Vascular aneurysms, characterized by abnormal dilations of blood vessel walls, represent localized structural anomalies that can lead to catastrophic rupture and high mortality rates. Early detection is crucial for timely diagnosis, follow-up, and prognosis. However, existing macro-scale diagnostic methods, including ultrasound, MRI, and CT, often lack sufficient sensitivity to detect small, early-stage anomalies deep within complex vascular topologies. This paper presents the novel distributed sensing framework of nanoparticle tomography (NPT) for identifying vascular abnormalities with Internet of Bio-Nano Things (IoBNT). We first develop a generic system model for NPT, treating the aneurysm as a perturbation in the transport channel, and propose the differencing-combining-rendering method for qualitative anomaly reconstruction. Subsequently, NPT is applied to detect internal carotid artery aneurysms by utilizing nanoparticle time-of-arrival data as the primary sensing metric. The accuracy of the proposed method is evaluated using both the COMSOL Multiphysics simulation platform and a physical phantom that synthesizes realistic vascular transport conditions. Simulation and phantom results show that NPT can recover aneurysm-related transport perturbations under controlled conditions. In the COMSOL evaluation, model-error compensation improves aneurysm-detection AUC from 0.733 to 0.916 under 5 ms TOA noise. In the phantom experiments, NPT achieves an overall success rate of 94.7% under the onegrid tolerance, with localization failures primarily occurring when physical perturbations shift the retained candidate regions outside the ground-truth tolerance window.