Re-Identification and growth detection of pulmonary nodules without image registration using 3D siamese neural networks

•Addressing automatic re-identification of pulmonary nodules with siamese networks.•Pretrained siamese networks allow successfully rank similarity between nodules.•Best settings with binary cross entropy, fully connected, and features from early layers.•Prior image registration is avoided providing computationally faster results.•Accurate detection of nodule growth embedding a nodule re-identification network.