Search

Significant advances in artificial intelligence (AI) over the past decade potentially may lead to dramatic effects on clinical practice. Digitized histology represents an area ripe for AI implementation. We describe several current needs within the world of gastrointestinal histopathology, and outline, using currently studied models, how AI potentially can address them. We also highlight pitfalls as AI makes inroads into clinical practice. [ABSTRACT FROM AUTHOR]

Total joint arthroplasty is becoming one of the most common surgeries within the United States, creating an abundance of analyzable data to improve patient experience and outcomes. Unfortunately, a large majority of this data is concealed in electronic health records only accessible by manual extraction, which takes extensive time and resources. Natural language processing (NLP), a field within artificial intelligence, may offer a viable alternative to manual extraction. Using NLP, a researcher can analyze written and spoken data and extract data in an organized manner suitable for future research and clinical use. This article will first discuss common subtasks involved in an NLP pipeline, including data preparation, modeling, analysis, and external validation, followed by examples of NLP projects. Challenges and limitations of NLP will be discussed, closing with future directions of NLP projects, including large language models. [ABSTRACT FROM AUTHOR]

The growth of artificial intelligence combined with the collection and storage of large amounts of data in the electronic medical record collection has created an opportunity for orthopedic research and translation into the clinical environment. Machine learning (ML) is a type of artificial intelligence tool well suited for processing the large amount of available data. Specific areas of ML frequently used by orthopedic surgeons performing total joint arthroplasty include tabular data analysis (spreadsheets), medical imaging processing, and natural language processing (extracting concepts from text). Previous studies have discussed models able to identify fractures in radiographs, identify implant type in radiographs, and determine the stage of osteoarthritis based on walking analysis. Despite the growing popularity of ML, there are limitations including its reliance on "good" data, potential for overfitting, long life cycle for creation, and ability to only perform one narrow task. This educational article will further discuss a general overview of ML, discussing these challenges and including examples of successfully published models. [Display omitted] [ABSTRACT FROM AUTHOR]

Image data has grown exponentially as systems have increased their ability to collect and store it. Unfortunately, there are limits to human resources both in time and knowledge to fully interpret and manage that data. Computer Vision (CV) has grown in popularity as a discipline for better understanding visual data. Computer Vision has become a powerful tool for imaging analytics in orthopedic surgery, allowing computers to evaluate large volumes of image data with greater nuance than previously possible. Nevertheless, even with the growing number of uses in medicine, literature on the fundamentals of CV and its implementation is mainly oriented toward computer scientists rather than clinicians, rendering CV unapproachable for most orthopedic surgeons as a tool for clinical practice and research. The purpose of this article is to summarize and review the fundamental concepts of CV application for the orthopedic surgeon and musculoskeletal researcher. [Display omitted] [ABSTRACT FROM AUTHOR]

Revision total hip arthroplasty (THA) requires preoperatively identifying in situ implants, a time-consuming and sometimes unachievable task. Although deep learning (DL) tools have been attempted to automate this process, existing approaches are limited by classifying few femoral and zero acetabular components, only classify on anterior-posterior (AP) radiographs, and do not report prediction uncertainty or flag outlier data. This study introduces Total Hip Arhtroplasty Automated Implant Detector (THA-AID), a DL tool trained on 241,419 radiographs that identifies common designs of 20 femoral and 8 acetabular components from AP, lateral, or oblique views and reports prediction uncertainty using conformal prediction and outlier detection using a custom framework. We evaluated THA-AID using internal, external, and out-of-domain test sets and compared its performance with human experts. THA-AID achieved internal test set accuracies of 98.9% for both femoral and acetabular components with no significant differences based on radiographic view. The femoral classifier also achieved 97.0% accuracy on the external test set. Adding conformal prediction increased true label prediction by 0.1% for acetabular and 0.7 to 0.9% for femoral components. More than 99% of out-of-domain and >89% of in-domain outlier data were correctly identified by THA-AID. The THA-AID is an automated tool for implant identification from radiographs with exceptional performance on internal and external test sets and no decrement in performance based on radiographic view. Importantly, this is the first study in orthopedics to our knowledge including uncertainty quantification and outlier detection of a DL model. [ABSTRACT FROM AUTHOR]

This study introduces THA-Net, a deep learning inpainting algorithm for simulating postoperative total hip arthroplasty (THA) radiographs from a single preoperative pelvis radiograph input, while being able to generate predictions either unconditionally (algorithm chooses implants) or conditionally (surgeon chooses implants). The THA-Net is a deep learning algorithm which receives an input preoperative radiograph and subsequently replaces the target hip joint with THA implants to generate a synthetic yet realistic postoperative radiograph. We trained THA-Net on 356,305 pairs of radiographs from 14,357 patients from a single institution's total joint registry and evaluated the validity (quality of surgical execution) and realism (ability to differentiate real and synthetic radiographs) of its outputs against both human-based and software-based criteria. The surgical validity of synthetic postoperative radiographs was significantly higher than their real counterparts (mean difference: 0.8 to 1.1 points on 10-point Likert scale, P <.001), but they were not able to be differentiated in terms of realism in blinded expert review. Synthetic images showed excellent validity and realism when analyzed with already validated deep learning models. We developed a THA next-generation templating tool that can generate synthetic radiographs graded higher on ultimate surgical execution than real radiographs from training data. Further refinement of this tool may potentiate patient-specific surgical planning and enable technologies such as robotics, navigation, and augmented reality (an online demo of THA-Net is available at: https://demo.osail.ai/tha_net). [ABSTRACT FROM AUTHOR]

Objective: To develop and validate a clinical classification system for urethral stricture disease (USD) based on the retrograde urethrogram (RUG), physical exam, and stricture-specific patient history.Materials and Methods: Three elements were chosen to be included in the classification system: 1) Length of urethral stricture (L); 2) Stricture segment/location (S); 3) Stricture Etiology (E) (LSE classification system). Each element was divided into clinically relevant sub-categories. A three-step development and validation process then ensued, culminating in an in-person Trauma and Urologic Reconstruction Network of Surgeons (TURNS) meeting, at which the final classification system was unanimously agreed upon by attendees based on interrater reliability data obtained from the classifying of 22 clinical vignettes. A final validation step involved retrospectively classifying cases in the TURNS database to determine if classification influenced surgical technique and was associated with presumed stricture etiology.Results: The final LSE classification system was found to have an interrater reliability of 0.79 (individual components 0.76, 0.70 and 0.93 respectfully). Retrospective classification of the 2162 TURNS strictures revealed the segment (S) to be strongly associated with urethroplasty type (p = 0.0005) and stricture etiology (E) (p = 0.0005).Conclusion: We developed and validated a novel, easy to use, urethral stricture classification system. The system's ability to aid in directing treatments, predict treatment outcomes, and facilitate collaborative research efforts will require further study. [ABSTRACT FROM AUTHOR]

Ultrasound is the most commonly used imaging modality in clinical practice because it is a nonionizing, low-cost, and portable point-of-care imaging tool that provides real-time images. Artificial intelligence (AI)-powered ultrasound is becoming more mature and getting closer to routine clinical applications in recent times because of an increased need for efficient and objective acquisition and evaluation of ultrasound images. Because ultrasound images involve operator-, patient-, and scanner-dependent variations, the adaptation of classical machine learning methods to clinical applications becomes challenging. With their self-learning ability, deep-learning (DL) methods are able to harness exponentially growing graphics processing unit computing power to identify abstract and complex imaging features. This has given rise to tremendous opportunities such as providing robust and generalizable AI models for improving image acquisition, real-time assessment of image quality, objective diagnosis and detection of diseases, and optimizing ultrasound clinical workflow. In this report, the authors review current DL approaches and research directions in rapidly advancing ultrasound technology and present their outlook on future directions and trends for DL techniques to further improve diagnosis, reduce health care cost, and optimize ultrasound clinical workflow. [ABSTRACT FROM AUTHOR]

In this work, we applied and validated an artificial intelligence technique known as generative adversarial networks (GANs) to create large volumes of high-fidelity synthetic anteroposterior (AP) pelvis radiographs that can enable deep learning (DL)-based image analyses, while ensuring patient privacy. AP pelvis radiographs with native hips were gathered from an institutional registry between 1998 and 2018. The data was used to train a model to create 512 × 512 pixel synthetic AP pelvis images. The network was trained on 25 million images produced through augmentation. A set of 100 random images (50/50 real/synthetic) was evaluated by 3 orthopaedic surgeons and 2 radiologists to discern real versus synthetic images. Two models (joint localization and segmentation) were trained using synthetic images and tested on real images. The final model was trained on 37,640 real radiographs (16,782 patients). In a computer assessment of image fidelity, the final model achieved an "excellent" rating. In a blinded review of paired images (1 real, 1 synthetic), orthopaedic surgeon reviewers were unable to correctly identify which image was synthetic (accuracy = 55%, Kappa = 0.11), highlighting synthetic image fidelity. The synthetic and real images showed equivalent performance when they were assessed by established DL models. This work shows the ability to use a DL technique to generate a large volume of high-fidelity synthetic pelvis images not discernible from real imaging by computers or experts. These images can be used for cross-institutional sharing and model pretraining, further advancing the performance of DL models without risk to patient data safety. Level III. [ABSTRACT FROM AUTHOR]

Deep learning (DL) is a popular method that is used to perform many important tasks in radiology and medical imaging. Some forms of DL are able to accurately segment organs (essentially, trace the boundaries, enabling volume measurements or calculation of other properties). Other DL networks are able to predict important properties from regions of an image-for instance, whether something is malignant, molecular markers for tissue in a region, even prognostic markers. DL is easier to train than traditional machine learning methods, but requires more data and much more care in analyzing results. It will automatically find the features of importance, but understanding what those features are can be a challenge. This article describes the basic concepts of DL systems and some of the traps that exist in building DL systems and how to identify those traps. [ABSTRACT FROM AUTHOR]

The risk of progression in Barrett's esophagus (BE) increases with development of dysplasia. There is a critical need to improve the diagnosis of BE dysplasia, given substantial interobserver disagreement among expert pathologists and overdiagnosis of dysplasia by community pathologists. We developed a deep learning model to predict dysplasia grade on whole-slide imaging. We digitized nondysplastic BE (NDBE), low-grade dysplasia (LGD), and high-grade dysplasia (HGD) histology slides. Two expert pathologists confirmed all histology and digitally annotated areas of dysplasia. Training, validation, and test sets were created (by a random 70/20/10 split). We used an ensemble approach combining a "you only look once" model to identify regions of interest and histology class (NDBE, LGD, or HGD) followed by a ResNet101 model pretrained on ImageNet applied to the regions of interest. Diagnostic performance was determined for the whole slide. We included slides from 542 patients (164 NDBE, 226 LGD, and 152 HGD) yielding 8596 bounding boxes in the training set, 1946 bounding boxes in the validation set, and 840 boxes in the test set. When the ensemble model was used, sensitivity and specificity for LGD was 81.3% and 100%, respectively, and >90% for NDBE and HGD. The overall positive predictive value and sensitivity metric (calculated as F1 score) was.91 for NDBE,.90 for LGD, and 1.0 for HGD. We successfully trained and validated a deep learning model to accurately identify dysplasia on whole-slide images. This model can potentially help improve the histologic diagnosis of BE dysplasia and the appropriate application of endoscopic therapy. [ABSTRACT FROM AUTHOR]

Rationale and Objectives: Inefficient transfer of personal health records among providers negatively impacts quality of health care and increases cost. This multicenter study evaluates the implementation of the first Internet-based image-sharing system that gives patients ownership and control of their imaging exams, including assessment of patient satisfaction.Materials and Methods: Patients receiving any medical imaging exams in four academic centers were eligible to have images uploaded into an online, Internet-based personal health record. Satisfaction surveys were provided during recruitment with questions on ease of use, privacy and security, and timeliness of access to images. Responses were rated on a five-point scale and compared using logistic regression and McNemar's test.Results: A total of 2562 patients enrolled from July 2012 to August 2013. The median number of imaging exams uploaded per patient was 5. Most commonly, exams were plain X-rays (34.7%), computed tomography (25.7%), and magnetic resonance imaging (16.1%). Of 502 (19.6%) patient surveys returned, 448 indicated the method of image sharing (Internet, compact discs [CDs], both, other). Nearly all patients (96.5%) responded favorably to having direct access to images, and 78% reported viewing their medical images independently. There was no difference between Internet and CD users in satisfaction with privacy and security and timeliness of access to medical images. A greater percentage of Internet users compared to CD users reported access without difficulty (88.3% vs. 77.5%, P < 0.0001).Conclusion: A patient-directed, interoperable, Internet-based image-sharing system is feasible and surpasses the use of CDs with respect to accessibility of imaging exams while generating similar satisfaction with respect to privacy. [ABSTRACT FROM AUTHOR]

Rationale and Objectives To examine the effect of changes in utilization and advances in cross-sectional imaging on radiologists' workload. Materials and Methods All computed tomography (CT) and magnetic resonance imaging (MRI) examinations performed at a single institution between 1999 and 2010 were identified and associated with the total number of images for each examination. Annual trends in institutional numbers of interpreted examinations and images were translated to changes in daily workload for the individual radiologist by normalizing to the number of dedicated daily CT and MRI work assignments, assuming a 255-day/8-hour work day schedule. Temporal changes in institutional and individual workload were assessed by Sen's slope analysis ( Q = median slope) and Mann–Kendall test ( Z = Z statistic). Results From 1999 to 2010, a total of 1,517,149 cross-sectional imaging studies (CT = 994,471; MRI = 522,678) comprising 539,210,581 images (CT = 339,830,947; MRI = 199,379,634) were evaluated at our institution. Total annual cross-sectional studies steadily increased from 84,409 in 1999 to 147,336 in 2010, representing a twofold increase in workload ( Q = 6465/year, Z = 4.2, P < .0001). Concomitantly, the number of annual departmental cross-sectional images interpreted increased from 9,294,140 in 1990 to 94,271,551 in 2010, representing a 10-fold increase ( Q = 8707876/year, Z = 4.5, P < .0001). Adjusting for staffing changes, the number of images requiring interpretation per minute of every workday per staff radiologist increased from 2.9 in 1999 to 16.1 in 2010 ( Q = 1.7/year, Z = 4.3, P < .0001). Conclusions Imaging volumes have grown at a disproportionate rate to imaging utilization increases at our institution. The average radiologist interpreting CT or MRI examinations must now interpret one image every 3–4 seconds in an 8-hour workday to meet workload demands. [ABSTRACT FROM AUTHOR]

Interoperability is a major focus of the quickly evolving world of Health IT. Easy, yet secure and confidential exchange of imaging exams and the associated reports must be a part of the solutions that are implemented. The availability of historical exams is essential in providing a quality interpretation and reducing inappropriate utilization of imaging services. Today, the exchange of imaging exams is most often achieved via a compact disc. We describe the virtues of this solution as well as challenges that have surfaced. Internet- and cloud-based technologies employed for many consumer services can provide a better solution. Vendors are making these solutions available. Standards for Internet-based exchange are emerging. Just as radiology converged on DICOM as a standard to store and view images, we need a common exchange standard. We will review the existing standards and how they are organized into useful workflows through Integrating the Healthcare Enterprise profiles. Integrating the Healthcare Enterprise and standards development processes are discussed. Health care and the domain of radiology must stay current with quickly evolving Internet standards. The successful use of the "cloud" will depend on both the technologies and the policies put into place around them, both of which we discuss. The radiology community must lead the way and provide a solution that works for radiologists and clinicians with use of the electronic medical record. We describe features we believe radiologists should consider when adding Internet-based exchange solutions to their practice. [ABSTRACT FROM AUTHOR]

Objective To evaluate the short- to medium-term outcomes after artificial urinary sphincter (AUS) placement from a large, multi-institutional, prospective, follow-up study. We hypothesize that along with radiation, patients with any history of a direct surgery to the urethra will have higher rates of eventual AUS explantation for erosion and/or infection. Materials and Methods A prospective outcome analysis was performed on 386 patients treated with AUS placement from April 2009 to December 2012 at 8 institutions with at least 3 months of follow-up. Charts were analyzed for preoperative risk factors and postoperative complications requiring explantation. Results Approximately 50% of patients were considered high risk. High risk was defined as patients having undergone radiation therapy, urethroplasty, multiple treatments for bladder neck contracture or urethral stricture, urethral stent placement, or a history of erosion or infection in a previous AUS. A total of 31 explantations (8.03%) were performed during the follow-up period. Overall explantation rates were higher in those with prior radiation and prior UroLume. Men with prior AUS infection or erosion also had a trend for higher rates of subsequent explantation. Men receiving 3.5-cm cuffs had significantly higher explantation rates than those receiving larger cuffs. Conclusion This outcomes study confirms that urethral risk factors, including radiation history, prior AUS erosion, and a history of urethral stent placement, increase the risk of AUS explantation in short-term follow-up. [ABSTRACT FROM AUTHOR]

Objective: To evaluate multi-institutional outcomes of bulbar urethroplasty utilizing a standardized cystoscopic follow-up protocol. Methods: Eight reconstructive surgeons prospectively enrolled urethral stricture patients in a multi-institutional study and performed postoperative cystoscopy at 3 and 12 months. Anatomic failure was defined as the inability to pass a flexible cystoscope without force. Functional failure was defined as the need for a secondary procedure. Men not compliant with the 12-month cystoscopy were called and asked if any interval secondary procedures had been performed. Patients with bothersome voiding complaints at cystoscopy were considered symptomatic. Results: Of 213 men in study, 136 underwent excisional urethroplasty (excision and primary anastomosis [EPA]) and 77 underwent repair with buccal grafts. Cystoscopy compliance was 79.8% at 3 months and 54.4% at 12 months. Anatomic success rates were higher at 3 vs 12 months for EPA repairs (97.2% [106 of 109] vs 85.5% [65 of 76; P = .003] but not buccal repairs (85.5% [53 of 62] vs 77.5% [31 of 40]; P = .30). Functional success rates at a year were higher but statistically similar to anatomical success rates (EPA—90.3% [93 of 103]; P = .33; buccal—87% [47 of 54]; P = .22). Of the 20 anatomic recurrences, only 13 (65%) were symptomatic at the time of cystoscopic diagnosis. Conclusion: Rates of success are lower when using the anatomic vs traditional definition. Of recurrences found by cystoscopy, only 65% were symptomatic. One-year patient cystoscopy compliance was poor and its ability to be used as the gold standard screening methodology for recurrence is questionable. [ABSTRACT FROM AUTHOR]

Objective: To examine initial treatments given to men with newly diagnosed lower urinary tract dysfunction (LUTD) within a large integrated health care system in the United States. Methods: We used data from 2003 to 2009 from the Veteran's Health Administration to identify newly diagnosed cases of LUTD using established ICD-9CM codes. Our primary outcome was initial LUTD treatment (3 months), categorized as watchful waiting (WW), medical therapy (MT), or surgical therapy (ST); our secondary outcome was pharmacotherapy class received. We used logistic regression models to examine patient, provider, and health system factors associated with receiving MT or ST when compared with WW. Results: There were 393,901 incident cases of LUTD, of which 58.0% initially received WW, 41.8% MT, and 0.2% ST. Of the MT men, 79.8% received an alpha-blocker, 7.7% a 5-alpha reductase inhibitor, 3.3% an anticholinergic, and 7.3% combined therapy (alpha-blocker and 5-alpha reductase inhibitor). In our regression models, we found that age (higher), race (white/black), income (low), region (northeast/south), comorbidities (greater), prostate-specific antigen (lower), and provider (nonurologist) were associated with an increased odds of receiving MT. We found that age (higher), race (white), income (low), region (northeast/south), initial provider (urologist), and prostate-specific antigen (higher) increased the odds of receiving ST. Conclusion: Most men with newly diagnosed LUTD in the Veteran's Health Administration receive WW, and initial surgical treatment is rare. A large number of men receiving MT were treated with monotherapy, despite evidence that combination therapy is potentially more effective in the long-term, suggesting opportunities for improvement in initial LUTD management within this population. [ABSTRACT FROM AUTHOR]

Objective: To analyze the practice patterns of recently fellowship-trained reconstructive urologists to help guide fellowship program curriculum development and to evaluate the impact that formal reconstructive urology training has on academic urology programs. Methods: We evaluated the case logs of 7 recently fellowship-trained reconstructive urologists affiliated with US academic institutions from August 2009 to August 2011 (median years in practice = 2, range 1-6 years). We categorized cases into endoscopic, oncological, female, general (nononcological), and reconstructive. Our primary outcome was the volume of reconstructive procedures as a percentage of all procedures. Our secondary outcome was the correlation between years in practice and reconstructive volume and case complexity. Results: A total of 3561 cases were analyzed, representing 12 surgeon-years. Endoscopic surgery was most common (42.7%), followed by reconstructive (36.1%), general urologic (10.5%), and oncological (3.7%). The most common type of reconstructive procedure performed was anterior urethroplasty (mean 42.8 per year) followed by bladder reconstruction (mean 17.7 per year). The percentage of yearly cases considered reconstructive was positively associated with total years in practice (r = .688, P = .013) as was the complexity of artificial urinary sphincter cases (r = .857, P = .0004), but not urethral reconstructive complexity (r = .40, P = .197). Conclusion: The demand for services delivered by fellowship-trained reconstructive urologists is high, as evidence by the large percentage of reconstructive procedures in this cohort even early in practice. With additional years in practice comes further specialization. [ABSTRACT FROM AUTHOR]

Objective: To evaluate the association between dietary quality and the prevalence of lower urinary tract symptoms (LUTS). Methods: We used urinary symptom and dietary data obtained from the 2000-2001 National Health and Nutrition Examination Survey (NHANES) for the study. Dietary quality was assessed using the 10-component United States Department of Agriculture (USDA) Healthy Eating Index (HEI). We used bivariate methods to examine rates of LUTS among men with poor versus good diets. Multivariable logistic regression was used to calculate odds ratios after applying sample weights and controlling for age, race/ethnicity, smoking status, diabetes, alcohol intake, and exercise. Results: Our study cohort consisted of 1385 men aged ≥40 years, of whom 279 (21.1%) reported LUTS. We found higher rates of LUTS among men with poor dietary intake of dairy (22.4% vs 16.4%, P = .013) and among men with poor intake of protein (24.6% vs 17.9%, P = .012) as well as among those with overall poor diet (25.8 vs 17.8%, P = .018) with little dietary variety (26.1 vs 17.6%, P = .001). On multivariate analysis, an unhealthy diet (odds ratios [OR] = 1.7; 95% confidence interval [CI] = 1.05-2.90) was associated with more LUTS, whereas alcohol intake was protective from LUTS (OR = 0.67; 95% CI = 0.48-0.93). Conclusion: In an analysis of NHANES data, we found that poor diet quality was independently associated with patient-reported LUTS. [ABSTRACT FROM AUTHOR]

Rationale: Both outcomes researchers and informaticians are concerned with information and data. As such, some of the central challenges to conducting successful comparative effectiveness research can be addressed with informatics solutions. Methods: Specific informatics solutions which address how data in comparative effectiveness research are enriched, stored, shared, and analyzed are reviewed. Results: Imaging data can be made more quantitative, uniform, and structured for researchers through the use of lexicons and structured reporting. Secure and scalable storage of research data is enabled through data warehouses and cloud services. There are a number of national efforts to help researchers share research data and analysis tools. Conclusion: There is a diverse arsenal of informatics tools designed to meet the needs of comparative effective researchers. [ABSTRACT FROM AUTHOR]

Abstract: Purpose: In newborn hydronephrosis (HN), the level of differential function (%df) measured by diuretic renography (DR) is used to judge the need for pyeloplasty. As DR testing is complex, we sought to determine if grading the level of HN (Society of Fetal Urology grade, SFU Gr) by a simple ultrasound correlates with percentage differential function (%df) and thereby obviates the need to perform DR. Materials and methods: Between 1990 and 2003 our institution prospectively enrolled all cases of fetal HN who showed unilateral newborn SFU Gr HN ≥3. The cases underwent standardized testing. DR was done using the method of Well-Tempered Renography which was then followed by ultrasound (US). The US studies were performed while the hydration induced by DR was in effect. The level of %df was categorized as preserved (≥40%) or reduced (<40%). Cases were excluded if there was an additional urological abnormality (e.g. ureterocele). Results: There were 71 cases that met our study criteria. The SFU Gr HN was 3 (n =33) or 4 (n =38). Kidneys with SFU Gr 3 HN showed preserved %df (33/33,100%) (mean=50.1±3.6) significantly more often than kidneys with SFU Gr 4 HN (27/38, 71%) (mean=42.2±13.9) (RR=1.41, 95% CI (1.15–1.72), p <0.001). Conclusion: In newborns with a history of fetal HN, the postnatal finding of SFU Gr 3 HN uniformly correlates with preserved %df. Standardized hydration prior to US study is done to assure consistency in measurement of the SFU Gr HN. Determining the duration of the relationship between SFU Gr 3 HN and preserved %df will require prospective, longitudinal studies. [Copyright &y& Elsevier]

Objectives: To determine whether the routine use of nasogastric tubes (NGTs) after bowel surgery for neurogenic bladder dysfunction improves outcomes. Methods: We retrospectively evaluated 54 consecutive patients (30 women, 24 men) with neurogenic bladder who underwent bladder reconstruction or replacement with bowel segments by one surgeon from December 2000 to August 2005. The first 32 [NGT(+)] had NGTs placed during the procedure, whereas the subsequent 22 [NGT(−)] did not. We compared short-term postoperative outcomes between groups. Results: Patient age ranged from 17 to 74 years (mean, 42.6 years). Procedures included augmentation cystoplasty with or without creation of catheterizable stoma (31), ileovesicostomy (13), and ileal conduit (9). Mean age or mean operative time did not differ between the NGT(+) and NGT(−) groups. The NGT(−) patients experienced less time to oral intake (3.1 versus 4.4 days, P <0.01), fewer days to flatus (2.9 versus 4.0 days, P = 0.01), and fewer days to first bowel movement (4.4 versus 5.9 days, P = 0.01). We found no statistical differences in the incidence of postoperative complications. Overall hospital days were less in the NGT(−) patients, but this did not reach statistical significance (9.9 versus 11.0, P = 0.2). Conclusions: Routine use of NGTs in patients undergoing bladder reconstruction or replacement for neurogenic bladder dysfunction seems to confer no benefit. The omission of NGTs in this population is possible without increasing overall morbidity. These findings parallel those previously reported in neurologically intact patients undergoing urinary diversion. [Copyright &y& Elsevier]

Abstract: Introduction: We report a novel surgical technique used to repair a rectourethral fistula associated with two short-segment urethral strictures located in the anterior and posterior segments of the urethra in a patient with prior unsuccessful repairs. Technical Considerations: The anterior urethral stricture was reconstructed with a ventral onlay of buccal mucosa in the exaggerated lithotomy position. In a modified prone position, the rectourethral fistula was repaired using the transrectal transsphincteric (York-Mason) technique and the posterior urethral stricture with a radial forearm fasciocutaneous free flap which was anastomosed to the inferior gluteal artery and vein. The coexistence of a rectourethral fistula and distal urethral stricture requires simultaneous repair, because the urethral pressure from the distal obstruction may compromise fistula closure. Reconstructive efforts should be tailored to minimize disruption of the urethral blood supply in patients with previous pelvic trauma. Rectal and urethral repairs should be separated by well-vascularized tissue to prevent fistula recurrence. Conclusions: The radial fasciocutaneous flap may offer the reconstructive surgeon another surgical option for complex urethral stricture and rectourethral fistula reconstruction when the local blood supply is in question. Longer follow-up and more cases are needed to further evaluate the continued use of this technique. [Copyright &y& Elsevier]

Abstract: Objectives: To compare postoperative intraoral morbidity after mucosal graft harvesting from the lower lip and inner cheek in men undergoing substitution urethroplasty. Methods: From 2000 to 2004, 40 patients underwent oral mucosal graft harvesting for urethral reconstruction. Of the 40 patients, 17 underwent harvest from the lower lip (group 1), 19 from the inner cheek (group 2), and 4 from both sites (group 3). We analyzed the postoperative differences in discomfort levels, neurosensory deficits, salivary flow, complications, and overall quality of life according to harvest site. Outcomes were assessed using analog scales and a questionnaire. Results: At longer term follow-up, the proportion of men with persistent postoperative discomfort, neurosensory deficits, and changes in salivary flow was greater in men whose grafts were harvested from the lower lip than in those whose grafts were harvested from the cheek. Contractures were noted in 1 (5.9%) of 17 and 5 (26.3%) of 19 men in groups 1 and 2, respectively. At the last follow-up visit, overall quality of life (measured by an analog scale of 0 [delighted] to 6 [terrible]) was not significantly different in patients with lower lip versus inner cheek harvest (1.35 versus 0.66, respectively, P = 0.079). Conclusions: Persistent postoperative discomfort, neurosensory deficits, and salivary flow changes were more common in men whose grafts were harvested from the lower lip. However, the overall quality of life in this group was not significantly different compared with men whose grafts were harvested from the inner cheek. From a morbidity standpoint, the inner cheek appears to be the preferable site of graft harvest. [Copyright &y& Elsevier]

Background: Inappropriate acetabular component angular position is believed to increase the risk of hip dislocation after total hip arthroplasty. However, manual measurement of these angles is time consuming and prone to interobserver variability. The purpose of this study was to develop a deep learning tool to automate the measurement of acetabular component angles on postoperative radiographs.Methods: Two cohorts of 600 anteroposterior (AP) pelvis and 600 cross-table lateral hip postoperative radiographs were used to develop deep learning models to segment the acetabular component and the ischial tuberosities. Cohorts were manually annotated, augmented, and randomly split to train-validation-test data sets on an 8:1:1 basis. Two U-Net convolutional neural network models (one for AP and one for cross-table lateral radiographs) were trained for 50 epochs. Image processing was then deployed to measure the acetabular component angles on the predicted masks for anatomical landmarks. Performance of the tool was tested on 80 AP and 80 cross-table lateral radiographs.Results: The convolutional neural network models achieved a mean Dice similarity coefficient of 0.878 and 0.903 on AP and cross-table lateral test data sets, respectively. The mean difference between human-level and machine-level measurements was 1.35° (σ = 1.07°) and 1.39° (σ = 1.27°) for the inclination and anteversion angles, respectively. Differences of 5⁰ or more between human-level and machine-level measurements were observed in less than 2.5% of cases.Conclusion: We developed a highly accurate deep learning tool to automate the measurement of angular position of acetabular components for use in both clinical and research settings.Level Of Evidence: III. [ABSTRACT FROM AUTHOR]

Background: Dislocation is a common complication following total hip arthroplasty (THA), and accounts for a high percentage of subsequent revisions. The purpose of this study is to illustrate the potential of a convolutional neural network model to assess the risk of hip dislocation based on postoperative anteroposterior pelvis radiographs.Methods: We retrospectively evaluated radiographs for a cohort of 13,970 primary THAs with 374 dislocations over 5 years of follow-up. Overall, 1490 radiographs from dislocated and 91,094 from non-dislocated THAs were included in the analysis. A convolutional neural network object detection model (YOLO-V3) was trained to crop the images by centering on the femoral head. A ResNet18 classifier was trained to predict subsequent hip dislocation from the cropped imaging. The ResNet18 classifier was initialized with ImageNet weights and trained using FastAI (V1.0) running on PyTorch. The training was run for 15 epochs using 10-fold cross validation, data oversampling, and augmentation.Results: The hip dislocation classifier achieved the following mean performance (standard deviation): accuracy = 49.5 (4.1%), sensitivity = 89.0 (2.2%), specificity = 48.8 (4.2%), positive predictive value = 3.3 (0.3%), negative predictive value = 99.5 (0.1%), and area under the receiver operating characteristic curve = 76.7 (3.6%). Saliency maps demonstrated that the model placed the greatest emphasis on the femoral head and acetabular component.Conclusion: Existing prediction methods fail to identify patients at high risk of dislocation following THA. Our radiographic classifier model has high sensitivity and negative predictive value, and can be combined with clinical risk factor information for rapid assessment of risk for dislocation following THA. The model further suggests radiographic locations which may be important in understanding the etiology of prosthesis dislocation. Importantly, our model is an illustration of the potential of automated imaging artificial intelligence models in orthopedics.Level Of Evidence: Level III. [ABSTRACT FROM AUTHOR]