Machine Learning Assisted Classification of Cell Lines and Cell States on Quantitative Phase Images

In this report, we present implementation and validation of machine-learning classifiers for distinguishing between cell types (HeLa, A549, 3T3 cell lines) and states (live, necrosis, apoptosis) based on the analysis of optical parameters derived from cell phase images. Validation of the developed classifier shows the accuracy for distinguishing between the three cell types of about 93% and between different cell states of the same cell line of about 89%. In the field test of the developed algorithm, we demonstrate successful evaluation of the temporal dynamics of relative amounts of live, apoptotic and necrotic cells after photodynamic treatment at different doses.

Significant difference in response of malignant tumor cells of individual patients to photodynamic treatment as revealed by digital holographic microscopy

The investigation of in-vitro response of cell cultures derived from tumor material of individual patients with similar tumor localizations to photodynamic treatment is presented. Tumor types included in the research were renal cell carcinoma, melanoma and alveolar, synovial, lypo- and osteo- sarcomas. Long-term observations of treatment-induced morphological changes in cells were performed by means of digital holographic microscopy. A substantial variance in response of cells of individual patients with similar tumor types and localizations to photodynamic treatment with the same dose has been observed. These peculiarities are indicative of the demand to personalized protocols of photodynamic treatment. The elevated resistance of cells of some patients to treatment at high doses highlights potential limitations of photodynamic therapy for some patients. Digital holographic microscopy is shown to be an informative label-free noninvasive tool allowing for long-term monitoring of cell samples in vitro and providing quantitative information on necrosis rate and loss of cellular dry mass. The developed methodology can be generalized for analysis of cellular response to various therapies.

Quantitative assessment of changes in cellular morphology at photodynamic treatment in vitro by means of digital holographic microscopy

Temporal dependence of changes in the morphological characteristics of cells of two cultured lines of cancer origin, HeLa and A549, induced by photodynamic treatment with Radachlorin photosensitizer, have been monitored using digital holographic microscopy during first two hours after short-term irradiation. The observed post-treatment early dynamics of the phase shift in the transmitted wavefront indicated several distinct scenarios of cell behavior depending upon the irradiation dose. In particular the phase shift increased at low doses, which can be associated with apoptosis, while at high doses it decreased, which can be associated with necrosis. As shown, the two cell types responded differently to similar irradiation doses. Although the sequence of death scenarios with the increase of the irradiation dose was the same, each scenario was realized at substantially different doses. These findings suggest that the average phase shift of the transmitted wavefront can be used for quantitative non-invasive cell death characterization. The conclusions made were cofirmed by commonly used test assays using confocal fluorescent microscopy.