Akito Yabu1, Masatoshi Hoshino2, Hitoshi Tabuchi3, Shinji Takahashi1, Hiroki Masumoto4, Masahiro Akada4, Shoji Morita5, Takafumi Maeno6, Masayoshi Iwamae6, Hiroyuki Inose7, Tsuyoshi Kato8, Toshitaka Yoshii7, Tadao Tsujio9, Hidetomi Terai1, Hiromitsu Toyoda1, Akinobu Suzuki1, Koji Tamai1, Shoichiro Ohyama1, Yusuke Hori1, Atsushi Okawa7, Hiroaki Nakamura1. 1. Department of Orthopaedic Surgery, Osaka City University Graduate School of Medicine, 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan. 2. Department of Orthopaedic Surgery, Osaka City University Graduate School of Medicine, 1-4-3 Asahi-machi, Abeno-ku, Osaka 545-8585, Japan. Electronic address: hoshino717@gmail.com. 3. Department of Ophthalmology, Tsukazaki Hospital, 68-1 Waku, Aboshi-ku, Himeji, Hyogo 671-1227, Japan; Department of Technology and Design Thinking for Medicine, Hiroshima University, 1-2-3 Kasumi, Minami-ku, Hiroshima City, Hiroshima 734-8551, Japan. 4. Department of Ophthalmology, Tsukazaki Hospital, 68-1 Waku, Aboshi-ku, Himeji, Hyogo 671-1227, Japan. 5. Graduate School of Engineering, University of Hyogo, 2167, Shosha, Himeji, Hyogo 671-2280, Japan. 6. Department of Orthopaedic Surgery, Ishikiriseiki Hospital, 18-28, Yayoi-machi, Higashiosaka, Osaka 579-8026, Japan. 7. Department of Orthopaedic Surgery, Tokyo Medical and Dental University, Graduate School, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8510, Japan. 8. Department of Orthopaedic Surgery, Ome municipal general Hospital, 4-16-5, Higashiome, Ome, Tokyo 198-0042, Japan. 9. Department of Orthopaedic Surgery, Shiraniwa Hospital, 6-10-1, Shiraniwadai, Ikoma, Nara 630-0136, Japan.
Abstract
BACKGROUND CONTEXT: Accurate diagnosis of osteoporotic vertebral fracture (OVF) is important for improving treatment outcomes; however, the gold standard has not been established yet. A deep-learning approach based on convolutional neural network (CNN) has attracted attention in the medical imaging field. PURPOSE: To construct a CNN to detect fresh OVF on magnetic resonance (MR) images. STUDY DESIGN/ SETTING: Retrospective analysis of MR images PATIENT SAMPLE: This retrospective study included 814 patients with fresh OVF. For CNN training and validation, 1624 slices of T1-weighted MR image were obtained and used. OUTCOME MEASURE: We plotted the receiver operating characteristic (ROC) curve and calculated the area under the curve (AUC) in order to evaluate the performance of the CNN. Consequently, the sensitivity, specificity, and accuracy of the diagnosis by CNN and that of the two spine surgeons were compared. METHODS: We constructed an optimal model using ensemble method by combining nine types of CNNs to detect fresh OVFs. Furthermore, two spine surgeons independently evaluated 100 vertebrae, which were randomly extracted from test data. RESULTS: The ensemble method using VGG16, VGG19, DenseNet201, and ResNet50 was the combination with the highest AUC of ROC curves. The AUC was 0.949. The evaluation metrics of the diagnosis (CNN/surgeon 1/surgeon 2) for 100 vertebrae were as follows: sensitivity: 88.1%/88.1%/100%; specificity: 87.9%/86.2%/65.5%; accuracy: 88.0%/87.0%/80.0%. CONCLUSIONS: In detecting fresh OVF using MR images, the performance of the CNN was comparable to that of two spine surgeons.
BACKGROUND CONTEXT: Accurate diagnosis of osteoporotic vertebral fracture (OVF) is important for improving treatment outcomes; however, the gold standard has not been established yet. A deep-learning approach based on convolutional neural network (CNN) has attracted attention in the medical imaging field. PURPOSE: To construct a CNN to detect fresh OVF on magnetic resonance (MR) images. STUDY DESIGN/ SETTING: Retrospective analysis of MR images PATIENT SAMPLE: This retrospective study included 814 patients with fresh OVF. For CNN training and validation, 1624 slices of T1-weighted MR image were obtained and used. OUTCOME MEASURE: We plotted the receiver operating characteristic (ROC) curve and calculated the area under the curve (AUC) in order to evaluate the performance of the CNN. Consequently, the sensitivity, specificity, and accuracy of the diagnosis by CNN and that of the two spine surgeons were compared. METHODS: We constructed an optimal model using ensemble method by combining nine types of CNNs to detect fresh OVFs. Furthermore, two spine surgeons independently evaluated 100 vertebrae, which were randomly extracted from test data. RESULTS: The ensemble method using VGG16, VGG19, DenseNet201, and ResNet50 was the combination with the highest AUC of ROC curves. The AUC was 0.949. The evaluation metrics of the diagnosis (CNN/surgeon 1/surgeon 2) for 100 vertebrae were as follows: sensitivity: 88.1%/88.1%/100%; specificity: 87.9%/86.2%/65.5%; accuracy: 88.0%/87.0%/80.0%. CONCLUSIONS: In detecting fresh OVF using MR images, the performance of the CNN was comparable to that of two spine surgeons.
Authors: Ahmed Benzakour; Pavlos Altsitzioglou; Jean Michel Lemée; Alaaeldin Ahmad; Andreas F Mavrogenis; Thami Benzakour Journal: Int Orthop Date: 2022-07-29 Impact factor: 3.479