Importance of electrode position for the distribution of tumor treating fields (TTFields) in a human brain. Identification of effective layouts through systematic analysis of array positions for multiple tumor locations

Tumor treating fields (TTFields) is a new modality used for the treatment of glioblastoma. It is based on antineoplastic low-intensity electric fields induced by two pairs of electrode arrays placed on the patient's scalp. The layout of the arrays greatly impacts the intensity (dose) of TTFields in the pathology. The present study systematically characterizes the impact of array position on the TTFields distribution calculated in a realistic human head model using finite element methods. We investigate systematic rotations of arrays around a central craniocaudal axis of the head and identify optimal layouts for a large range of (nineteen) different frontoparietal tumor positions. In addition, we present comprehensive graphical representations and animations to support the users' understanding of TTFields. For most tumors, we identified two optimal array positions. These positions varied with the translation of the tumor in the anterior-posterior direction but not in the left-right direction. The two optimal directions were oriented approximately orthogonally and when combining two pairs of orthogonal arrays, equivalent to clinical TTFields therapy, we correspondingly found a single optimum position. In most cases, an oblique layout with the fields oriented at forty-five degrees to the sagittal plane was superior to the commonly used anterior-posterior and left-right combinations of arrays. The oblique configuration may be used as an effective and viable configuration for most frontoparietal tumors. Our results may be applied to assist clinical decision-making in various challenging situations associated with TTFields. This includes situations in which circumstances, such as therapy-induced skin rash, scar tissue or shunt therapy, etc., require layouts alternative to the prescribed. More accurate distributions should, however, be based on patient-specific models. Future work is needed to assess the robustness of the presented results towards variations in conductivity.