Real-Time Monitoring Of Microwave Ablation Using Artificial Neural Network And Ultra-Thin Soft Temperature Sensors

Microwave ablation(MWA)is a commonly used minimally invasive thermal ablation technique for tumors in clinical practice.The principle is to use one or multiple microwave antennas to radiate electromagnetic waves in the tumor lesion area,causing polar molecules(such as water molecules)in the area to oscillate to generate heat,causing tissue proteins to become inactive due to high-temperature denaturation,leading to coagulation necrosis of tumor cells.To avoid insufficient or excessive ablation,the real-time monitoring of tissue temperature in the target area is necessary during the treatment process.However,existing microwave ablation systems cannot provide the real-time temperature information during the treatment process,and only determine the ablation zone of the tissue through postoperative imaging(i.e.,ultrasound,CT,or MRI).The additional temperature probes found in literature can be inserted to monitor ablation temperature,but this method can cause secondary harm to the patient and the position for temperature monitoring is not accessible in some circumstances.Therefore,in order to address the above issue,this study proposed a real-time monitoring technology for MWA by using the artificial neural network technique and ultra-thin soft temperature sensors;The proposed method in this work can achieve the real-time monitoring of temperature and ablation zone without an additional invasion to the human body.The main research objectives of this work include:(1)to design an ultra-thin and soft temperature sensor array that can be installed on a microwave antenna probe without significantly increasing the antenna diameter.The sensor array mainly consists of several temperature measurement units and interface units,which has an extremely high temperature sensitivity and provides the real-time temperature information of tissues along the antenna probe.(2)To build a finite element model of MWA for liver cancer and create an artificial neural network(ANN)model between the temperature information obtained by the temperature sensor array and the ablation zone under different MW powers,ablation durations,and measured temperatures.(3)to validate the proposed ANN model by using an ex vivo MWA experiment with egg white.The temperatures of the tissues along the antenna probe,input MW powers,ablation durations were taken as the inputs of the ANN model,while the long and short diameters of ablation zone were taken as the outputs of the ANN model.The ANN model was created by using the data generated in the computer simulation with liver cancer and then validated in the ex vivo experiment with egg white.The relevant data obtained based on simulation and in vitro experiments indicated that the use of ultra-thin flexible sensors can achieve real-time monitoring of tissue temperatures along the antenna probe without leading to extra invasiveness.Combined with the ANN prediction model,the temperature-sensor-array installed MW antenna can be used to achieve the real-time monitoring of MWA in the treatment of liver tumors.