A random-effects meta-analysis was conducted based on 9 original studies with a total 1558 patients (seven photon, n = 1488 patients; two PT, n = 70 patients) and data from a proton registry. The 1-, 2- and 3-year OS rates for the photon group, PT group and registry group were 55, 65 and 72%; 26, 38 and 38%; 12, 35 and 18%. The only statistically significant difference is the 1-year OS between the photon group and the registry group. The authors concluded that the efficacy of PT was not superior to that of photon RT during this meta-analysis.