Development of Machine Learning Models for Predicting the 1-Year Risk of Reoperation After Lower Limb Oncological Resection and Endoprosthetic Reconstruction Based on Data From the PARITY Trial.

BACKGROUND: Oncological resection and reconstruction involving the lower extremities commonly lead to reoperations that impact patient outcomes and healthcare resources. This study aimed to develop a machine learning (ML) model to predict this reoperation risk. METHODS: This study was conducted according to TRIPOD + AI. Data from the PARITY trial was used to develop ML models to predict the 1-year reoperation risk following lower extremity oncological resection and reconstruction. Six ML algorithms were tuned and calibrated based on fivefold cross-validation. The best-performing model was identified using classification and calibration metrics. RESULTS: The polynomial support vector machine (SVM) model was chosen as the best-performing model. During internal validation, the SVM exhibited an AUC-ROC of 0.73 and a Brier score of 0.17. Using an optimal threshold that balances all quadrants of the confusion matrix, the SVM exhibited a sensitivity of 0.45 and a specificity of 0.81. Using a high-sensitivity threshold, the SVM exhibited a sensitivity of 0.68 and a specificity of 0.68. Total operative time was the most important feature for reoperation risk prediction. CONCLUSION: The models may facilitate reoperation risk stratification, allowing for better patient counseling and for physicians to implement measures that reduce surgical risks.

Prevalence of gender-based and sexual harassment within orthopedic surgery in Canada.

BACKGROUND: Gender-based and sexual harassment are prevalent in the medical profession. We aimed to quantify the prevalence of such behaviours within orthopedic surgery in Canada and to identify any risk factors for experiencing gender-based or sexual harassment in the workplace. METHODS: In collaboration with the Canadian Orthopaedic Association, we conducted a Canada-wide email questionnaire survey in June 2019 of all orthopedic surgeons registered with the Canadian Orthopaedic Association and the Canadian Orthopaedic Residents' Association. The development of our questionnaire was informed by a review of the literature and published surveys on gender-based and sexual harassment, and consultation with researchers in intimate partner violence. We conducted a multivariable logistic regression analysis to identify risk factors for harassment. RESULTS: Of the 1783 surgeons invited to participate, 465 returned the questionnaire (response rate 26.1%); the response rate was 48.1% for females and 22.1% for males. Overall, 331/433 respondents (76.4%, 95% confidence interval [CI] 72%-80%) and 315/423 respondents (74.5%, 95% CI 70%-78%) reported having experienced at least 1 occurrence of gender-based and sexual harassment, respectively. Women were significantly more likely than men to have experienced both gender-based and sexual harassment (odds ratio [OR] 16.2, 95% CI 4.8-54.0, and OR 2.2, 95% CI 1.2-4.0, respectively). Respondents who identified as nonwhite were significantly less likely than those who identified as white to have experienced gender-based harassment (OR 0.5, 95% CI 0.3-0.99). CONCLUSION: The prevalence of gender-based and sexual harassment is high within Canadian orthopedic surgery, and women are at highest risk for experiencing harassment. The results may provide the impetus for orthopedic societies to develop action plans and to re-examine and enforce policies to address these damaging behaviours appropriately.

Short-term outcomes vary by surgical approach in total hip arthroplasty: a network meta-analysis.

BACKGROUND: The direct anterior approach (DAA) has increased in popularity in recent years. Proponents cite its muscle-sparing approach and purported reduction in pain and improvement in function when compared to the traditional surgical approaches. There remains controversy surrounding the validity of these claims. The objective of this study was to compare the common total hip surgical approaches in terms of pain scores, functional outcomes, opioid use and complications within the first 12 weeks postoperatively. METHODS: A network meta-analysis of randomized controlled trials (RCT) comparing postoperative outcomes of different surgical approaches in primary THA up to 12 weeks was performed. PubMed, MEDLINE, Embase, Web of Science and SCOPUS were systematically searched from inception to May 2020. Outcomes included pain scores, functional outcome scores, length of stay (LOS), complications and opioid consumption. RESULTS: Twenty-five RCTs (n = 2339) were included. The DAA demonstrated statistically significant improvement in Harris Hip Scores at 6 weeks when compared to the posterior and direct lateral approaches. The DAA reduced pain scores on postoperative day 2 and at 2 weeks compared to the direct lateral approach. The anterolateral approach was found to have a significantly shorter LOS compared to the other major surgical approaches. The differences in functional outcomes or pain scores did not surpass conventional cutoffs for a minimal clinically important difference. CONCLUSION: The DAA led to functional improvements at 6 weeks compared to the posterior and direct lateral approaches and reduced postoperative pain compared to the direct lateral approach. However, these improvements failed to reach clinical significance. All major surgical approaches led to large improvements in function by 12 weeks with relatively low complication rates. Whether a short-term statistically significant improvement in function is sufficiently patient important to recommend DAA as a standard remains uncertain.

Predicting survival after radical prostatectomy: Variation of machine learning performance by race.

BACKGROUND: Robust prediction of survival can facilitate clinical decision-making and patient counselling. Non-Caucasian males are underrepresented in most prostate cancer databases. We evaluated the variation in performance of a machine learning (ML) algorithm trained to predict survival after radical prostatectomy in race subgroups. METHODS: We used the National Cancer Database (NCDB) to identify patients undergoing radical prostatectomy between 2004 and 2016. We grouped patients by race into Caucasian, African-American, or non-Caucasian, non-African-American (NCNAA) subgroups. We trained an Extreme Gradient Boosting (XGBoost) classifier to predict 5-year survival in different training samples: naturally race-imbalanced, race-specific, and synthetically race-balanced. We evaluated performance in the test sets. RESULTS: A total of 68,630 patients met inclusion criteria. Of these, 57,635 (84%) were Caucasian, 8173 (12%) were African-American, and 2822 (4%) were NCNAA. For the classifier trained in the naturally race-imbalanced sample, the F1 scores were 0.514 (95% confidence interval: 0.513-0.511), 0.511 (0.511-0.512), 0.545 (0.541-0.548), and 0.378 (0.378-0.389) in the race-imbalanced, Caucasian, African-American, and NCNAA test samples, respectively. For all race subgroups, the F1 scores of classifiers trained in the race-specific or synthetically race-balanced samples demonstrated similar performance compared to training in the naturally race-imbalanced sample. CONCLUSIONS: A ML algorithm trained using NCDB data to predict survival after radical prostatectomy demonstrates variation in performance by race, regardless of whether the algorithm is trained in a naturally race-imbalanced, race-specific, or synthetically race-balanced sample. These results emphasize the importance of thoroughly evaluating ML algorithms in race subgroups before clinical deployment to avoid potential disparities in care.

What Is the Relative Effectiveness of the Various Surgical Treatment Options for Distal Radius Fractures? A Systematic Review and Network Meta-analysis of Randomized Controlled Trials.

BACKGROUND: Many acceptable treatment options exist for distal radius fractures (DRFs); however, a simultaneous comparison of all methods is difficult using conventional study designs. QUESTIONS/PURPOSES: We performed a network meta-analysis of randomized controlled trials (RCTs) on DRF treatment to answer the following questions: Compared with nonoperative treatment, (1) which intervention is associated with the best 1-year functional outcome? (2) Which intervention is associated with the lowest risk of overall complications? (3) Which intervention is associated with the lowest risk of complications requiring operation? METHODS: Ten databases were searched from inception to July 25, 2019. Search and analysis reporting adhered to Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines. Included studies were English-language RCTs that assessed at least one surgical treatment arm for adult patients with displaced DRFs, with less than 20% loss to follow-up. We excluded RCTs reporting on patients with open fractures, extensive bone loss, or ipsilateral upper extremity polytrauma. Seventy RCTs (n = 4789 patients) were included. Treatments compared were the volar locking plate, bridging external fixation, nonbridging external fixation, dynamic external fixation, percutaneous pinning, intramedullary fixation, dorsal plating, fragment-specific plating, and nonoperative treatment. Subgroup analyses were conducted for intraarticular fractures, extraarticular fractures, and patients with an average age greater than 60 years. Mean (range) patient age was 59 years (56 to 63) and was similar across all treatment groups except for dynamic external fixation (44 years) and fragment-specific plating (47 years). Distribution of intraarticular and extraarticular fractures was approximately equal among the treatment groups other than that for intramedullary fixation (73% extraarticular), fragment-specific plating (66% intraarticular) [13, 70], and dorsal plating (100% intraarticular). Outcomes were the DASH score at 1 year, total complications, and reoperation. The minimum clinically important different (MCID) for the DASH score was set at 10 points. The analysis was performed using Bayesian methodology with random-effects models. Rank orders were generated using surface under the cumulative ranking curve values. Evidence quality was assessed using Grades of Recommendation, Assessment, Development and Evaluation (GRADE) methodology. Most studies had a low risk of bias due to randomization and low rates of incomplete follow-up, unclear risk of bias due to selective reporting, and high risk of bias due to lack of patient and assessor blinding. Studies assessing bridging external fixation and/or nonoperative treatment arms had a higher overall risk of bias while studies with volar plating and/or percutaneous pinning treatment arms had a lower risk of bias. RESULTS: Across all patients, there were no clinically important differences in terms of the DASH score at 1 year; although differences were found, all were less than the MCID of 10 points. Volar plating was ranked the highest for DASH score at 1 year (mean difference -7.34 [95% credible interval -11 to -3.7) while intramedullary fixation, with low-quality evidence, also showed improvement in DASH score (mean difference -7.75 [95% CI -14.6 to -0.56]). The subgroup analysis revealed that only locked volar plating was favored over nonoperative treatment for patients older than 60 years of age (mean difference -6.4 [95% CI -11 to -2.1]) and for those with intraarticular fractures (mean difference -8.4 [95% CI -15 to -2.0]). However, its clinical importance was uncertain as the MCID was not met. Among all patients, intramedullary fixation (odds ratio 0.09 [95% CI 0.02 to 0.84]) and locked volar plating (OR 0.14 [95% CI 0.05 to 0.39]) were associated with a lower complication risk compared with nonoperative treatment. For intraarticular fractures, volar plating was the only treatment associated with a lower risk of complications than nonoperative treatment (OR 0.021 [95% CI < 0.01 to 0.50]). For extraarticular fractures, only nonbridging external fixation was associated with a lower risk of complications than nonoperative treatment (OR 0.011 [95% CI < 0.01 to 0.65]), although the quality of evidence was low. Among all patients, the risk of complications requiring operation was lower with intramedullary fixation (OR 0.06 [95% CI < 0.01 to 0.85) than with nonoperative treatment, but no treatment was favored over nonoperative treatment when analyzed by subgroups. CONCLUSION: We found no clinically important differences favoring any surgical treatment option with respect to 1-year functional outcome. However, relative to the other options, volar plating was associated with a lower complication risk, particularly in patients with intraarticular fractures, while nonbridging external fixation was associated with a lower complication risk in patients with extraarticular fractures. For patients older than 60 years of age, nonoperative treatment may still be the preferred option because there is no reliable evidence showing a consistent decrease in complications or complications requiring operation among the other treatment options. Particularly in this age group, the decision to expose patients to even a single surgery should be made with caution. LEVEL OF EVIDENCE: Level I, therapeutic study.

Which Bone-Modifying Agent is Associated with Better Outcomes in Patients with Skeletal Metastases from Lung Cancer? A Systematic Review and Network Meta-analysis.

BACKGROUND: Lung cancer is one of the most commonly diagnosed cancers and is the leading cause of cancer-related deaths. Metastatic bone disease occurs in 20% to 40% of patients with lung cancer, and these patients often present with pain or skeletal-related events (SREs) that are associated with decreased survival. Bone-modifying agents such as denosumab or bisphosphonates are routinely used; however, to our knowledge, there has been no quantitative synthesis of randomized controlled trial data to determine the most effective pharmacologic treatment of metastatic bone disease because of lung cancer. QUESTIONS/PURPOSES: We aimed to perform a network meta-analysis of randomized trials to identify the bone-modifying agent that is associated with the (1) highest overall survival, (2) longest time to SRE, (3) lowest SRE incidence, and (4) greatest likelihood of pain resolution. METHODS: We conducted our study according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses protocol and pre-registered the analysis on PROSPERO (ID: CRD42019124364). We performed a librarian-assisted search of MEDLINE, PubMed, EMBASE, Cochrane Library, and Chinese databases including China National Knowledge Infrastructure and Wanfang Data. We included randomized controlled trials reporting outcomes specifically for patients with lung cancer treated with a bisphosphonate or denosumab. SREs included pathologic fractures, spinal cord compression, hypercalcemia of malignancy, or pain resulting in surgical intervention or radiation therapy. We excluded trials exclusively reporting surrogate outcomes such as changes in bone turnover markers. Screening, data extraction, risk of bias evaluation, and Grading of Recommendations Assessment, Development, and Evaluation evaluations were performed in duplicate. We included 131 randomized controlled trials that evaluated 11,105 patients with skeletal metastases from lung cancer. The network meta-analysis was performed using a frequentist model and the R statistical software. Results are reported as relative risks or mean differences, and the I2 value is reported for heterogeneity. The P-score, a measure of ranking certainty that accounts for standard error, is reported for each outcome. Heterogeneity in the network was considered moderate for overall survival and time to SRE, mild for the incidence of SRE, and low for pain resolution. RESULTS: For overall survival, denosumab was ranked above zoledronic acid and estimated to confer a mean of 3.3 months (95% CI 0.3-6.3) of increased overall survival compared with untreated patients (P-score = 89%). For the time to SRE, denosumab was ranked first with a mean of 9.1 additional SRE-free months (95% CI 6.7-11.5) compared with untreated patients (P-score = 99%), while zoledronic acid conferred an additional 4.8 SRE-free months (95% CI 3.6-6.1). Reduction in the incidence of SREs was not different between patients treated with denosumab (relative risk 0.54; 95% CI 0.33-0.87) and those treated with zoledronic acid (relative risk 0.56; 95% CI 0.46-0.67). Patients treated with the combination of ibandronate and systemic therapy were more likely to experience successful pain resolution than untreated patients (relative risk 2.4; 95% CI 1.8-3.2). CONCLUSION: In this comprehensive synthesis of all available randomized controlled trial evidence guiding the pharmacologic treatment of bone metastases from lung cancer, denosumab was ranked above zoledronic acid for overall survival and time to SRE and was not different for reducing the incidence of SRE. Both were superior to no treatment for each of these outcomes. Given this, we encourage physicians to consider the use of denosumab or zoledronic acid in treating this patient population. The combination of ibandronate and systemic therapy was the most effective at reducing pain because of metastases. No cost-effectiveness analysis has yet been performed for denosumab and zoledronic acid on patients with metastatic lung cancer, and this represents an avenue for future research. LEVEL OF EVIDENCE: Level I, therapeutic study.

Intra-articular saline injection is as effective as corticosteroids, platelet-rich plasma and hyaluronic acid for hip osteoarthritis pain: a systematic review and network meta-analysis of randomised controlled trials.

OBJECTIVE: Intra-articular (IA) injections represent a commonly used modality in the treatment of hip osteoarthritis (OA). Commonly used injections include corticosteroids (CCS), hyaluronic acid (HA) and platelet-rich plasma (PRP). A network meta-analysis allows for comparison among more than two treatment arms and uses both direct and indirect comparisons between interventions. The objective of this network meta-analysis is to compare the efficacy of the various IA injectable treatments in treating hip OA at up to 6 months of follow-up. DESIGN: This is a systematic review and network meta-analysis. Bayesian random-effects model was performed to assess the direct and indirect comparisons of all treatment options. DATA SOURCES: PubMed, Embase, Cochrane Central Register of Controlled Trials, Scopus and Web of Science, from inception to October 2019. ELIGIBILITY CRITERIA FOR SELECTED STUDIES: Randomised controlled trials assessing the efficacy of CCS, HA, PRP and placebo in the form of IA saline injection for patients with hip OA. RESULTS: Eleven randomised controlled trials comprising 1353 patients were included. For pain outcomes at both 2-4 and 6 months, no intervention significantly outperformed placebo IA injection. For functional outcomes at both 2-4 and 6 months, no intervention significantly outperformed placebo IA injection. Regarding change from baseline at 2-4 months and 6 months, pooled data demonstrated that all interventions (including placebo), with the exception of HA+PRP, led to a clinically important improvement in both pain, exceeding the minimal clinically important difference. CONCLUSION: Evidence suggests that IA hip saline injections performed as well as all other injectable options in the management of hip pain and functional outcomes.

What Is the Best Evidence for Management of Displaced Midshaft Clavicle Fractures? A Systematic Review and Network Meta-analysis of 22 Randomized Controlled Trials.

BACKGROUND: Displaced mid-third clavicle fractures are common, and their management remains unclear. Although several meta-analyses have compared specific operative techniques with nonoperative management, it is not possible to compare different operative constructs with one another using a standard meta-analysis. Conversely, a network meta-analysis allows comparisons among more than two treatment arms, using both direct and indirect comparisons between interventions across many trials. To our knowledge, no network meta-analysis has been performed to compare the multiple treatment options for displaced clavicle fractures. QUESTIONS/PURPOSES: We performed a network meta-analysis of randomized, controlled trials (RCTs) to determine from among the approaches used to treat displaced midshaft clavicle fractures: (1) the intervention with the highest chance of union at 1 year, (2) the intervention with the lowest risk of revision surgery, and (3) the intervention with the highest functional outcome scores. Secondarily, we also (4) compared the surgical subtypes in the available RCTs on the same above endpoints. METHODS: MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials were reviewed for relevant randomized controlled trials published up to July 25, 2018. Two hundred and eighty four papers were reviewed, with 22 meeting inclusion criteria of RCTs with appropriate randomization techniques, adult population, minimum of 1 year follow-up and including at least one operative treatment arm. In total, 1002 patients were treated with a plate construct, 378 with an intramedullary device, and 585 patients were managed nonoperatively. Treatment subtypes included locked intramedullary devices (56), unlocked intramedullary devices (322), anterior plating (89), anterosuperior plating (150), superior plating (449) or plating not otherwise specified (314). We performed a network meta-analysis to compare and rank the treatments for displaced clavicle fractures. We considered the following outcomes: union achievement, revision surgery risk and functional outcomes (DASH and Constant Scores). The minimal clinically important difference (MCID) was considered for both Constant and DASH scores to be at 8 points, representing the average of MCID scores reported for both DASH and Constant in the evidence, respectively. RESULTS: Union achievement was lower in patients treated nonoperatively (88.9%), and higher in patients treated operatively (96.7%, relative risk [RR] 1.128 [95% CI 1.1 to 1.17]; p < 0.001), Number needed to treat (NNT) = 10). Union achievement increased with any plate construct (97.8%, RR 1.13 [95% CI 1.1 to 1.7]; p < 0.0001, NNT = 9) and with anterior or anterosuperior plates (99.3%, RR 1.14 [95% CI 1.1 to 1.8]; p < 0.0001, NNT = 8). Risk of reoperation, when considering planned removal of hardware, was similar across all treatment arms. Lastly, operative treatment outperformed nonoperative treatment with minor improvements in DASH and Constant scores, though not approaching the MCID. At the subtype level, anterosuperior plating ranked highest in DASH and Constant functional scores with mean differences reaching 10-point improvement for Constant scores (95% CI 4.4 to 2.5) and 7.6 point improvement for DASH (95% CI 5.2 to 20). CONCLUSIONS: We found that surgical treatment led to a greater likelihood of union at 1 year of follow-up among adult patients with displaced mid-third clavicle fractures. In aggregate, surgical treatment did not increase functional scores by amounts that patients were likely to consider clinically important. Use of specific subtypes of plating (anterior, anterosuperior) resulted in improvements in the Constant score that were slightly above the MCID but did not reach the MCID for the DASH score, suggesting that any outcomes-score benefits favoring surgery were likely to be imperceptible or small. In light of these findings, we believe patients can be informed that surgery for this injury can increase the likelihood of union incrementally (about 10 patients would need to undergo surgery to avoid one nonunion), but they should not expect better function than they would achieve without surgery; most patients can avoid surgery altogether with little absolute risk of nonunion. Patients who opt for surgery must be told that the decision should be weighed against complications and the possibility of undergoing a second procedure for hardware removal. Patients opting not to have surgery for acute midshaft clavicle fractures can be told that nonunion occurs in slightly more than 10% of patients, and that these can be more difficult to manage than acute fractures. LEVEL OF EVIDENCE: Level I, therapeutic study.

Deep Learning and Multimodal Artificial Intelligence in Orthopaedic Surgery.

This review article focuses on the applications of deep learning with neural networks and multimodal neural networks in the orthopaedic domain. By providing practical examples of how artificial intelligence (AI) is being applied successfully in orthopaedic surgery, particularly in the realm of imaging data sets and the integration of clinical data, this study aims to provide orthopaedic surgeons with the necessary tools to not only evaluate existing literature but also to consider AI's potential in their own clinical or research pursuits. We first review standard deep neural networks which can analyze numerical clinical variables, then describe convolutional neural networks which can analyze image data, and then introduce multimodal AI models which analyze various types of different data. Then, we contrast these deep learning techniques with related but more limited techniques such as radiomics, describe how to interpret deep learning studies, and how to initiate such studies at your institution. Ultimately, by empowering orthopaedic surgeons with the knowledge and know-how of deep learning, this review aspires to facilitate the translation of research into clinical practice, thereby enhancing the efficacy and precision of real-world orthopaedic care for patients.

Distraction Osteogenesis Reconstruction Following Resection of Bone Sarcomas: Surgical, Functional, and Oncological Outcomes from a Prospective Trial Analysis.

BACKGROUND: While sustainable long-term function has been established for biological reconstruction with distraction osteogenesis (DO) following osseous resections, there is a paucity of published data informing surgeons and patients on important milestones in the reconstructive process. The objectives of this study were to determine when to expect complete bone healing and full weight-bearing as well as to quantify the influence of chemotherapy on the osseous regeneration process. METHODS: Prospectively, pathological and clinical data were collected for 30 consecutive patients who underwent primary or secondary DO-based reconstruction following osseous resection from 2018 to 2021. Serial radiographs indicated the times to cortex formation and full union. An unpaired t test was used to compare the time required for full bone remodeling of segments transported with and without concurrent chemotherapy. RESULTS: The average resection length was 13.6 cm (range, 4 to 22 cm). Patients underwent an average of 6.1 procedures (range, 1 to 14 procedures). Half (50%) of all procedures were planned, while half were unplanned procedures. All patients achieved full, independent weight-bearing at a median of 12 months (interquartile range [IQR], 9 to 16 months). For the 34 segments transported concurrently with chemotherapy, the mean bone healing index (BHI) was 2.3 ± 0.7, and the mean BHI was 1.2 ± 0.4 for the 25 segments without chemotherapy at any point during their transport (p < 0.0001). CONCLUSIONS: All 30 patients achieved full bone healing and independent weight-bearing at a median of 1 year postoperatively and continued to show functional improvement afterward. Surgeons and patients can expect bone healing to be nearly twice as fast for segments transported after completion of systemic chemotherapy compared with segments transported concurrently with adjuvant chemotherapy. LEVEL OF EVIDENCE: Therapeutic Level II . See Instructions for Authors for a complete description of levels of evidence.

The Impact of Adjunct Medical Therapy on Survival after Spine Metastasis: A Systematic Review and Pooled Data Analysis.

Targeted therapy has greatly improved the outlook for patients with spinal metastatic cancers. Scoring systems like the Tokuhashi or Tomita scores are commonly used to predict prognosis and inform surgical decisions, but they are outdated and fail to consider recent advancements. We aimed to investigate the current state of the literature and treatment options pertaining to advancements in targeted therapy compared to other forms of medical management for metastatic spinal tumors. This study represents the first comprehensive systematic review that encompasses the most common primary cancers that metastasize to the spine and evaluates the median overall survival (mOS) across five different medical treatment modalities as well as surgical intervention. Additionally, our study analyzes the tumor receptor status in conjunction with these treatments. A PubMed search was conducted, and according to the PRISMA guidelines, 28 articles out of 1834 met the inclusion criteria. The pooled data analysis highlighted the superior efficacy of targeted therapy, evidenced by a significant improvement in the mOS and lower hazard ratios in patients with lung and breast cancers who received targeted therapy compared to those who did not. Our study provides valuable insights into the recent advancements in the medical management of metastatic spinal tumors. Future indications include incorporating this literature into personalized treatment approaches for metastatic spinal tumors.

A multimodal neural network with gradient blending improves predictions of survival and metastasis in sarcoma.

The objective of this study is to develop a multimodal neural network (MMNN) model that analyzes clinical variables and MRI images of a soft tissue sarcoma (STS) patient, to predict overall survival and risk of distant metastases. We compare the performance of this MMNN to models based on clinical variables alone, radiomics models, and an unimodal neural network. We include patients aged 18 or older with biopsy-proven STS who underwent primary resection between January 1st, 2005, and December 31st, 2020 with complete outcome data and a pre-treatment MRI with both a T1 post-contrast sequence and a T2 fat-sat sequence available. A total of 9380 MRI slices containing sarcomas from 287 patients are available. Our MMNN accepts the entire 3D sarcoma volume from T1 and T2 MRIs and clinical variables. Gradient blending allows the clinical and image sub-networks to optimally converge without overfitting. Heat maps were generated to visualize the salient image features. Our MMNN outperformed all other models in predicting overall survival and the risk of distant metastases. The C-Index of our MMNN for overall survival is 0.77 and the C-Index for risk of distant metastases is 0.70. The provided heat maps demonstrate areas of sarcomas deemed most salient for predictions. Our multimodal neural network with gradient blending improves predictions of overall survival and risk of distant metastases in patients with soft tissue sarcoma. Future work enabling accurate subtype-specific predictions will likely utilize similar end-to-end multimodal neural network architecture and require prospective curation of high-quality data, the inclusion of genomic data, and the involvement of multiple centers through federated learning.

Operative Treatment and Outcomes of Pediatric Patients with an Extremity Bone Tumor: A Secondary Analysis of the PARITY Trial Data.

BACKGROUND: Osteosarcoma and Ewing sarcoma are the 2 most common primary bone sarcomas, occurring predominantly in pediatric patients, with the incidence of osteosarcoma correlating with periods of peak bone-growth velocity. Although survival outcomes have plateaued over the past several decades, ongoing treatment advances have improved function, decreased infection rates, and improved other clinical outcomes in patients with bone tumors. Recently, the Prophylactic Antibiotic Regimens in Tumor Surgery (PARITY) trial addressed the serious problem of surgical site infection (SSI) and the lack of consensus regarding the appropriate prophylactic postoperative antibiotic regimen. The objective of the present secondary analysis of the PARITY trial was to characterize the modern treatment and surgical and oncologic outcomes of pediatric patients with bone tumors at 1 year postoperatively. METHODS: The PARITY trial included patients ≥12 years old with a bone tumor or soft-tissue sarcoma that was invading the femur or tibia, necessitating osseous resection and endoprosthetic reconstruction. This pediatric subanalysis of the PARITY trial data included all PARITY patients ≤18 years old. As in the main PARITY study, patients were randomized to either a 5-day or 1-day course of postoperative antibiotic prophylaxis. The primary outcome measure was the development of an SSI within 1 year, and secondary outcomes included antibiotic-related adverse events, unplanned additional operations, local recurrence, metastasis, and death. RESULTS: A total of 151 patients were included. An adjudicated SSI occurred in 27 patients (17.9%). There was no difference in the rate of any SSI between the 5-day and 1-day antibiotic groups (hazard ratio [HR], 0.92; 95% confidence interval [CI], 0.4 to 1.9; p = 0.82). Antibiotic-related complications occurred in 13 patients (8.6%), with no difference noted between groups (HR, 0.46; 95% CI, 0.2 to 1.4; p = 0.18). A total of 45 patients (29.8%) required a return to the operating room within the first postoperative year, which corresponded with a 68.8% reoperation-free rate of survival at 1 year when accounting for competing risks. The most common reason for reoperation was infection (29 of 45; 64.4%). A total of 7 patients (4.6%) required subsequent amputation of the operative extremity, and an additional 6 patients (4.0%) required implant revision within 12 months. A total of 36 patients (23.8%) developed metastases, and 6 patients (4.0%) developed a local recurrence during the first postoperative year. A total of 11 patients (7.3%) died during the study period. There were no significant differences in oncologic outcomes between the 5-day and 1-day antibiotic groups (HR, 0.97; 95% CI, 0.5-1.8; p = 0.92). CONCLUSIONS: There were no significant differences in surgical or oncologic outcomes between pediatric patients who underwent a 1-day versus 5-day antibiotic regimen following endoprosthetic reconstruction in the PARITY trial. Surgeons should be aware of and counsel patients and caregivers regarding the 30% rate of reoperation and the risks of infection (17.9%), death (7.3%), amputation (4.6%), and implant revision (4%) within the first postoperative year. LEVEL OF EVIDENCE: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.

Perpetrators of Gender-Based and Sexual Harassment in the Field of Orthopaedic Surgery.

The prevalence of gender-based and sexual harassment in the field of orthopaedic surgery in Canada is high. Previous research in other jurisdictions has identified the most common perpetrators of harassment to be senior surgeons or directors. We aimed to identify the most frequent perpetrators of gender-based and sexual harassment in orthopaedic surgery in Canada. METHODS: We conducted a Canada-wide survey of all orthopaedic surgeons registered with the Canadian Orthopaedic Association and the Canadian Orthopaedic Residents' Association. The development of our 116-item questionnaire was informed by a review of the literature and other published gender-based and sexual harassment surveys. Descriptive analyses, including frequency counts with associated 95% confidence intervals (CIs), are reported for all data. RESULTS: Of the 465 survey respondents, the median age was 43 years (interquartile range, 35 to 59) and respondents were most commonly male (72%), White (81%), married (77%), and staff orthopaedic surgeons (68%). Peers were identified as the most common perpetrators of gender-based harassment (55%, 95% CI, 50 to 59), and patients were identified as the most common perpetrators of sexual harassment (48%, 95% CI, 43 to 52). Women were more likely to report direct supervisors or patients as the perpetrators of gender-based and sexual harassment, and men reported peers as the most common perpetrators. CONCLUSION: Orthopaedic surgery peers and patients are the most commonly reported perpetrators of gender-based and sexual harassment in Canada. The results of this study may be helpful to institutions in designing and focusing educational programs and/or policies and procedures to help reduce harassment incidents in the training and work environment.

A machine learning approach to predict progression on active surveillance for prostate cancer.

PURPOSE: Robust prediction of progression on active surveillance (AS) for prostate cancer can allow for risk-adapted protocols. To date, models predicting progression on AS have invariably used traditional statistical approaches. We sought to evaluate whether a machine learning (ML) approach could improve prediction of progression on AS. PATIENTS AND METHODS: We performed a retrospective cohort study of patients diagnosed with very-low or low-risk prostate cancer between 1997 and 2016 and managed with AS at our institution. In the training set, we trained a traditional logistic regression (T-LR) classifier, and alternate ML classifiers (support vector machine, random forest, a fully connected artificial neural network, and ML-LR) to predict grade-progression. We evaluated model performance in the test set. The primary performance metric was the F1 score. RESULTS: Our cohort included 790 patients. With a median follow-up of 6.29 years, 234 developed grade-progression. In descending order, the F1 scores were: support vector machine 0.586 (95% CI 0.579 - 0.591), ML-LR 0.522 (95% CI 0.513 - 0.526), artificial neural network 0.392 (95% CI 0.379 - 0.396), random forest 0.376 (95% CI 0.364 - 0.380), and T-LR 0.182 (95% CI 0.151 - 0.185). All alternate ML models had a significantly higher F1 score than the T-LR model (all p <0.001). CONCLUSION: In our study, ML methods significantly outperformed T-LR in predicting progression on AS for prostate cancer. While our specific models require further validation, we anticipate that a ML approach will help produce robust prediction models that will facilitate individualized risk-stratification in prostate cancer AS.

Recruitment patterns in a large international randomized controlled trial of perioperative care in cancer patients.

INTRODUCTION: The Prophylactic Antibiotic Regimens in Tumor Surgery (PARITY) randomized controlled trial (RCT) was the first study to prospectively enroll and randomize orthopedic oncology patients in multiple centers internationally. The objective of this study was to describe recruitment patterns, to examine the differences in enrollment across different PARITY sites, and to identify variables associated with differing levels of recruitment. METHODS: Data from this study was obtained from the PARITY trial Methods Center and records of correspondence between the Methods Center and recruiting sites. We performed descriptive statistics to report the recruitment patterns over time. We compared recruitment, time to set up, and time to enroll the first patient between North American and international sites, private and public healthcare models, and the presence or absence of research personnel. Two-tailed non-paired t tests were performed to test average monthly recruitment rates between groups. RESULTS: A total of 602 patients from 36 North American and 12 international sites were recruited from 2013 to 2019. North American sites were able to become fully enrollment-ready at an average of 19.5 months and international sites at an average of 27 months. Once enrolling, international sites were able to enroll 0.59 patients per/month whereas North American sites averaged a monthly recruitment rate of 0.2 patients/month once enrolling. Sites with research personnel reached enrollment-ready status at an average of 19.3 months and sites without research support at an average of 30.3 months. Once enrolling, the recruitment rate was 0.28 patients/month and 0.2 patients per month for sites with and without research support, respectively. Publicly funded sites had a monthly enrollment of 0.4 patients/month whereas privately funded sites had a monthly enrollment rate of 0.17 patients/month. CONCLUSIONS: As a collaborative group, the PARITY investigators increased the pace of recruitment throughout the trial, likely by increasing the number of active sites. The longer time to start-up at international sites may be due to the complex governing regulations of pharmaceutical trials. Nevertheless, international sites should be considered essential as they recruited significantly more patients per month once active. The absence of research support personnel may lead to delays in the time to start-up. The results of the current study will provide guidance for choosing which sites to recruit for participation in future collaborative clinical trials in orthopedic oncology and other surgical specialties. TRIAL REGISTRATION: ClinicalTrials.gov NCT01479283 . Prospectively registered on November 24, 2011.

Unicompartmental Knee Arthroplasty: Survivorship and Risk Factors for Revision: A Population-Based Cohort Study with Minimum 10-Year Follow-up.

BACKGROUND: Operative management for end-stage knee osteoarthritis (OA) primarily consists of arthroplasty. Total knee arthroplasty (TKA) is the so-called gold standard for multicompartmental OA. In selected patients, specifically those with unicompartmental OA and no definite contraindications, realignment osteotomies and unicompartmental knee arthroplasty (UKA) are viable options. UKA offers several advantages over TKA, including preservation of both cruciate ligaments, faster recovery, and less osseous resection. The purpose of this study was to determine the survivorship of UKAs and risk factors for all-cause revision surgery in patients with a minimum follow-up of 10 years in a large population-based database. METHODS: We performed a population-based cohort study using linked administrative databases in Ontario, Canada. We identified all UKAs performed from January 2002 through December 2006. All patients had a minimum 10-year follow-up. Demographic data and outcomes were summarized using descriptive statistics. We used a Cox proportional hazards model with the Fine and Gray method accounting for competing risks such as death to analyze the effect of surgical and patient factors on the risk of revision of UKA. RESULTS: A total of 4,385 patients were identified as having undergone primary UKA from 2002 through 2006 in Ontario. Of these patients, 779 underwent revision surgery, for a cumulative risk for all-cause revision of 16.5% at 10 years (95% confidence interval [CI]: 15.4% to 17.7%). Mechanical loosening was the reason for 83.4% of all revisions. Male sex, diabetes, an age younger than 50 years, and cementless implants were significant risk factors for revision surgery. Residence in a rural or urban setting, income quintile, discharge destination, and type of OA were not associated with the risk of revision. Survivorship at 1, 5, 10, and 15 years was 97.2%, 90.5%, 83.5%, and 81.9%, respectively. CONCLUSIONS: This study demonstrated a >80% long-term survivorship for UKA based on a large administrative database sample and identified important risk factors for failure including male sex, younger age, diabetes, and cementless fixation. This study found rates of survivorship similar to those previously reported in the literature, while identifying key risk factors for failure. Patient selection has always been identified as a key factor in ensuring successful UKA, and the risk factors identified in this study may help guide decision-making for surgeons and patients dealing with unicompartmental knee OA. LEVEL OF EVIDENCE: Therapeutic Level IV. See Instructions for Authors for a complete description of levels of evidence.

Primary sarcomas of the spine: population-based demographic and survival data in 107 spinal sarcomas over a 23-year period in Ontario, Canada.

BACKGROUND CONTEXT: Spinal sarcomas are a rare, heterogeneous group of mesenchymal tumors. Current literature reporting demographic variables and survival information is limited to small case series, and a single registry with variable treatment modalities and time periods. PURPOSE: We report on population-level data regarding all spinal sarcomas diagnosed over a 23-year period in Ontario, Canada, for the purposes of calculating incidence and prevalence of these tumors. Secondarily, survival is assessed by tumor type as well as adjuvant therapies during this time period. STUDY DESIGN: Retrospective Cohort Study PATIENT SAMPLE: Population-based data from the Institute for Clinical Evaluative Sciences (ICES) between 1993 and 2015. OUTCOME MEASURES: Outcome measures include incidence and prevalence of spinal osteosarcoma, Ewing's sarcoma, and chondrosarcoma of the spine, as well as 2-, 5-, 10- and 15-year survival and prevalence of adjuvant therapies. METHODS: Utilizing population-based data from the Institute for Clinical Evaluative Sciences (ICES) between 1993 and 2015, ICD codes were searched and available data extracted for the purposes of reporting basic demographic information and calculation of Kaplan Meyer survival curves. Databases include the Ontario Cancer Registry, Discharge Abstract Database, Ontario Health Insurance Plan, National Ambulatory Care Reporting System, Registered Persons DataBase (death) were analyzed. RESULTS: One hundred and seven spinal sarcomas were identified, with a mean incidence was 0.38 sarcomas per million population per year, that was stable over time. The mean prevalence was 8.1 sarcomas per million population. The most common diagnosis was Ewing's sarcoma (48 [44.9%] patients), followed by chondrosarcoma (33 [30.8%] patients), and osteosarcoma (26 [24.3%] patients). Chondrosarcoma had the highest survival rates with 77.2% and 64.2% 5- and 10-year survival rates, respectively, followed by Ewing's sarcoma with 48.1% and 44.9% 5 and 10-year survival and osteosarcoma with 36.0% and 30.9% 5- and 10-year survival. CONCLUSIONS: Spinal sarcoma is a rare disease with variable survival depending on the histologic diagnosis. This population-level study involves a heterogeneous group of patients with variable stages of disease at presentation and variable treatments. Our data fit with the published literature for survival for those treated conservatively and surgically. Our data show considerable improvement in 5- and 10-year mortality when compared with previous population level studies on earlier patient cohorts, likely reflecting improvements in systemic and surgical treatments.