RESUMO
OBJECTIVE: The objective of the present study was to employ high throughput image analysis to detect necrosis and apoptosis. Specific markers were replaced by morphological parameters of cells and nuclei. METHOD: Fresh blood was taken from a healthy female and given a treatment to induce cell necrosis and apoptosis. Afterward, the samples were stained with AnnexinV-FITC, DRAQ5 and DAPI. Slides were made and analyzed using the cytometer iCys. Pictures were scanned. The analyzed sample consisted of 73 sets of images of DAPI, DRAQ5 and AnnexinV-FITC, respectively. For image analysis and subsequent statistical processing, the CellProfiler and CellProfilerAnalyst were used. Each sample was analyzed twice. The first analysis was conducted using the markers (DAPI, DRAQ5 and Annexin) for an unequivocal identification and subsequent count of necrotic, apoptotic and live cells (gold standard). Thereafter, a second analysis was performed for the nuclear morphology and texture (morphometric analysis). After the machine learning process was completed, the software calculated the quantity of cells in each of the three groups. A comparison between the result of the gold standard and the morphometric analysis was performed using linear regression and a Bland-Altman test. RESULTS: The linear regression between the two compared analyses was r(2)=0.57 for apoptosis, r(2)=0.84 for necrosis and r(2)=0.79 for living cells. CONCLUSION: It may be concluded that it is possible to replace specific markers against morphology without losing the reproducible high-throughput character of a cytometric analysis.