Biology subjects

Jovic, A.

Publications and source records attributed to Jovic, A..

3 recordsLinked to original sources

bioRxiv2025

Deep learning-driven morphology analysis enables label-free classification of therapeutic agent- naive versus resistant cancer cells

Therapeutic drug treatments of solid tumors are often undermined by various resistance mechanisms. Identification of drug-resistance phenotypes at the single cell level is challenging because conventional molecular methods are cell-destructive, labor-intensive, and cost-prohibitive. To overcome these challenges, we developed an orthogonal approach to drug-resistance phenotyping, through the use of deep-learning-driven morphology analysis of single, high resolution cell images. Specifically, we trained deep learning-based drug resistance classifiers using cell images from 5 different cell lines that were rendered resistant to 5 different therapeutic agents, using a foundation model framework. With high accuracy, the classifier correctly predicted naive or resistance phenotypes across all cancer types and across all the therapeutic agent types (chemotherapeutic, targeted) tested. These results showed that morphology can capture complex phenotype information in the context of drug treatment. To demonstrate the potential clinical utility of the drug resistance classifier, it was applied to a dissociated tumor biopsy and the resulting phenotype predictions were in close concordance with scRNASeq analysis of the biopsy. Our study highlights the potential of deep-learning-driven morphology analysis to provide complex phenotype information, and ultimately shape oncology drug treatment strategies at the patient-level in a clinical context.

cancer biology↗

bioRxiv2024

Label-Free Melanoma Phenotype Classification Using Artificial Intelligence-Based Morphological Profiling

Melanomas are the deadliest skin cancers, in part due to cellular plasticity and heterogeneity. Intratumoral heterogeneity drives varied mutable phenotypes, specifically "melanocytic" and "mesenchymal" cell states, which result in differential functional properties and drug responses. Definitive and rigorous classification of these phenotypic states has been challenging with conventional biomarker-based methods, and high-parameter molecular methods are cell-destructive, labor-intensive, and time-consuming. To overcome these technical and practical limitations, we utilized label-free artificial intelligence-based morphological profiling to classify live melanoma cells into melanocytic and mesenchymal phenotypes based on high resolution imaging of single cells. To predict the phenotypes of single melanoma cells based on morphology alone, we developed the AI-based Melanoma Phenotype Classifier trained with 19 patient-derived cell lines with known melanocytic or mesenchymal transcriptional profiles. To link phenotypic state with high-dimensional morphological profiles, cells were subjected to genetic and chemical perturbations known to shift phenotypic states. The AI classifier successfully predicted phenotypic shifts which were confirmed by single-cell RNA-Seq (scRNA-Seq). These results demonstrate that correlations between melanoma cell phenotypes and morphological changes are detectable by AI. Additionally, the Melanoma Phenotype Classifier was applied to dissociated tumor biopsy samples and characterization of phenotypic heterogeneity was supported by scRNA-Seq transcriptional profiles. This work establishes a link between cell morphology and melanoma phenotypes, laying the groundwork for the use of a label-free morphology-based method for phenotyping live melanoma cells combined with additional analyses.

cancer biology↗

bioRxiv2022

Realtime morphological characterization and sorting of unlabeled viable cells using deep learning

Phenotyping of single cells has dramatically lagged advances in molecular characterization and remains a manual, subjective, and destructive process. We introduce COSMOS, a platform for phenotyping and enrichment of cells based on deep learning interpretation of high-content morphology data in realtime. By training models on an atlas of >1.5 billion images, we demonstrate enrichment of unlabeled cells up to 33,000 fold. We apply COSMOS to multicellular tissue biopsy samples demonstrating that it can identify malignant cells with similar accuracy to molecular approaches while enriching viable cells for functional evaluation. We show high-dimensional embedding vectors of morphology generated by COSMOS without any need for complex sample pre-processing, gating, or bioinformatics capabilities, which enables discovery of cellular phenotypes, and integration of morphology into multi-dimensional analyses. One sentence summaryA novel platform capable of high-throughput imaging and gently sorting cells using deep morphological assessment.

cell biology↗