What is the long-term clinical outcome after fragility fractures of the pelvis? - A CT-based cross-sectional study.

BACKGROUND: Recently, Rommens and Hoffman introduced a CT-based classification system for fragility fractures of the pelvis (FFP). Although fracture characteristics have been described, the relationship with clinical outcome is lacking. The purpose of this study was to get insight into the type of treatment and subsequent clinical outcome after all types of FFP. METHODS: A cross-sectional cohort study was performed including all elderly patients (≥ 65 years) with a CT-diagnosed FFP, between 2007-2019 in two level 1 trauma centers. Data regarding treatment, mortality and clinical outcome was gathered from the electronic patient files. Patients were asked to complete patient-reported outcome measures (PROMs) regarding physical functioning (SMFA) and quality of life (EQ-5D). Additionally, a standardized multidisciplinary treatment algorithm was constructed. RESULTS: A total of 187 patients were diagnosed with an FFP of whom 117 patients were available for follow-up analysis and 58 patients responded. FFP type I was most common (60%), followed by type II (27%), type III (8%) and type IV (5%). Almost all injuries were treated non-operatively (98%). Mobility at six weeks ranged from 50% (type III) to 80% type II). Mortality at 1 year was respectively 16% (type I and II), 47% (type III) and 13% (type IV). Physical functioning (SMFA function index) ranged from 62 (type III and IV) to 69 (type II) and was significantly decreased (P=<0.001) compared to the age-matched general population. Quality of life was also significantly decreased, ranging from 0.26 (type III) to 0.69 (type IV). CONCLUSIONS: FFP type I and II are most common. Treatment is mainly non-operative, resulting in good mobility after six weeks, especially for patients with FFP type I and II. Mortality rates at one year were substantial in all patients. Physical functioning and quality of life was about 20-30% decreased compared to the general population.

The accuracy of gap and step-off measurements in acetabular fracture treatment.

The assessment of gaps and steps in acetabular fractures is challenging. Data from various imaging techniques to enable accurate quantification of acetabular fracture displacement are limited. The aim of this study was to assess the accuracy of pelvic radiographs, intraoperative fluoroscopy, and computed tomography (CT) in detecting gaps and step-offs in acetabular fractures. Sixty patients, surgically treated for acetabular fractures, were included. Five observers (5400 measurements) measured the gaps and step-offs on radiographs and CT scans. Intraoperative fluoroscopy images were reassessed for the presence of gaps and/or step-offs. Preoperatively, 25% of the gaps and 40% of the step-offs were undetected on radiographs compared to CT. Postoperatively, 52% of the gaps and 80% of the step-offs were missed on radiographs compared to CT. Radiograph analysis led to a significantly smaller gap and step-off compared to the CT measurements, an underestimation by a factor of two. Approximately 70% of the residual gaps and step-offs was not detected using intraoperative fluoroscopy. Gaps and step-offs that exceed the critical cut-off indicating worse prognosis often remained undetected on radiographs compared to CT scans. Less-experienced observers tend to overestimate gaps and step-offs compared to the more-experienced observers. In acetabular fracture treatment, gaps and step-offs were often undetected and underestimated on radiographs and intraoperative fluoroscopy in comparison with CT scans. This means that CT is superior to radiographs in detecting acetabular fracture displacement, which is clinically relevant for patient counselling regarding treatment decisions and prognosis.

Machine learning based natural language processing of radiology reports in orthopaedic trauma.

OBJECTIVES: To compare different Machine Learning (ML) Natural Language Processing (NLP) methods to classify radiology reports in orthopaedic trauma for the presence of injuries. Assessing NLP performance is a prerequisite for downstream tasks and therefore of importance from a clinical perspective (avoiding missed injuries, quality check, insight in diagnostic yield) as well as from a research perspective (identification of patient cohorts, annotation of radiographs). METHODS: Datasets of Dutch radiology reports of injured extremities (n = 2469, 33% fractures) and chest radiographs (n = 799, 20% pneumothorax) were collected in two different hospitals and labeled by radiologists and trauma surgeons for the presence or absence of injuries. NLP classification was applied and optimized by testing different preprocessing steps and different classifiers (Rule-based, ML, and Bidirectional Encoder Representations from Transformers (BERT)). Performance was assessed by F1-score, AUC, sensitivity, specificity and accuracy. RESULTS: The deep learning based BERT model outperforms all other classification methods which were assessed. The model achieved an F1-score of (95 ± 2)% and accuracy of (96 ± 1)% on a dataset of simple reports (n= 2469), and an F1 of (83 ± 7)% with accuracy (93 ± 2)% on a dataset of complex reports (n= 799). CONCLUSION: BERT NLP outperforms traditional ML and rule-base classifiers when applied to Dutch radiology reports in orthopaedic trauma.

Persistent posterior seroma after laparoscopic repair of ventral abdominal wall hernias with expanded polytetrafluoroethylene mesh: prevalence, independent predictors and detached tacks : Retrospective review.

PURPOSE: A persistent seroma located posterior to a mesh (PPS) remains a little known complication after laparoscopic ventral hernia repair (LVHR). The aim of this large case series was to analyse the prevalence and clinical course as well as identify related factors and independent predictors of PPS. METHODS: All 1288 adult patients who underwent a LVHR with an expanded polytetrafluoroethylene mesh (ePTFE) between January 2003 and July 2014 were reviewed. Those who underwent an abdominal computed tomography (CT) scan more than 3 months afterwards (n = 166) were included and their scans were analysed. The primary outcome measure was the prevalence of a PPS and its characteristics. The secondary outcome measures were identification of significantly related factors and independent predictors of PPS. RESULTS: A PPS was observed in 14 of 166 analysed CT scans (8.4%). Eleven patients were symptomatic; conservative treatment (wait-and-see policy) was successful in eight. Three underwent relaparoscopy with removal of a thick neoperitoneum. Several instances of tack and/or mesh detachment were identified on CT scans and during relaparoscopy. Independent predictors were: > 3 trocars (RR 5.0, 95% CI 1.6-15.8) and use of a mesh larger than > 300 cm2 (RR 9.9, 95% CI 1.9-51.2). CONCLUSIONS: A PPS is a relatively common complication after LVHR with an ePTFE mesh of usually larger hernias. A "wait-and-see" approach seems justified in most cases. Some require laparoscopic excision of the thick neoperitoneum. A PPS can cause tack and mesh detachment but the clinical consequences are unclear. Recurrences have not been observed in this series.