Machine learning based natural language processing of radiology reports in orthopaedic trauma.

OBJECTIVES: To compare different Machine Learning (ML) Natural Language Processing (NLP) methods to classify radiology reports in orthopaedic trauma for the presence of injuries. Assessing NLP performance is a prerequisite for downstream tasks and therefore of importance from a clinical perspective (avoiding missed injuries, quality check, insight in diagnostic yield) as well as from a research perspective (identification of patient cohorts, annotation of radiographs). METHODS: Datasets of Dutch radiology reports of injured extremities (n = 2469, 33% fractures) and chest radiographs (n = 799, 20% pneumothorax) were collected in two different hospitals and labeled by radiologists and trauma surgeons for the presence or absence of injuries. NLP classification was applied and optimized by testing different preprocessing steps and different classifiers (Rule-based, ML, and Bidirectional Encoder Representations from Transformers (BERT)). Performance was assessed by F1-score, AUC, sensitivity, specificity and accuracy. RESULTS: The deep learning based BERT model outperforms all other classification methods which were assessed. The model achieved an F1-score of (95 ± 2)% and accuracy of (96 ± 1)% on a dataset of simple reports (n= 2469), and an F1 of (83 ± 7)% with accuracy (93 ± 2)% on a dataset of complex reports (n= 799). CONCLUSION: BERT NLP outperforms traditional ML and rule-base classifiers when applied to Dutch radiology reports in orthopaedic trauma.

Ossification of abdominal scar tissue: a case series with a translational review on its development.

PURPOSE: Bone formation in abdominal scar tissue is a form of heterotopic ossification. It is a rare and underreported phenomenon following abdominal surgery. Heterotopic ossification (HO) is the formation of bone where normally no bone is present and two theories on its pathogenesis prevail: (1) physical dislocation of small bony fragments from the xyphoid process or os pubis into the wound, (2) differentiation of locally available multipotent mesenchymal stromal cells into osteoblasts resulting in the calcification of extracellular matrix. Multipotent mesenchymal stromal cells can differentiate into different cell types by exposing them to different stimuli. We hypothesize that pro-osteogenic signals derived from e.g., macrophages steers multipotent mesenchymal stromal cells involved in the wound healing towards osteogenesis. METHODS: In a retrospective case study we analyzed ossified tissue, patient demographics, medical history, number of laparotomies, scar location, indication for surgery and time in which HO occurred. RESULTS: Ten (8 male, 2 female) patients had proven HO. The mean age was 62 (46-80) years. The mean time for HO to occur was 99 (24-382) days after the previous laparotomy. The mean number of relaparotomies was 3 (1-9). CONCLUSION: We conclude that ossification of abdominal scar tissue is a rare but innocent finding and provides interesting leads to other fields of research.