Radiotherapy treatment of moving lesions is a challenging task in which different strategies can be used. The clinical implementation of the variety of complex technical solutions pursued to treat cancer, while sparing healthy tissues, requires accurate quality assurance tests and adequate phantoms. The aim of this work is to report on the anthropomorphic dynamic breathing model (ADAM) and on its applications to test image quality and dose delivery for four-dimensional (4D) techniques used to compensate for breathing-induced tumor motion. ADAM reproduces a male torso containing realistic ribs and spinal cord. Lungs move along circular or elliptical paths lying on sagittal planes, while the anterior chest surface moves independently up and down. Lungs host radiochromic films, a diamond detector and objects of known shape and dimensions to test 4D scanners. Markers are embedded around a target and a specific tool, hosting orthogonal radiochromic films, has been provided to perform end-to-end tests of tracking systems. To demonstrate ADAM's performance in testing techniques and methods used to image and treat moving lesions we report the results of three experiments performed to evaluate the accuracy of 4D computed tomography reconstructed volumes, the repeatability of measured dose in volumetric modulated arc treatments and tracking performances of a robotic system. The results obtained show that the percentage differences of reconstructed volumes, with respect to the known volume, depend on the breathing pattern and the pitch size (maximum 5% for breathing rate > 10 breaths per minute). Repeatability of measured dose maps obtained comparing radiochromic films was demonstrated (GI3%/2 mm > 99%). Differences between planned and delivered point dose measurements were <1%. Tracking errors were all below the tolerance level (range 0.6-0.9 mm). The results of this study demonstrate that ADAM is suitable to test techniques and methods used to image and treat moving lesions.