We present our AKARI study of massive star forming regions where a large-scale cloud-cloud collision possibly contributes to massive star formation. Our targets are Spitzer bubbles, which consist of two types of bubbles, closed and broken ones; the latter is a candidate of the objects created by cloud-cloud collisions. We performed mid- and far-infrared surface photometry toward Spitzer bubbles to obtain the relationship between the total infrared luminosity, LIR, and the bubble radius, R. As a result, we nd that LIR is roughly proportional to R where  = 2:10:4. Broken bubbles tend to have larger radii than closed bubbles for the same LIR.