Drug screening using in-vitro primary tumor cells offers a direct approach to assessing drug sensitivity, playing a crucial role in personalized cancer diagnosis and precision medicine. However, existing tumor drug screening methods encounter a common challenge: the presence of normal cells alongside cancer cells in patient tumor samples complicates the accurate evaluation of drug effects on cancer cells alone. In this study, we introduce a real-time impedance sensing method based on an integrated circuit (IC) chip to distinguish between cancer and normal cells at the single-cell level, enabling individual analysis of their sensitivity to drugs. This innovative approach integrates impedance calculation circuits directly beneath an array of 352 single-cell-sized sensing electrodes, effectively reducing the high background noise that typically hampers the detection of ultra-weak signals. We validated this method across various cancer types, demonstrating its capability for impedance-based classification on IC chips. Our findings highlight the potential of this real-time, non-invasive, and label-free single-cell distinguishing method for precision medicine, laying the foundation for the development of ultra-miniaturized biological signal sensors and the realization of lab-on-IC technology.