Keel bone damage, include deviations and fractures, is common in both white and brown laying hens, regardless of the housing system. Radiography for assessing birds' keel bones is was proposed by previous studies. However, radiographs show only 2 out of 3 dimensions of the dissected keel bones. The current study aimed to (1) investigate the association of radiographic optical density (keel and tibiotarsal) and geometry (keel) with dissected keel bone pathology. Previous studies suggested that keel bone fractures may result from internal pressure exerted by pelvic cavity contents. The current study also aimed to (2) investigate the potential associations between pelvic dimensions and measures of keel bone damage. A sample of 200 laying hens on a commercial farm were radiographed at 16, 29, 42, 55, and 68 weeks, and culled at the end of the laying period (week 74). The birds were examined post-mortem for pelvic dimensions and underwent whole-body radiography, followed by keel and tibiotarsal bone dissection and radiography, and keel bone scoring. The radiographs were used to estimate radiographic optical density (keel and tibiotarsal bone) and keel bone geometry (ratio of keel bone length to mid-depth). The method for on-farm radiography of laying hens, including live bird restraint, positioning for live keel imaging, and post-imaging measurements, was developed, tested, and found to be reproducible. The radiographs (1,116 images of 168 birds) and the respective measurements and post-mortem scores of keel bones are also provided for further development of radiographic metrics relevant to keel bone damage. Some longitudinal radiographic measurements of keel geometry (ratio of length to mid-depth) and optical density (keel and tibiotarsal) showed associations with the damage (deviations/fractures) observed on the dissected keel bones. The associations of keel damage were clearer with the radiographic keel geometry than with keel and tibiotarsal optical density, also clearer for the keel deviations than for keel fractures. The higher radiography ratio of keel length to mid-depth at weeks 42, 55 and 68 of age, the larger deviations size observed on the dissected keels at age of 74 weeks. The higher the tibiotarsal radiographic optical density at week 55 of age, the lower deviations size and fractures count observed on the dissected keels at age of 74 weeks. Pelvic dimensions showed a positive correlation with body weight, but a larger pelvic cavity was associated with increased keel bone damage. These findings lay the foundations for future use of on-farm radiography in identifying appropriate phenotypes for genetic selection for keel bone health.