Novel, multi-object X-ray computed tomography (CT) methodologies can individually analyse vacuum-packed meat samples scanned in batches of three or more, saving money and time compared to scanning live animals. If intramuscular fat (IMF), as a proxy for meat quality, can be predicted with similar accuracies as in live lambs, this method could be used to grade on quality, or to inform breeding programmes. Lamb loin cuts from commercial carcasses (n=303), varying in fat and conformation grade, were vacuum-packed and CT scanned, then tested for meat quality traits and by a trained taste panel. Tissue density values measured by CT, alongside carcass and loin weights, predicted IMF with moderate accuracy (R 0.36), but did not accurately predict shear force or sensory traits. Juiciness and flavour increased linearly with IMF, whilst texture and overall liking increased to an optimum between 4 and 5% IMF. Samples predicted by CT as having >3% IMF scored significantly higher for sensory traits, than those predicted as <3% IMF.