Recommender-based bone tumour classification with radiographs-a link to the past.

OBJECTIVES: To develop an algorithm to link undiagnosed patients to previous patient histories based on radiographs, and simultaneous classification of multiple bone tumours to enable early and specific diagnosis. MATERIALS AND METHODS: For this retrospective study, data from 2000 to 2021 were curated from our database by two orthopaedic surgeons, a radiologist and a data scientist. Patients with complete clinical and pre-therapy radiographic data were eligible. To ensure feasibility, the ten most frequent primary tumour entities, confirmed histologically or by tumour board decision, were included. We implemented a ResNet and transformer model to establish baseline results. Our method extracts image features using deep learning and then clusters the k most similar images to the target image using a hash-based nearest-neighbour recommender approach that performs simultaneous classification by majority voting. The results were evaluated with precision-at-k, accuracy, precision and recall. Discrete parameters were described by incidence and percentage ratios. For continuous parameters, based on a normality test, respective statistical measures were calculated. RESULTS: Included were data from 809 patients (1792 radiographs; mean age 33.73 ± 18.65, range 3-89 years; 443 men), with Osteochondroma (28.31%) and Ewing sarcoma (1.11%) as the most and least common entities, respectively. The dataset was split into training (80%) and test subsets (20%). For k = 3, our model achieved the highest mean accuracy, precision and recall (92.86%, 92.86% and 34.08%), significantly outperforming state-of-the-art models (54.10%, 55.57%, 19.85% and 62.80%, 61.33%, 23.05%). CONCLUSION: Our novel approach surpasses current models in tumour classification and links to past patient data, leveraging expert insights. CLINICAL RELEVANCE STATEMENT: The proposed algorithm could serve as a vital support tool for clinicians and general practitioners with limited experience in bone tumour classification by identifying similar cases and classifying bone tumour entities. KEY POINTS: • Addressed accurate bone tumour classification using radiographic features. • Model achieved 92.86%, 92.86% and 34.08% mean accuracy, precision and recall, respectively, significantly surpassing state-of-the-art models. • Enhanced diagnosis by integrating prior expert patient assessments.

SAM-X: sorting algorithm for musculoskeletal x-ray radiography.

OBJECTIVE: To develop a two-phased deep learning sorting algorithm for post-X-ray image acquisition in order to facilitate large musculoskeletal image datasets according to their anatomical entity. METHODS: In total, 42,608 unstructured and pseudonymized radiographs were retrieved from the PACS of a musculoskeletal tumor center. In phase 1, imaging data were sorted into 1000 clusters by a self-supervised model. A human-in-the-loop radiologist assigned weak, semantic labels to all clusters and clusters with the same label were merged. Three hundred thirty-two non-musculoskeletal clusters were discarded. In phase 2, the initial model was modified by "injecting" the identified labels into the self-supervised model to train a classifier. To provide statistical significance, data split and cross-validation were applied. The hold-out test set consisted of 50% external data. To gain insight into the model's predictions, Grad-CAMs were calculated. RESULTS: The self-supervised clustering resulted in a high normalized mutual information of 0.930. The expert radiologist identified 28 musculoskeletal clusters. The modified model achieved a classification accuracy of 96.2% and 96.6% for validation and hold-out test data for predicting the top class, respectively. When considering the top two predicted class labels, an accuracy of 99.7% and 99.6% was accomplished. Grad-CAMs as well as final cluster results underlined the robustness of the proposed method by showing that it focused on similar image regions a human would have considered for categorizing images. CONCLUSION: For efficient dataset building, we propose an accurate deep learning sorting algorithm for classifying radiographs according to their anatomical entity in the assessment of musculoskeletal diseases. KEY POINTS: • Classification of large radiograph datasets according to their anatomical entity. • Paramount importance of structuring vast amounts of retrospective data for modern deep learning applications. • Optimization of the radiological workflow and increase in efficiency as well as decrease of time-consuming tasks for radiologists through deep learning.

Two-Phase Deep Learning Algorithm for Detection and Differentiation of Ewing Sarcoma and Acute Osteomyelitis in Paediatric Radiographs.

BACKGROUND/AIM: Ewing sarcoma is a highly malignant tumour predominantly found in children. The radiological signs of this malignancy can be mistaken for acute osteomyelitis. These entities require profoundly different treatments and result in completely different prognoses. The purpose of this study was to develop an artificial intelligence algorithm, which can determine imaging features in a common radiograph to distinguish osteomyelitis from Ewing sarcoma. MATERIALS AND METHODS: A total of 182 radiographs from our Sarcoma Centre (118 healthy, 44 Ewing, 20 osteomyelitis) from 58 different paediatric (≤18 years) patients were collected. All localisations were taken into consideration. Cases of acute, acute on chronic osteomyelitis and intraosseous Ewing sarcoma were included. Chronic osteomyelitis, extra-skeletal Ewing sarcoma, malignant small cell tumour and soft tissue-based primitive neuroectodermal tumours were excluded. The algorithm development was split into two phases and two different classifiers were built and combined with a Transfer Learning approach to cope with the very limited amount of data. In phase 1, pathological findings were differentiated from healthy findings. In phase 2, osteomyelitis was distinguished from Ewing sarcoma. Data augmentation and median frequency balancing were implemented. A data split of 70%, 15%, 15% for training, validation and hold-out testing was applied, respectively. RESULTS: The algorithm achieved an accuracy of 94.4% on validation and 90.6% on test data in phase 1. In phase 2, an accuracy of 90.3% on validation and 86.7% on test data was achieved. Grad-CAM results revealed regions, which were significant for the algorithms decision making. CONCLUSION: Our AI algorithm can become a valuable support for any physician involved in treating musculoskeletal lesions to support the diagnostic process of detection and differentiation of osteomyelitis from Ewing sarcoma. Through a Transfer Learning approach, the algorithm was able to cope with very limited data. However, a systematic and structured data acquisition is necessary to further develop the algorithm and increase results to clinical relevance.