AFM-based robust image analysis to contrast reversal effects in cell-cerium oxide nanoparticles interactions

Atomic force microscopy is gaining interest as a technique to quantitatively study biological samples in native environment. However, the measuring principles behind may cause the presence of different sources of artifacts and image degradations. In this work, we present an AFM image analysis tool able to recognize morphological alterations in human leukemia cells after 3 h incubation with the antioxidant cerium oxide nanoparticles. To demonstrate the robustness of the approach to a particular artifact called contrast reversal (CR) effect, consisting in an unexpected switching between repulsive and attractive tip regime during scanning, we present a technique to artificially inject the artefact on the image and then apply the tool. Maximum area under the ROC (AUC) curve results of 0.91 (0.11) in the discrimination between exposed and unexposed cells confirm the validity of the approach and its applicability in AFM-based cell studies.