Utilizing 18F-NaF PET/CT to Inform Pathologic and Radiographic Progression of an Ultra-rare Case of Familial Tumoral Calcinosis (FTC): plus a case for synergistic information

Introduction: Hyperphosphatemic Familial Tumoral Calcinosis (FTC) is an ultra-rare autosomal recessive disorder due to a functional deficiency of fibroblast growth factor 23 (FGF23) that leads to elevated serum phosphate and ectopic calcification. FTC manifests with sometimes massive, highly morbid, tumor-like calcific lesions. The pathophysiology and heterogeneity of this process is poorly characterized. LIke many bone disorders, CT is used to detect these calcified lesions. However, ¹⁸F-NaF PET/CT offers a powerful tool in studying the metabolic properties of these mineralized tissues, particularly their dynamic properties. The goal of this study was to apply ¹⁸F-NaF PET/CT imaging to study the nature and progression of ectopic calcification, using data acquired from the only known FTC patient with serial timepoint ¹⁸F-NaF PET/CT imaging to our knowledge. We also explore whether advanced voxel-level radiomics (CT-based) can explain intra-lesion metabolic activity by ¹⁸F-NaF PET, compared to structural information alone.

Methods: One patient (11F), who did not receive potentially disease-modifying treatment, underwent serial ¹⁸F-NaF PET/CT scans approximately 2 years apart (11 and 13 yr old). MIM Software (version 7.2.3, MIM Software Inc.) was used for PET/CT image registration and labeling of lesions. The voxel-level data (3D coordinates, HU values, and SUVs) were exported into MATLAB (version 2022b, MathWorks) for analysis. The volume (cm³), Total Lesion Activity (TLA (SUV*cm³)), Total Lesion Density (TLD (∑HU)), and standard deviation (SD) of HUs were calculated for the lesion at both timepoints to understand progression. Additionally, voxel-level scatter plots of SUV vs. HU were created to infer how intra-lesion metabolic activity relates to density so as to understand how this relation may progress over time. More advanced, voxel-level radiomic features (103 features) were calculated for the lesion using the structural (CT-based) data using 9x9x9 voxel neighborhoods. To test the relationship between structure and function, multiple regression models were trained to fit the metabolic activity (SUV)/voxel. The coefficient of determination (R²) was computed to inform how the variance in structure explains changes in SUV.

Results: The patient’s calcific lesion markedly increased over 2 years from 393.7 cm³ to 597.1 (52%). The TLA slightly decreased from 5,873 to 5,606 SUVcm³ (4.5%). Whereas the TLD increased from 5.9(10⁷) to 8.84(10⁷) ∑HU (49%). Additionally, the SD of HUs increased from 166.5 to 239.3 (43.7%), indicating an increase in intra-lesion heterogeneity. Upon analysis of the scatter plot of SUV vs. HU for both timepoints (Figure 1), 2 unique populations of voxels were discernible: (1) low density/high metabolism and (2) high density/low metabolism. From timepoint 1 to 2, population (1) decreased by 1.3%, and population (2) increased by 7.2%, indicating that highly active regions progress to mature marocalcifications. To further test the relationship between structure (CT) and metabolism on ¹⁸F-NaF PET imaging, multiple regression models were trained to fit the CT-based radiomic features to the SUVs. A decision-tree (coarse, 32-leaf) regression model had the best fit , with an r = 0.80 and R² = 0.64. Suggesting that structural characteristics of the calcific lesion only explains 64% of the variability in metabolic activity.

Conclusions: As is common in FTC, this patient’s lesion progressed clinically over the two-year period. Analyses demonstrated the lesion consolidated, becoming overall denser but remained metabolically active. This pattern suggests that very active regions with low density represent early-stage microcalcifications that will eventually mature into macro-calcifications. Further, from our analysis of structure and metabolism, metabolic activity cannot be purely explained by structure alone, arguing for the necessity of metabolic imaging (PET) when studying this complex disease.

• Download figure
• Open in new tab
• Download powerpoint