Relationship of Antenna Work and Ablation Cavity Volume Following Percutaneous Microwave Ablation of Hepatic Tumors.

PURPOSE: To formulate a statistical model relating ablation time, power, and work with posttreatment cavity volume following percutaneous microwave ablation of hepatic tumors in vivo. MATERIALS AND METHODS: A retrospective review (October 2015 to October 2018) yielded 122 hepatic tumors treated with microwave ablation. Ablation cavity dimensions were measured at 1-month follow-up examination and calculated using an ellipsoid volume formula. The antenna manufacturer (Neuwave Medical, Madison, Wisconsin) provided the activation time and energy used to calculate the antenna work. Generalized estimating equations with ordinary least-squares regression models were obtained to relate tumor volume with cumulative antenna work. Coefficient of determination (R2) and mean square error were used as statistical measures of model prediction performance. RESULTS: There is a logarithmic relationship between postablation cavity volume (cm3) and cumulative work (kJ), represented by the formula: log10 cm3 = -0.4583 + 0.9887 × cumulative work (log10 kJ) (R2 = 0.41, mean square error, 0.102). Ablation volumes were predicted as a function of antenna work, calculated using an antilog transformation. When a single antenna was used, ablation cavity volume was predicted using a generalized estimating equation ordinary least-squares regression model of power and time: log10cm3= -0.0546 + 0.0485 × total time (min) + 0.0107 × power (W) (R2 = 0.30; mean square error, 0.106). Using this model, a nomogram was developed to predict the postablation cavity volume based on total activation time and target power. CONCLUSION: There is a logarithmic relationship between the ablation work and posttreatment ablation cavity volume, which can be expressed in a nomogram when using a single probe.

Automatic data analysis of real-time song and locomotor activity in zebra finches.

The zebra finch is a superb natural animal model to study cognition and as such could contribute to the further understanding of nicotinic intervention therapies for patients suffering from cognitive impairment as observed in neurodegenerative disorders. Manual analysis of data produced by this model is extremely labour intensive, error-prone, and typically takes weeks to complete. We designed data acquisition methods, selected analysis algorithms, and developed software to efficiently and accurately automate the detection and classification of song production (cognitive functioning) and locomotor activity (physical condition). Our custom-designed software accurately classifies song and locomotor activities. After classification, the reduced data sets can be further analysed with popular tools, such as 'R'.