INVESTIGATING INTRATUMOR HETEROGENEITY OF CANCER METASTASIS USING AN INTEGRATIVE MICROFLUIDIC DETERMINISTIC LATERAL DISPLACEMENT (I-DLD) ARRAY

Cancer is a highly heterogeneous disease, and this complexity has rendered one-size-fits-all therapies largely ineffective. For instance, in oral squamous cell carcinoma (OSCC), the 5-year survival rate remains below 60% despite aggressive multimodal treatments, including surgery, radiotherapy, and chemotherapy. This limited efficacy highlights a critical gap: although numerous genetic mutations have been identified, a lack of mechanistic insight into intratumor heterogeneity (ITH)—particularly the metastatic functions of individual cells—continues to hinder the discovery of predictive biomarkers essential for precision oncology ^[1].
Flow cytometry rapidly sorts cells (~10³/sec) but lacks single-cell isolation, allowing cell interactions to mask true behaviors. Intracellular cytokine staining (ICS) introduces artifacts, distorting cellular responses. Microwell systems isolate cells in nanoliter wells for drug testing, with secreted proteins detected by antibody arrays. However, their low capacity (~100 wells/run) cannot handle the 10⁵–10⁶ cells in tumor samples, limiting scalability for rare subpopulation analysis.
Droplet microfluidics offers scalable single-cell analysis by encapsulating cells in tiny (~7 pL) droplets, preventing cross-contamination ^[2]. Proteins are detected via fluorescent sensors at ~102 cells/second ^[3]. However, poor nutrient supply and waste removal restrict cell viability to under 24 hours, making long-term studies (e.g., metastasis) impossible.
New droplet washing techniques, like pico-washers and dielectric polarization (DEP) washer ^[4], use electric fields to exchange fluids, achieving 80% efficiency. Yet, these methods risk droplet instability without surfactants, and electric fields may alter cell behavior. Currently, no technology provides stable, high-throughput droplet medium exchange for prolonged single-cell analysis.
To address this unmet need, we developed a novel automatic droplet medium refreshing device integrating a deterministic lateral displacement (i-DLD) array ^[5,6] (Figure 1), designed to establish stable multiple-phase lamellar flows for continuous droplet washing (Figure 2). An AI-based droplet sorter (Figure 3) was implemented to remove empty droplets post-washing, ensuring efficient processing. Using fluorene dye–loaded droplets, we validated that the i-DLD array achieves near-complete (100%) droplet fluid exchange efficiency (Figure 4, Figure 5). This system enables long-term single-cell incubation by refreshing the droplet medium every 12 hours for up to 5 days. We applied this platform to study OSCC cells, continuously profiling their metastatic behavior through protease secretion. Protease sensors will be anchored directly to the cell surfaces, allowing real-time phenotypic assessment. The system operated at a throughput of ~10³ cells per second, making it possible to identify and isolate rare, sustainable metastatic subpopulations (Figure 6). Unlike prior systems, this platform supports both high-throughput screening and extended functional assays, enabling new insights into the mechanisms of cancer metastasis. The isolated metastatic cells were analyzed via sequencing to identify metastatic biomarkers to guide patient-specific therapy (Figure 7). In summary, this high-throughput i-DLD device represents a breakthrough in long-term single- cell metastatic profiling. It offers a scalable and precise solution to dissect ITH and drive the development of personalized cancer therapies in precision oncology.