Advances in cancer therapy have prolonged patient survival even in the presence of disseminated disease and an increasing number of cancer patients are living with metastatic bone disease (MBD). The proximal femur is the most common long bone involved in MBD and pathologic fractures of the femur are associated with significant morbidity, mortality and loss of quality of life (QoL). Successful prophylactic surgery for an impending fracture of the proximal femur has been shown in multiple cohort studies to result in longer survival, preserved mobility, lower transfusion rates and shorter post-operative hospital stays. However, there is currently no optimal method to predict a pathologic fracture. The most well-known tool is Mirel's criteria, established in 1989 and is limited from guiding clinical practice due to poor specificity and sensitivity. The ideal clinical decision support tool will be of the highest sensitivity and specificity, non-invasive, generalizable to all patients, and not a burden on hospital resources or the patient's time. Our research uses novel machine learning techniques to develop a model to fill this considerable gap in the treatment pathway of MBD of the femur. The goal of our study is to train a convolutional neural network (CNN) to predict fracture risk when metastatic bone disease is present in the proximal femur. Our fracture risk prediction tool was developed by analysis of prospectively collected data of consecutive MBD patients presenting from 2009–2016. Patients with primary bone tumors, pathologic fractures at initial presentation, and hematologic malignancies were excluded. A total of 546 patients comprising 114 pathologic fractures were included. Every patient had at least one Anterior-Posterior X-ray and clinical data including patient demographics, Mirel's criteria, tumor biology, all previous radiation and chemotherapy received, multiple pain and function scores, medications and time to fracture or time to death. We have trained a convolutional neural network (CNN) with AP X-ray images of 546 patients with metastatic bone disease of the proximal femur. The digital X-ray data is converted into a matrix representing the color information at each pixel. Our CNN contains five convolutional layers, a fully connected layers of 512 units and a final output layer. As the information passes through successive levels of the network, higher level features are abstracted from the data. The model converges on two fully connected deep neural network layers that output the risk of fracture. This prediction is compared to the true outcome, and any errors are back-propagated through the network to accordingly adjust the weights between connections, until overall prediction accuracy is optimized. Methods to improve learning included using stochastic gradient descent with a learning rate of 0.01 and a momentum rate of 0.9. We used average classification accuracy and the average F1 score across five test sets to measure model performance. We compute F1 = 2 x (precision x recall)/(precision + recall). F1 is a measure of a model's accuracy in binary classification, in our case, whether a lesion would result in pathologic fracture or not. Our model achieved 88.2% accuracy in predicting fracture risk across five-fold cross validation testing. The F1 statistic is 0.87. This is the first reported application of convolutional neural networks, a machine learning algorithm, to this important Orthopaedic problem. Our neural network model was able to achieve reasonable accuracy in classifying fracture risk of metastatic proximal femur lesions from analysis of X-rays and clinical information. Our future work will aim to externally validate this algorithm on an international cohort

Advances in cancer therapy have prolonged cancer patient survival even in the presence of disseminated disease and an increasing number of cancer patients are living with metastatic bone disease (MBD). The proximal femur is the most common long bone involved in MBD and pathologic fractures of the femur are associated with significant morbidity, mortality and loss of quality of life (QoL). Successful prophylactic surgery for an impending fracture of the proximal femur has been shown in multiple cohort studies to result in patients more likely to walk after surgery, longer survival, lower transfusion rates and shorter post-operative hospital stays. However, there is currently no optimal method to predict a pathologic fracture. The most well-known tool is Mirel's criteria, established in 1989 and is limited from guiding clinical practice due to poor specificity and sensitivity. The goal of our study is to train a convolutional neural network (CNN) to predict fracture risk when metastatic bone disease is present in the proximal femur. Our fracture risk prediction tool was developed by analysis of prospectively collected data for MBD patients (2009–2016) in order to determine which features are most commonly associated with fracture. Patients with primary bone tumors, pathologic fractures at initial presentation, and hematologic malignancies were excluded. A total of 1146 patients comprising 224 pathologic fractures were included. Every patient had at least one Anterior-Posterior X-ray. The clinical data includes patient demographics, tumor biology, all previous radiation and chemotherapy received, multiple pain and function scores, medications and time to fracture or time to death. Each of Mirel's criteria has been further subdivided and recorded for each lesion. We have trained a convolutional neural network (CNN) with X-ray images of 1146 patients with metastatic bone disease of the proximal femur. The digital X-ray data is converted into a matrix representing the color information at each pixel. Our CNN contains five convolutional layers, a fully connected layers of 512 units and a final output layer. As the information passes through successive levels of the network, higher level features are abstracted from the data. This model converges on two fully connected deep neural network layers that output the fracture risk. This prediction is compared to the true outcome, and any errors are back-propagated through the network to accordingly adjust the weights between connections. Methods to improve learning included using stochastic gradient descent with a learning rate of 0.01 and a momentum rate of 0.9. We used average classification accuracy and the average F1 score across test sets to measure model performance. We compute F1 = 2 x (precision x recall)/(precision + recall). F1 is a measure of a test's accuracy in binary classification, in our case, whether a lesion would result in pathologic fracture or not. Five-fold cross validation testing of our fully trained model revealed accurate classification for 88.2% of patients with metastatic bone disease of the proximal femur. The F1 statistic is 0.87. This represents a 24% error reduction from using Mirel's criteria alone to classify the risk of fracture in this cohort. This is the first reported application of convolutional neural networks, a machine learning algorithm, to an important Orthopaedic problem. Our neural network model was able to achieve impressive accuracy in classifying fracture risk of metastatic proximal femur lesions from analysis of X-rays and clinical information. Our future work will aim to validate this algorithm on an external cohort

The presence of metastatic bone disease (MBD) often necessitates major orthopaedic surgery. Patients will enter surgical care either through emergent or electively scheduled care pathways. Patients in a pain crisis or with an acute fracture are generally admitted via emergent care pathways whereas patients with identified high-risk bone lesions are often booked for urgent yet scheduled elective procedures. The purpose of this study is to compare the post-operative outcomes of patients who present through emergent or electively scheduled care pathways in patients in a Canadian health care system. We have conducted a retrospective, multicenter cohort study of all patients presenting for surgery for MBD of the femur, humerus, tibia or pelvis in southern Alberta between 2006 and 2021. Patients were identified by a search query of all patients with a diagnosis of metastatic cancer who underwent surgery for an impending or actual pathologic fracture in the Calgary, South and Central Alberta Zones. Subsequent chart reviews were performed. Emergent surgeries were defined by patients admitted to hospital via urgent care mechanisms and managed via unscheduled surgical bookings (“on call list”). Elective surgeries were defined by patients seen by an orthopaedic surgeon at least once prior to surgery, and booked for a scheduled urgent, yet elective procedure. Outcomes include overall survival from the time of surgery, hospital length of stay, and 30-day hospital readmission rate. We have identified 402 patients to date for inclusion. 273 patients (67.9%) underwent surgery through emergent pathways and 129 patients (32.1%) were treated through urgent, electively scheduled pathways. Lung, prostate, renal cell, and breast cancer were the most common primary malignancies and there was no significant difference in these primaries amongst the groups (p=0.06). Not surprisingly, emergent patients were more likely to be treated for a pathologic fracture (p<0.001) whereas elective patients were more likely to be treated for an impending fracture (p<0.001). Overall survival was significantly shorter in the emergent group (5.0 months, 95%CI: 4.0-6.1) compared to the elective group (14.9 months 95%CI: 10.4-24.6) [p<0.001]. Hospital length of stay was significantly longer in the emergent group (13 days, 95%CI: 12-16 versus 5 days, 95%CI: 5-7 days). There was a significantly greater rate of 30-day hospital readmission in the emergent group (13.3% versus 7.8%) [p=0.01]. Electively managed MBD has multiple benefits including longer post-operative survival, shorter length of hospital stay, and a lower rate of 30-day hospital readmission. These findings from a Canadian healthcare system demonstrate clinical value in providing elective orthopaedic care when possible for patients with MBD. Furthermore, care delivery interventions capable of decreasing the footprint of emergent surgery through enhanced screening or follow-up of patients with MBD has the potential to significantly improve clinical outcomes in this population. This is an ongoing study that will justify refinements to the current surgical care pathways for MBD in order to identify patients prior to emergent presentation. Future directions will evaluate the costs associated with each care delivery method to provide opportunity for health economic efficiencies

Metastatic bone disease (MBD) is a significant contributor to diminished quality of life in cancer patients, often leading to pathologic fractures, hypercalcaemia, intractable bone pain, and reduced functional independence. Standard of care management for MBD patients undergoing orthopaedic surgery is multi-disciplinary, includes regular surgical follow-up, case by case assessment for use of bone protective medications, and post-operative radiation therapy to the operative site. The number of patients in southern Alberta receiving standard of care post-operative management is currently unclear. Our aim is to develop a database of all patients in southern Alberta undergoing orthopaedic surgery for MBD and to assess for deficiencies and opportunities to ensure standard of care for this complex patient population. Patients were identified for database inclusion by a search query of the Alberta Cancer Registry of all patients with a diagnosis of metastatic cancer who underwent surgery for an impending or pathologic fracture in the Calgary, South and Central Alberta Zones. Demographic information, primary cancer history, previous local and systemic treatments, anatomical location of MBD event(s), surgical fixation techniques, and post-operative care details were collected. The rate of standard of care post-operative treatment was evaluated. A comparison of outcomes between tertiary urban centres and rural centres was also completed. Survival was calculated from time of first operation to date of death. Univariate and multivariate analyses were performed to identify the impact of post-operative care variables on survival amongst patients surviving longer than one month. We identified 402 patients who have undergone surgical treatment for MBD in southern Alberta from 2006-2018. Median age at time of surgery was 66.3 years and 52.7% of patients were female. Breast, lung, prostate, renal cell and multiple myeloma were the most common primary malignancies (n=328, 81.6%). Median post-operative survival was 6.8 months (95%CI: 5.7-8.3). 203 patients (52.5%) were treated with post-operative radiotherapy and 159 patients (50.8%) had post-operative surgical follow-up. Only 39 patients (11.3%) received bone protective agents in the peri-operative period. On multivariate survival analysis, post-operative surgical follow-up was associated with improved survival (p<0.001). Patients were treated at nine hospitals across southern Alberta with most patients treated in an urban center (65.9%). Post-operative survival was significantly longer amongst patients treated in an urban center (9.0 months, 95%CI: 6.9-12.3 versus 4.3 months, 95%CI: 3.4-5.6, p<0.001). The burden of MBD is significant and increasing. With treatment occurring at multiple provincial sites, there is a need for standardized, primary disease-specific peri- and post-operative protocols to ensure quality and efficacious patient care. To provide evidence informed treatment recommendations, we have developed a database of all patients in southern Alberta undergoing orthopaedic surgery for MBD. Our results demonstrate that many patients were not treated according to post-operative standard of care recommendations. Notably, half of the included patients did not have documented surgical follow-up, post-operative radiation treatment was low and only 11% were actively treated with bone protective agents. This data justifies the need for established surgical MBD care pathways and provides reference data to benchmark prospective QA and QI outcomes in this patient population

Surgical management for acute or impending pathologic fractures in metastatic bone disease (MBD) places patients at high-risk for post-operative venous thromboembolism (VTE). Due to the combination of malignancy, systemic cancer treatment, and surgical treatment, VTE-risk is increased 7-fold in patients with MBD compared to non-cancer patients undergoing the same procedure. The extent and duration of post-operative hypercoagulability in patients with MBD remains unknown and thromboprophylaxis guidelines were developed for non-cancer patients, limiting their applicability to address the elevated VTE-risk in cancer patients. Thrombelastography (TEG) analysis is a point-of-care test that measures clot formation, stabilization, and lysis in whole blood samples. The TEG parameter, maximal amplitude (MA), indicates clot strength and the threshold of ≥65 mm has been used to define hypercoagulability and predict VTE events in non-cancer patients requiring orthopaedic surgery. Therefore, this study aims to quantify the extent and duration of post-operative hypercoagulability in patients with MBD using serial TEG analysis. Consecutive adults (≥18 years) with MBD who required orthopaedic surgery for acute or impending pathologic fractures were enrolled into this single-centre, prospective cohort study. Serial TEG analysis was performed onsite using a TEG®6s haemostasis analyzer (Haemonetics Corporation, Boston, MA) on whole blood samples collected at seven timepoints: pre-operatively; on post-operative day (POD) 1, 3, and 5; and at 2-, 6-, and 12-weeks post-operatively. Hypercoagulability was defined as MA ≥65 mm. Participants received standardized thromboprophylaxis for four weeks and patient-reported compliance with thromboprophylaxis was recorded. VTE was defined as symptomatic DVT or PE, or asymptomatic proximal DVT, and all participants underwent a screening post-operative lower extremity Doppler ultrasound on POD3. Descriptive statistics were performed and difference between pre-operative MA values of participants with VTE versus no VTE was evaluated using Student's t-test (p≤0.05). Twenty-one participants (10 female; 47.6%) with a mean age of 70 ± 12 years were enrolled. Nine different primary cancers were identified amongst participants, with breast (23.8%), colorectal (19.0%), and lung cancer (14.3%) most frequently reported. Most participants (57.1%) were hypercoagulable pre-operatively, and nearly half remained hypercoagulable at 6- and 12-weeks post-operatively (47.1 and 46.7%, respectively). VTE occurred in 5 patients (23.8%) and mean MA was 68.1 ± 4.6 mm at the time of diagnosis. Mean pre-operative MA values were significantly higher (p=0.02) in patients who experienced VTE (68.9 ± 3.5 mm) compared to those who did not (62.7 ± 6.5 mm). VTE incidence was highest in the first week post-operatively, during which time four VTE events (80%) occurred. The proportion of patients in a hypercoagulable state increased at three consecutive timepoints, beginning on POD3 (85.0%), increasing on POD5 (87.5%), and peaking at 2-weeks post-operatively (88.9%). Current thromboprophylaxis guidelines do not consider cancer-associated risk factors that contribute to increased VTE incidence and prescription duration may be inadequate to address prolonged post-operative hypercoagulability in patients with MBD. The high rate of VTE events observed and sustained hypercoagulable state indicate that thromboprophylaxis may be prematurely terminated while patients remain at high risk for VTE. Therefore, extending thromboprophylaxis duration beyond 4-weeks post-operatively in patients with MBD warrants further investigation

The burden of metastatic bone disease (MBD) in our Canadian cancer population continues to increase. MBD has a significant effect on patient morbidity, mortality, and health-related quality of life (HRQOL). There are various technical options used to surgically stabilize MBD lesions, surgical decision-making is variable and largely dependent on anatomic and surgeon-based factors. There is a paucity of research examining how surgical decision-making for MBD can be modified or individualized to improve quality of life (QOL) and functional outcomes, while more accurately aligning with patient-reported goals and expectations. The objective of this study was tosurvey MBD patients, support persons, physicians, and allied health care providers (HCP) with the goal of identifying 1) important contributors to HRQOL, 2) discordance in peri-operative expectations, and 3) perceived measures of success in the surgical management of MBD. This project is a longitudinal patient-engaged research initiative in MBD. A survey was developed based on HRQOL themes in the literature and based on feedback from our patient research partners. Participants were asked to identify 1) important contributors to HRQOL and 2) perceived measures of success relevant to the surgical management of MBD. Participants were asked to rank themes from ‘extremely important’ to ‘not important at all’. Using open-ended questions, participants were asked to identify areas of improvement. Responses from the open-ended questions were analyzed by an experienced qualitative researcher using conventional content analysis. Participant's demographics were calculated using descriptive statistics. Concordance or discordance of perceived measure of success was assessed via a Chi-Square test of independence. All statistical analyses were performed using IBM SPSS® software. Nine patients, seven support persons, 23 orthopaedic surgeons, 11 medical oncologists, 16 radiation oncologists, 16 nurses, and eight physiotherapists completed the survey. Regarding perceived measures of success, increased life expectancy (p Two main themes emerged around the timeliness of surgical care and the coordination of multidisciplinary care from patients and support persons. Patients and support persons expressed a sense of urgency in progressing to surgery/treatment, and frustration at perceived delays in treatment. Within coordination of care, patients and support persons would like clearer communication from the health care team. There is discordance between patient/support person goals compared to physicians/HCP goals in the surgical management of MBD. Surgical decision-making and operative techniques that minimize disease progression and improve survival are important to MBD patients. Timely access to surgery/surgical consultation and improved multidisciplinary communication is important to patients. This data suggests improved peri-operative communication and education is needed for MBD patients. Furthermore, future research evaluating how modern orthopaedic surgical techniques influence survival and disease progression in MBD is highly relevant and important to patients with MBD

The management of pathological fractures due to Metastatic Bone Disease (MBD) and Primary Bone Tumours (PBTs) has implications for the Trauma service due to the extra pressures on staff, service delivery and budgets. We undertook an analysis of a cohort of patients presenting with MBD and PBTs. A retrospective chart review of all cases with MBD and PBTs admitted to a 40-bed Trauma Unit between 2005 and 2009 was conducted. The study looked at frequency, primary pathology, and site of pathology/fracture, time from primary diagnosis to referral, subsequent interventions and others. The results identified 34 patients, 21 females (62%) and 13 males (38%) (mean age: 64.6 years) with MBD or PBTs. Metastases secondary to breast cancer (n=13, 38%) and Myeloma (n=5, 15%) were the most common with the majority being found in the femur (n=22, 65%) and the Humerus (n=6, 18%). The mean time from primary tumour diagnosis to fracture referral was 29.6 months with 27 (79%) patients undergoing definitive surgical management within the unit. The conclusions of the study demonstrate that a wide variety of pathology presented to the unit over a 5 year period. Considerable variation was noted in the time from primary tumour diagnosis to presentation with a fracture. This could be due to improvements in treatments of specific cancers or a lack of understanding of what an Orthopaedic surgeon can offer the cancer patient. No definitive increase in pathological fractures was seen. The consensus opinion is that prompt and appropriate management of pathological fractures in cancer patients is cost effective. Management of these injuries, in a Trauma Unit, represents a small, but significant part of the annual work-load. While no significant trend has been seen, with respect to an increased incidence, it is noted that a proportion of these patients were a number of years from their initial diagnosis. With improvements in the survivorship of cancer patients, close scrutiny will be required to determine whether this ultimately translates into an increased fracture burden