Patients with cancer and bone metastases can have an increased risk of fracturing their femur. Treatment is based on the impending fracture risk: patients with a high fracture risk are considered for prophylactic surgery, whereas low fracture risk patients are treated conservatively with radiotherapy to decrease pain. Current clinical guidelines suggest to determine fracture risk based on axial cortical involvement of the lesion on conventional radiographs, but that appears to be difficult. Therefore, we developed a patient-specific finite element (FE) computer model that has shown to be able to predict fracture risk in an experimental setting and in patients. The goal of this study was to determine whether patient-specific finite element (FE) computer models are better at predicting fracture risk for femoral bone metastases compared to clinical assessments based on axial cortical involvement on conventional radiographs, as described in current clinical guidelines.

45 patients (50 affected femurs) affected with predominantly lytic bone metastases who were treated with palliative radiotherapy for pain were included. CT scans were made and patients were followed for six months to determine whether or not they fractured their femur. Non-linear isotropic FE models were created with the patient-specific geometry and bone density obtained from the CT scans. Subsequently, an axial load was simulated on the models mimicking stance. Failure loads normalized for bodyweight (BW) were calculated for each femur. High and low fracture risks were determined using a failure load of 7.5 × BW as a threshold. Experienced assessors measured axial cortical involvement on conventional radiographs. Following clinical guidelines, patients with lesions larger than 30 mm were identified as having a high fracture risk. FE predictions were compared to clinical assessments by means of diagnostic accuracy values (sensitivity, specificity and positive (PPV) and negative predictive values (NPV)).

Seven femurs (14%) fractured during follow-up. Median time to fracture was 8 weeks. FE models were better at predicting fracture risk in comparison to clinical assessments based on axial cortical involvement (sensitivity 100% vs. 86%, specificity 74% vs. 42%, PPV 39% vs. 19%, and NPV 100% vs. 95%, for the FE computer model vs. axial cortical involvement, respectively). We concluded that patient-specific FE computer models improve fracture risk predictions of femoral bone metastases in advanced cancer patients compared to clinical assessments based on axial cortical involvement, which is currently used in clinical guidelines. Therefore, we are initiating a pilot for clinical implementation of the FE model.

Patients with advanced cancer can develop bone metastases in the femur which are often painful and increase the risk of pathological fracture. Accurate segmentation of bone metastases is, amongst others, important to improve patient-specific computer models which calculate fracture risk, and for radiotherapy planning to determine exact radiation fields. Deep learning algorithms have shown to be promising to improve segmentation accuracy for metastatic lesions, but require reliable segmentations as training input. The aim of this study was to investigate the inter- and intra-operator reliability of manual segmentation of femoral metastatic lesions and to define a set of lesions which can serve as a training dataset for deep learning algorithms. F

CT-scans of 60 advanced cancer patients with a femur affected with bone metastases (20 osteolytic, 20 osteoblastic and 20 mixed) were used in this study. Two operators were trained by an experienced radiologist and then segmented the metastatic lesions in all femurs twice with a four-week time interval. 3D and 2D Dice coefficients (DCs) were calculated to quantify the inter- and intra-operator reliability of the segmentations. We defined a DC>0.7 as good reliability, in line with a statistical image segmentation study.

Mean first and second inter-operator 3D-DCs were 0.54 (±0.28) and 0.50 (±0.32), respectively. Mean intra-operator I and II 3D-DCs were 0.56 (±0.28) and 0.71 (±0.23), respectively. Larger lesions (>60 cm³) scored higher DCs in comparison with smaller lesions.

This study reveals that manual segmentation of metastatic lesions is challenging and that the current manual segmentation approach resulted in dissatisfying outcomes, particularly for lesions with small volumes. However, segmentation of larger lesions resulted in a good inter- and intra-operator reliability. In addition, we were able to select 521 slices with good segmentation reliability that can be used to create a training dataset for deep learning algorithms. By using deep learning algorithms, we aim for more accurate automated lesion segmentations which might be used in computer modelling and radiotherapy planning.

Objectives

In this prospective cohort study, we investigated whether patient-specific finite element (FE) models can identify patients at risk of a pathological femoral fracture resulting from metastatic bone disease, and compared these FE predictions with clinical assessments by experienced clinicians.