Impact of segmentation detail in hyperthermia treatment planning: comparison between detailed and clinical tissue segmentation

Treatment planning for deep pelvic hyperthermia is currently based on tissue models comprising four tissue categories. For head and neck hyperthermia, we earlier found that more tissues are required for an accurate representation. Hence, we studied the accuracy of the clinical tissue segmentation (4 tissues) using a full detailed tissue list segmentation (80 tissues) as benchmark. The SAR and temperature distributions were evaluated and relevant differences were found. Also, the large and unknown variation in blood perfusion results in a large uncertainty in the predicted temperature distributions. In summary, this study showed that the number of tissues segmented is relevant for both SAR and the temperature prediction accuracy.