Abstract
Currently there is no in vitro testing of glioblastoma biopsy material to assess tumour sensitivity to radiation, which could form a basis for personalised treatment plans for patients. In this work, a model to predict sensitivity to radiation, via the micronucleus assay, is set out and a proof of concept is presented where numbers of micronuclei in a glioblastoma cell line, LN18 is predicted. One key requirement for the model is that any in vitro testing needs to yield results within a few days, as the timeline for glioblastoma patients from diagnosis to treatment is short. In order to achieve this, a flow cytometry technique is assessed against traditional fluorescence microscopy for detection of micronuclei. Flow cytometry was completed using an in vitro Microflow kit from Litron Laboratories. There was no previous experience using this kit in cancerous cell lines and limited experience in cell lines that adhere to their flask’s surface as the kit is used mostly in peripheral blood lymphocytes. The flow cytometry technique can be completed within the required time frame and is much less labour intensive that fluorescence microscopy. However, there is a significant amount of variance between samples which makes the microscopy results more useful for fitting the modelling work to. It is expected with further experience and use of the supplied template, as this was incompatible at the time of the experiments, will play a role in reducing some of the variance. The model, in its current state of development, is able to predict numbers of micronuclei in a cohort of cells following doses of radiation between 1 and 3 Gy. The numerical solution is based on a decision tree structure where each double strand break that would be caused by radiation is run through the tree. The tree is traversed populated using probabilities for each decision, such as the success of a repair pathway, and Monte Carlo methods for predicting the cohort response to radiation. These probabilities are fitted to experimental data. The prediction of micronuclei is the first step for the MiNiMUS model. Future work should prioritise incorporating cell death into the model and further assessing the suitability of flow cytometry for rapid micronuclei detection.