Verification of self-calibration algorithms for phased array applicators

Other conference contribution, 2018

Introduction
In microwave hyperthermia (MW-HT), treatment planning determines the steering parameters for a phased array to yield appropriate tumor coverage and hot-spot suppression. In real HT systems, however, such arrays are subjected to mismatches, which might not be considered in the models used in treatment planning. While certain mismatches can be addressed via channel calibration, those occurring inside the array are more difficult to predict as they can vary during the treatment session itself. The effect of such mismatches can be as relevant as to disrupt the interference pattern.
Objectives
This contribution proposes self-calibration (SC) as a solution for real-time compensation of various types of mismatches, such as different cable lengths, manufacturing tolerances, patient misplacement and air bubbles in the water bolus. Two SC algorithms have been designed for use with applicator arrays of arbitrary shapes.
Materials & Methods
The algorithms are based on comparison of simulated and measured S-matrices of the phased array. The extra time delays caused by various mismatches at each channel are then compensated accordingly. The verification of both algorithms includes virtual and experimental models of our neck applicator used in a setup with a patient model and a muscle phantom. The accuracy has been evaluated numerically by comparing the ideal E-field distributions with those obtained by introducing a set of randomly distributed mismatches to the applicator model. The proof-of-principle has then been demonstrated experimentally by means of temperature measurements.
Results
Results indicate that at least one of the tested SC algorithms converge to the correct compensation solution with performances largely comparable and sometimes even exceeding those typical of an external calibration. Antenna offsets of ±5 mm and air bubbles about 1 cm big are well handled. Improvements can be done with respect to patient misplacement, which is compensated by the algorithm up to ±1 mm. Experimental results confirm the ability of the algorithm to restore focus shape.
Conclusion
Self-calibration can be a valid solution for mismatch compensation in MW-HT. The potential real-time application of SC makes it a desirable candidate for use in clinical settings.