3D Biodistribution and dosimetry within tumours using a silicon-strip beta imager for serial digital autoradiography

Abstract

94

Objectives: In order to optimize radioimmunotherapy, high spatial resolution images of the distributions of antibody, radioactivity and the absorbed dose within tumours and tissues are a prerequisite. An imaging device is then required that has sufficiently high resolution and that allows for imaging of suitable radioisotopes within a reasonable timeframe. The aim of this study was to evaluate the feasibility of using a real-time silicon-strip beta-imager (Biomolex 700 Imager, Biomolex A/S, Oslo) with a pixel-size of 50 µm to image sections of animal tissue and to develop prototype software to align the imaged sections. The resulting activity volume can then be convolved with dose-kernels to display 3D dose-rate distributions.

Methods: Nude mice, implanted with CD20+ tumours, were injected with either intact anti-CD20 rituximab (150 kDa; Genentech Inc., San Francisco, CA) or a smaller recombinant rituximab fragment [minibody of 80 kDa; (scFv-C_H3)₂], both radiolabeled with ¹³¹I. Following 48 h and 11 h after administration of antibody and minibody respectively, tumours were excised, frozen, fixed and sliced into 100 µm sections that were imaged in the beta-imager. Software was developed to correct for image artefacts and to align the imaged sections into a volume. For each section initial x-y translations and in-plane rotation are determined by thresholding the images and performing a least square error minimization. Fine adjustment of translations and rotations is then made using mutual information registration of unthresholded images. Lastly a manual correction can be made if necessary and the software can interpolate missing sections.

Results: The spatial resolution of the beta-imager vary with beta energy and was found to be 115 µm (FWHM) for ¹⁴C. The efficiency was 27% for ¹⁸F and 10% for ¹²⁴I. The count rate response was linear for count rates below 400 cps. The background was 0.7 cps over the entire imaging area of 2.8x6cm². The software was implemented with one of the minibody datasets and an activity volume has been obtained which after convolved with beta dose-point kernels also provided dose-rate distribution volumes for ⁶⁴Cu, ¹³¹I, ³²P and ⁹⁰Y. A similar set of data is currently being analyzed for the intact antibody and will be presented.

Conclusions: These results confirm the viability of the method, giving the possibility to acquire a 3D distribution of radiopharmaceutical concentration and dose rate in tissues and tumours with high spatial resolution.