Learning Morphological, Spatial, and Dynamic Models of Cellular Components.

In this chapter, we describe protocols for using the CellOrganizer software on the Jupyter Notebook platform to analyze and model cell and organelle shape and spatial arrangement. CellOrganizer is an open-source system for using microscope images to learn statistical models of the structure of cell components and how those components are organized relative to each other. Such models capture the statistical variation in the organization of cellular components by jointly modeling the distributions of their number, shape, and spatial distributions. These models can be created for different cell types or conditions and compared to reflect differences in their spatial organizations. The models are also generative, in that they can be used to synthesize new cell instances reflecting what a model learned and to provide well-structured cell geometries that can be used for biochemical simulations.

3DCellComposer - A Versatile Pipeline Utilizing 2D Cell Segmentation Methods for 3D Cell Segmentation.

Background: Cell segmentation is crucial in bioimage informatics, as its accuracy directly impacts conclusions drawn from cellular analyses. While many approaches to 2D cell segmentation have been described, 3D cell segmentation has received much less attention. 3D segmentation faces significant challenges, including limited training data availability due to the difficulty of the task for human annotators, and inherent three-dimensional complexity. As a result, existing 3D cell segmentation methods often lack broad applicability across different imaging modalities. Results: To address this, we developed a generalizable approach for using 2D cell segmentation methods to produce accurate 3D cell segmentations. We implemented this approach in 3DCellComposer, a versatile, open-source package that allows users to choose any existing 2D segmentation model appropriate for their tissue or cell type(s) without requiring any additional training. Importantly, we have enhanced our open source CellSegmentationEvaluator quality evaluation tool to support 3D images. It provides metrics that allow selection of the best approach for a given imaging source and modality, without the need for human annotations to assess performance. Using these metrics, we demonstrated that our approach produced high-quality 3D segmentations of tissue images, and that it could outperform an existing 3D segmentation method on the cell culture images with which it was trained. Conclusions: 3DCellComposer, when paired with well-trained 2D segmentation models, provides an important alternative to acquiring human-annotated 3D images for new sample types or imaging modalities and then training 3D segmentation models using them. It is expected to be of significant value for large scale projects such as the Human BioMolecular Atlas Program.

Expanding the coverage of spatial proteomics: a machine learning approach.

MOTIVATION: Multiplexed protein imaging methods use a chosen set of markers and provide valuable information about complex tissue structure and cellular heterogeneity. However, the number of markers that can be measured in the same tissue sample is inherently limited. RESULTS: In this paper, we present an efficient method to choose a minimal predictive subset of markers that for the first time allows the prediction of full images for a much larger set of markers. We demonstrate that our approach also outperforms previous methods for predicting cell-level protein composition. Most importantly, we demonstrate that our approach can be used to select a marker set that enables prediction of a much larger set than could be measured concurrently. AVAILABILITY AND IMPLEMENTATION: All code and intermediate results are available in a Reproducible Research Archive at https://github.com/murphygroup/CODEXPanelOptimization.

Intraoperative Radiation Exposure in Adolescent Idiopathic and Neuromuscular Scoliosis.

BACKGROUND: Intraoperative imaging is often used to aid pedicle screw placement during scoliosis operations. Higher rates of cancer and death have been observed in orthopaedic surgeons and radiation technologists, including a fourfold higher rate of breast cancer in female orthopaedic surgeons. The purpose of this study was to evaluate variability in intraoperative radiation during spinal fusions for both adolescent idiopathic scoliosis (AIS) and neuromuscular scoliosis (NMS). METHODS: A retrospective review of posterior spinal fusion and segmental spinal instrumentation for scoliosis performed by pediatric orthopaedic surgeons from 2017 to 2019 at a single institution was performed. Inclusion criteria included: a diagnosis of AIS or NMS and patients between 8 and 18 years of age. Exclusion criteria included: revision surgery, use of intraoperative navigation, and patients younger than 10 at the time of scoliosis onset within the AIS cohort. Data collected included: preoperative curve, body mass index (BMI), number of levels fused, number of Ponte osteotomies, and fluoroscopy time. One-way analysis of variance tests, Bonferroni post hoc tests, independent t tests, and Pearson correlations were utilized with significance determined at the 95% confidence level ( a = 0.05). RESULTS: A total of 148 patients were included in the study. The average fluoroscopy time was 143 ± 67 seconds. Patients with NMS had higher average fluoroscopy times (193 ± 75 s) compared with patients with AIS (129 ± 58 s, P < 0.001). In patients with AIS, fluoroscopy time correlated to the patient's preoperative curve ( r = 0.182, P = 0.050). Patients with AIS with fewer than 12 levels fused had significantly less radiation exposure than those with 12 or more levels fused ( P = 0.01). When controlling for the number of levels fused, patients with AIS with higher BMIs had significantly greater fluoroscopy times ( P = 0.001). In patients with NMS, fluoroscopy time negatively correlated with BMI ( r = -0.459, P = 0.009) and positively correlated with a preoperative curve ( r = 0.475, P = 0.007). CONCLUSION: Fluoroscopy times vary greatly during adolescent spinal fusions for scoliosis. Longer fluoroscopy times are correlated with: NMS diagnosis, larger preoperative curve, BMI, and number of levels fused. Surgeons' knowledge of factors affecting fluoroscopy time will increase awareness and may be the first step in decreasing intraoperative radiation risks. LEVEL OF EVIDENCE: Level III; Therapeutic-a retrospective study.

Five Inhibitory Receptors Display Distinct Vesicular Distributions in Murine T Cells.

T cells can express multiple inhibitory receptors. Upon induction of T cell exhaustion in response to a persistent antigen, prominently in the anti-tumor immune response, many are expressed simultaneously. Key inhibitory receptors are CTLA-4, PD-1, LAG3, TIM3, and TIGIT, as investigated here. These receptors are important as central therapeutic targets in cancer immunotherapy. Inhibitory receptors are not constitutively expressed on the cell surface, but substantial fractions reside in intracellular vesicular structures. It remains unresolved to which extent the subcellular localization of different inhibitory receptors is distinct. Using quantitative imaging of subcellular distributions and plasma membrane insertion as complemented by proximity proteomics and biochemical analysis of the association of the inhibitory receptors with trafficking adaptors, the subcellular distributions of the five inhibitory receptors were discrete. The distribution of CTLA-4 was most distinct, with preferential association with lysosomal-derived vesicles and the sorting nexin 1/2/5/6 transport machinery. With a lack of evidence for the existence of specific vesicle subtypes to explain divergent inhibitory receptor distributions, we suggest that such distributions are driven by divergent trafficking through an overlapping joint set of vesicular structures. This extensive characterization of the subcellular localization of five inhibitory receptors in relation to each other lays the foundation for the molecular investigation of their trafficking and its therapeutic exploitation.

A Comparative Analysis of Revision Surgery Before or After 2 Years After Graduation From Growth-friendly Surgery for Early Onset Scoliosis.

INTRODUCTION: After discontinuation of growth-friendly (GF) surgery for early onset scoliosis, patients are termed graduates: they undergo a spinal fusion, are observed after final lengthening with GF implant maintenance, or are observed after GF implant removal. The purpose of this study was to compare the rates of and reasons for revision surgery in two cohorts of GF graduates: before or after 2 years of follow-up from graduation. METHODS: A pediatric spine registry was queried for patients who underwent GF spine surgery with a minimum of 2 years of follow-ups after graduation by clinical and/or radiographic evidence. Scoliosis etiology, graduation strategy, number of, and reasons for revision surgery were queried. RESULTS: There were 834 patients with a minimum of 2-year follow-up after graduation who were analyzed. There were 241 (29%) congenital, 271 (33%) neuromuscular, 168 (20%) syndromic, and 154 (18%) idiopathic. 803 (96%) had traditional growing rod/vertical expandable titanium rib as their GF construct and 31 (4%) had magnetically controlled growing rod. Five hundred ninety-six patients (71%) underwent spinal fusion at graduation, 208 (25%) had GF implants retained, and 30 (4%) had GF implants removed.In the entire cohort, there were 108/834 (13%) patients who underwent revision surgery. Of the revisions, 71/108 (66%) occurred as acute revisions (ARs) between 0 and 2 years from graduation (mean 0.6 y), and the most common AR indication was infection (26/71, 37%). The remaining 37/108 (34%) patients underwent delayed revision (DR) surgery >2 years (mean 3.8 y) from graduation, and the most common DR indication was implant issues (17/37, 46%).Graduation strategy affected revision rates. Of the 596 patients with spinal fusion as a graduation strategy, 98/596 (16%) underwent revision, compared with only 8/208 (4%) patients who had their GF implants retained, and 2/30 (7%) that had their GF implants removed ( P ≤ 0.001).A significantly higher percentage of the ARs had a spinal fusion as the graduation strategy (68/71, 96%) compared with 30/37 DRs, (81%, P = 0.015). In addition, the 71 patients who underwent AR undergo more revision surgeries (mean: 2, range: 1 to 7) than 37 patients who underwent DR (mean: 1, range: 1 to 2) ( P = 0.001). CONCLUSION: In this largest reported series of GF graduates to date, the overall risk of revision was 13%. Patients who undergo a revision at any time, as well as ARs in particular, are more likely to have a spinal fusion as their graduation strategy. Patients who underwent AR, on average, undergo more revision surgeries than patients who underwent DR. LEVEL OF EVIDENCE: Level III, comparative.

Five inhibitory receptors display distinct vesicular distributions in T cells.

T cells can express multiple inhibitory receptors. Upon induction of T cell exhaustion in response to persistent antigen, prominently in the anti-tumor immune response, many are expressed simultaneously. Key inhibitory receptors are CTLA-4, PD-1, LAG3, TIM3 and TIGIT, as investigated here. These receptors are important as central therapeutic targets in cancer immunotherapy. Inhibitory receptors are not constitutively expressed on the cell surface, but substantial fractions reside in intracellular vesicular structures. It remains unresolved to which extent the subcellular localization of different inhibitory receptors is distinct. Using quantitative imaging of subcellular distributions and plasma membrane insertion as complemented by proximity proteomics and a biochemical analysis of the association of the inhibitory receptors with trafficking adaptors, the subcellular distributions of the five inhibitory receptors were discrete. The distribution of CTLA-4 was most distinct with preferential association with lysosomal-derived vesicles and the sorting nexin 1/2/5/6 transport machinery. With a lack of evidence for the existence of specific vesicle subtypes to explain divergent inhibitory receptor distributions, we suggest that such distributions are driven by divergent trafficking through an overlapping joint set of vesicular structures. This extensive characterization of the subcellular localization of five inhibitory receptors in relation to each other lays the foundation for the molecular investigation of their trafficking and its therapeutic exploitation.

Definition of Tweener: Consensus Among Experts in Treating Early-onset Scoliosis.

BACKGROUND: The term "Tweener" is colloquially used to refer to early-onset scoliosis (EOS) patients whose age and development make them candidates for multiple surgical options. The purpose of this study was to establish expert consensus on a definition to formally characterize the Tweener population. METHODS: A 3-round survey of surgeons in an international EOS study group was conducted. Surgeons were provided with various patient characteristics and asked if each was part of their definition for Tweener patients. Responses were analyzed for consensus (≥70%), near-consensus (60% to 69%), and no consensus (<60%). RESULTS: Consensus was reached (89% of respondents) for including chronological age in the Tweener definition; 8 to 10 years for females and 9 to 11 years for males. Surgeons agreed for inclusion of Sanders score, particularly Sanders 2 (86.0%). Patients who have reached Sanders 4, postmenarche, or have closed triradiate cartilage should not be considered Tweeners. Bone age range of 8 years and 10 months to 10 years and 10 months for females (12 y for males) could be part of the Tweener definition. CONCLUSIONS: This study suggests that the Tweener definition could be the following: patients with open triradiate cartilage who are not postmenarche and have not reached Sanders 4, and if they have one of the following: Sanders 2 or chronological age 8 to 10 years for females (9 to 11 y for males) or bone age 8 years and 10 months to 10 years and 10 months for females (12 y for males). This definition will allow for more focused and comparative research on this population. LEVEL OF EVIDENCE: Level V-expert opinion.

Basal body organization and cell geometry during the cell cycle in Tetrahymena thermophila.

Tetrahymena thermophila possesses arrays of motile cilia that promote fluid flow for cell motility. These consist of intricately organized basal bodies (BBs) that nucleate and position cilia at the cell cortex. Tetrahymena cell geometry and spatial organization of BBs play important roles in cell size, swimming, feeding, and division. How cell geometry and BB organization are established and maintained remains poorly understood, and prior studies have been limited due to difficulties in accurate BB identification and small sample size. We therefore developed an automated image processing pipeline that segments single cells, distinguishes unique BB populations, assigns BBs into distinct ciliary rows, and distinguishes new from mature BBs. We identified unique features to describe the variation of cell shape and BB spatial organization in unsynchronized single-cell images. The results reveal asymmetries in BB distribution and ingression of the cytokinetic furrow within the cell. Moreover, we establish novel spatial and temporal waves in new BB assembly through the cell cycle. Finally, we used measurements from single cells across the cell cycle to construct a generative model that allows synthesis of movies depicting single cells progressing through the cell cycle. Our approach is expected to be of particular value for characterizing Tetrahymena mutants.

Evaluation of cell segmentation methods without reference segmentations.

Cell segmentation is a cornerstone of many bioimage informatics studies, and inaccurate segmentation introduces error in downstream analysis. Evaluating segmentation results is thus a necessary step for developing segmentation methods as well as for choosing the most appropriate method for a particular type of sample. The evaluation process has typically involved comparison of segmentations with those generated by humans, which can be expensive and subject to unknown bias. We present here an approach to evaluating cell segmentation methods without relying upon comparison to results from humans. For this, we defined a number of segmentation quality metrics that can be applied to multichannel fluorescence images. We calculated these metrics for 14 previously described segmentation methods applied to datasets from four multiplexed microscope modalities covering five tissues. Using principal component analysis to combine the metrics, we defined an overall cell segmentation quality score and ranked the segmentation methods. We found that two deep learning-based methods performed the best overall, but that results for all methods could be significantly improved by postprocessing to ensure proper matching of cell and nuclear masks. Our evaluation tool is available as open source and all code and data are available in a Reproducible Research Archive.

Improving and evaluating deep learning models of cellular organization.

MOTIVATION: Cells contain dozens of major organelles and thousands of other structures, many of which vary extensively in their number, size, shape and spatial distribution. This complexity and variation dramatically complicates the use of both traditional and deep learning methods to build accurate models of cell organization. Most cellular organelles are distinct objects with defined boundaries that do not overlap, while the pixel resolution of most imaging methods is n sufficient to resolve these boundaries. Thus while cell organization is conceptually object-based, most current methods are pixel-based. Using extensive image collections in which particular organelles were fluorescently labeled, deep learning methods can be used to build conditional autoencoder models for particular organelles. A major advance occurred with the use of a U-net approach to make multiple models all conditional upon a common reference, unlabeled image, allowing the relationships between different organelles to be at least partially inferred. RESULTS: We have developed improved Generative Adversarial Networks-based approaches for learning these models and have also developed novel criteria for evaluating how well synthetic cell images reflect the properties of real images. The first set of criteria measure how well models preserve the expected property that organelles do not overlap. We also developed a modified loss function that allows retraining of the models to minimize that overlap. The second set of criteria uses object-based modeling to compare object shape and spatial distribution between synthetic and real images. Our work provides the first demonstration that, at least for some organelles, deep learning models can capture object-level properties of cell images. AVAILABILITY AND IMPLEMENTATION: http://murphylab.cbd.cmu.edu/Software/2022_insilico. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Trends in the Utilization of Implants in Index Procedures for Early Onset Scoliosis From the Pediatric Spine Study Group.

INTRODUCTION: Because of the relative rarity of Early Onset Scoliosis (EOS) cases, patient registries were developed to combine clinical information from multiple institutions to maximize patient care and outcomes. This study examines the history and trends regarding the use of growth-friendly devices for index surgical procedures in EOS patients within the Pediatric Spine Study Group database. METHODS: All index growth-friendly implants were queried from registry inception until October 2020. EOS etiology, device/implant type, and geographic area/institution for each procedure were recorded. RESULTS: From 1994 to 2020, 2786 patients underwent index surgery at a mean age of 6.2±2.9 years. There were 908 traditional growing rods (TGR) (32.3%), 922 vertical expandable prosthetic titanium rib devices (VEPTR) (33.1%), 5 hybrid VEPTR/TGR (0.18%), and 951 magnetically controlled growing rods (MCGR) (34.2%) index implants. Fifty-six different institutions reported an index implant, and 5 accounted for 823 (30%) of the cases during the study period. Institutions in the Northeast accounted for more index implants than other regions of the United States. There was a 40% increase in index implant insertions annually when comparing 1994 (3 implants/1 center) to 2018 (234 implants/56 centers), ( P <0.001). Beginning in 2009, there was a 90.9% decrease in the number of TGR/VEPTR procedures (2009: 156 implants/32 centers; 2019: 22 implants/49 centers P =0.001), and a 479% increase in MCGR (2009: 1 implant/1 center; 2018: 197 implants/34 centers ( P =0.005). The overall number of growth-friendly index procedures performed in 2019 (150/49 centers) decreased 34.5% when compared to 2018 (234/48 centers). CONCLUSION: The number of growth-friendly implants reported in the Pediatric Spine Study Group registry as the initial surgical management of EOS increased markedly over the past 20 years. MCGR is currently the predominant type of device utilized for index surgical procedures by group members, surpassing the use of VEPTR and TGR in 2014. There was a significant decrease in index growth-friendly procedures in 2019 compared to 2018. LEVEL OF EVIDENCE: Level IV.

Contraindications to magnetically controlled growing rods: consensus among experts in treating early onset scoliosis.

PURPOSE: The purpose of this study was to describe contraindications to the magnetically controlled growing rod (MCGR) in patients with early onset scoliosis (EOS) by establishing consensus amongst expert surgeons who treat these patients frequently. METHODS: Nine pediatric spine surgeons from an international EOS study group participated in semi-structured interviews via email to identify factors that influence decision making in the use of MCGR. A 39-question survey was then developed to specify these factors as contraindications for MCGR-these included patient age and size, etiology, medical comorbidities, coronal and sagittal curve profiles, and skin and soft tissue characteristics. Pediatric spine surgeons from the EOS international study group were invited to complete the survey. A second 29-item survey was created to determine details and clarify results from the first survey. Responses were analyzed for consensus (> 70%), near consensus (60-69%), and no consensus/variability (< 60%) for MCGR contraindication. RESULTS: 56 surgeons of 173 invited (32%) completed the first survey, and 64 (37%) completed the second survey. Responders had a mean of over 15 years in practice (range 1-45) with over 6 years of experience with using MCGR (range 2-12). 71.4% of respondents agreed that patient size characteristics should be considered as contraindications, including BMI (81.3%) and spinal height (84.4%), although a specific BMI range or a specific minimum spinal height were not agreed upon. Among surgeons who agreed that skin and soft tissue problems were contraindications (78.6%), insufficient soft tissue (98%) and skin (89%) to cover MCGR were specified. Among surgeons who reported curve stiffness as a contraindication (85.9%), there was agreement that this curve stiffness should be defined by clinical evaluation (78.2%) and by traction films (72.3%). Among surgeons who reported sagittal curve characteristics as contraindications, hyperkyphosis (95.3%) and sagittal curve apex above T3 (70%) were specified. Surgeons who indicated the need for repetitive MRI as a contraindication (79.7%) agreed that image quality (72.9%) and not patient safety (13.6%) was the concern. In the entire cohort, consensus was not achieved on the following factors: patient age (57.4%), medical comorbidities (46.4%), etiology (53.6%), and coronal curve characteristics (58.9%). CONCLUSION: Surgeon consensus suggests that MCGR should be avoided in patients who have insufficient spinal height to accommodate the MCGR, have potential skin and soft tissue inadequacy, have too stiff a spinal curve, have too much kyphosis, and require repetitive MRI, particularly of the spine. Future data-driven studies using this framework are warranted to generate more specific criteria (e.g. specific degrees of kyphosis) to facilitate clinical decision making for EOS patients. LEVEL OF EVIDENCE: Level V-expert opinion.

Pelvic fixation is not always necessary in children with cerebral palsy scoliosis treated with growth-friendly instrumentation.

PURPOSE: A previous study showed that patients with neuromuscular scoliosis who underwent fusion to L5 had excellent coronal curve correction and improvement in pelvic obliquity (PO) when preoperative L5 tilt was < 15°. Our purpose was to identify indications to exclude the pelvis in children with cerebral palsy (CP) scoliosis treated with growing-friendly instrumentation. METHODS: In a retrospective cohort study, children with CP scoliosis treated with TGR, MCGR, or VEPTR with minimum 2-year follow-up were identified from a multicenter database. RESULTS: 27 patients with distal spine anchors (DSA) and 71 patients with distal pelvic anchors (DPA) placed at the index surgery were analyzed. The DSA group had a lower pre-index PO (9° vs 16°, P = 0.0001). Most recent radiographic data were similar except the DSA patients had a smaller major curve (47° vs 58°, P = 0.038). 6 (22%) DSA patients underwent extension of the instrumentation to the pelvis (DSA-EXT), most commonly at final fusion (5 patients). DSA-EXT patients had a higher pre-index L5 tilt than patients who did not require extension (DSA-NO EXT) (19° vs 10°, P = 0.009). Sub-analysis showed a lower major curve at most recent follow-up in the DSA-EXT group compared to the DPA group (33° vs 58°, P = 0.021). The DSA-EXT group had a higher number of complications per patient compared to the DSA-NO EXT group (2.3 vs 1.1, P = 0.029). CONCLUSION: Pre-index L5 tilt ≤ 10° and PO < 10° may be indications to exclude the pelvis in children with CP scoliosis treated with growth-friendly instrumentation. DSA may provide better long-term control of the major curve than DPA.

Baclofen Pump Use: Complications After Growth-friendly Instrumentation for Early-onset Scoliosis.

BACKGROUND: Patients with early-onset scoliosis (EOS) and spasticity may receive treatment with an intrathecal baclofen pump. We assessed how baclofen pumps are associated with the odds of complications and secondary interventions after growth-friendly (GF) spine surgery for EOS and analyzed infectious complications within the pump cohort. METHODS: Using a prospectively maintained, international multicenter database, we studied patients with neuromuscular EOS with baclofen pumps who underwent GF spine surgery from 2002 through 2019 (n=25). Baclofen pumps were implanted before GF instrumentation in 18 patients, during in 2 patients, and after in 5 patients. Patients with existing pumps at initial GF spine surgery were matched 1:3 with 54 patients (control group) without pumps according to treatment center, year of surgery, diagnosis, surgery type, and preoperative curve magnitude. Univariate analysis and multivariate logistic regression were performed to compare complications and secondary interventions between the 2 cohorts. RESULTS: Patients with baclofen pumps had 4.8 times the odds [95% confidence interval (CI): 1.5-16] of experiencing any complication within 1 year after initial GF spine surgery compared with controls. During mean follow-up of 6.9±4.3 years, they had 4.7 times the odds (95% CI: 1.3-16) of deep surgical site infection and 5.6 times the odds (95% CI: 1.2-26) of spinal rod removal after any complication. Differences in rates of mechanical complication, such as rod migration and breakage, were nonsignificant between the 2 groups. For the 9 patients (50%) with pumps who experienced infections, the most common microorganisms were Staphylococcus aureus (4 patients) and Pseudomonas aeruginosa (2). The pump/catheter was revised or removed, in addition to antibiotic therapy or surgical irrigation and debridement, in 2 patients. CONCLUSIONS: Among patients with neuromuscular EOS, those with baclofen pumps are much more likely to experience complications within 1 year after GF spine surgery. They are also more likely to have deep surgical site infections, with S. aureus and P. aeruginosa being the most common causative organisms, and to require spinal rod removal. LEVEL OF EVIDENCE: Level III-retrospective comparative study.

Mechanical Failure of 2 Cannulated Screw Fixation for Unstable SCFE: A Case Report.

CASE: A 13-year-old obese boy presented with an acute-on-chronic unstable left slipped capital femoral epiphysis (SCFE). He underwent in situ surgical fixation with two 6.5-mm fully threaded cannulated screws. At 6 months, he presented with mechanical failure of both screws. He underwent screw removal, revision in situ fixation, a peritrochanteric flexion and internal rotational osteotomy, and an open femoroplasty. The osteotomy healed at 6 weeks. The femoral physis took an additional year to close. CONCLUSION: This case highlights an uncommon complication of in situ pinning of SCFE, discusses revision fixation options, and suggests possible prolonged physeal closure in severe slips.

Evaluation of categorical matrix completion algorithms: toward improved active learning for drug discovery.

MOTIVATION: High throughput and high content screening are extensively used to determine the effect of small molecule compounds and other potential therapeutics upon particular targets as part of the early drug development process. However, screening is typically used to find compounds that have a desired effect but not to identify potential undesirable side effects. This is because the size of the search space precludes measuring the potential effect of all compounds on all targets. Active machine learning has been proposed as a solution to this problem. RESULTS: In this article, we describe an improved imputation method, Impute by Committee, for completion of matrices containing categorical values. We compare this method to existing approaches in the context of modeling the effects of many compounds on many targets using latent similarities between compounds and conditions. We also compare these methods for the task of driving active learning in well-characterized settings for synthetic and real datasets. Our new approach performed the best overall both in the accuracy of matrix completion itself and in the number of experiments needed to train an accurate predictive model compared to random selection of experiments. We further improved upon the performance of our new method by developing an adaptive switching strategy for active learning that iteratively chooses between different matrix completion methods. AVAILABILITY AND IMPLEMENTATION: A Reproducible Research Archive containing all data and code is available at http://murphylab.cbd.cmu.edu/software. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Analysis of Author Gender in the Pediatric Orthopaedic Literature from 2011 to 2020.

INTRODUCTION: Orthopaedic surgery remains a male-dominated specialty. To date there has not been a focused analysis of gender in authorship within the pediatric orthopaedic literature. METHODS: The electronic table of contents from 2011 to 2020 of 3 major pediatric orthopaedic journals [Journal of Children's Orthopaedics (JCO), Journal of Pediatric Orthopaedics (JPO), and Journal of Pediatric Orthopaedics Part B (JPO-B)] were reviewed. Publications were reviewed for the number of articles with at least 1 female author and the number of articles with women listed as first authors. These were compared over the 10-year study period, and by individual year of publication. Statistical analysis included a general linear model with factorial one-way anslysis of variance and Bonferroni post hoc testing. RESULTS: A total of 4097 articles were reviewed. In 2020, there was a significantly higher percentage of articles with a female author when compared with 2011 (64% to 42%, P=0.010). A female was listed as first author in significantly more publications in 2020 as compared with 2011 (23% to 10%, P=0.031). During the 10-year study period, the highest mean proportion of articles with at least 1 female author was seen in JPO (60%), with similar findings in JCO (55%). Significantly fewer articles in JPO-B contained a female author (37%, P=0.001). The highest percentage of publications with a female first author across 10 years was in JCO (22%), followed by JPO (20%). Significantly fewer articles with a female first author were found in JPO-B (9%, P=0.001). CONCLUSION: There is an increasing proportion of publications in the pediatric orthopaedic literature with female authors and female first authors from 2011 to 2020. In addition, there was a statistical difference in female authorship when comparing specific publications, which should be investigated further. LEVEL OF EVIDENCE: Level IV.

A Standardized Order-Set Improves Variability in Opioid Discharge Prescribing Patterns After Surgical Fixation of Pediatric Supracondylar Humerus Fractures.

OBJECTIVE: To evaluate institutional opioid prescribing patterns following percutaneous fixation of pediatric supracondylar humerus fractures before and after implementation of a standardized discharge order set. DESIGN: A retrospective review of patients who underwent closed reduction and percutaneous skeletal fixation of a Type II or III supracondylar humerus fracture in 2017 (prior to pain protocol implementation) and again in 2019 (after pain protocol implementation) SETTING: Single Tertiary Care Children's Hospital PARTICIPANTS: In total, 106 patients met inclusion criteria between years 2017 (n = 49) and 2019 (n = 57). Exclusion criteria included miscoded patients, open fractures, patients who presented with vascular injury or nerve palsy, polytrauma patients with multiple fractures in the same upper extremity, and supracondylar humerus fractures that underwent an open procedure. RESULTS: There were no significant differences between inpatient pain scores (p = 0.91) and MDE prescribed (p = 0.75) between the 2 cohorts. In 2017, large variability was noted in day supply of opioids (0-11.4 days) and MDE (0-8.45 mg/kg), with significant differences between prescribing patterns of junior and senior level residents (mean day supply of opioids (p = 0.045), mean MDE prescribed on discharge (p = 0.001)). After implementation of a standardized opioid discharge order set, there was a tenfold increase in the number of patients discharged without an opioid prescription (2017: 4%, 2019: 44%). Additionally, any discrepancies between prescribing practices of junior and senior level residents were eliminated (mean day supply of opioids (p = 0.65), mean MDE prescribed on discharge (p = 0.69)). CONCLUSIONS: The introduction of a standardized post-operative opioid discharge order set led to a 10-fold increase in the number of patients discharged without an opioid prescription. Additionally, the order set decreased the variability in the prescribing patterns of discharge opioid medications without change in pain control. The resident prescribing variability based upon level of experience resolved with the use of the order set.

Evaluation and Management of Common Accessory Ossicles of the Foot and Ankle in Children and Adolescents.

Accessory ossicles are a common radiographic finding about the foot and ankle in children and adolescents. They are often noted incidentally during evaluation of foot and ankle injuries, and most can be managed nonsurgically. Although over 20 accessory ossicles have been described around the foot and ankle, five specific structures generate the most concern in pediatric patients. An accessory navicular presents commonly with medial midfoot pain and may require surgical intervention after failure of nonsurgical treatment. Although an accessory navicular can be treated surgically with simple excision, there is some recent evidence that supports concomitant reconstruction of associated flatfoot deformities. Os trigonum, an ossicle posterior to the talus, is also commonly asymptomatic. However, os trigonum may be associated with persistent posterior ankle pain, and open and endoscopic resection techniques are successful. Os subfibulare is an uncommon ossicle that may be associated with recurrent ankle sprains. Recent literature reports successful return to activities after ossicle excision and ligament reconstruction. Os subtibiale may be confused with a medial malleolar fracture in skeletally immature patients. Os peroneum may contribute to lateral midfoot pain.