Radiomics and Machine Learning Differentiate Soft-Tissue Lipoma and Liposarcoma Better than Musculoskeletal Radiologists

Abstract

Distinguishing lipoma from liposarcoma is challenging on conventional MRI examination. In case of uncertain diagnosis following MRI, further invasive procedure (percutaneous biopsy or surgery) is often required to allow for diagnosis based on histopathological examination. Radiomics and machine learning allow for several types of pathologies encountered on radiological images to be automatically and reliably distinguished. The aim of the study was to assess the contribution of radiomics and machine learning in the differentiation between soft-tissue lipoma and liposarcoma on preoperative MRI and to assess the diagnostic accuracy of a machine-learning model compared to musculoskeletal radiologists. 86 radiomics features were retrospectively extracted from volume-of-interest on T1-weighted spin-echo 1.5 and 3.0 Tesla MRI of 38 soft-tissue tumors (24 lipomas and 14 liposarcomas, based on histopathological diagnosis). These radiomics features were then used to train a machine-learning classifier to distinguish lipoma and liposarcoma. The generalization performance of the machine-learning model was assessed using Monte-Carlo cross-validation and receiver operating characteristic curve analysis (ROC-AUC). Finally, the performance of the machine-learning model was compared to the accuracy of three specialized musculoskeletal radiologists using the McNemar test. Machine-learning classifier accurately distinguished lipoma and liposarcoma, with a ROC-AUC of 0.926. Notably, it performed better than the three specialized musculoskeletal radiologists reviewing the same patients, who achieved ROC-AUC of 0.685, 0.805, and 0.785. Despite being developed on few cases, the trained machine-learning classifier accurately distinguishes lipoma and liposarcoma on preoperative MRI, with better performance than specialized musculoskeletal radiologists.

Sarcoma treatment in the era of molecular medicine

“…The ultimate goal of personalized medicine is to be able to integrate clinical, genomic, transcriptomic, and epigenomic data to increase the accuracy of diagnosis and prognosis, and to identify the most effective therapy for treatment (Burdach et al , 2018; Salgado et al , 2018; Gargallo et al , 2020). Recent advances in machine learning‐based methods for analysis of histology and radiography imaging may also play an increasingly important role (Blackledge et al , 2019; Wang et al , 2019; Malinauskaite et al , 2020). For instance, clinical investigations into immune checkpoint therapy have designated UPS, myxofibrosarcoma, and similar genomically complex histotypes as “UPS” (Que et al , 2017), making comparisons with other studies difficult.…”

Manual versus semiautomatic segmentation of soft-tissue sarcomas on magnetic resonance imaging: evaluation of similarity and comparison of segmentation times

“…In a search of the literature, we identified various studies that analyzed MRI scans of soft-tissue sarcomas using manual segmentation ( 8 - 15 ) , semiautomatic segmentation ( 16 , 17 ) , or automatic segmentation ( 18 ) . Only one of those studies evaluated interobserver variability ( 13 ) , and none of them evaluated intraobserver variability.…”

Imaging of Bone Sarcomas and Soft-Tissue Sarcomas

“…The introduction of artificial intelligence (AI) in MSK radiology and the employment of radiomic texture analysis have resulted in the development of radiomic MRI-based models that could help distinguish low-grade from high-grade sarcomas [47,48]. In a study by Malinauskaite et al, radiomics incorporated with machine-learning methods performed better than specialized MSK radiologists, reaching a diagnostic accuracy of 94.7 % [49]. Additionally, radiomic features extracted from MRI show promise as biomarkers for predicting overall survival in patients with STS.…”