|
You want to find out more about research in our department? Ok. We asked members from our staff to give a short description of what they are working on at the moment. Here are three of these descriptions:
Markus Alber:
Central area of expertise is the development of biological models for IMRT optimisation and the mathematics
of optimisation. This went into the basic design of the treatment planning system Hyperion as a modular tool
for clinical demands. Hyperion allows to employ both biological and physical objectives. It permits to optimise
treatment plans with Monte-Carlo dose computation including head-scatter modelling and enhanced practicality of
compensators or dMLC treatments.  The optimisation of
beam portal directions is in the stage of implementation. As a sideline, the development of NTCP models including
temporal aspects is pursued.
The project is supported by the Deutsche Krebshilfe.
Interesting clinical example (502 kB)
Daniela Thorwarth:
Hypoxiesensitive Fmiso PET-Bildgebung zur biologisch adaptiven
Bestrahlungsplanung
Tumor hypoxia (undersupply of oxygen) leads to a reduced sensitivity regarding
radiotherapy treatment. In order to control this by a factor two to three increased
resistance, imaging modalities to visualize hypoxia in the first place are
necessary to be subsequently able to design individually adapted treatment strategies.
[18-F]-Fluoromisonidazole (Fmiso) is a positron emission tomography (PET)
tracer that selectively binds in hypoxic tumor areas.
The identification of hypoxia in tumor areas is performed by a kinetic analysis
of dynamically acquired Fmiso PET scans. This method allows to determine
characteristic parameters for the structural architecture of the underlying
tumor tissue. The determined parameter values for tissue hypoxia and also
perfusion hint at the probability of success of a conventional treatment.
The investigated kinetic analysis of Fmiso PET scans might in the future
provide a basis to select patients that could maximally profit of an
individually adapted radiotherapy treatment. Additionally, the resulting
model parameters may allow to determine individual dose escalation factors to
overcome the radiation resistance induced by tumor hypoxia.
 Poster Fortüne.
Talk Estro'04.
This project is supported by the Deutsche
Forschungsgemeinschaft.
Matthias Söhn
Geometrical uncertainties in radiotherapy
The aim of radiotherapy is to apply a dose as high as possible to the
region of a tumor, without affecting too much adjecent healthy tissue.
The accuracy here is essentially limited by so-called geometric uncertainties. As an
example, these may result from positioning errors of the patient on the
treatment table or from internal changes of position and shape of
organs between treatment sessions. However, due to e.g. breathing,
tumor motion may be present during the irradiation session as well.
In this project methods for the quantitative characterisation of organ
motion are developed. In future versions of the treatment planning
software Hyperion this will
be used for individual, probability based IMRT optimization, which
allows a higher tumor control probability without increasing side
effects in healthy tissues.
An algorithm based on methods from multivariate statistics was developed which quantifies the
statistically most relevant modes of patient-specific geomteric change.
This method was successfully applied to patient data of the prostatic region.
Moreover, an automatic algorithm for fast deformable registration was developed.
The algorithm can extract deformation fields and thus the patient specific tumor motion from
computertomographic (CT) data (or 3D-datasets of other imaging modalities). A typical example for an
application of the algorithm are time-resolved thoraic CTs ("4D-CTs"), for which the deformation field of the lung
due to breathing and the tumor motion can be calculated.
Söhn, Birkner, Yan and Alber 2005, PMB 50(24), p. 5893-5908
Presentation given at ESTRO'07.
Söhn, Birkner, Chi, Wang, Yan, Berger and Alber 2008, MedPhys 35(3), p. 866-878
Presentation given at ESTRO'08.
Clinical studies about (side)effects of Radiotherapy
Profound knowledge about dose-effect relationships is a central element of modern, evidence-based radiotherapy.
For this project, a number of clinical studies on data of prostate and rectal cancer patients were carried out, which
allowed quantitative determination of parameters for different
biological
normal tissue complication models. This allows more specific targeting with a better control of side-effects for future treatments
of these rather common types of cancer.
Söhn, Yan, Liang, Meldolesi, Vargas and Alber 2007, IJROBP 67(4), p. 1066-1073
Söhn, Alber and Yan 2007, IJROBP 69(1), p. 230-239
Poster at ASTRO'07
The projects are supported by the Deutschen Krebshilfe
and the Deutschen Forschungsgemeinschaft.
Niklas Rehfeld:
Compression of a Monte-Carlo bases system matrix for positron emission tomography
High scatter fractions of modern PET scanners require an accurate
treatment of scatter. This scatter can be calculated with Monte-Carlo methods
up to arbitrary accuracy if modeled correctly. To achieve this, many simulated particle histories are averaged. In principle it is
therefore possible to calculate the system matrix very precisely.
3D PET scanners (with much higher scatter fractions than 2D scanners), however, are modeled by system matrices
that have approximately 1 mio times more elements than a recent computer can store (RAM). In addition, a correct treatment of scatter
demands a great many of histories. This results in a very long lasting computer simulation.
For this reasons the topic of this work is the compression of system matrices of PET scanners. A Monte-Carlo programm for photon tracing in the patient is under development. Since a direct storage of the system matrix of a 3D scanner
is prohibitive, compressed and uncompressed system matrices of a 2D scanner are simulated and the reconstructed images are compared.
Poster IEEE Medical Imaging Conference 2004.
Poster IEEE Medical Imaging Conference 2005.
Poster EuroMedIm 2006.
This project is supported by the Deutsche
Forschungsgemeinschaft.
Martin Soukup
Dose computation for treatment planning in proton therapy
Proton therapy is due to the advantageous properties of the proton dose
distribution (Bragg-peak) an interesting alternative to
photon or electron radiotherapy. An especially promising technique is
IMPT (Intensity modulated proton
therapy). With this technique, weights of typically several thousands
of narrow proton
beams are optimized to spare healthy tissue as much as possible
with simultaneous killing of cancer cells.
For clinical use in IMPT an accurate pencil beam algorithm is being
developed and implemented in
Hyperion treatment planning system.
 
To compute dose distributions in radiotherapy most precisely, Monte
Carlo
methods are prime choice but unfortunately, they are usually quite
slow. Monte
Carlo based codes VMCpro, Geant4 and Fluka are used for several tasks
connected with proton therapy (validations of pencil beam algorithm,
investigation of important effects like dose due to products of nuclear
interactions etc). Fast Monte Carlo code VMCpro that is specially
designed
for purposes of treatment planning in
proton therapy is implemented into Hyperion.
Monte Carlo methods are of course widely used also in photon therapy.
Geant4 allows simulations of complicated geometries like precise model
of multi-leaf collimator used in IMRT. It is therefore tested for
reliability of simulations of linear accelerator treatment
head.
The importance of nuclear interactions for dose
calculations in proton therapy - ESTRO'03
Geant4 in Photon Radiotherapy -
Comparison with BEAM and Experimental Results - Poster Montreal'04
The project is supported by the Deutsche
Forschungsgemeinschaft
Urszula Jelen
Pencil Beam dose computation algorithm for IMRT
The primary requirements for any dose computation algorithm are accuracy, speed and simplicity of the commissioning method. For new and complex irradiation technique like intensity modulated radiotherapy (IMRT), requiring a high spatial accuracy, an important factor is the proper consideration of a beam penumbra, as a dose gradient misplacement may cause the overdosage of critical structures adjacent to the target volume.
One approach to IMRT optimisation starts with the discretization of the radiation fields into small elements, called beamlets or pencil beams. Dose is then calculated as a superposition of a number of such finite-size pencil beams, and intensity modulation is obtained by weighting of the fluence assigned to them. In the final step the dose distribution created by modulated fields is recalculated with Monte Carlo method. This concept of pre-computation and storage of beamlet dose distributions as well as the close proximity between optimized (from fsPB) and final dose (from Monte Carlo) is a crucial element in improving and accelerating IMRT routine planning.
A finite size pencil beam (fsPB) algorithm was designed and commissioned specifically for the purpose of beamlet-based IMRT. The algorithm employs an analytical function for the cross-profiles of the beamlets which is based on the assumption of self-consistency, i.e. the requirement that an arbitrary superposition of abutting beamlets should add up to a homogeneous field. By choosing the size of the beamlets appropriately and small enough, arbitrary MLC field shapes can be created.
Inclusion of lateral density correction into the algorithm is the subject of current research.
Breast IMRT - treatment planning study
A standard approach for the adjuvant whole breast irradiation is to deliver the prescribed dose by means of two tangential fields. This may however produce inhomogeneous dose distributions due to the variations in thickness across the target volume and parts of the underlying lung, left ventricle and ribs may be included in the treatment fields.
The new radiation techniques such as IMRT offer the possibilities to improve the dose homogeneity and reduce the normal tissue complication probability employing inverse planning methods and biological complication models for heart and lungs.
A potential of inverse treatment planning system Hyperion for whole breast irradiation was investigated. For breast cancer patients who underwent the breast conserving radiotherapy IMRT plans were prepared with an objective to obtain the high target conformity and dose homogeneity and maintaining the irradiation of organs at risk: ipsilateral lung, heart and contralateral breast below certain limits. In order to establish the class
solution for clinical implementation of IMRT for breast cancer cases beam setup choice was investigated.
The project is supported by Fortüne.
| 
Welcome!
Markus Alber
webmaster