Pharmacokinetic digital phantoms for accuracy assessment of image-based dosimetry in¹⁷⁷Lu-DOTATATE peptide receptor radionuclide therapy

Abstract

Patient-specific image-based dosimetry is considered to be a useful tool to limit toxicity associated with peptide receptor radionuclide therapy (PRRT). To facilitate the establishment and reliability of absorbed-dose response relationships, it is essential to assess the accuracy of dosimetry in clinically realistic scenarios. To this end, we developed pharmacokinetic digital phantoms corresponding to patients treated with 177 Lu-DOTATATE. Three individual voxel phantoms from the XCAT population were generated and assigned a dynamic activity distribution based on a compartment model for 177 Lu-DOTATATE, designed specifically for this purpose. The compartment model was fitted to time-activity data from 10 patients, primarily acquired using quantitative scintillation camera imaging. S values for all phantom source-target combinations were calculated based on Monte-Carlo simulations. Combining the S values and time-activity curves, reference values of the absorbed dose to the phantom kidneys, liver, spleen, tumours and whole-body were calculated. The phantoms were used in a virtual dosimetry study, using Monte-Carlo simulated gamma-camera images and conventional methods for absorbed-dose calculations. The characteristics of the SPECT and WB planar images were found to well represent those of real patient images, capturing the difficulties present in image-based dosimetry. The phantoms are expected to be useful for further studies and optimisation of clinical dosimetry in 177 Lu PRRT.

A method for tumor dosimetry based on hybrid planar‐SPECT/CT images and semiautomatic segmentation

“…Three computer phantoms from the XCAT family representing the anatomies of two females and one male were used to obtain Monte Carlo simulated planar gamma‐camera and SPECT images. The phantoms were voxelized with a voxel size of 2.5 × 2.5 × 2.5 mm, and a pharmacokinetic compartment model was used to obtain activity concentration curves for all phantom regions . Each phantom contained two or three tumors (eight in total) with volumes between 2.75 ml and 45.5 ml and were located in the liver and abdomen.…”

“…The phantoms were voxelized with a voxel size of 2.5 9 2.5 9 2.5 mm 3 , and a pharmacokinetic compartment model was used to obtain activity concentration curves for all phantom regions. 33 Each phantom contained two or three tumors (eight in total) with volumes between 2.75 ml and 45.5 ml and were located in the liver and abdomen. Tumor locations are illustrated in Brolin et al 33 .…”

Comparison of commercial dosimetric software platforms in patients treated with 177Lu‐DOTATATE for peptide receptor radionuclide therapy

“…[1][2][3][4][5] Different approaches to clinical dosimetry have been proposed, based on whole body (WB) planar images, [6][7][8][9] single photon emission computed tomography/ computed tomography (SPECT/CT) images, [10][11][12][13] and hybrid methods by combining WB planar images with one or two SPECT/CT scans. [14][15][16][17][18] Imaging and data processing methodology have historically been predominantly specific to each institution, because there were no commercial software applications available to perform all aspects of clinical dosimetry. Therefore, academic/research institutions/hospitals developed in-house dosimetric software, most of the time available only locally and not registered as medical devices (i.e., used in a research context only): DOSIMG, 19 MABDOSE, 20 DOSE3D, 21 RMDP, 22 VoxelDose, 23 MrVoxel, 24 OEDIPE, 25 MINERVA, 26 3D-RD, 27 RAYDOSE, 28 and NUKDOS, 29 to name a few, belong to that category.…”

The Effect of Calibration Factors and Recovery Coefficients on 177Lu SPECT Activity Quantification Accuracy: A Monte Carlo Study

“…All simulations were conducted with a high number of photon histories to ensure data sets with low simulation noise. Sixty (60) projections with an equivalent of 45 s acquisition time were simulated and the image data were stored in a 128 × 128 matrix with a pixel size of 4.8 × 4.8 mm [16,49].…”