoulos Y, Hammer B, Iliadis L, Maglogiannis I, editors. Lecture notes in computer science. Artificial neural networks and machine learning - ICANN 2018. Cham: Springer International Publishing; 2018. p. 127-36. doi: 10.1007/978- 3-030-01421-6_13 47. Choy G, Khalilzadeh O, Michalski M, Do S, Samir AE, Pianykh OS, et al. Current applications and future impact of machine learning in radiology. Radiology 2018; 288: 318-28. doi: 10.1148/radiol.2018171820 48. Gorelik N, Chong J, Lin DJ. Pattern recognition in musculoskeletal imaging using artificial intelligence. Semin Musculoskelet Radiol 2020; 24: 38-49. doi: 10.1055/s-0039-3400266 49. Burns JE, Yao J, Summers RM. Artificial intelligence in musculoskeletal imaging: a paradigm shift. J Bone Miner Res 2020; 35: 28-35. doi: 10.1002/ jbmr.3849 50. Jokar N, Velez E, Shooli H, Dadgar H, Sadathosseini S, Assadi M, et al. Advanced modalities of molecular imaging in precision medicine for mus- culoskeletal malignancies. World J Nucl Med 2019; 18: 345. doi: 10.4103/ wjnm.WJNM_119_18 51. Hinohara K, Polyak K. Intratumoral heterogeneity: more than just muta- tions. Trends Cell Biol 2019; 29: 569-79. doi: 10.1016/j.tcb.2019.03.003 52. Acharya UR, Hagiwara Y, Sudarshan VK, Chan WY, Ng KH. Towards precision medicine: from quantitative imaging to radiomics. J Zhejiang Univ Sci B 2018; 19: 6-24. doi: 10.1631/jzus.B1700260