A Model-Based Approach To Detection Of The Circulating Melanoma Cells From The Photoacoustic Cytometry Data

Background ― Effective treatment of melanoma is only possible during the early stages of the disease. However, current techniques of cancer diagnostics are not sensitive enough to be reliable during the early stages. Photoacoustic cytometry provides a very perspective method for early diagnostics, which, in principle, can detect even a single melanoma circulating melanoma cell. However, the photoacoustic monitoring system needs to be paired with an effective automated algorithm, which will detect acoustic responses from melanoma cells among normal blood cells. Objective ― The study aims to compare the currently adopted algorithm for the detection of circulating melanoma cells based on a calculation of the standard deviation of the acoustic responses with the new model-based algorithm. Methods ― We tested and compared two different methods of automated detection of melanoma acoustic responses. The first approach was based on a calculation of the deviation of the acoustic responses. The second approach was based on the subtraction of the ensemble average normal acoustic response from melanoma acoustic responses. The resulting differential signals contained positive spikes, which were detected. Results ― The model-based approach demonstrated better sensitivity and specificity when detecting circulating melanoma cells in comparison to the approach based on the calculation of standard deviation of the acoustic responses. Conclusions ― Obtained results suggest that the model-based approach is a more reliable basis for an automated algorithm for the detection of circulating melanoma cells. The model-based approach has an additional advantage of being simpler to calculate and requires less computational power to run in real-time.