In oncology trials, progression-free survival (PFS) is often used in phase 2 trials as a measure of the efficacy of a new compound. PFS is often used because it requires less time to measure. Similarly, objective response rate (ORR) is often used for similar reasons. However, the objective of the development usually requires an assess of overall survival in the phase 3 trial. But the use of these surrogate endpoints is helpful only if they are correlated to OS. To investigate the correlation, results from a sample of 153 oncology clinical trials were downloaded from clinicaltrials.gov. We found that that PFS is correlated with OS, as is ORR. The correlation also depends on the class of compounds. Models for predicting OS are also presented.