In my dissertation, I present the development and application of nanofluidic diagnostics using specific subpopulations of extracellular vesicles (EVs) towards multiple clinical questions in both neurological disease and cancer. EVs contain multiple protein and nucleic acid cargoes which are reflective of their cells of origin but circulate in high concentrations in peripheral bodily fluids (blood, saliva, urine, etc.). However, the isolation of subpopulations of EVs specific to a given disease process is made challenging by the high circulating background concentrations of non-disease EVs (>10^11 EVs/mL in plasma) and their nanoscale size (< 800 nm). Conventional EV isolation methods such as ultracentrifugation and size-exclusion chromatography lack the capability and throughput to perform surface-marker-specific labeling and isolation of EVs from complex samples such as plasma. Previous microfluidic approaches have been considered for surface-marker-specific EV isolation, but these are restricted by low throughput, lack of resilience to clogging, or reliance on single biomarker targets for isolation. The technologies that I have developed in my PhD use millions of parallelized magnetic nanopores to perform the immunomagnetic sorting and isolation of EV subpopulations in high throughput for applications in cancer and neurological disease. To this end, I have developed an optimized version of our previous track-etch nanopore (TENPO) platform with >10x improvements in specificity while maintaining sensitivity. I have applied this technology in mapping miRNAs from tumor-derived EVs in lung cancer immunotherapy responders compared to non-responders, and have found multiple markers with statistically-significant differential expression. Using this technology, I then performed the first mapping of miRNAs from neuron-derived and astrocyte-derived EV subpopulations in the plasma of patients with of multiple types of dementia (Alzheimer’s disease, Lewy body dementia, frontotemporal dementia) versus clinically-normal controls. Lastly, I have developed a next-generation immunomagnetic EV sorting platform which uses 3-dimensional inverse-opal nanomaterials to deliver ~75x increases in EV capacity and improved robustness to clogging for the isolation of EV subpopulations from complex background. By leveraging parallelization in advanced materials towards the specific isolation of disease-related EVs, this work offers a set of approaches and candidate biomarkers for the development of robust and non-invasive diagnostic assays in multiple diseases.