Irreversible electroporation (IRE), a relatively new energy-based tumor ablation technology, has shown itself in the last decade to be able to safely ablate tumors with favorable clinical outcomes, yet little work has been done on optimizing the IRE protocol to variously sized tumors. Incomplete tumor ablation has been shown to be the main reason leading to the local recurrence and thus treatment failure. The goal of this study was to develop a general optimization approach to optimize the IRE protocol for cervical tumors in different sizes, while minimizing the damage to normal tissues. This kind of approach can lay a foundation for future personalized treatment of IRE. First, a statistical IRE cervical tumor death model was built using previous data in our group. Then, a multi-objective optimization problem model was built, in which the decision variables are five IRE-setting parameters, namely, the pulse strength (U), the length of active tip (H), the number of pulses delivered in one round between a pair of electrodes (A), the distance between electrodes (D), and the number of electrodes (N). The domains of the decision variables were determined based on the clinical experience. Finally, the problem model was solved by using nondominated sorting genetic algorithms II (NSGA-II) algorithm to give respective optimal protocol for three sizes of cervical tumors. Every protocol was assessed by the evaluation criterion established in the study to show the efficacy in a more straightforward way. The results of the study demonstrate this approach can theoretically provide the optimal IRE protocol for different sizes of tumors and may be generalizable to other types, sizes, and locations of tumors.

Background: The purpose of this project is to identify prognostic features in resectable pancreatic head adenocarcinoma and use these features to develop a machine learning algorithm that prognosticates survival for patients pursuing pancreaticoduodenectomy. Methods: A retrospective cohort study of 93 patients who underwent a pancreaticoduodenectomy was performed. The patients were analyzed in 2 groups: Group 1 (n = 38) comprised of patients who survived < 2 years, and Group 2 (n = 55) comprised of patients who survived > 2 years. After comparing the two groups, 9 categorical features and 2 continuous features (11 total) were selected to be statistically significant (p < .05) in predicting outcome after surgery. These 11 features were used to train a machine learning algorithm that prognosticates survival. Results: The algorithm obtained 75% accuracy, 41.9% sensitivity, and 97.5% specificity in predicting whether survival is less than 2 years after surgery. Conclusion: A supervised machine learning algorithm that prognosticates survival can be a useful tool to personalize treatment plans for patients with pancreatic cancer.