Daniel Baumgarten

Wir präsentieren einen automatischen Algorithmus zur Registrierung und überlagerung von Fundusbildern zu gro flächigen Kompositionsaufnahmen. Das Verfahren kombiniert flächenbasierte und punktbasierte Ansätze. Als ähnlichkeitsma normierte Korrelationskoeffizient, der sich im Vergleich zur Transinformation als robuster erwies und schneller zu berechnen ist. Den Transformationen der Bilder liegt ein quadratisches Modell zugrunde, das die annähernd sphärische Oberfläche der Retina berücksichtigt und anhand visueller Bewertung ausgewählt wurde. Bei der Validierung an realen klinischen Daten erwies sich der vorgestellte Algorithmus als robust und genau. Die Grenzen des Verfahrens bilden sehr unscharfe Bilder und solche, die nur sehr wenig relevante Strukturen enthalten.

Magnetic marker monitoring (MMM) is a technique to determine the motility of the gastrointestinal tract and to observe the dissolution of pharmaceutical compounds. Today's magnetic markers usually consist of magnetized magnetite. Because of their weak magnetic fields, highly sensitive sensor systems are required. For a wider class of applications, stronger markers and more flexible measurement setups are necessary. In this paper, a novel marker design is introduced. This marker comprises one permanent magnet and a compartment of iron powder in a magnetically unstable configuration. During dissolution of the pharmaceuticals, the powder is redistributed around the magnet, thereby altering the externally measured magnetic induction. Based on this design, magnetically marked tablets and capsules were prepared and their magnetic field during dissolution was observed. Magnetic induction values were between 16 and 0.2 μT at distances of 5-30 cm, which is considerably higher compared to the pico-Tesla range of conventional markers. During dissolution, the magnetic induction decreased by between 14% and 27%. These values could be confirmed in detailed finite element method simulations. In conclusion, the present results indicate that our novel marker design is well suited for MMM with more flexible sensor technologies, such as magnetoresistive sensors.

Magnetic nanoparticles can be employed for a broad range of biomedical applications where the knowledge of the distribution of the magnetic nanoparticles is of importance for efficacy, patient's safety, etc. The need exists to have an as accurate as possible quantification of the unknown particles distribution. Magnetorelaxometry (MRX) measurements are able to measure the magnetic induction originating from a certain distribution of magnetically activated nanoparticles. Starting from these measurements it is possible to determine the distribution using a minimum norm estimation technique. This approach is however ill-posed. We sequentially activate the magnetic nanoparticles through the use of excitation coil arrays with the aim to reduce the ill-posedness. This paper presents some advancements in magnetic nanoparticle reconstruction in terms of reconstruction quality using numerical simulations. The results show that inhomogeneous sequential activation is a proper alternative to homogeneous activation with Helmholtz coils since an increase in accuracy with a factor ranging from 1.5 until 2 is obtained. The presented numerical techniques coupled to MRX measurements can be of significant aid so to have more quantitative knowledge of the biodistribution.

In magnetic nanoparticle imaging, magnetic nanoparticles are coated and functionalized to bind to specific targets. After measuring their magnetic relaxation or remanence, their distribution can be determined by means of inverse methods. The reconstruction algorithm presented in this paper includes first a dipole fit using a Levenberg–Marquardt optimizer to determine the reconstruction plane. Secondly, a minimum norm estimate is obtained on a regular grid placed in that plane. Computer simulations involving different parameter sets and conditions show that the used approach allows for the reconstruction of distributed sources, although the reconstructed shapes are distorted by blurring effects. The reconstruction quality depends on the signal-to-noise ratio of the measurements and decreases with larger sensor-source distances and higher grid spacings. In phantom measurements, the magnetic remanence of nanoparticle columns with clinical relevant sizes is determined with two common measurement systems. The reconstructions from these measurements indicate that the approach is applicable for clinical measurements. Our results provide parameter sets for successful application of minimum norm approaches to Magnetic Nanoparticle Imaging.

Abstract Magnetic nanoparticles have a variety of applications in medicine. One of these applications is magnetic imaging, where functionalized nanoparticles bind to specific targets in the human body. The distribution of the particles can be reconstructed from the magnetic field ...