AI for detection, classification and prediction of loss of alignment of distal radius fractures; a systematic review.

PURPOSE: Early and accurate assessment of distal radius fractures (DRFs) is crucial for optimal prognosis. Identifying fractures likely to lose threshold alignment (instability) in a cast is vital for treatment decisions, yet prediction tools' accuracy and reliability remain challenging. Artificial intelligence (AI), particularly Convolutional Neural Networks (CNNs), can evaluate radiographic images with high performance. This systematic review aims to summarize studies utilizing CNNs to detect, classify, or predict loss of threshold alignment of DRFs. METHODS: A literature search was performed according to the PRISMA. Studies were eligible when the use of AI for the detection, classification, or prediction of loss of threshold alignment was analyzed. Quality assessment was done with a modified version of the methodologic index for non-randomized studies (MINORS). RESULTS: Of the 576 identified studies, 15 were included. On fracture detection, studies reported sensitivity and specificity ranging from 80 to 99% and 73-100%, respectively; the AUC ranged from 0.87 to 0.99; the accuracy varied from 82 to 99%. The accuracy of fracture classification ranged from 60 to 81% and the AUC from 0.59 to 0.84. No studies focused on predicting loss of thresholds alignement of DRFs. CONCLUSION: AI models for DRF detection show promising performance, indicating the potential of algorithms to assist clinicians in the assessment of radiographs. In addition, AI models showed similar performance compared to clinicians. No algorithms for predicting the loss of threshold alignment were identified in our literature search despite the clinical relevance of such algorithms.

Traditional radiography versus computed tomography to assess reduced distal radius fractures.

INTRODUCTION: This study compares computed tomography (CT) with plain radiography in its ability to assess distal radius fracture (DRF) malalignment after closed reduction and cast immobilization. METHODS: Malalignment is defined as radiographic fracture alignment beyond threshold values according to the Dutch guideline encompassing angulation, inclination, positive ulnar variance and intra-articular step-off or gap. After identifying 96 patients with correct alignment on initial post-reduction radiographs, we re-assessed alignment on post-reduction CT scans. RESULTS: Significant discrepancies were found between radiographs and CT scans in all measurement parameters. Notably, intra-articular step-off and gap variations on CT scans led to the reclassification of the majority of cases from correct alignment to malalignment. CT scans showed malalignment in 53% of cases, of which 73% underwent surgery. CONCLUSION: When there is doubt about post-reduction alignment based on radiograph imaging, additional CT scanning often reveals malalignment, primarily due to intra-articular incongruency.

Automatic identification of radius and ulna bone landmarks on 3D virtual models.

BACKGROUND: For bone morphology and biomechanics analysis, landmarks are essential to define position, orientation, and shape. These landmarks define bone and joint coordinate systems and are widely used in these research fields. Currently, no method is known for automatically identifying landmarks on virtual 3D bone models of the radius and ulna. This paper proposes a knowledge-based method for locating landmarks and calculating a coordinate system for the radius, ulna, and combined forearm bones, which is essential for measuring forearm function. This method does not rely on pre-labeled data. VALIDATION: The algorithm is validated by comparing the landmarks placed by the algorithm with the mean position of landmarks placed by a group of experts on cadaveric specimens regarding distance and orientation. RESULTS: The median Euclidean distance differences between all the automated and reference landmarks range from 0.4 to 1.8 millimeters. The median angular differences of the coordinate system of the radius and ulna range from -1.4 to 0.6 degrees. The forearm coordinate system's median errors range from -0.2 to 2.0 degrees. The median error in calculating the rotational position of the radius relative to the ulna is 1.8 degrees. CONCLUSION: The automatic method's applicability depends on the use context and desired accuracy. However, the current method is a validated first step in the automatic analysis of the three-dimensional forearm anatomy.

Redisplacement of reduced distal radius fractures in adults: does the type of casting play a role? The CAST study, a multicentre cluster randomized controlled trial.

Aims: It is not clear which type of casting provides the best initial treatment in adults with a distal radial fracture. Given that between 32% and 64% of adequately reduced fractures redisplace during immobilization in a cast, preventing redisplacement and a disabling malunion or secondary surgery is an aim of treatment. In this study, we investigated whether circumferential casting leads to fewer the redisplacement of fewer fractures and better one-year outcomes compared with plaster splinting. Methods: In a pragmatic, open-label, multicentre, two-period cluster-randomized superiority trial, we compared these two types of casting. Recruitment took place in ten hospitals. Eligible patients aged ≥ 18 years with a displaced distal radial fracture, which was acceptably aligned after closed reduction, were included. The primary outcome measure was the rate of redisplacement within five weeks of immobilization. Secondary outcomes were the rate of complaints relating to the cast, clinical outcomes at three months, patient-reported outcome measures (PROMs) (using the numerical rating scale (NRS), the abbreviated version of the Disabilities of the Arm, Shoulder and Hand (QuickDASH), and Patient-Rated Wrist/Hand Evaluation (PRWHE) scores), and adverse events such as the development of compartment syndrome during one year of follow-up. We used multivariable mixed-effects logistic regression for the analysis of the primary outcome measure. Results: The study included 420 patients. There was no significant difference between the rate of redisplacement of the fracture between the groups: 47% (n = 88) for those treated with a plaster splint and 49% (n = 90) for those treated with a circumferential cast (odds ratio 1.05 (95% confidence interval (CI) 0.65 to 1.70); p = 0.854). Patients treated in a plaster splint reported significantly more pain than those treated with a circumferential cast, during the first week of treatment (estimated mean NRS 4.7 (95% CI 4.3 to 5.1) vs 4.1 (95% CI 3.7 to 4.4); p = 0.014). The rate of complaints relating to the cast, clinical outcomes and PROMs did not differ significantly between the groups (p > 0.05). Compartment syndrome did not occur. Conclusion: Circumferential casting did not result in a significantly different rate of redisplacement of the fracture compared with the use of a plaster splint. There were comparable outcomes in both groups.

Normative ultrasound values for Achilles tendon thickness in the general population and patients with Achilles tendinopathy: A large international cross-sectional study.

The objective of the study was to obtain adjusted ultrasonographic reference values of the Achilles tendon thickness (maximum anterior-posterior distance) in adults without (previous) Achilles tendinopathy (AT) and to compare these reference values with AT patients. Six hundred participants were consecutively included, comprising 500 asymptomatic individuals and 100 patients with clinically diagnosed chronic AT. The maximum tendon thickness was assessed using Ultrasound Tissue Characterization. A multiple quantile regression model was developed, incorporating covariates (personal characteristics) that were found to have a significant impact on the maximum anterior-posterior distance of the Achilles tendon. A 95% reference interval (RI) was derived (50th, 2.5th-97.5th percentile). In asymptomatic participants median (95% RI) tendon thickness was 4.9 (3.8-6.9) mm for the midportion region and 3.7 (2.8-4.8) mm for the insertional region. Age, height, body mass index, and sex had a significant correlation with maximum tendon thickness. Median tendon thickness for the midportion region was calculated with the normative equation -2.1 + AGE × 0.021 + HEIGHT × 0.032+ BMI × 0.028 + SEX × 0.05. For the insertional region, the normative equation was -0.34 + AGE × 0.010+ HEIGHT × 0.018 + BMI × 0.022 + SEX × -0.05. In the equations, SEX is defined as 0 for males and 1 for females. Mean (95% CI) difference in tendon thickness compared to AT patients was 2.7 mm (2.3-3.2, p < 0.001) for the midportion and 1.4 mm (1.1-1.7, p < 0.001) for the insertional region. Compared to the asymptomatic population 73/100 (73%) AT patients exhibited increased tendon thickening, with values exceeding the 95% RI. This study presents novel reference values for the thickness of midportion and insertional region of the Achilles tendon, which were adjusted for personal characteristics. Our novel web-based openly accessible calculator for determining normative Achilles tendon thickness (www.achillestendontool.com) will be a useful resource in the diagnostic process. Trial registration number: This trial is registered in the Netherlands Trial Register (NL9010).

Which radial head fractures are best treated surgically?

Despite the common occurrence of radial head fractures, there is still a lack of consensus on which radial head fractures should be treated surgically. The radial head is an important secondary stabilizer in almost all directions. An insufficient radial head can lead to increased instability in varus-valgus and posterolateral rotatory directions, especially in a ligament-deficient elbow. The decision to perform surgery is often not dictated by the fracture pattern alone but also by the presence of associated injury. Comminution of the radial head and complete loss of cortical contact of at least one fracture fragment are associated with a high occurrence of associated injuries. Nondisplaced and minimally displaced radial head fractures can be treated non-operatively with early mobilization. Displacement (>2 mm) of fragments in radial head fractures without a mechanical block to pronation/supination is not a clear indication for surgery. Mechanical block to pronation/supination and comminution of the fracture are indications for surgery. The following paper reviews the current literature and provides state-of-the-art guidance on which radial head fractures should be treated surgically.

Does Early Conversion to Below-elbow Casting for Pediatric Diaphyseal Both-bone Forearm Fractures Adversely Affect Patient-reported Outcomes and ROM?

BACKGROUND: For distal forearm fractures in children, it has been shown that a below-elbow cast is an adequate treatment that overcomes the discomfort of an above-elbow cast and unnecessary immobilization of the elbow. For reduced diaphyseal both-bone forearm fractures, our previous randomized controlled trial (RCT)-which compared an above-elbow cast with early conversion to a below-elbow cast-revealed no differences in the risk of redisplacement or functional outcomes at short-term follow-up. Although studies with a longer follow-up after diaphyseal both-bone forearm fractures in children are scarce, they are essential, as growth might affect the outcome. QUESTIONS/PURPOSES: In this secondary analysis of an earlier RCT, we asked: (1) Does early conversion from an above-elbow to a below-elbow cast in children with reduced, stable diaphyseal forearm fractures result in worse clinical and radiological outcome? (2) Does a malunion result in inferior clinical outcomes at 7.5 years of follow-up? METHODS: In this study, we evaluated children at a minimum of 5 years of follow-up who were included in a previous RCT. The median (range) duration of follow-up was 7.5 years (5.2 to 9.9). The patients for this RCT were included from the emergency departments of four different urban hospitals. Between January 2006 and August 2010, we treated 128 patients for reduced diaphyseal both-bone forearm fractures. All 128 patients were eligible; 24% (31) were excluded because they were lost before the minimum study follow-up or had incomplete datasets, leaving 76% (97) for secondary analysis. The loss in the follow-up group was comparable to the included population. Eligible patients were invited for secondary functional and radiographic assessment. The primary outcome was the difference in forearm rotation compared with the uninjured contralateral arm. Secondary outcomes were the ABILHAND-kids and QuickDASH questionnaire, loss of flexion and extension of the elbow and wrist compared with the contralateral forearm, JAMAR grip strength ratio, and radiological assessment of residual deformity. The study was not blinded regarding the children, parents, and clinicians. RESULTS: At 7.5-year follow-up, there were no differences in ABILHAND-kids questionnaire score (above-elbow cast: 41 ± 2.4 versus above/below-elbow cast: 41.7 ± 0.7, mean difference -0.7 [95% confidence interval (CI) -1.4 to 0.04]; p = 0.06), QuickDASH (above-elbow cast: 5.8 ± 9.6 versus 2.9 ± 6.0 for above-/below-elbow cast, mean difference 2.9 [95% CI -0.5 to 6.2]; p = 0.92), and grip strength (0.9 ± 0.2 for above-elbow cast versus 1 ± 0.2 for above/below-elbow cast, mean difference -0.04 [95% CI -1 to 0.03]; p = 0.24). Functional outcomes showed no difference (loss of forearm rotation: above-elbow cast 7.9 ± 17.7 versus 4.1 ± 6.9 for above-/below-elbow cast, mean difference 3.8 [95% CI -1.7 to 9.4]; p = 0.47; arc of motion: above-elbow cast 152° ± 21° versus 155° ± 11° for the above/below-elbow cast group, mean difference -2.5 [95% CI -9.3 to -4.4]; p = 0.17; loss of wrist flexion-extension: above-elbow cast group 1.0° ± 5.0° versus 0.6° ± 4.2° for above/below-elbow cast, mean difference 0.4° [95% CI -1.5° to 2.2°]; p = 0.69). The secondary follow-up showed improvement in forearm rotation in both groups compared with the rotation at 7 months. For radiographical analysis, the only difference was in AP ulna (above-elbow cast: 6° ± 3° versus above/below-elbow cast: 5° ± 2°, mean difference 1.8° [0.7° to 3°]; p = 0.003), although this is likely not clinically relevant. There were no differences in the other parameters. Thirteen patients with persistent malunion at 7-month follow-up showed no clinically relevant differences in functional outcomes at 7.5-year follow-up compared with children without malunion. The loss of forearm rotation was 5.5ׄ° ± 9.1° for the malunion group compared with 6.0° ± 13.9° in the no malunion group, with a mean difference of 0.4 (95% CI of -7.5 to 8.4; p = 0.92). CONCLUSION: In light of these results, we suggest that surgeons perform an early conversion to a below-elbow cast for reduced diaphyseal both-bone forearm fractures in children. This study shows that even in patients with secondary fracture displacement, remodeling occurred. And even in persistent malunion, these patients mostly showed good-to-excellent final results. Future studies, such as a meta-analysis or a large, prospective observational study, would help to establish the influence of skeletal age, sex, and the severity and direction of malunion angulation of both the radius and ulna on clinical result. Furthermore, a similar systematic review could prove beneficial in clarifying the acceptable angulation for pediatric lower extremity fractures. LEVEL OF EVIDENCE: Level I, therapeutic study.

Shedding Light on Pediatric Fractures: Bridging the Knowledge Gap.

After the great success of the printed edition of the Special Issue "Pediatric Fractures-Volume I", which was published in 2023 containing 24 high-quality papers [...].

Should we always perform scaphoid nonunion surgery in patients with minor preoperative symptoms?

The objective of this study was to assess the downsides of surgical treatment of scaphoid fracture nonunion in patients with minor preoperative symptoms. Patients were classified with minor symptoms based on the Patient-Rated Hand/Wrist Evaluation questionnaire. Of the 35 included patients, most patients encountered problems with patient-specific activities; 9% reported worse postoperative functional outcomes, 34% were not satisfied with the treatment and 9% were reoperated. The risk of a worse functional outcome after surgery with the need for further operations and the chance of developing wrist osteoarthritis, along with the possibility of poor patient satisfaction and ongoing daily functional impairment, should be considered during preoperative counselling.Level of evidence: III.

Corrective Osteotomy in a Patient With Congenital Absence of Pronation Based on Three-Dimensional Statistical Shape Modeling.

We present a new indication of a three-dimensional statistical shape model (SSM): a patient with bilateral impaired forearm rotation due to a congenital variance in bone shape. A corrective osteotomy was planned and performed to best match the SSM created by computed tomography (CT) scans of 18 peers. Postoperatively, pronation increased by 70°, and the patient was pain-free. A CT scan showed accurate correction of the deformity and union of all osteotomies. This technique offers opportunities for patients with bilateral nontraumatic osseous forearm pathology.

Persistent elbow dislocation.

Acute elbow dislocation is a common injury with an incidence in the general population estimated at around 5/100,000. Persistent (or static) elbow dislocation is a relatively rare problem but might occur due to inappropriate assessment or treatment of acute simple or complex elbow dislocations. Persistent elbow dislocation can be an invalidating and painful condition with a more ominous prognosis than an acute elbow dislocation with appropriate treatment. Surgical treatment of persistent elbow dislocation is a complex intervention that requires extended surgical exposure and arthrolysis in combination with circumferential ligamentous and osseous stabilization. Satisfactory results are described, but complication and reintervention rates are high. After-treatment with a dynamic external fixator is often necessary.

A comparison of 3-D CT and 2-D plain radiograph measurements of the wrist in extra-articular malunited fractures of the distal radius.

Two-dimensional (2-D) plain radiographs may be insufficient for the evaluation of distal radial malunion, as it is a three-dimensional (3-D) deformity. This study introduced a 3-D measuring method that outputs radial inclination, ulnar variance, palmar tilt and axial rotation. To this end, a standardized and clearly defined coordinate system was constructed that allowed 3-D measurements closely resembling the conventional 2-D method in 35 patients. Mean differences between 3-D and 2-D measurements in affected wrists were 1.8° for radial inclination, 0.8 mm for ulnar variance and 3.7° for palmar tilt. In addition, inter- and intra-observer reproducibility of all 3-D and 2-D measurements were good or excellent (intraclass correlation coefficient >0.75), with 3-D reproducibility always better than 2-D. Axial rotation was present in all patients with a mean of 7.9° (SD 6.9). Although the differences between 2-D and 3-D measurements were small, 3-D evaluation enables the assessment of axial rotation and brings us closer to a routine 3-D evaluation of malunion.Level of evidence: III.

Improved forearm rotation even after early conversion to below-elbow cast for non-reduced diaphyseal both-bones forearm fractures in children: a secondary 7.5-year follow up of a randomized trial.

BACKGROUND AND PURPOSE: previous RCT compared short-term results of above-elbow cast (AEC) with early conversion to below-elbow cast (BEC) in children with non-reduced diaphyseal both-bone forearm fractures. After 7 months both groups had comparable function. Our primary aim was to investigate whether forearm rotation improves or worsens over time. Secondary aims were loss of flexion and extension of the elbow and wrist, patient-reported outcomes measures, grip strength ratio, and radiographic assessment. PATIENTS AND METHODS: We performed long-term follow-up (FU) of a previous RCT. All patients were invited again for the long-term FU measurements. Primary outcome was limitation of forearm rotation. Secondary outcomes were loss of flexion and extension of the elbow and wrist compared with the contralateral forearm, the ABILHAND-Kids questionnaire and the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire, grip strength ratio, and radiographic assessment. RESULTS: The mean FU was 7.5 (4.4-9.6) years. Of the initial 47 children, 38 (81%) participated. Rotation improved in both groups over time, with no significant difference in the final forearm rotation: 8° (SD 22) for the AEC group and 8° (SD 15) for the BEC group with a mean difference of 0° (95% confidence interval -13 to 12). Secondary outcomes showed no statistically significant differences. Finally, children < 9 years almost all have full recovery of function. CONCLUSION: Long-term follow-up showed that loss of forearm rotation after a non-reduced diaphyseal both-bone forearm fracture improved significantly compared with that at 7 months, independent of the initial treatment and children aged < 9 will have almost full recovery of function. This substantiates that the remaining growth behaves like a "friend" at long-term follow-up.

Lateral Humeral Condyle Fractures in Pediatric Patients.

Lateral humeral condyle fractures are frequently seen in pediatric patients and have a high risk of unfavorable outcomes. A fall on the outstretched arm with supination of the forearm is the most common trauma mechanism. A physical examination combined with additional imaging will confirm the diagnosis. Several classifications have been described to categorize these fractures based on location and comminution. Treatment options depend on the severity of the fracture and consist of immobilization in a cast, closed reduction with percutaneous fixation, and open reduction with fixation. These fractures can lead to notable complications such as lateral condyle overgrowth, surgical site infection, pin tract infections, stiffness resulting in decreased range of motion, cubitus valgus deformities, 'fishtail' deformities, malunion, non-union, avascular necrosis, and premature epiphyseal fusion. Adequate follow-up is therefore warranted.

Impact of bone and cartilage segmentation from CT and MRI on both bone forearm osteotomy planning.

INTRODUCTION: The use of MRI scans for pre-operative surgical planning of forearm osteotomies provides additional information of joint cartilage and soft tissue structures and reduces radiation exposure in comparison with the use of CT scans. In this study, we investigated whether using 3D information obtained from MRI with and without cartilage information leads to a different outcome of pre-operative planning. METHODS: Bilateral CT and MRI scans of the forearms of 10 adolescent and young adult patients with a unilateral bone deformation were acquired in a prospective study. The bones were segmented from CT and MRI, and cartilage only from MRI. The deformed bones were virtually reconstructed, by registering the joint ends to the healthy contralateral side. An optimal osteotomy plane was determined that minimized the distance between the resulting fragments. This process was performed in threefold: using the CT and MRI bone segmentations, and the MRI cartilage segmentations. RESULTS: Comparison of bone segmentation from MRI and CT scan resulted in a 0.95 ± 0.02 Dice Similarity Coefficient and 0.42 ± 0.07 mm Mean Absolute Surface Distance. All realignment parameters showed excellent reliability across the different segmentations. However, the mean differences in translational realignment between CT and MRI bone segmentations (4.5 ± 2.1 mm) and between MRI bone and MRI bone and cartilage segmentations (2.8 ± 2.1 mm) were shown to be clinically and statistically significant. A significant positive correlation was found between the translational realignment and the relative amount of cartilage. CONCLUSION: This study indicates that although bone realignment remained largely similar when using MRI with and without cartilage information compared to using CT, the small differences in segmentation could induce statistically and clinically significant differences in the osteotomy planning. We also showed that endochondral cartilage might be a non-negligible factor when planning osteotomies for young patients.

Management of the stiff elbow: a literature review.

The elbow is prone to stiffness due to its unique anatomy and profound capsular reaction to inflammation. The resulting movement impairment may significantly interfere with a patient's activities of daily living. Trauma (including surgery for trauma), posttraumatic arthritis, and heterotopic ossification (HO) are the most common causes of elbow stiffness. In stiffness caused by soft tissue contractures, initial conservative treatment with physiotherapy (PT) and splinting is advised. In cases in which osseous deformities limit range of motion (e.g. malunion, osseous impingement, or HO), early surgical intervention is recommended. Open and arthroscopic arthrolysis are the primary surgical options. Arthroscopic arthrolysis has a lower complication and revision rate but has narrower indications. Early active mobilization using PT after surgery is recommended in postoperative rehabilitation and may be complemented by splinting or continuous passive motion therapy. Most results are gained within the first few months but can continue to improve until 12 months. This paper reviews the current literature and provides state-of-the-art guidance on the management regarding prevention, evaluation, and treatment of elbow stiffness.

Indications and Timing of Guided Growth Techniques for Pediatric Upper Extremity Deformities: A Literature Review.

Osseous deformities in children arise due to progressive angular growth or complete physeal arrest. Clinical and radiological alignment measurements help to provide an impression of the deformity, which can be corrected using guided growth techniques. However, little is known about timing and techniques for the upper extremity. Treatment options for deformity correction include monitoring of the deformity, (hemi-)epiphysiodesis, physeal bar resection, and correction osteotomy. Treatment is dependent on the extent and location of the deformity, physeal involvement, presence of a physeal bar, patient age, and predicted length inequality at skeletal maturity. An accurate estimation of the projected limb or bone length inequality is crucial for optimal timing of the intervention. The Paley multiplier method remains the most accurate and simple method for calculating limb growth. While the multiplier method is accurate for calculating growth prior to the growth spurt, measuring peak height velocity (PHV) is superior to chronological age after the onset of the growth spurt. PHV is closely related to skeletal age in children. The Sauvegrain method of skeletal age assessment using elbow radiographs is possibly a simpler and more reliable method than the method by Greulich and Pyle using hand radiographs. PHV-derived multipliers need to be developed for the Sauvegrain method for a more accurate calculation of limb growth during the growth spurt. This paper provides a review of the current literature on the clinical and radiological evaluation of normal upper extremity alignment and aims to provide state-of-the-art directions on deformity evaluation, treatment options, and optimal timing of these options during growth.

Morphological risk factors for scaphoid fracture: a case-control study.

PURPOSE: Most patients with a clinically suspected scaphoid fracture and normal initial radiograph are unnecessarily treated. Previously developed prediction rules using demographic and clinical risk are unable to accurately predict occult fractures. Adding other risk factors could enhance this. Therefore, we aim to explore if there are morphological risk factors of the wrist for sustaining a scaphoid fracture. METHODS: We retrospectively included adult patients with a clinically suspected scaphoid fracture between 2013 and 2019 in our case-control study. There were 82 patients with a scaphoid fracture and 158 patients with a wrist contusion. Morphological risk factors were identified using statistical shape modelling (SSM) and linear measurements. Independent wrist shape variations on posteroanterior and lateral radiographs were captured in modes using SSM. Associations between outcomes and a scaphoid fracture were explored using logistic regression and the reliability was assessed. RESULTS: Of the 15 posteroanterior modes and 8 lateral modes identified and linear measurements performed, 1 PA mode was associated with a scaphoid fracture (PA mode 4; OR 1.40, CI 1.04-1.93, p = 0.031). We described this mode as an ulna plus and narrower distal radius with more volar tilt and radial inclination. The reliability of the posteroanterior modes and linear measurements was mostly good/excellent and moderate/poor for the lateral modes. CONCLUSION: There was one complex wrist shape significantly associated with a scaphoid fracture. Since the association was weak and the shape is difficult to identify radiographs, we believe this morphological risk factor would not enhance identifying occult scaphoid fractures in the future.

Artificial intelligence fracture recognition on computed tomography: review of literature and recommendations.

PURPOSE: The use of computed tomography (CT) in fractures is time consuming, challenging and suffers from poor inter-surgeon reliability. Convolutional neural networks (CNNs), a subset of artificial intelligence (AI), may overcome shortcomings and reduce clinical burdens to detect and classify fractures. The aim of this review was to summarize literature on CNNs for the detection and classification of fractures on CT scans, focusing on its accuracy and to evaluate the beneficial role in daily practice. METHODS: Literature search was performed according to the PRISMA statement, and Embase, Medline ALL, Web of Science Core Collection, Cochrane Central Register of Controlled Trials and Google Scholar databases were searched. Studies were eligible when the use of AI for the detection of fractures on CT scans was described. Quality assessment was done with a modified version of the methodologic index for nonrandomized studies (MINORS), with a seven-item checklist. Performance of AI was defined as accuracy, F1-score and area under the curve (AUC). RESULTS: Of the 1140 identified studies, 17 were included. Accuracy ranged from 69 to 99%, the F1-score ranged from 0.35 to 0.94 and the AUC, ranging from 0.77 to 0.95. Based on ten studies, CNN showed a similar or improved diagnostic accuracy in addition to clinical evaluation only. CONCLUSIONS: CNNs are applicable for the detection and classification fractures on CT scans. This can improve automated and clinician-aided diagnostics. Further research should focus on the additional value of CNN used for CT scans in daily clinics.