Development and validation of an artificial intelligence model for the classification of hip fractures using the AO-OTA framework.

BACKGROUND AND PURPOSE: Artificial intelligence (AI) has the potential to aid in the accurate diagnosis of hip fractures and reduce the workload of clinicians. We primarily aimed to develop and validate a convolutional neural network (CNN) for the automated classification of hip fractures based on the 2018 AO-OTA classification system. The secondary aim was to incorporate the model's assessment of additional radiographic findings that often accompany such injuries. METHODS: 6,361 plain radiographs of the hip taken between 2002 and 2016 at Danderyd University Hospital were used to train the CNN. A separate set of 343 radiographs representing 324 unique patients was used to test the performance of the network. Performance was evaluated using area under the curve (AUC), sensitivity, specificity, and Youden's index. RESULTS: The CNN demonstrated high performance in identifying and classifying hip fracture, with AUCs ranging from 0.76 to 0.99 for different fracture categories. The AUC for hip fractures ranged from 0.86 to 0.99, for distal femur fractures from 0.76 to 0.99, and for pelvic fractures from 0.91 to 0.94. For 29 of 39 fracture categories, the AUC was ≥ 0.95. CONCLUSION: We found that AI has the potential for accurate and automated classification of hip fractures based on the AO-OTA classification system. Further training and modification of the CNN may enable its use in clinical settings.

Automating classification of osteoarthritis according to Kellgren-Lawrence in the knee using deep learning in an unfiltered adult population.

BACKGROUND: Prevalence for knee osteoarthritis is rising in both Sweden and globally due to increased age and obesity in the population. This has subsequently led to an increasing demand for knee arthroplasties. Correct diagnosis and classification of a knee osteoarthritis (OA) are therefore of a great interest in following-up and planning for either conservative or operative management. Most orthopedic surgeons rely on standard weight bearing radiographs of the knee. Improving the reliability and reproducibility of these interpretations could thus be hugely beneficial. Recently, deep learning which is a form of artificial intelligence (AI), has been showing promising results in interpreting radiographic images. In this study, we aim to evaluate how well an AI can classify the severity of knee OA, using entire image series and not excluding common visual disturbances such as an implant, cast and non-degenerative pathologies. METHODS: We selected 6103 radiographic exams of the knee taken at Danderyd University Hospital between the years 2002-2016 and manually categorized them according to the Kellgren & Lawrence grading scale (KL). We then trained a convolutional neural network (CNN) of ResNet architecture using PyTorch. We evaluated the results against a test set of 300 exams that had been reviewed independently by two senior orthopedic surgeons who settled eventual interobserver disagreements through consensus sessions. RESULTS: The CNN yielded an overall AUC of more than 0.87 for all KL grades except KL grade 2, which yielded an AUC of 0.8 and a mean AUC of 0.92. When merging adjacent KL grades, all but one group showed near perfect results with AUC > 0.95 indicating excellent performance. CONCLUSION: We have found that we could teach a CNN to correctly diagnose and classify the severity of knee OA using the KL grading system without cleaning the input data from major visual disturbances such as implants and other pathologies.

Artificial intelligence for the classification of fractures around the knee in adults according to the 2018 AO/OTA classification system.

BACKGROUND: Fractures around the knee joint are inherently complex in terms of treatment; complication rates are high, and they are difficult to diagnose on a plain radiograph. An automated way of classifying radiographic images could improve diagnostic accuracy and would enable production of uniformly classified records of fractures to be used in researching treatment strategies for different fracture types. Recently deep learning, a form of artificial intelligence (AI), has shown promising results for interpreting radiographs. In this study, we aim to evaluate how well an AI can classify knee fractures according to the detailed 2018 AO-OTA fracture classification system. METHODS: We selected 6003 radiograph exams taken at Danderyd University Hospital between the years 2002-2016, and manually categorized them according to the AO/OTA classification system and by custom classifiers. We then trained a ResNet-based neural network on this data. We evaluated the performance against a test set of 600 exams. Two senior orthopedic surgeons had reviewed these exams independently where we settled exams with disagreement through a consensus session. RESULTS: We captured a total of 49 nested fracture classes. Weighted mean AUC was 0.87 for proximal tibia fractures, 0.89 for patella fractures and 0.89 for distal femur fractures. Almost ¾ of AUC estimates were above 0.8, out of which more than half reached an AUC of 0.9 or above indicating excellent performance. CONCLUSION: Our study shows that neural networks can be used not only for fracture identification but also for more detailed classification of fractures around the knee joint.

No effect of risedronate on femoral periprosthetic bone loss following total hip arthroplasty. A 4-year follow-up of 61 patients in a double-blind, randomized placebo-controlled trial.

BACKGROUND AND PURPOSE: We have previously shown that during the first 2 years after total hip arthroplasty (THA), periprosthetic bone resorption can be prevented by 6 months of risedronate therapy. This follow-up study investigated this effect at 4 years. PATIENTS AND METHODS: A single-center, double-blind, randomized placebo-controlled trial was carried out from 2006 to 2010 in 73 patients with osteoarthritis of the hip who were scheduled to undergo THA. The patients were randomly assigned to receive either 35 mg risedronate or placebo orally, once a week, for 6 months postoperatively. The primary outcome was the percentage change in bone mineral density (BMD) in Gruen zones 1 and 7 in the proximal part of the femur at follow-up. Secondary outcomes included migration of the femoral stem and clinical outcome scores. RESULTS: 61 of the 73 patients participated in this 4-year (3.9- to 4.1-year) follow-up study. BMD was similar in the risedronate group (n = 30) and the placebo group (n = 31). The mean difference was -1.8% in zone 1 and 0.5% in zone 7. Migration of the femoral stem, the clinical outcome, and the frequency of adverse events were similar in the 2 groups. INTERPRETATION: Although risedronate prevents periprosthetic bone loss postoperatively, a decrease in periprosthetic BMD accelerates when therapy is discontinued, and no effect is seen at 4 years. We do not recommend the use of risedronate following THA for osteoarthritis of the hip.

The etiological factors of recurrence after tracheal resection and reconstruction in post-intubation stenosis.

We assessed several factors which might be responsible for the recurrence of post-intubation airway stenosis in a large group of patients who underwent resection and reconstruction surgery by one surgical team. Four hundred and ninety-four patients underwent reconstruction of post-intubation airway stenosis during 1995-2006. The case group comprised patients who had developed recurrence, while controls had no recurrence. The diagnosis of the recurrence was made based on the presence of clinical signs or symptoms and bronchoscopic evaluation. The following variables were compared in both groups: age, sex, duration of intubation, reason for intubation, period of time between intubation and surgery, history of previous tracheotomy, previous therapeutic interventions, subglottic involvement, length of resection, presence of unusual tension at the site of anastomosis and anastomotic infection. Fifty-two patients (10.5%) developed recurrence. Lengthy resection, presence of tension at the site of anastomosis, anastomotic infection and subglottic involvement were significantly higher in the case group. Logistic regression model showed that the three main predictors are anastomotic infection (OR=3.44), subglottic involvement (OR=2.43), and presence of tension (OR=1.97), respectively. It is concluded that the surgeon can play an important role in avoiding recurrence by decreasing tension, preventing infection, and preserving subglottic structure.