Compilation by Category of Confounding Factors Influencing SUV Determination of Defined Tissue Type and State for Defined Population of Patients
| Factor* | Comment | |
|---|---|---|
| Tissue activity factors | ||
| ROI shape within which to average | Either pixels or voxels can be used. Specific criteria for shape of outer boundary have precedents that include visual judgment, noise-affected maximum pixel, fixed size, contour defined by some fraction of maximum pixel, and others. Considerations also involve the character of heterogeneity encountered. | |
| Partial-volume and spillover effects | Factors obtained in small phantom data allow observed ROI activity to be corrected to that truly present (14). There is dependency on reconstructed resolution, size and geometry, and the ratio of activities in ROI’s region and surrounding region. Interrelated is motion blurring (e.g., from diaphragm) that undesirably averages pixel intensities. | |
| Attenuation correction | Always required, methodology (e.g., γ-energy used, counts obtained, CT contrast agent usage, algorithm approximations) affects both absolute accuracy and noise. | |
| Reconstruction method and parameters for scanner type | Techniques used in reconstruction—including spatial filter values, total number of pixels, and other parameters (e.g., number of iterations in some algorithms)—influence noise and resolution. | |
| Counts’ noise bias effect | Total number of noise equivalent counts from acquisition and analysis system, for a decayed dose, affects pixel randomness. ROIs based on maximum pixel value have pixel averages affected. | |
| Tissue state† factors amenable for corrections | ||
| Time of SUV evaluation | Injection-to-midacquisition time for SUV determination is a characterizing parameter (15). Traditionally, this time interval is chosen so that (dSUV/dt)/SUV is typically not excessive during acquisition. Corrections for time are possible (16,17). | |
| Competing transport effects | In facilitated transport of amino acids or glucose, there is competition between variable serum concentrations of these and their tracer analogs. Where justified by data, correction for this effect can sometimes be appropriate. | |
| Normalization factor | ||
| Body size | In the (injected dose)/size denominator of SUV calculation, precedents for size are body weight, lean body mass, and body surface area. For 18F-FDG, evidence shows the latter 2 generally reduce variability by more consistently describing a body volume into which tracer distributes (18). | |
-
↵* Factors can be interrelated. Thus, some aspect of one (e.g., induce more noise or influence resolution) may be listed elsewhere in table as a factor directly affecting SUV.
-
↵† Many biologic factors determine a particular tissue’s uptake (19), such as kind and extent of disease, vascularity, organ usage, urine management policy, population characteristic, and so forth. Among such conditions at scan time, only 2 are singled out here as candidates for SUV corrections.
