Abstract
Objective
To predict the local recurrence of giant cell bone tumors (GCTB) on MR features and the clinical characteristics after curettage using a deep convolutional neural network (CNN).
Methods
MR images were collected from 56 patients with histopathologically confirmed GCTB after curettage who were followed up for 5.8 years (range, 2.0 to 9.5 years). The inception v3 CNN architecture was fine-tuned by two categories of the MR datasets (recurrent and non-recurrent GCTB) obtained through data augmentation and was validated using fourfold cross-validation to evaluate its generalization ability. Twenty-eight cases (50%) were chosen as the training dataset for the CNN and four radiologists, while the remaining 28 cases (50%) were used as the test dataset. A binary logistic regression model was established to predict recurrent GCTB by combining the CNN prediction and patient features (age and tumor location). Accuracy and sensitivity were used to evaluate the prediction performance.
Results
When comparing the CNN, CNN regression, and radiologists, the accuracies of the CNN and CNN regression models were 75.5% (95% CI 55.1 to 89.3%) and 78.6% (59.0 to 91.7%), respectively, which were higher than the 64.3% (44.1 to 81.4%) accuracy of the radiologists. The sensitivities were 85.7% (42.1 to 99.6%) and 87.5% (47.3 to 99.7%), respectively, which were higher than the 58.3% (27.7 to 84.8%) sensitivity of the radiologists (p < 0.05).
Conclusion
The CNN has the potential to predict recurrent GCTB after curettage. A binary regression model combined with patient characteristics improves its prediction accuracy.
Key Points
• Convolutional neural network (CNN) can be trained successfully on a limited number of pre-surgery MR images, by fine-tuning a pre-trained CNN architecture.
• CNN has an accuracy of 75.5% to predict post-surgery recurrence of giant cell tumors of bone, which surpasses the 64.3% accuracy of human observation.
• A binary logistic regression model combining CNN prediction rate, patient age, and tumor location improves the accuracy to predict post-surgery recurrence of giant cell bone tumors to 78.6%.
Similar content being viewed by others
Abbreviations
- CNN:
-
Convolutional neural network
- GCTB:
-
Giant cell tumor of bone
- PMMA:
-
Polymethylmethacrylate
- RMSProp:
-
Root mean square prop
- t-SNE:
-
t-Distributed stochastic neighbor embedding
References
van der Heijden L, Dijkstra PD, van de Sande MA et al (2014) The clinical approach toward giant cell tumor of bone. Oncologist 19:550–561
Ghert MA, Rizzo M, Harrelson JM, Scully SP (2002) Giant-cell tumor of the appendicular skeleton. Clin Orthop Relat Res 400:201–210
van der Heijden L, Dijkstra PD, Campanacci DA, Gibbons CL, van de Sande MA (2013) Giant cell tumor with pathologic fracture: should we curette or resect? Clin Orthop Relat Res 471:820–829
Klenke FM, Wenger DE, Inwards CY, Rose PS, Sim FH (2011) Giant cell tumor of bone: risk factors for recurrence. Clin Orthop Relat Res 469:591–599
Arbeitsgemeinschaft Knochentumoren, Becker WT, Dohle J et al (2008) Local recurrence of giant cell tumor of bone after intralesional treatment with and without adjuvant therapy. J Bone Joint Surg Am 90:1060–1067
He Y, Zhang J, Ding X (2017) Prognosis of local recurrence in giant cell tumour of bone: what can we do? Radiol Med 122:505–519
Abat F, Almenara M, Peiro A, Trullols L, Bague S, Gracia I (2015) Giant cell tumour of bone: a series of 97 cases with a mean follow-up of 12 years. Rev Esp Cir Ortop Traumatol 59:59–65
Chen L, Ding XY, Wang CS et al (2014) In-depth analysis of local recurrence of giant cell tumour of bone with soft tissue extension after intralesional curettage. Radiol Med 119:861–870
Teixeira LE, Vilela JC, Miranda RH, Gomes AH, Costa FA, de Faria VC (2014) Giant cell tumors of bone: nonsurgical factors associated with local recurrence. Acta Orthop Traumatol Turc 48:136–140
Siddiqui MA, Seng C, Tan MH (2014) Risk factors for recurrence of giant cell tumours of bone. J Orthop Surg (Hong Kong) 22:108–110
Turcotte RE, Wunder JS, Isler MH et al (2002) Giant cell tumor of long bone: a Canadian Sarcoma Group study. Clin Orthop Relat Res 397:248–258
Gouin F, Dumaine V, French Sarcoma and Bone Tumor Study Groups GSF-GETO (2013) Local recurrence after curettage treatment of giant cell tumors in peripheral bones: retrospective study by the GSF-GETO (French Sarcoma and Bone Tumor Study Groups). Orthop Traumatol Surg Res 99:S313–S318
Klenke FM, Wenger DE, Inwards CY, Rose PS, Sim FH (2011) Recurrent giant cell tumor of long bones: analysis of surgical management. Clin Orthop Relat Res 469:1181–1187
Wang H, Wan N, Hu Y (2012) Giant cell tumour of bone: a new evaluating system is necessary. Int Orthop 36:2521–2527
Sanduleanu S, Woodruff HC, de Jong EEC et al (2018) Tracking tumor biology with radiomics: a systematic review utilizing a radiomics quality score. Radiother Oncol 127:349–360
Huang YQ, Liang CH, He L et al (2016) Development and validation of a radiomics nomogram for preoperative prediction of lymph node metastasis in colorectal cancer. J Clin Oncol 34:2157–2164
He Y, Wang J, Rui W et al (2018) Retrospective investigation of “paint brush borders” sign in association with local recurrence of giant cell tumor of bone after intralesional curettage. J Bone Oncol 10:41–48
He Y, Wang J, Zhang J, Yuan F, Ding X (2017) A prospective study on predicting local recurrence of giant cell tumour of bone by evaluating preoperative imaging features of the tumour around the knee joint. Radiol Med 122:546–555
LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444
Mnih V, Kavukcuoglu K, Silver D et al (2015) Human-level control through deep reinforcement learning. Nature 518:529–533
Silver D, Huang A, Maddison CJ et al (2016) Mastering the game of Go with deep neural networks and tree search. Nature 529:484–489
Silver D, Schrittwieser J, Simonyan K et al (2017) Mastering the game of Go without human knowledge. Nature 550:354–359
Becker AS, Marcon M, Ghafoor S, Wurnig MC, Frauenfelder T, Boss A (2017) Deep learning in mammography: diagnostic accuracy of a multipurpose image analysis software in the detection of breast cancer. Invest Radiol 52:434–440
Yasaka K, Akai H, Abe O, Kiryu S (2017) Deep learning with convolutional neural network for differentiation of liver masses at dynamic contrast-enhanced CT: a preliminary study. Radiology 286:887–896
Anthimopoulos M, Christodoulidis S, Ebner L, Christe A, Mougiakakou S (2016) Lung pattern classification for interstitial lung diseases using a deep convolutional neural network. IEEE Trans Med Imaging 35:1207–1216
Banerjee I, Crawley A, Bhethanabotla M, Daldrup-Link HE, Rubin DL (2018) Transfer learning on fused multiparametric MR images for classifying histopathological subtypes of rhabdomyosarcoma. Comput Med Imaging Graph 65:167–175
Chartrand G, Cheng PM, Vorontsov E et al (2017) Deep learning: a primer for radiologists. Radiographics 37:2113–2131
Li XL, Zhang H, Zhang XL, Liu H, Xie GT (2017) Exploring transfer learning for gastrointestinal bleeding detection on small-size imbalanced endoscopy images. Conf Proc IEEE Eng Med Biol Soc 2017:1994–1997
Choi JY, Yoo TK, Seo JG, Kwak J, Um TT, Rim TH (2017) Multi-categorical deep learning neural network to classify retinal images: a pilot study employing small database. PLoS One 12:e0187336
Lausten GS, Jensen PK, Schiodt T, Lund B (1996) Local recurrences in giant cell tumour of bone. Long-term follow up of 31 cases. Int Orthop 20:172–176
Kivioja AH, Blomqvist C, Hietaniemi K et al (2008) Cement is recommended in intralesional surgery of giant cell tumors: a Scandinavian Sarcoma Group study of 294 patients followed for a median time of 5 years. Acta Orthop 79:86–93
Chakarun CJ, Forrester DM, Gottsegen CJ, Patel DB, White EA, Matcuk GR Jr (2013) Giant cell tumor of bone: review, mimics, and new developments in treatment. Radiographics 33:197–211
James G, Witten D, Hastie T, Tibshirani R (2013) An introduction to statistical learning: with applications in R. Springer, Berlin
Esteva A, Kuprel B, Novoa RA et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115–118
Hohmann E, Wetzler MJ, D’Agostino RB Jr (2017) Research pearls: the significance of statistics and perils of pooling. Part 2: predictive modeling. Arthroscopy 33:1423–1432
DeLong ER, DeLong DM, Clarke-Pearson DL (1988) Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44:837–845
Richardson ML (2016) The Zombie plot: a simple graphic method for visualizing the efficacy of a diagnostic test. AJR Am J Roentgenol 207:W43–W52
Li D, Zhang J, Li Y et al (2016) Surgery methods and soft tissue extension are the potential risk factors of local recurrence in giant cell tumor of bone. World J Surg Oncol 14:114
Acknowledgements
We thank Dr. Xiaochun Yuan, Dr. Qin Chen, Dr. Gehong Yao, and Dr. Lin Zhang in Shanghai General Hospital for analyzing the images.
Funding
This study was sponsored by the National Natural Science Foundation of China (project no. 81471662), Ministry of Science and Technology of China (2016YFE0103000), Science and Technology Commission of Shanghai Municipality (16411968500 and 16410722300), Shanghai Jiao Tong University (ZH2018ZDB10), Shanghai Jiao Tong University School of Medicine - Gaofeng Clinical Medicine Grant Support (20181814), and Clinical Research Innovation Plan of Shanghai General Hospital (CTCCR-2018B04).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Guarantor
The scientific guarantor of this publication is Prof. Xueqian Xie, MD PhD, Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine
Conflict of interest
The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.
Statistics and biometry
One of the co-authors, Dr. Jiapan Guo, PhD, from the University Medical Center Groningen, The Netherlands, kindly provided statistical advice and IT supports for this manuscript.
Informed consent
Written informed consent was waived by the Institutional Review Board.
Ethical approval
Institutional Review Board approval was obtained.
Methodology
• Retrospective
• Diagnostic study
• Performed at one institution
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
He, Y., Guo, J., Ding, X. et al. Convolutional neural network to predict the local recurrence of giant cell tumor of bone after curettage based on pre-surgery magnetic resonance images. Eur Radiol 29, 5441–5451 (2019). https://doi.org/10.1007/s00330-019-06082-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00330-019-06082-2