The 5-year survival rate is the most useful parameter for evaluating the effect of management on most malignant tumors. Recurrence after a curative resection for gastric cancer occurs mostly within 3 years of the operation, which caused us to evaluate whether a 3-year disease-free survival (3DFS) can be substituted for a 5-year overall survival (5OS).

We reviewed the medical records of 656 consecutive patients who had undergone a curative resection for gastric cancer. To assess whether 3DFS represents 5OS, we used a simple linear regression with survival probability calculated by using the survival function.

Recurrence was found in 175 cases during the follow-up periods. The accumulative frequencies of recurrence at postoperative 1 year, 3 years, and 5 years were 46% (81 cases), 89% (156 cases), and 97% (170 cases), respectively. The correlation coefficient (r) and the coefficient of determination (r²) between 3DFS and 5OS were 0.87 and 0.76, respectively, and the regression equation was 5OS=0.18+(0.80 × 3DFS). The r and R² according to the type of recurrence were 0.89 and 0.80 in peritoneal seeding, 0.88 and 0.78 in hematogeneous metastasis, and 0.86 and 0.73 in local recurrence, respectively. The r (0.77) and r² (0.60) were relatively lower in low stages (stage I and II) compared to r (0.88) and r² (0.77) in high stages (stage III and IV).

The 3DFS is an excellent predictor of 5OS. Therefore, if we use the former as the treatment evaluating method, 2-year time reduction in assessing and reporting treatment results is expected.