Texture Analysis Using CT and Chemical Shift Encoding-Based Water-Fat MRI Can Improve Differentiation Between Patients With and Without Osteoporotic Vertebral Fractures

Abstract

PurposeOsteoporosis is a highly prevalent skeletal disease that frequently entails vertebral fractures. Areal bone mineral density (BMD) derived from dual-energy X-ray absorptiometry (DXA) is the reference standard, but has well-known limitations. Texture analysis can provide surrogate markers of tissue microstructure based on computed tomography (CT) or magnetic resonance imaging (MRI) data of the spine, thus potentially improving fracture risk estimation beyond areal BMD. However, it is largely unknown whether MRI-derived texture analysis can predict volumetric BMD (vBMD), or whether a model incorporating texture analysis based on CT and MRI may be capable of differentiating between patients with and without osteoporotic vertebral fractures.Materials and MethodsTwenty-six patients (15 females, median age: 73 years, 11 patients showing at least one osteoporotic vertebral fracture) who had CT and 3-Tesla chemical shift encoding-based water-fat MRI (CSE-MRI) available were analyzed. In total, 171 vertebral bodies of the thoracolumbar spine were segmented using an automatic convolutional neural network (CNN)-based framework, followed by extraction of integral and trabecular vBMD using CT data. For CSE-MRI, manual segmentation of vertebral bodies and consecutive extraction of the mean proton density fat fraction (PDFF) and T2* was performed. First-order, second-order, and higher-order texture features were derived from texture analysis using CT and CSE-MRI data. Stepwise multivariate linear regression models were computed using integral vBMD and fracture status as dependent variables.ResultsPatients with osteoporotic vertebral fractures showed significantly lower integral and trabecular vBMD when compared to patients without fractures (p<0.001). For the model with integral vBMD as the dependent variable, T2* combined with three PDFF-based texture features explained 40% of the variance (adjusted R2[Ra2] = 0.40; p<0.001). Furthermore, regarding the differentiation between patients with and without osteoporotic vertebral fractures, a model including texture features from CT and CSE-MRI data showed better performance than a model based on integral vBMD and PDFF only (Ra2 = 0.47 vs. Ra2 = 0.81; included texture features in the final model: integral vBMD, CT_Short-run_emphasis, CT_Varianceglobal, and PDFF_Variance).ConclusionUsing texture analysis for spine CT and CSE-MRI can facilitate the differentiation between patients with and without osteoporotic vertebral fractures, implicating that future fracture prediction in osteoporosis may be improved.

CT image-based biomarkers for opportunistic screening of osteoporotic fractures: a systematic review and meta-analysis

“…Case: vertebral fracture; Control: age- and sex- matched and no vertebral fracture L1-L4 18 (10) Case: 75.4 (10.2) Control: 71.4 (10.1) Vertebral strength Johannesdottir, 2021 [ 35 ] (Iceland) P (CC) 5 Participants from the AGES-Reykjavik cohort who underwent CT scans of the chest and abdomen, with an average of 5 years between baseline and follow-up scans. Case: vertebral fracture; Control: age- and sex- matched and no vertebral fracture T8, L2 401 (272) Case: 76.3 (5.4) Control: 75.7 (5.2) CT volumetric BMD; Vertebral strength Sollmann, 2021 [ 15 ] (Germany) CS 0 Patients with CT and chemical shift encoded-MRI within a maximum of six months and overlap in spatial coverage between CT and chemical shift encoded -MRI of at least two consecutive vertebral bodies Variable, from T1-L5 26 (15) VF: 75.1 No VF: 71.5 CT volumetric BMD; CT texture features Biamonte, 2022 [ 86 ] (Italy) CS 0 Patients aged > 18 years with lumbar spine CT scans L1-L5 240 (110) 60.4 (15.4) Radiomic features; DXA areal BMD Dieckmeyer, 2022 [ 87 ] (Germany) R (CC) 1.51 (mean) Patients with thoraco-lumbar routine CT, mainly for oncologic staging, restaging, and follow-up. Case: vertebral fracture; Control: age- and sex- matched and no vertebral fracture T1-L5 105 (69) Case: 67.4 (11.1) Control: 66.7 (9.4) CT volumetric BMD Li, 2022 [ 88 ] (China) CS (CC) 0 Patients aged ≥ 50 years with CT scans without m...…”

“…SI1 ), and those were assessed to have low risk of bias. The remaining nine studies did not detail on their fracture ascertainment and were considered as having unclear risk of bias [ 15 , 33 , 37 , 40 , 58 , 71 – 74 ]. Eleven studies included less than 90% of eligible participants in the analysis (Fig.…”

“…However, like attenuation, opportunistic extraction of volumetric BMD from clinical routine CT is sensitive to contrast agents that can impact the accuracy of measures [ 92 ]. Integrating texture features of trabecular bone using support vector machines, which has been shown to have very good discriminative ability (accuracy of 83%) [ 72 ], could further leverage the opportunistic use of CT data even beyond the use of standard measurements such as volumetric BMD [ 15 ]. Moreover, the extraction of geometric features such as the anterior–posterior ratio and middle-posterior ratio of vertebral bodies could help to detect vertebral fractures and could also be useful in the prediction of incidental fractures [ 79 ].…”

“…Clinical routine CT examinations obtained for clinical indications other than dedicated assessment of osteoporosis (e.g., lung, cardiovascular, malignancy-related indications) can be used for opportunistic screening for osteoporosis and fractures [ 9 – 11 ]. While especially non-contrast-enhanced CT examinations may be used for the purpose of opportunistic assessments, methods are currently being developed that may deliver accurate measures of areal and/or volumetric BMD even in cases of contrast-enhanced scans or inhomogeneous scanning protocols between CT scanners or sites [ 12 – 15 ]. Given that such imaging data used for this kind of screening are initially procured for non-osteoporosis-related assessments, the implementation comes at low cost without the need for additional imaging exams or radiation exposure to the patient [ 16 ].…”

Radiomic analysis of the proximal femur in osteoporosis women using 3T MRI

“…Another study used opportunistic abdominal CT to retrospectively compute 41 radiomic features in the proximal femur of 500 patients (24) to predict OP status and found an AUROC of 0.96 to predict OP. More recent studies combine radiomic features computed using CT and MRI scans of the lumbar spine (25). They notably used chemical shift encoded MRI methods to separate bone marrow from MR images.…”