SUMMARY When selecting cancer therapy, physicians generally begin with first-line treatment options and monitor patient progress on a watch-and-wait basis, following a set of guidelines based on clinical trials from a large patient population. But this traditional method has been questioned on whether it provides individual patients with the optimal treatment. To better find a matching treatment individually, a concept called precision cancer medicine or personalized cancer medicine has been studied. Among other approaches, functional precision medicine directly tests chemotherapy options on tumor cells biopsied from a patient to find the best matching treatment for the specific patient. This promising approach, however, has not been widely adopted by clinicians because the tumor microenvironment in a lab differed from the one within the patient’s body, leading to inconsistent drug responses between the sample and patient, and the quantity of biopsied cells is generally insufficient for a reliable number of options to be tested. The first problem is being addressed by recent advances in three- dimensional (3D) cell culture techniques, which better mimic the body’s microenvironment in a lab. But the second problem, the limited number of testable options, is mainly due to limitations in the current assay techniques that assess chemosensitivity in 3D culture. With most current assays, a sample can only be tested once, and multiple drugs with different mechanisms of action cannot be simultaneously tested by a single assay. Combined, these limitations exponentially reduce the number of testable options when involving multiple assessment time points to design a sequential therapy or when increasing the number of drugs to test a combination therapy. Here, we will develop a new technique for the assessment of chemosensitivity in 3D culture, by maximizing the potential of a label-free 3D microscopy technology, called optical coherence tomography (OCT). The majority of prior OCT research measured only one or two types of signals and showed the signals corresponding to only a single type of cell viability disruption process in each study. But this approach has led to a concern about specificity (i.e., other types of processes than the one tested in the study can generate similar OCT signals). This low specificity, along with unclear mechanisms of viability assessment, have prevented OCT methods from being adopted for the promising concept of functional precision medicine. Therefore, we will develop at least 18 different types of OCT signals and establish their sensitivity and specificity to four major types of viability disruption processes. The feasibility of this approach has been strongly supported by a pilot study where we imaged and analyzed more than 6,000 3D-cultured cell spheroids. This R01 project will image and analyze up to 100,000+ spheroids for an unprecedentedly systemic investigation of the comprehensive range of OCT signal types.