Timo Paulus

Evaluation of a compartmental model for estimating tumor hypoxia via FMISO dynamic PET imaging.

This paper systematically evaluates a pharmacokinetic compartmental model for identifying tumor hypoxia using dynamic positron emission tomography (PET) imaging with 18F-fluoromisonidazole (FMISO). A generic irreversible one-plasma two-tissue compartmental model was used. A dynamic PET image dataset was simulated with three tumor regions-normoxic, hypoxic and necrotic-embedded in a normal-tissue background, and with an image-based arterial input function. Each voxelized tissues time activity curve (TAC) was simulated with typical values of kinetic parameters, as deduced from FMISO-PET data from nine head-and-neck cancer patients. The dynamic dataset was first produced without any statistical noise to ensure that correct kinetic parameters were reproducible. Next, to investigate the stability of kinetic parameter estimation in the presence of noise, 1000 noisy samples of the dynamic dataset were generated, from which 1000 noisy estimates of kinetic parameters were calculated and used to estimate the sample mean and covariance matrix. It is found that a more peaked input function gave less variation in various kinetic parameters, and the variation of kinetic parameters could also be reduced by two region-of-interest averaging techniques. To further investigate how bias in the arterial input function affected the kinetic parameter estimation, a shift error was introduced in the peak amplitude and peak location of the input TAC, and the bias of various kinetic parameters calculated. In summary, mathematical phantom studies have been used to determine the statistical accuracy and precision of model-based kinetic analysis, which helps to validate this analysis and provides guidance in planning clinical dynamic FMISO-PET studies.

Phantom validation of quantitative Y-90 PET/CT-based dosimetry in liver radioembolization

BackgroundPET/CT has recently been shown to be a viable alternative to traditional post-infusion imaging methods providing good quality images of 90Y-laden microspheres after selective internal radiation therapy (SIRT). In the present paper, first we assessed the quantitative accuracy of 90Y-PET using an anthropomorphic phantom provided with lungs, liver, spine, and a cylindrical homemade lesion located into the hepatic compartment. Then, we explored the accuracy of different computational approaches on dose calculation, including (I) direct Monte Carlo radiation transport using Raydose, (II) Kernel convolution using Philips Stratos, (III) local deposition algorithm, (IV) Monte Carlo technique (MCNP) considering a uniform activity distribution, and (V) MIRD (Medical Internal Radiation Dose) analytical approach. Finally, calculated absorbed doses were compared with those obtained performing measurements with LiF:Mg,Cu,P TLD chips in a liquid environment.ResultsOur results indicate that despite 90Y-PET being likely to provide high-resolution images, the 90Y low branch ratio, along with other image-degrading factors, may produce non-uniform activity maps, even in the presence of uniform activity. A systematic underestimation of the recovered activity, both for the tumor insert and for the liver background, was found. This is particularly true if no partial volume correction is applied through recovery coefficients. All dose algorithms performed well, the worst case scenario providing an agreement between absorbed dose evaluations within 20%. Average absorbed doses determined with the local deposition method are in excellent agreement with those obtained using the MIRD and the kernel-convolution dose calculation approach.Finally, absorbed dose assessed with MC codes are in good agreement with those obtained using TLD in liquid solution, thus confirming the soundness of both calculation approaches. This is especially true for Raydose, which provided an absorbed dose value within 3% of the measured dose, well within the stated uncertainties.ConclusionsPatient-specific dosimetry is possible even in a scenario with low true coincidences and high random fraction, as in 90Y–PET imaging, granted that accurate absolute PET calibration is performed and acquisition times are sufficiently long. Despite Monte Carlo calculations seeming to outperform all dose estimation algorithms, our data provide a strong argument for encouraging the use of the local deposition algorithm for routine 90Y dosimetry based on PET/CT imaging, due to its simplicity of implementation.

TH-C-351-06: Evaluation of a Compartmental Model for Estimating Tumor Hypoxia Via FMISO Dynamic PET Imaging

Purpose: To evaluate a pharmacokinetic compartmental model for identifying intra‐tumor hypoxia using dynamic positron‐emission‐tomography (PET) imaging with 18F‐fluoromisonidazole (FMISO) radiotracer.Method and Materials: The compartmental model used for this work is an irreversible generic two‐tissue type implemented within a pharmacokinetic modeling program called Voxulus by Philips Research. A dynamic PET image dataset (spatial and time) was simulated with 3 tissue regions: normoxia, hypoxia and necrosis, and with an image‐based arterial input function. Each voxelized tissue time‐activity‐curve (TAC) simulation used typical kinetic parameters, generalized from 6 head‐and‐neck cancer patient FMISO‐PET data. The dynamic image was first produced without any statistical noise, to ensure that correct kinetic parameters were reproducible by Voxulus. Next, to investigate the stability of kinetic parameter estimation in the presence of noise, 1000 noisy samples of the dynamic image were generated, from which 1000 noisy samples of kinetic parameters were calculated, and used to estimate sample mean and covariance matrix. To further investigate how bias in the arterial input function affected the kinetic parameter estimation, a shift error was introduced in the peak amplitude, peak location and tail amplitude of the input TAC, and the bias of various kinetic parameters computed. Results: Without noise, the estimated kinetic parameters matched their true values perfectly. With noise, the hypoxia rate constant k3 had more variation than other parameters. The plasma‐to‐tissue and tissue‐to‐plasma rate constants (k1 and k2) for diffusible compartment, and vascular density β were highly correlated with each other; while k3 had no correlation with others. Voxulus was applied to estimate parametric image maps of hypoxia for 6 head‐and‐neck cancer patients. Conclusion: Mathematical phantom studies have been used to determine the statistical accuracy of Voxulus, which provides us guidance and confidence in clinical dynamic FMISO‐PET data analysis.