| 1. |
- Meaney, Paul M, 1960, et al.
(författare)
-
y Two-step inversion with a logarithmic transformation for microwave breast imaging
- 2017
-
Ingår i: Medical Physics. - : Wiley. - 2473-4209 .- 0094-2405. ; 44:8, s. 4239-4251
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: The authors have developed a new two-step microwave tomographic image reconstruction process specifically designed to incorporate logarithmic transformed microwave imaging algorithms as a means of significantly improving spatial resolution and target property recovery. Log transform eliminates the need for a priori information, but spatial filtering often integrated as part of the regularization required to stabilize image recovery, generally smooths image features and reduces object definition. The new implementation begins with this smoothed image as the first step, but then utilizes it as the starting estimate for a second step which continues the iterative process with a standard weighted Euclidean distance regularization. The penalty term of the latter restricts the new image to a multi-dimensional location close to the original but allows the algorithm to optimize the image without excessive smoothing. Methods: The overall approach is based on a Gauss-Newton iterative scheme which incorporates a log transformation as a way of making the reconstruction more linear. It has been shown to be robust and not require a priori information as a condition for convergence, but does produce somewhat smoothed images as a result of associated regularization. The new two-step process utilizes the previous technique to generate a smoothed initial estimate and then uses the same reconstruction process with a weighted Euclidean distance penalty term. A simple and repeatable method has been implemented to determine the weighting factor without significant computational burden. The reconstructions are assessed according to conventional parameter estimation metrics. Results: We apply the approach to phantom experiments using large, high contrast canonical shapes followed by a set of images recovered from an actual patient exam. The image improvements are substantial in regards to improved property recovery and feature delineation without inducing unwanted artifacts. Analysis of the residual vector after the reconstruction process further emphasizes that the minimization criterion is efficient with minimal biases. Conclusions: The outcome is a novel synergism of an established stable reconstruction algorithm with a conventional regularization technique. It maintains the ability to recover high quality microwave tomographic images without the bias of a priori information while substantially improving image quality. The results are confirmed on both phantom experiments and patient exams.
|
|
| 2. |
- Uusijärvi, Helena, 1979, et al.
(författare)
-
Dosimetric characterization of radionuclides for systemic tumor therapy: influence of particle range, photon emission, and subcellular distribution.
- 2006
-
Ingår i: Medical physics. - : Wiley. - 0094-2405 .- 2473-4209. ; 33:9, s. 3260-9
-
Tidskriftsartikel (refereegranskat)abstract
- Various radionuclides have been proposed for systemic tumor therapy. However, in most dosimetric analysis of proposed radionuclides the charged particles are taken into consideration while the potential photons are ignored. The photons will cause undesirable irradiation of normal tissue, and increase the probability of toxicity in, e.g., the bone marrow. The aim of this study was to investigate the dosimetric properties according to particle range, photon emission, and subcellular radionuclide distribution, of a selection of radionuclides used or proposed for radionuclide therapy, and to investigate the possibility of dividing radionuclides into groups according to their dosimetric properties. The absorbed dose rate to the tumors divided by the absorbed dose rate to the normal tissue (TND) was estimated for different tumor sizes in a mathematical model of the human body. The body was simulated as a 70-kg ellipsoid and the tumors as spheres of different sizes (1 ng-100 g). The radionuclides were either assumed to be uniformly distributed throughout the entire tumor and normal tissue, or located in the nucleus or the cytoplasm of the tumor cells and on the cell membrane of the normal cells. Fifty-nine radionuclides were studied together with monoenergetic electrons, positrons, and alpha particles. The tumor and normal tissue were assumed to be of water density. The activity concentration ratio between the tumor and normal tissue was assumed to be 25. The radionuclides emitting low-energy electrons combined with a low photon contribution, and the alpha emitters showed high TND values for most tumor sizes. Electrons with higher energy gave reduced TND values for small tumors, while a higher photon contribution reduced the TND values for large tumors. Radionuclides with high photon contributions showed low TND value for all tumor sizes studied. The radionuclides studied could be divided into four main groups according to their TND values: beta emitters, Auger electron emitters, photon emitters, and alpha emitters. The TND values of the beta emitters were not affected by the subcellular distribution of the radionuclide. The TND values of the Auger electron emitters were affected by the subcellular radionuclide distribution. The photon emitters showed low TND values that were only slightly affected by the subcellular radionuclide distribution. The alpha emitters showed high TND values that were only slightly affected by the subcellular radionuclide distribution. This dosimetric characterization of radionuclides may be valuable in choosing the appropriate radionuclides for specific therapeutic applications.
|
|
| 3. |
- Zanoli, Massimiliano, 1989, et al.
(författare)
-
Suitability of eigenvalue beam-forming for discrete multi-frequency hyperthermia treatment planning
- 2021
-
Ingår i: Medical Physics. - : Wiley. - 2473-4209 .- 0094-2405. ; 48:11, s. 7410-7426
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: Thermal dose delivery in microwave hyperthermia for cancer treatment is expected to benefit from the introduction of ultra-wideband (UWB)-phased array applicators. A full exploitation of the combination of different frequencies to improve the deposition pattern is, however, a nontrivial problem. It is unclear whether the cost functions used for hyperthermia treatment planning (HTP) optimization in the single-frequency setting can be meaningfully extended to the UWB case. Method: We discuss the ability of the eigenvalue (EV) and a novel implementation of iterative-EV (i-EV) beam-forming methods to fully exploit the available frequency spectrum when a discrete set of simultaneous operating frequencies is available for treatment. We show that the quadratic power deposition ratio solved by the methods can be maximized by only one frequency in the set, therefore rendering EV inadequate for UWB treatment planning. We further investigate whether this represents a limitation in two realistic test cases, comparing the thermal distributions resulting from EV and i-EV to those obtained by optimizing for other nonlinear cost functions that allow for multi-frequency. Results: The classical EV-based single-frequency HTP yields systematically lower target SAR deposition and temperature values than nonlinear HTP. In a larynx target, the proposed single-frequency i-EV scheme is able to compensate for this and reach temperatures comparable to those given by global nonlinear optimization. In a meninges target, the multi-frequency setting outperforms the single-frequency one, achieving better target coverage and (Formula presented.) higher (Formula presented.) in the tumor than single-frequency-based HTP. Conclusions: Classical EV performs poorly in terms of resulting target temperatures. The proposed single-frequency i-EV scheme can be a viable option depending on the patient and tumor to be treated, as long as the proper operating frequency can be selected across a UWB range. Multi-frequency HTP can bring a considerable benefit in regions typically difficult to treat such as the brain.
|
|
