[LIRADS] Unequivocal decrease in the diameter of an observation, measured on examinations performed on different dates and in the absence of treatment, which is not attributable to artifact, differences in technique between the two examinations, or measurement error