Multi-Class Deep Learning Model for Detecting Pediatric Distal Forearm Fractures Based on the AO/OTA Classification.

Common pediatric distal forearm fractures necessitate precise detection. To support prompt treatment planning by clinicians, our study aimed to create a multi-class convolutional neural network (CNN) model for pediatric distal forearm fractures, guided by the AO Foundation/Orthopaedic Trauma Association (AO/ATO) classification system for pediatric fractures. The GRAZPEDWRI-DX dataset (2008-2018) of wrist X-ray images was used. We labeled images into four fracture classes (FRM, FUM, FRE, and FUE with F, fracture; R, radius; U, ulna; M, metaphysis; and E, epiphysis) based on the pediatric AO/ATO classification. We performed multi-class classification by training a YOLOv4-based CNN object detection model with 7006 images from 1809 patients (80% for training and 20% for validation). An 88-image test set from 34 patients was used to evaluate the model performance, which was then compared to the diagnosis performances of two readers-an orthopedist and a radiologist. The overall mean average precision levels on the validation set in four classes of the model were 0.97, 0.92, 0.95, and 0.94, respectively. On the test set, the model's performance included sensitivities of 0.86, 0.71, 0.88, and 0.89; specificities of 0.88, 0.94, 0.97, and 0.98; and area under the curve (AUC) values of 0.87, 0.83, 0.93, and 0.94, respectively. The best performance among the three readers belonged to the radiologist, with a mean AUC of 0.922, followed by our model (0.892) and the orthopedist (0.830). Therefore, using the AO/OTA concept, our multi-class fracture detection model excelled in identifying pediatric distal forearm fractures.

Two-stage revision for treatment of tuberculous prosthetic hip infection: an outcome analysis.

OBJECTIVES: Prosthetic joint infections (PJI) and especially tuberculosis (TB) PJI are rare diseases and hard to cure. The effectiveness of treatments for tuberculous PJI still remains a problem. The objective of this research was to indicate the success of two-stage revision replacement and also giving the associated criteria. METHODS: From 2015 to 2020, five patients with tuberculous PJI were treated with two-stage revision at Cho Ray hospital, Vietnam. We collected the dataset which included demographic data, the interval from the time of joint replacement to reported infection, records of tuberculous PJI, administration of anti-TB medications (duration, months), history of operation(s), duration of follow-up, and specific type(s) of antibiotics loaded in bone cement. The approval for this study was made by the institutional review board from Cho Ray Hospital, Vietnam. We conducted a literature review based on the keywords "PJI" and "TB" on PubMed. RESULTS: Five patients [median age 66 years (range 35-84)] had found tuberculous PJI. The median time from arthroplasty to diagnosis was 19 months (range 4-48). The diagnosis was confirmed by joint aspirates or synovial tissue. Positive PCR was also reported in all cases. The average duration of anti-tuberculosis polytherapy administration was 14.4 months. The operative techniques on five patients included debridement and using spacer loaded with 2 g streptomycin (and 2 g vancomycin if they got a coinfection) for 1 pack of bone cement, and revision arthroplasty. In most cases, the outcome of treatment using two-stage revision replacement was 80%. Overall, the auxiliary bacterial infections were recognized in three patients with tuberculous PJI and Staphylococcus aureus. Streptomycin and vancomycin were loaded in a cement spacer to increase the success rate, and tuberculous PJI was controlled for all patients. CONCLUSION: Tuberculous PJI can be controlled with two-stage revision replacement with an antibiotic-loaded cement spacer that is molded intraoperatively with custom mold and prolonged anti-tuberculosis treatment in all cases. LEVEL OF EVIDENCE: IV.

Linear and nonlinear transfer functions of single mode fiber for optical transmission systems.

The transmittance transfer function of single mode optical fibers operating in both linear and nonlinear regions is presented. For the linear domain, Fresnel sine and cosine integrals are obtained via the Fourier transform. In the nonlinear region dominated by self-phase-modulation effects, the Volterra series is essential to obtain the nonlinear transfer function. A convergence criterion for the Volterra series transfer function (VSTF) approach is described for solving the nonlinear Schrödinger wave propagation equation. Soliton transmission over single fibers is demonstrated as a case study of the application of the VSTF and a modified VSTF with a number of segmented steps whose distance is within the limit of the convergence of the VSTF.

Tunable photonic filters: a digital signal processing design approach.

Digital signal processing techniques are used for synthesizing tunable optical filters with variable bandwidth and centered reference frequency including the tunability of the low-pass, high-pass, bandpass, and bandstop optical filters. Potential applications of such filters are discussed, and the design techniques and properties of recursive digital filters are outlined. The basic filter structures, namely, the first-order all-pole optical filter (FOAPOF) and the first-order all-zero optical filter (FOAZOF), are described, and finally the design process of tunable optical filters and the designs of the second-order Butterworth low-pass, high-pass, bandpass, and bandstop tunable optical filters are presented. Indeed, we identify that the all-zero and all-pole networks are equivalent with well known principles of optics of interference and resonance, respectively. It is thus very straightforward to implement tunable optical filters, which is a unique feature.

New approach for the design of an optical square pulse generator.

What is believed to be a new approach for the design and analysis of a reconfigurable optical square pulse generator using the concept of temporal optical integration and the digital signal processing method is presented. The reconfigurable square pulse generator is synthesized using compact active semiconductor-based waveguide technology, and it consists simply of the cascade of a tunable microring resonator (or a tunable all-pole filter) and a tunable asymmetrical Mach-Zehnder interferometer (or a tunable all-zero filter). The reconfigurable generator can convert an input picosecond pulse (i.e., soliton or Gaussian pulse) into an optical square pulse. The pulse width of the generated square pulse can be adjusted by controlling the time delay of a variable delay element in the tunable all-zero filter. The reconfigurable generator can convert an input picosecond pulse train into return-to-zero (RZ) and non-return-to-zero (NRZ) signals with square pulse shapes. The repetition rates of the generated RZ and NRZ signals can be varied by adjusting the bit period of the input picosecond pulse train, the input pulse width, and the time delay of the variable delay element. The effect of the deviation of the parameter values on the generator performance is also studied.