ESS - RADT - Livro, parte de livro ou capítulo de livro
Permanent URI for this collection
Browse
Browsing ESS - RADT - Livro, parte de livro ou capítulo de livro by Author "Costa Ferreira, Brigida"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
- Automated Radiotherapy Treatment Planning Using Fuzzy Inference SystemsPublication . Dias, Joana; Rocha, Humberto; Ventura, Tiago; Costa Ferreira, Brigida; Lopes, Maria do CarmoRadiotherapy is one of the treatments available for cancer patients, aiming to irradiate the tumor while preserving healthy structures. The planning of a treatment is a lengthy trial and error procedure, where treatment parameters are iteratively changed and the delivered dose is calculated to see whether it complies with the desired medical prescription. In this paper, a procedure based on fuzzy inference systems (FIS) for automated treatment planning is developed, allowing the calculation of high quality treatment plans without requiring human intervention. The procedure is structured in two different phases, incorporating the automatic selection of the best set of equidistant beam irradiation directions by an enumeration procedure. The developed method is extensively tested using ten head-and-neck cancer cases.
- Comparison of Combinatorial and Continuous Frameworks for the Beam Angle Optimization Problem in IMRTPublication . Rocha, Humberto; Dias, Joana; Ventura, Tiago; Costa Ferreira, Brigida; do Carmo Lopes, MariaRadiation therapy (RT) is used nowadays for the majority of cancer patients. A technologically advanced type of RT is IMRT – intensity-modulated radiation therapy. With this RT modality the cancerous cells of the patient can be irradiated using non-uniform radiation maps delivered from different beam directions. Although non-uniform radiation maps allow, by themselves, an enhanced sparing of the neighboring healthy organs while properly irradiating the tumor with the prescribed dose, selection of appropriate irradiation directions play a decisive role on these conflicting tasks: deliver dose to the tumor while preventing (too much) dose to be deposited in the surrounding tissues. This paper focus on the problem of choosing the best set of irradiation directions, known as beam angle optimization (BAO) problem. Two completely different mathematical formulations of this problem can be found in the literature. A combinatorial formulation, widely used and addressed by many different algorithms and strategies, and a continuous formulation proposed by the authors and addressed by derivative-free algorithms. In this paper, a comparison of two of the most successful strategies to address each one of these formulations is done resorting to a set of ten clinical nasopharyngeal tumor cases already treated at the Portuguese Institute of Oncology of Coimbra.
- Feature selection in small databases: A medical-case studyPublication . Soares, Inês; Dias, Joana; Rocha, Humberto; Lopes, Maria do Carmo; Costa Ferreira, BrigidaPredictions made by using machine learning classification models are recurrent in many research fields for a variety of reasons. In some cases, feature selection can effi- ciently improve the accuracy of classifications, while reducing the computational requirements. However, some predictive studies are characterized by a high dimensionality or based on small datasets.
- A heuristic based on fuzzy inference systems for multiobjective IMRT treatment planningPublication . Dias, Joana; Rocha, Humberto; Ventura, Tiago; Costa Ferreira, Brigida; Lopes, Maria do CarmoRadiotherapy is one of the treatments used against cancer. Each treatment has to be planned considering the medical prescription for each specific patient and the information contained in the patient’s medical images. The medical prescription usually is composed by a set of dosimetry constraints, imposing maximum or minimum radiation doses that should be satisfied. Treatment planning is a trial-and-error time consuming process, where the planner has to tune several parameters (like weights and bounds) until an admissible plan is found. Radiotherapy treatment planning can be interpreted as a multiobjective optimization problem, because besides the set of dosimetry constraints there are also several conflicting objectives: maximizing the dose deposited in the volumes to treat and, at the same time, minimizing the dose delivered to healthy cells. In this paper we present a new multiobjective optimization procedure that will, in an automated way, calculate a set of potential non-dominated treatment plans. It is also possible to consider an interactive procedure whenever the planner wants to explore new regions in the non-dominated frontier. The optimization procedure is based on fuzzy inference systems. The new methodology is described and it is applied to a head-and-neck cancer case.
- Semi-supervised Self-training Approaches in Small and Unbalanced Datasets: Application to Xerostomia Radiation Side-EffectPublication . Soares, Inês; Dias, Joana; Rocha, Humberto; Khouri, Leila; Lopes, Maria do Carmo; Costa Ferreira, BrigidaSupervised learning algorithms have been widely used as predictors and applied in a myriad of studies. The accuracy of the classification algorithms is strongly dependent on the existence of large and balanced training sets. The existence of a reduced number of labeled data can deeply affect the use of supervised approaches. In these cases, semi-supervised learning algorithms can be a way to circumvent the problem.