结肠癌最新进展（重复）课件.pptVIP

Slide 8. 3-Year DFS vs 5-Year OS We then looked at the simple correlation between disease-free survival at 3 years and overall survival at 5 years for each arm of each trial. As you can see from this slide, the correlation is tight. The correlation coefficient is 0.90, indicating a very high correlation. The regression equation is shown on the bottom of the slide, and it shows that 5-year overall survival equals 0.0 + 1 x 3-year disease-free survival, indicating a very tight relationship between these 2 endpoints, which is consistent across the range of disease-free and overall survivals for all of the studies. Slide 9. Conclusion (1) On the basis of these data we conclude, first, that on a patient-by-patient basis 3 years seems a reasonable time point to examine. The rate of recurrence is high for the first 3 years and then falls off quickly. In addition, the median survival following a patient recurrence is in the order of 1 year. These 2 factors indicate that patients who are alive and disease-free at 3 years are very likely to be alive at 5 years. In addition, on an arm-by-arm basis, 3-year disease-free survival is an excellent predictor of 5-year overall survival. Slide 10. Assessment of Treatment Effect We then turned to this question: does a comparison of arms within trial using disease-free survival reach the same conclusion as if we used overall survival? Slide 11. Hazard Ratios: DFS vs OS Shown on this slide are hazard ratios comparing the control arm with the experimental arm for each trial included. On the x-axis is the hazard ratio comparing treatment with control for disease-free survival. On the y-axis is the same comparison for overall survival. Again, we can see a very tight correlation between these 2 endpoints with a correlation coefficient of 0.91, which is indicating a very high correlation. The regression equation indicates that the hazard ratio for overall survival is approximately 0.1 + 90% times the disease-free survival hazard ratio. The