Development and validation of a deep learning model for radiographic classification of pediatric femoral neck fractures.

Journal and index pages often block iframe embedding. This reader keeps the evidence details in Orthonotes and leaves the source page one click away.

Abstract

Conflict of interest statement: Declarations. Ethics approval and consent to participate: This study was approved by Ethics Committee of the Fuzhou Second General Hospital (ethical no. 2025228). Consent to participate was not applicable. Consent for publication: Not applicable. Competing interests: The work is original and has not been previously published in whole or submitted elsewhere for publication. All authors of the present paper qualify with COPE criteria for authorship. All authors have given their final approval of the version to be published. The authors declared that they have no competing interests. 3. Orthop Traumatol Surg Res. 2026 Jul 11:104789. doi: 10.1016/j.otsr.2026.104789. Online ahead of print. Development and validation of a multimodal deep learning model for the radiographic classification of pediatric femoral neck fractures. Wang W(1), Canavese F(2), Tang S(3), Zhao C(4), Liu H(5), Andreacchio A(6), Michelis MB(7), Arrigoni C(8), Catena N(8), Gregorio S(8), Marengo L(7), Chen S(9), Mei Q(10), Li H(1). Author information: (1)Department of Orthopedics, Shenzhen Traditional Chinese Medicine Hospital, ShenZhen, GuangDong, China; Fourth Clinical Medical College of Guangzhou University of Chinese Medicine, ShenZhen, GuangDong, China. (2)Orthopedic and Traumatology Department, IRCCS Istituto Giannina Gaslini, Via Gerolamo Gaslini 5, Genoa, Italy; DISC-Dipartimento di scienze chirurgiche e diagnostiche integrate, University of Genova, Viale Benedetto XV N°6, Genova, Italy. (3)School of Integrated Circuits (International School of Microelectronics), Dongguan University of Technology, DongGuan, GuangDong, China. (4)Department of Pediatric Orthopedics, Foshan Hospital of Traditional Chinese Medicine, FoShan, GuangDong, China. (5)Department of Pediatric Orthopedics, Children's Hospital of Chongqing Medical University, ChongQing, China. (6)Department of Pediatric Orthopedics, Vittore Buzzi Children's Hospital, Milan, 20154, Italy. (7)DISC-Dipartimento di scienze chirurgiche e diagnostiche integrate, University of Genova, Viale Benedetto XV N°6, Genova, Italy; School of Integrated Circuits (International School of Microelectronics), Dongguan University of Technology, DongGuan, GuangDong, China. (8)DISC-Dipartimento di scienze chirurgiche e diagnostiche integrate, University of Genova, Viale Benedetto XV N°6, Genova, Italy. (9)Department of Pediatric Orthopedics, Fuzhou Second General Hospital, FuZhou, FuJian, China. Electronic address: csy508@163.com. (10)Department of Pediatric Orthopedics, Shenzhen Children's Hospital, ShenZhen, GuangDong, China. BACKGROUND: An alternative classification system for categorizing pediatric femoral neck fractures (PFNFs) has been shown to have significant clinical values in guiding treatment. This system, which is based on the direction of translation and the presence of posteromedial column comminution, is an alternative to the classic Delbet-Colonna system. However, this system relies on the user's clinical experience. Due to the rarity of PFNFs, can a deep learning (DL) model trained based on limited samples accurately classify PFNFs types according to the novel classification thereby minimizing the dependence on experience? HYPOTHESIS: Despite the scarcity of PFNFs, the multimodal DL model trained on the radiograph-text paired dataset can still accurately categorize PFNFs types according to the novel classification and outperform the unimodal DL model solely trained on radiographic dataset as well as the experienced pediatric orthopedic surgeons (POS). PATIENTS AND METHODS: This multicenter study retrospectively reviewed 711 anteroposterior (AP) and 674 lateral X-rays from 711 patients (mean age: 10.7 ± 3.7 years) with PFNFs treated surgically. AP X-rays were divided into 2 types (576 PFNFs without and 135 with medial comminutions), and lateral X-rays were classified into 4 types (132 Type I, 329 Type II, 91 Type III PFNFs, and 122 PFNFs with comminuted posterior columns). X-rays-text paired datasets were developed to include each subtype and its corresponding textual description. The datasets were randomly divided into the training and the testing datasets in an approximately 4:1 ratio. The multimodal and unimodal DL models were trained based on the X-rays-text paired datasets and X-rays, respectively. Four junior-level, 4 mid-level and 4 senior-level POS were asked to assess the X-ray subtypes of the testing datasets. The rating performance of multimodal, unimodal DL models and POS with various experience were evaluated. RESULTS: When classifying the AP X-ray subtypes in the testing dataset, the multimodal DL model outperformed the unimodal DL model in terms of accuracy (96.5% versus 90.1 %), recall (92.4% versus 85.8%), positive predictive value (96.3% versus 84.0%), negative predictive value (96.3% versus 84.0%), F1 score (94.2% versus 84.8%), and the area under the curve (95.2% versus 91.6%). Similar results were found when comparing the performance of the two models in classifying the lateral X-ray subtypes of the testing dataset. According to the testing dataset, one-way analysis of variance results indicated that senior-level POS show significantly superior accuracy in classifying AP (p