Ultrasonic assessment of death in HEp2 cells using spectral and wavelet based analysis of backscattered RF-signals

The efficacy of anticancer therapies cannot be assessed easily during the treatment yet it is an important determinant of eventual treatment success. It has been observed that ultrasound backscatter from collections of acute myeloid leukaemia cells that are undergoing cell death due to chemo- or radiation therapy show increased intensity and changes in spectral characteristics. However, similar experiments have not been performed on epithelial cells, of greater relevance to solid tumors. High frequency ultrasound in the range between 10 and 35 MHz was used to investigate the treatment dependent changes in ultrasound backscatter characteristics of epithelial cells (HEp2, larynx carcinoma cells). The data collection was performed at 7 time points during the treatment, with a control sample for each time point, using an ultrasound bio-microscope in combination with a 20MHz f# 2.35 transducer. Normalized power spectra and wavelet coefficients (using a Gaussian mother wavelet) were computed from the backscattered radio frequency ultrasound signals within a defined region of interest. Spectral slopes, integrated backscatter coefficients and wavelet parameters were estimated as a function of treatment time to monitor changes of the acoustical properties during apoptosis. The percentage of cells that were undergoing apoptosis was estimated using acridine orange-ethium bromide staining combined with fluorescence microscopy. Changes in spectral parameters were detected 17 hours post treatment and coincided with changes in the corresponding histology. Between untreated cells and 48 hours of treatment spectral slope values decreased by 30% within the frequency range analyzed. Backscatter (as measured by the mid-band fit) increased by 80% over the same time period. At these times, the histology indicated that 40% of the cells were undergoing apoptosis. Wavelet parameters showed similar trends compared with the spectral analysis. The detected changes are linked to structural cell alterations when undergoing apoptosis. We have shown that for epithelial cancer cells the structural changes during apoptosis can be imaged in the cell pellet model. This study contributes towards a non-invasive method for estimating the success of anticancer treatment responses.