RESUMEN
MATERIALS: (1) Screening test: The study utilized data from a regional cancer-screening center from April 1, 1997, to December 31, 1998. A total of 51,700 women underwent screening at this center during this period. All incident cases of breast cancer in these women were derived from the files of the Yamagata Prefectural Cancer registry and sensitivity, specificity and positive predictive value were then estimated. False negative cases were defined as occurring in women who tested negative but were registered as having breast cancer in the cancer registry within 12 months after screening. (2) Effectiveness of screening: The target population comprised female cases with breast cancer registered in the Yamagata Prefectural Cancer Registry from January 1, 1989 to December 31, 1998. During this period, a total of 2,746 cases were registered. Survival probabilities for breast cancer cases were estimated according to method of detection using Kaplan-Meier method. Overall survival probability of the screen-detected group was compared with the not screen-detected group using the log-rank test. Point estimates of 5- and 9-year survival rates between groups were compared using z statistics. RESULTS: (1) Screening detected 27 breast cancer cases, while 31 false negative cases were identified. Sensitivity, specificity and positive predictive value were 46.6% (95% confidence interval (95% CI), 33.3-60.1%), 97.3% (95% CI, 97.2-97.5%) and 1.9% (95% CI, 1.3-2.8%), respectively. (2) Overall survival rate for the screen-detected group was significantly higher than that for the not screen-detected group (P<0.001), with 11.6% (95% CI, 8.0-15.2%) and 13.1% (95% CI: 5.3-20.8%) differences in 5-year and 9-year survival probabilities, respectively. CONCLUSION: Cancer screening by clinical breast examination in Yamagata has improved survival from breast cancer for females. However, the screening sensitivity is insufficient for effective mass screening purposes.