A muscle-equivalent phantom with a four-point thermocouple temperature sensor inserted in the center of the central axis was sandwiched by overlay boluses from above and below. Electrodes with a diameter of 30 cm were further sandwiched from above and below. It was heated on a hyperthermic treatment table. After injecting 0, 150 or 300 ml of air into the upper and lower overlay boluses, heating was started.

　　As a result, 1) when there is no air entrained in the overlay bolus, the higher the output power during heating, the higher the heating efficiency, and 2) the higher the amount of air entrained in the overlay bolus, the lower the heating efficiency. 3) there was a tendency that the greater the output power during heating and the greater the amount of air entrained in the overlay bolus, the greater the variability in each temperature measurement value obtained from the multipoint thermometer.

　　Consequently, air entrapped in the overlay bolus should be removed during the pre-work inspection on the day of hyperthermia treatment to prevent reducing heating efficiency and to heat the target lesion uniformly.