Contrast enhanced-magnetic resonance imaging as a surrogate to map verteporfin delivery in photodynamic therapy

Abstract

The use of in vivo contrast-enhanced magnetic resonance (MR) imaging as a surrogate for photosensitizer (verteporfin) dosimetry in photodynamic therapy of pancreas cancer is demonstrated by correlating MR contrast uptake to ex vivo fluorescence images on excised tissue. An orthotopic pancreatic xenograft mouse model was used for the study. A strong correlation (r = 0.57) was found for bulk intensity measurements of T1-weighted gadolinium enhancement and verteporfin fluorescence in the tumor region of interest. The use of contrast-enhanced MR imaging shows promise as a method for treatment planning and photosensitizer dosimetry in human photodynamic therapy (PDT) of pancreas cancer.

Fig. 1

Magnetic resonance imaging (MRI) scans of orthotopic AsPC-1 tumors. (a) A T2W coronal scan was used to determine tumor position and align axial slice markers for tumor slices of interest. The bold dashed line (white) is the “tumor origin” and the smaller dashed lines (red) are a representation of the 1-mm image slices, not drawn to scale. The axial image slices are used for analysis, and a single slice is shown for each scan type: (b) T1W pre-gadolinium (Gd) administration, (c) T1W post-Gd, (d) T1W contrast difference, and (e) T2W.

Fig. 2

A representative tumor slice is shown to demonstrate region of interest (ROI, red dashed line) selection. A 4-μm H&E section (a) is used to distinguish the tumor and then transferred to the ex vivo fluorescence image of verteporfin (b) and one of the four MR images, T1W post-Gd is used for demonstration (c).

Fig. 3

Linear regressions and Pearson correlation coefficients ($r$) indicate that BPD fluorescence has no correlation with (a) T1W pre-gadolinium (Gd), (b) a strong, positive correlation with T1W post-Gd, (c) a medium positive correlation with T1W contrast difference, and (d) a weak negative correlation with T2W. Root-mean squared error (RMSE) is reported in arbitrary units of intensity. Black squares: individual slices from each animal; solid red line: best-fit line; dashed green lines: 95% confidence limit; and dotted blue line: 95% prediction limit.