The paper introduces SAM-Road, an adaptation of the Segment Anything Model (SAM) for extracting large-scale, vectorized road network graphs from satellite imagery. SAM-Road is designed to predict both the geometry and topology of road networks, leveraging the strengths of SAM for dense segmentation and a lightweight transformer-based graph neural network for topology prediction. The model directly predicts graph vertices and edges, avoiding complex post-processing heuristics, and can build complete road network graphs spanning multiple square kilometers in seconds. Compared to state-of-the-art methods like RNGDet++, SAM-Road achieves comparable accuracy while being 40 times faster on the City-scale dataset. The paper also discusses related works, including SAM and its applications, road network graph prediction, and graph representation and learning. Experiments on the City-scale and SpaceNet datasets demonstrate the effectiveness of SAM-Road, showing high accuracy and efficiency. The code for SAM-Road is available at <https://github.com/hcr/sam_road>.The paper introduces SAM-Road, an adaptation of the Segment Anything Model (SAM) for extracting large-scale, vectorized road network graphs from satellite imagery. SAM-Road is designed to predict both the geometry and topology of road networks, leveraging the strengths of SAM for dense segmentation and a lightweight transformer-based graph neural network for topology prediction. The model directly predicts graph vertices and edges, avoiding complex post-processing heuristics, and can build complete road network graphs spanning multiple square kilometers in seconds. Compared to state-of-the-art methods like RNGDet++, SAM-Road achieves comparable accuracy while being 40 times faster on the City-scale dataset. The paper also discusses related works, including SAM and its applications, road network graph prediction, and graph representation and learning. Experiments on the City-scale and SpaceNet datasets demonstrate the effectiveness of SAM-Road, showing high accuracy and efficiency. The code for SAM-Road is available at <https://github.com/hcr/sam_road>.