Injectable fillers are widely used for facial rejuvenation, correction of disabling volumetric fat loss in HIV-associated facial lipoatrophy, Romberg disease, and post-traumatic facial disfiguring. The purpose of this article is to acquaint the reader with the anatomy of facial fat compartments, as well as with the properties and key imaging features of commonly used facial fillers, filler-related complications, interpretation pitfalls, and dermatologic conditions mimicking filler-related complications. The distribution of facial fillers is characteristic and depends on the anatomy of the superficial fat compartments. Silicone has signature MRI features, calcium hydroxyapatite has characteristic calcifications, whereas other injectable fillers have overlapping imaging features. Most fillers (hyaluronic acid, collagen, and polyalkylimide–polyacrylamide hydrogels) have signal intensity patterns compatible with high water content. On PET-CT, most fillers show physiologic high FDG uptake, which should not be confounded with pathology. Abscess, cellulitis, non-inflammatory nodules, and foreign body granulomas are the most common filler-related complications, and imaging can help in the differential diagnosis. Diffusion weighted imaging helps in detecting a malignant lesion masked by injected facial fillers. Awareness of imaging features of facial fillers and their complications helps to avoid misinterpretation of MRI, and PET-CT scans and facilitates therapeutic decisions in unclear clinical cases. Key points • Facial fillers are common incidental findings on MRI and PET-CT scans. • They have a characteristic appearance and typical anatomic distribution • Although considered as safe, facial filler injections are associated with several complications • As they may mask malignancy, knowledge of typical imaging features is mandatory. • MRI is a problem-solving tool for unclear cases.

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.

BACKGROUND AND PURPOSE: SyMRI is a technique developed to perform quantitative MR imaging. Our aim was to analyze its potential use for measuring relaxation times of normal components of the spine and to compare them with values found in the literature using relaxometry and other techniques. MATERIALS AND METHODS: Thirty-two spine MR imaging studies (10 cervical, 5 dorsal, 17 lumbosacral) were included. A modified multiple-dynamic multiple-echo sequence was added and processed to obtain quantitative T1 (millisecond), T2 (millisecond), and proton density (percentage units [pu]) maps for each patient. An ROI was placed on representative areas for CSF, spinal cord, intervertebral discs, and vertebral bodies, to measure their relaxation. RESULTS: Relaxation time means are reported for CSF (T1 = 4273.4 ms; T2 = 1577.6 ms; proton density = 107.5 pu), spinal cord (T1 = 780.2 ms; T2 = 101.6 ms; proton density = 58.7 pu), normal disc (T1 = 1164.9 ms; T2 = 101.9 ms; proton density = 78.9 pu), intermediately hydrated disc (T1 = 723 ms; T2 = 66.8 ms; proton density = 60.8 pu), desiccated disc (T1 = 554.4 ms; T2 = 55.6 ms; proton density = 47.6 ms), and vertebral body (T1 = 515.3 ms; T2 = 100.8 ms; proton density = 91.1 pu). Comparisons among the mean T1, T2, and proton density values showed significant differences between different spinal levels (cervical, dorsal, lumbar, and sacral) for CSF (proton density), spinal cord (T2 and proton density), normal disc (T1, T2, and proton density), and vertebral bodies (T1 and proton density). Significant differences were found among mean T1, T2, and proton density values of normal, intermediately hydrated, and desiccated discs. CONCLUSIONS: Measurements can be easily obtained on SyMRI and correlated with previously published values obtained using conventional relaxometry techniques.