ESTIMATIVE OF THE IMPACT OF BETTER TREATMENT AND EARLIER DIAGNOSIS ON 5-YEAR SURVIVAL IN BREAST CANCER: A SEER ANALYSIS COMPARING 1988 AND 2011

Introduction: Breast cancer (BC) is the most frequent among women excluding non-melanoma skin cancer. Its 5-year overall survival (OS) has increased in the last 20 years. One of the reasons is that new therapies have emerged since then. Also, the diagnosis is being made at earlier stages. However, we dont know how much of the increase in OS is due to better treatment and how much is due to earlier diagnosis. The idea of this work is to answer this question. Methods: We performed a retrospective analysis of the Surveillance, Epidemiology and End Results (SEER) stat database and compared overall survival and stage distribution for the years of 1988 and 2011. We first calculated the 5-year BC mortality for the years of 1988 and 2011. We subtracted the 5-year OS of 2011 by the 1988 and got the real increase in 5 years OS. After that, we separated the mortality by stage and the stage distribution in each of these years. Then, we simulated how would be the 5-year OS in 2011 if the distribution of stages were the same as in 1988. It gave us the estimated 5-yeas OS for 2011 with constant distribution of stages of 1988. We then subtracted this estimated 5-year OS for 2011 by the actual one of 1988 and got the estimated increase in OS due only to better treatment. Finally, we divided this estimated increase in OS by the real increase in OS and found the percentage of increase due to the development of the treatment. The remaining part must be due to earlier diagnosis. Results: In 1988, 76.23% was the real 5-year OS of women with BC. In 2011, the real 5 years OS increased to 84.37%. This represents an absolute increase of 8.14%. We then applied the 5-year OS of 2011 by stage to the distribution of stages of 1988, which indicated that the estimated 5-year OS should be 83.22%, representing an estimated absolute increase of 6.99%. This means that of the 8.14% increase in 5 years OS, 6.99% are due to better treatment, representing a weight of 85.84%. The remaining part must be due to earlier diagnosis, which results in 1.15% of absolute increase and a weight of 14.16%. Discussion: Our analysis must be considered a hypothesis generator, regarding to the importance of both increase of treatment efficacy and early diagnosis. Its limitations are the different methods of diagnosis of the periods, the increase in non-breast cancer life expectation and others arising from the SEER database, such as the limited populations origin (USA) and the participants classified as unknown.