Image-based biomarkers in clinical practice

Abstract

The growth of functional and metabolically informative imaging is eclipsing anatomic imaging alone in clinical practice. The recognition that magnetic resonance (MR) and positron emission tomography (PET)-based treatment planning and response assessment are essential components of clinical practice and furthermore offer the potential of quantitative analysis being important. Extracting the greatest benefit from these imaging techniques will require refining the best combinations of multimodality imaging through well-designed clinical trials that use robust image-analysis tools and require substantial computer based infrastructure. Through these changes and enhancements, image-based biomarkers will enhance clinical decision making and accelerate the progress that is made through clinical trial research.

Figure 1

Comparison of conventional CT and MR-based gross tumor volume (GTV), clinical target volume (CTV) with MRSI derived Choline-to-NAA maps. MRSI fused with CT (left), T1-contrast enhanced MRI (middle), and T2-MRI of rim enhancing GBM lesion (pt 3). Contours: Purple, Cho:NAA = 2.0; Red, GTV; Green, CTV. Note a significant portion of metabolically active tumor falls outside both GTV and CTV.

Figure 2

Fluorothymidine (FLT)-PET to assess tumor response after the first week of chemoradiotherapy. Proliferation within the tumor is shown to be reduced at an early time point and may be used to assess outcome and/or modify therapy.

Figure 3

Metabolite ratio (Cho/NAA) color maps and corresponding MR spectral array from a representative case.

Figure 4

The pulmonary function change compared to the planned radiation dose distribution. The dose map, pulmonary function, and pulmonary function change are overlaid on the free breathing CT. The first column is the pulmonary function before RT. The second column is the pulmonary function after RT. The third column is the pulmonary function change from the subtraction of the previous two images. The fourth column is the planned radiation dose distribution. In the third column, the red arrows show regions with decreased pulmonary function, and the blue arrows show regions with increased pulmonary function.

Figure 5

Informatics system architecture as related to the Quantitative Imaging Network (QIN).

Figure 6

Mean SUV within a cylindrical liquid phantom filled with known concentrations of FDG. Data were acquired monthly; each data point represents the slice-specific SUV averaged over 14 months of data collection.