臨床整形外科57巻10号

2022年10月発行

特集 整形外科領域における人工知能の応用 各論

深層学習を用いたMRIでの脊髄腫瘍自動位置検出システムの構築

著者： 伊藤定之¹ 中島宏彰¹ 町野正明¹ 世木直喜¹ 小田昌宏² 大内田隼¹ 森下和明¹ 森健策² 今釜史郎¹

所属機関： ¹名古屋大学大学院医学系研究科整形外科学 ²名古屋大学大学院情報学研究科知能システム学専攻

ページ範囲：P.1189 - P.1195

深層物体検出手法のYou Only Look Once version3(Yolov3)を用いて脊椎MRIから神経鞘種の自動位置検出を可能とするシステムを構築した．T1強調画像を用いた検討では122枚中94枚(77.0％)，T2強調画像では122枚中111枚(91.0％)で神経鞘腫の位置を同定可能で，T1，2を統合したシステムでは122枚中114枚(93.4％)であった．2人の脊椎外科医はそれぞれ122枚中110枚(90.2％)，122枚中109枚(89.3％)で神経鞘腫の位置を同定可能であった．今回構築したシステムは脊椎外科医と同等の精度であり，スクリーニング検査として期待が持てる精度であると考えられる．

参考文献

1) Dai LY, Ding WG, Wang XY, et al. Assessment of ligamentous injury in patients with thoracolumbar burst fractures using MRI. J Trauma 2009;66(6):1610-5.

2) Radcliff K, Su BW, Kepler CK, et al. Correlation of posterior ligamentous complex injury and neurological injury to loss of vertebral body height, kyphosis, and canal compromise. Spine (Phila Pa 1976) 2012;37(13):1142-50.

3) Lin HH, Chou PH, Wang ST, et al. Determination of the painful level in osteoporotic vertebral fractures--Retrospective comparison between plain film, bone scan, and magnetic resonance imaging. J Chin Med Assoc 2015;78(12):714-8.

4) Takashima H, Takebayashi T, Yoshimoto M, et al. Differentiating spinal intradural-extramedullary schwannoma from meningioma using MRI T2 weighted images. Br J Radiol 2018;91(1092):20180262. doi:10.1259/bjr.20180262.

5) Cowley P. Neuroimaging of spinal canal stenosis. Magn Reson Imaging Clin N Am 2016;24(3):523-39.

6) el-Mahdy W, Kane PJ, Powell MP, et al. Spinal intradural tumours:part I--extramedullary. Br J Neurosurg 1999;13(6):550-7.

7) Kane PJ, el-Mahdy W, Singh A, et al. Spinal intradural tumours:part II--intramedullary. Br J Neurosurg 1999;13(6):558-63.

8) Balériaux DL. Spinal cord tumors. Eur Radiol 1999;9(7):1252-8.

9) Kato M, Nakamura H, Terai H, et al. Why does delay exist in the diagnosis of intradural spinal cord tumor despite the availability of MRI? J Clin Neurosci 2008;15(8):880-5.

10) LeCun Y, Bengio Y, Hinton G. Deep learning. Nature 2015;521(7553):436-44.

11) Kraus OZ, Ba JL, Frey BJ. Classifying and segmenting microscopy images with deep multiple instance learning. Bioinformatics 2016;32(12):i52-i9.

12) Yang G, Yang J, Sheng W, et al. Convolutional neural network-based embarrassing situation detection under camera for social robot in smart homes. Sensors (Basel) 2018;18(5):1530.

13) Russakovsky O, Deng J, Su H, et al. ImageNet large scale visual recognition challenge. 2014. http://arxivorg/abs/1409.0575 2014(2019年5月10日アクセス)

14) McBee MP, Awan OA, Colucci AT, et al. Deep learning in radiology. Acad Radiol 2018;25(11):1472-80.

15) Redmon J, Farhadi A. YOLOv3:an incremental improvement. 2018. https://arxiv.org/abs/1804.02767(2022年6月20日アクセス)

16) Pang S, Ding T, Qiao S, et al. A novel YOLOv3-arch model for identifying cholelithiasis and classifying gallstones on CT images. PLoS One 2019;14(6):e0217647. doi:10.1371/journal.pone.0217647.

17) Hirano K, Imagama S, Sato K, et al. Primary spinal cord tumors:review of 678 surgically treated patients in Japan. A multicenter study. Eur Spine J 2012;21(10):2019-26.

18) Tzutalin D. LabelImg. Git code 2015. https://github.com/tzutalin/labelImg(2022年6月20日アクセス)

19) Maki S, Furuya T, Horikoshi T, et al. A deep convolutional neural network with performance comparable to radiologists for differentiating between spinal schwannoma and meningioma. Spine (Phila Pa 1976) 2019;45(10):694-700.

20) Dillon WP, Norman D, Newton TH, et al. Intradural spinal cord lesions:Gd-DTPA-enhanced MR imaging. Radiology 1989;170(1 Pt 1):229-37.

21) Liu WC, Choi G, Lee SH, et al. Radiological findings of spinal schwannomas and meningiomas:focus on discrimination of two disease entities. Eur Radiol 2009;19(11):2707-15.

22) Iwata E, Shigematsu H, Yamamoto Y, et al. Preliminary algorithm for differential diagnosis between spinal meningioma and schwannoma using plain magnetic resonance imaging. J Orthop Sci 2018;23(2):408-13.

23) Ando K, Imagama S, Ito Z, et al. How do spinal schwannomas progress? The natural progression of spinal schwannomas on MRI. J Neurosurg Spine 2016;24(1):155-9.

24) Kobayashi K, Imagama S, Ando K, et al. Contrast MRI findings for spinal schwannoma as predictors of tumor proliferation and motor status. Spine (Phila Pa 1976) 2017;42(3):E150-E5. doi:10.1097/BRS.0000000000001732.

25) Friedman DP, Tartaglino LM, Flanders AE. Intradural schwannomas of the spine:MR findings with emphasis on contrast-enhancement characteristics. AJR Am J Roentgenol 1992;158(6):1347-50.