The study evaluates the performance of computer-aided detection (CAD) and volumetry software using a Kyoto Kaguku Lungman phantom with 3D-printed nodules as the ground truth (GT). Six nodule diameters (4-9 mm) and three morphologies (smooth, lobulated, spiculated) were scanned under varying CT radiation dose levels. The nodules were reconstructed using iterative and deep learning algorithms with soft and hard kernels. The AI-based algorithm, AVIEW LCS +, was assessed for detection, volumetric, and density accuracy against GT obtained via micro-CT. The results showed high detection sensitivity (≥ 83%) and precision (≥ 91%), unaffected by dose or reconstruction algorithm. Detection was significantly associated with nodule diameter (p < 0.0001), while volumetric accuracy was robust for nodules >6 mm. Nodule diameter and morphology significantly influenced volumetric accuracy (p < 0.001). Density characterization was unaffected by acquisition parameters. The findings confirm AVIEW's reliability and advocate for similar phantom setups in CAD quality assurance.