Background: The advancements in automated High Content Analysis (HCA) for Circulating Tumor Cells (CTCs) detection necessitate the integration of high-throughput screening (HTS) and Artificial Intelligence (AI) to improve cancer diagnosis efficiency and accuracy. Traditional approaches to cancer detection have faced limitations in speed, cost, and analysis depth, driving the need for innovative solutions in assay development and execution. Methods: We obtained peripheral blood from 150 asymptomatic individuals and 63 newly diagnosed cases of solid organ cancers including Lung, Breast, Head and Neck, Colorectal, Pancreas, Esophagus, Gynecologic cancers. We used differential apoptosis based negative enrichment technique and HCS fluorescent imaging for identification of CTCs and CTC like immune cells from the relevant blood samples. We devised AutoML algorithms which were trained on a cohort of 4,135 cropped region of interest (ROI) images of true CTCs and CTC-mimicking immune cells. These images were annotated with the coordinates of the top-left and bottom-right corners of rectangles to specify ROIs. This method employs a combination of automated sample processing, integration with the Laboratory Information Management System (LIMS) integration, digital filtration, and decision matrix algorithms for sample classification and smooth data flow. The image quality criteria were established prior to dataset analysis. The algorithms underwent training, testing, and validation based on expert recommendations, with dataset splitting of 80% for training, 10% for testing, and 10% for validation phases. Results: The automated deep learning model achieved remarkable performance metrics, presenting an accuracy of 96.66% (95% CI, 96.07%-97.19%), a specificity of 95.34% (95% CI, 94.37%-96.18%), and a sensitivity of 98.14% (95% CI, 97.44%-98.69%) in detecting true CTCs. Additional analysis included the area under the curve and confusion matrix evaluations, indicating the model's robustness in distinguishing between true positive and true negative specimens. This efficiency and precision in data handling, processing, and analysis demonstrate significant advancements in HCA screening, offering enhanced assay development, execution flexibility, and the facilitation of detailed cell-by-cell analysis for cancer detection from a large set of Peripheral Blood Mononuclear Cells (PBMCs) from samples obtained for screening or diagnosis. Conclusions: This study successfully evaluated a deep learning model for the detection of circulating tumor cells (CTCs) in peripheral blood mononuclear cells (PBMCs) using image data from HTS platform. The model achieved superior performance, demonstrating its potential for cancer detection from blood samples.