1
|
Wang HC, Wang SC, Yan JL, Ko LW. Artificial Intelligence Model Trained with Sparse Data to Detect Facial and Cranial Bone Fractures from Head CT. J Digit Imaging 2023; 36:1408-1418. [PMID: 37095310 PMCID: PMC10407005 DOI: 10.1007/s10278-023-00829-6] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/24/2022] [Revised: 02/15/2023] [Accepted: 03/31/2023] [Indexed: 04/26/2023] Open
Abstract
The presence of cranial and facial bone fractures is an important finding on non-enhanced head computed tomography (CT) scans from patients who have sustained head trauma. Some prior studies have proposed automatic cranial fracture detections, but studies on facial fractures are lacking. We propose a deep learning system to automatically detect both cranial and facial bone fractures. Our system incorporated models consisting of YOLOv4 for one-stage fracture detection and improved ResUNet (ResUNet++) for the segmentation of cranial and facial bones. The results from the two models mapped together provided the location of the fracture and the name of the fractured bone as the final output. The training data for the detection model were the soft tissue algorithm images from a total of 1,447 head CT studies (a total of 16,985 images), and the training data for the segmentation model included 1,538 selected head CT images. The trained models were tested on a test dataset consisting of 192 head CT studies (a total of 5,890 images). The overall performance achieved a sensitivity of 88.66%, a precision of 94.51%, and an F1 score of 0.9149. Specifically, the cranial and facial regions were evaluated and resulted in a sensitivity of 84.78% and 80.77%, a precision of 92.86% and 87.50%, and F1 scores of 0.8864 and 0.8400, respectively. The average accuracy for the segmentation labels concerning all predicted fracture bounding boxes was 80.90%. Our deep learning system could accurately detect cranial and facial bone fractures and identify the fractured bone region simultaneously.
Collapse
Affiliation(s)
- Huan-Chih Wang
- Division of Neurosurgery, Department of Surgery, National Taiwan University Hospital, Chungshan Rd, No. 7, Taipei City 100, Taiwan
- Division of Neurosurgery, Department of Surgery, National Taiwan University Hospital Hsinchu Branch, Hsinchu, Taiwan
- Department of Biological Science & Technology, College of Biological Science and Technology, National Yang Ming Chiao Tung University, Hsinchu, Taiwan
| | - Shao-Chung Wang
- Department of Medical Imaging and Intervention, Gung Medical Foundation, New Taipei Municipal Tucheng Hospital, Chang
, New Taipei City, Taiwan
| | - Jiun-Lin Yan
- Department of Neurosurgery, Keelung Chang Gung Memorial Hospital, Keelung, Taiwan
- College of Medicine, Chang Gung University, Taoyuan, Taiwan
| | - Li-Wei Ko
- Department of Biological Science & Technology, College of Biological Science and Technology, National Yang Ming Chiao Tung University, Hsinchu, Taiwan
- Present Address: Institute of Electrical and Control Engineering, College of Electrical and Computer Engineering, National Yang Ming Chiao Tung University, Hsinchu, Taiwan
| |
Collapse
|
2
|
Sommer F, Brand M, Scheithauer MO, Hoffmann TK, Theodoraki MN, Weber R. [Diagnosis and Treatment in frontobasal fractures]. HNO 2023; 71:35-47. [PMID: 36525033 DOI: 10.1007/s00106-022-01256-9] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Accepted: 10/18/2022] [Indexed: 12/23/2022]
Abstract
Traumatic brain injury can result in frontobasal fractures (FBF). The goals of treatment for FBF are to eliminate primary morbidity and/or prevent secondary morbidity. Of particular importance in this regard is the proximity of important sensory organs for hearing, vision, smell, and taste, as well as their supplying nervous structures. Medical history, clinical findings, or CT scan are necessary and should lead to an individual evaluation. Depending on the severity of the fractures, the following disciplines may be involved in the treatment of FBF: neurosurgery, plastic surgery, oral and maxillofacial surgery, and/or otorhinolaryngology. Particularly less invasive endoscopic endonasal therapy is a specialty of otorhinolaryngologic surgeons and has not been widely established in other disciplines. The present work provides an overview of the current state of the art in terms of the following aspects, taking into account the current literature: anatomic principles, classification of fractures, diagnostics (in particular clinical examination, imaging, and laboratory chemistry tests), clinical symptoms, and treatment.
Collapse
Affiliation(s)
- F Sommer
- Universitätsklinik für Hals‑, Nasen‑, Ohrenheilkunde, Kopf- und Hals-Chirurgie, Universität Ulm, Frauensteige 12, 89075, Ulm, Deutschland.
| | - M Brand
- Universitätsklinik für Hals‑, Nasen‑, Ohrenheilkunde, Kopf- und Hals-Chirurgie, Universität Ulm, Frauensteige 12, 89075, Ulm, Deutschland
| | - M O Scheithauer
- Universitätsklinik für Hals‑, Nasen‑, Ohrenheilkunde, Kopf- und Hals-Chirurgie, Universität Ulm, Frauensteige 12, 89075, Ulm, Deutschland
| | - T K Hoffmann
- Universitätsklinik für Hals‑, Nasen‑, Ohrenheilkunde, Kopf- und Hals-Chirurgie, Universität Ulm, Frauensteige 12, 89075, Ulm, Deutschland
| | - M-N Theodoraki
- Universitätsklinik für Hals‑, Nasen‑, Ohrenheilkunde, Kopf- und Hals-Chirurgie, Universität Ulm, Frauensteige 12, 89075, Ulm, Deutschland
| | - R Weber
- Hals-Nasen-Ohrenklinik des Städtischen Klinikums Karlsruhe, Karlsruhe, Deutschland
| |
Collapse
|
3
|
Facial fractures: classification and highlights for a useful report. Insights Imaging 2020; 11:49. [PMID: 32193796 PMCID: PMC7082488 DOI: 10.1186/s13244-020-00847-w] [Citation(s) in RCA: 24] [Impact Index Per Article: 6.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/02/2019] [Accepted: 02/06/2020] [Indexed: 11/16/2022] Open
Abstract
In patients with facial trauma, multidetector computed tomography is the first-choice imaging test because it can detect and characterize even small fractures and their associated complications quickly and accurately. It has helped clinical management and surgical planning, so radiologists must communicate their findings to surgeons effectively. In Le Fort fractures, there is a breach between the pterygoid plates and the posterior maxilla. These fractures are classified in three basic patterns that can be combined and associated with various complications. Conceptualized when low-speed trauma was predominant, the Le Fort classification system has become less relevant giving more importance on maxillary occlusion-bearing segments. The classification of naso-orbito-ethmoid depends on the extent of injury to the attachment of the medial canthal tendon, with possible complications like nasofrontal duct disruption. Displaced fractures of the zygomaticomaxillary complex often widen the angle of the lateral orbital wall, resulting in increased orbital volume and sometimes in enophthalmos. Severe comminution or angulation can lead to wide surgical exposure. In orbital fractures, entrapment of the inferior rectus muscles can lead to diplopia, so it is important to assess its positioning and morphology. Orbital fractures can also result in injuries to the globe or infraorbital nerve. Frontal sinus fractures that extend through the posterior sinus wall can create a communication with the anterior cranial fossa resulting in leakage of cerebrospinal fluid, intracranial bleeding. It is essential to categorize fracture patterns and highlight features that may affect fracture management in radiology reports of facial trauma.
Collapse
|
4
|
Sivanandapanicker J, Nagar M, Kutty R, Sunilkumar BS, Peethambaran A, Rajmohan BP, Asher P, Shinihas VP, Mohandas K, Jain S, Sharma S. Analysis and Clinical Importance of Skull Base Fractures in Adult Patients with Traumatic Brain Injury. J Neurosci Rural Pract 2019; 9:370-375. [PMID: 30069094 PMCID: PMC6050782 DOI: 10.4103/jnrp.jnrp_38_18] [Citation(s) in RCA: 4] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/05/2022] Open
Abstract
Aims and Objectives: Basal skull fracture (BSF) is rare in head injury (HI) patients and occasionally goes unnoticed which may lead to cerebrospinal fluid (CSF) fistula. With changing trends in HI, there is a need to reassess incidence and pattern of BSF pattern, CSF leak, meningitis, and management protocol, especially in this part of the world where detailed literature is lacking. Subjects and Methods: We closely followed adult patients admitted with BSF in our institute between January 2013 and December 2014. Associated clinical features were recorded. In case of CSF leak, detailed CSF study was done and patients were managed accordingly. Patients with persistent CSF leak were managed surgically. Results: During the study period, 194 of 5041 HI patients had evidence of BSF (3.85%). BSF was most commonly associated with moderate-to-severe HI (73.19%). About 81.44% patients were male and 29.9% were <30 years. Most common cause was road traffic accident (84.54%). Isolated anterior cranial fossa (ACF) fracture was most common (50%). About 63.92% patients had raccoon eyes. Forty-three patients had CSF leak with CSF rhinorrhea being more common. Culture of only 5 patients suggested bacterial meningitis. CSF leak lasted for more than 10 days in 8 patients, of which 4 patients required surgical repair. Conclusions: BSF is rare in victims of HI. It is more common in young adult males. ACF fractures are most common in our setup. Clinical signs of BSF are supportive but not definitive; high resolution computed tomography head is gold standard to detect BSF. CSF leak is rare and most of the cases can be managed conservatively. Leak persisting more than 7–10 days has high risk of developing meningitis and likely to need surgical intervention.
Collapse
Affiliation(s)
| | - Milesh Nagar
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - Raja Kutty
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - B S Sunilkumar
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - Anilkumar Peethambaran
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - B P Rajmohan
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - Prasanth Asher
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - V P Shinihas
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - K Mohandas
- Department of General Surgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - Sourabh Jain
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| | - Saurabh Sharma
- Department of Neurosurgery, Government Medical College, Thiruvananthapuram, Kerala, India
| |
Collapse
|
5
|
Hooper T, Eccles G, Milliken T, Mathieu-Burry JR, Reed W. Dose reduction in CT imaging for facial bone trauma in adults: A narrative literature review. J Med Radiat Sci 2019; 66:122-132. [PMID: 30706691 PMCID: PMC6545476 DOI: 10.1002/jmrs.319] [Citation(s) in RCA: 10] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/06/2017] [Revised: 12/15/2018] [Accepted: 12/18/2018] [Indexed: 01/08/2023] Open
Abstract
Trauma to the facial area accounts for a significant number of admissions to the emergency department. Diagnostic imaging is almost always required, and is critical in determining patient management. Multi‐detector computed tomography (MDCT) appears consistently in the literature as the gold‐standard imaging modality for facial bones, but results in a high radiation dose to the patient. This makes the application and advancement of dose reduction and dose optimisation methods vital. This narrative review presents a critical analysis of the literature concerning diagnostic imaging of facial bone trauma, with an emphasis on dose reduction methods for MDCT. Databases including Pubmed, Medline, Web of Science and Scopus were used to investigate this topic, with the key words: facial bone trauma, computed tomography (CT) imaging and dose reduction. Exclusion criteria included studies on nasal bone fracturing, dental imaging, elective surgeries and paediatric imaging. The literature shows overwhelming support for MDCT, given its accuracy, efficiency and ease of operation. Noise reducing reconstruction algorithms show promise as a successful method of dose reduction in facial bone imaging. Investigations of more innovative techniques also appear within the literature, including diagnostic cone‐beam CT (CBCT), intraoperative CBCT and dual‐source CT (DSCT), but further research is required to confirm their clinical value.
Collapse
Affiliation(s)
- Tayla Hooper
- Discipline of Medical Radiation Sciences, the University of Sydney, Lidcombe, Australia
| | - Grace Eccles
- Discipline of Medical Radiation Sciences, the University of Sydney, Lidcombe, Australia
| | - Talia Milliken
- Discipline of Medical Radiation Sciences, the University of Sydney, Lidcombe, Australia
| | | | - Warren Reed
- Discipline of Medical Radiation Sciences, the University of Sydney, Lidcombe, Australia
| |
Collapse
|
6
|
The effect of direction of force to the craniofacial skeleton on the severity of brain injury in patients with a fronto-basal fracture. Int J Oral Maxillofac Surg 2016; 45:872-7. [DOI: 10.1016/j.ijom.2016.01.016] [Citation(s) in RCA: 5] [Impact Index Per Article: 0.6] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/02/2015] [Revised: 12/17/2015] [Accepted: 01/26/2016] [Indexed: 11/24/2022]
|