Hyperthermia treatment

Hyperthermia treatment planning for head and neck cancer: improving patient models using advanced MRI characteristics.

Information Technology Foundation STW project 10846


Treatment of cancer in the head and neck region is difficult and the side-effects of currently standard treatment modalities (radiotherapy and chemotherapy) are severe, i.e. patients often loose swallowing function and/or saliva production. The effectiveness of these treatments is greatly enhanced when a hyperthermia treatment, i.e. increase of tumour temperature to the range of 39-44°C, is added while side-effects remain the same. Heating tumours in the head and neck is challenging, so we developed specific heating equipment. Simultaneously, hyperthermia treatment planning is developed to exploit the specific heat-focusing capabilities of the new equipment. The hyperthermia treatment planning consists of two models: electromagnetic (EM) models and thermal models. The input for these models is a 3d patient specific model, usually segmented from computed tomography (CT). Currently, both the EM and thermal models are either non-specific or require a labour-intensive delineation of tissues on image data.

The purpose of this project is to facilitate accurate and efficient model generation for developing patient specific EM and thermal hyperthermia treatment planning. Tools will be developed, validated and clinically integrated to generate patient specific EM models. Furthermore, tools will be developed and validated for the computation and integration of vascular tissue characteristics in patient-specific thermal models based on MRI data. By these improved patient models, head and neck tumours can potentially be heated more accurately leading to a more effective treatment.

Plan of investigation
We will develop an automatic (atlas-based) segmentation algorithm to speed up the tissue delineation process. A combination of both CT and MRI images will be used to improve the accuracy of the delineations.
The most predictive thermal model will be assessed by correlating the temperature distribution from the model with the temperature measurements during treatment. The required perfusion parameters and vascular information for these models will be determined using CT and/or MRI techniques. The end goal is to get the thermal modeling ready for a clinical validation study.

The current procedure of patient model generation for H&N HTP has a number of drawbacks. Creating and EM model is time consuming and prone to large inter- and intra-rater variability. Currently, the Rotterdam HT group is one of the few centers that uses patient specific EM-based HTP in the clinic. For thermal modeling, the situation is even worse. Worldwide, no institute is using thermal simulations in the clinical routine due to time-requirements and accuracy problems of patient-specific model generation. Currently, HT is controlled using glass-fiber thermometers inserted in tissue. HTP has a great potential to replace such measurements, which is specifically important for the H&N area where such measurements are often not feasible. Faster and more accurate HTP will facilitate more accurate application of HT and, hence, their clinical use will increase significantly. Therefore, all institutes and manufacturers working in the HT field will benefit from faster and more-accurate segmentation algorithms, e.g. SPEAG: the developers of the most advanced HTP program (SEMCAD X). The results from this project can be extended for patient-specific model generation for other HT sites and will be a powerful tool in the development of MR equipment and safety procedures.