Ten Commandments for the Diagnosis of Bone Tumors

Exploiting biochemical data to improve osteosarcoma diagnosis with deep learning

Early and accurate diagnosis of osteosarcomas (OS) is of great clinical significance, and machine learning (ML) based methods are increasingly adopted. However, current ML-based methods for osteosarcoma diagnosis consider only X-ray images, usually fail to generalize to new cases, and lack explainability. In this paper, we seek to explore the capability of deep learning models in diagnosing primary OS, with higher accuracy, explainability, and generality. Concretely, we analyze the added value of integrating the biochemical data, i.e., alkaline phosphatase (ALP) and lactate dehydrogenase (LDH), and design a model that incorporates the numerical features of ALP and LDH and the visual features of X-ray imaging through a late fusion approach in the feature space. We evaluate this model on real-world clinic data with 848 patients aged from 4 to 81. The experimental results reveal the effectiveness of incorporating ALP and LDH simultaneously in a late fusion approach, with the accuracy of the considered 2608 cases increased to 97.17%, compared to 94.35% in the baseline. Grad-CAM visualizations consistent with orthopedic specialists further justified the model's explainability.

Outcomes of management of primary benign aggressive or malignant bone tumors around the elbow by limb-salvage surgery

PURPOSE

Bone tumors around the elbow are rare, with frequently delayed diagnosis. The current study aimed to assess the functional and oncological outcomes of limb salvage surgery for primary benign aggressive or malignant bone tumors around the elbow.

METHODS

We conducted a retrospective review of patients with primary aggressive benign and malignant bone tumors around the elbow treated with limb salvage surgery between 1995 and 2020 at a single musculoskeletal oncology center. The minimum follow-up period was 24 months. Functional results were assessed using the Musculoskeletal Tumor Society (MSTS) scoring system at the last follow-up visit. Local recurrence, chest metastasis, and complications were recorded.

RESULTS

This study included 30 patients, 19 males and 11 females, with a mean age of 25.4 ± 14.2 years. The tumor location was the distal humerus (n = 21), proximal radius (n = 5), and proximal ulna (n = 4). Reconstruction was done by elbow fusion using fibular graft (n = 10), mobile endoprosthesis (n = 9), excision arthroplasty (n = 7), and extracorporeal freezing and reimplantation (n = 4). The mean follow-up period was 36.2 ± 21.3 months. The median follow-up MSTS score was 27 [Interquartile range (IQR): 26-30]. Skeletally immature patients had a significantly higher MSTS score. The rate of postoperative complications was 26.7%.

CONCLUSION

Limb salvage surgery with different reconstructive options for benign aggressive and malignant bone tumors around the elbow can achieve good functional and oncological outcomes.

LEVEL OF EVIDENCE

Level IV.

"Do-not-touch" lesions of bone revisited

Incidental bone lesions are encountered frequently in day-to-day practice. Many of these lesions are indeterminate requiring referral to specialist centres for further characterisation with or without biopsy; however, as biopsy has its own drawbacks, not all lesions can be subjected to biopsy. The primary role of a radiologist in these situations is to characterise these lesions based on their imaging appearances into aggressive lesions requiring biopsy and non-aggressive lesions that do not require a biopsy. The term "do-not-touch lesion" is used to describe a lesion with typical radiographic appearances that can be characterised based on radiographic appearances alone without needing a biopsy. With recent advances in imaging, many incidental lesions can be characterised into do-not-touch lesions based on their imaging appearances alone using a single imaging technique or using a combination of imaging techniques and, less frequently, with the additional help of serological investigations, without the need for biopsy. Hence, the definition of do-not-touch lesions of bone needs a revisit. In this article, we attempt to redefine do-not-touch lesions of bone and propose an imaging-based classification for characterisation of these lesions.