Authors: Olivotto IA, Gomi A, Bancej C, Brisson J, Tonita J, Kan L, Mah Z, Harrison M, Shumak R
Title: Influence of delay to diagnosis on prognostic indicators of screen-detected breast carcinoma.
Journal: Cancer 94(8):2143-50
Date: 2002 Apr 15
Abstract: BACKGROUND: Although delay to diagnosis after a breast screening abnormality causes anxiety, its effect on prognosis is unknown. METHODS: Using pooled data from five Canadian organized breast cancer screening programs, the authors used unconditional logistic regression to evaluate the effect of delay to diagnosis on prognostic indicators among 4465 women with invasive breast carcinoma diagnosed in the ipsilateral breast within 3 years of an abnormal screen performed during 1990-1996. RESULTS: Women with high-suspicion screens (n = 1569) compared with those without (n = 2896) were more promptly investigated (median days from screen to diagnosis, 31 vs. 47; P < or = 0.0001), had larger tumors (79.4% vs. 55.9% > 10 mm; P < or = 0.0001), and were more likely to be lymph node positive (33.9% vs. 17.3%; P < or = 0.0001). For delays beyond > 12 to < or = 20 weeks, a linear trend of increased tumor size and lymph node positivity began to emerge. Controlling for suspicion, the authors found that odds ratios for tumor size greater than 10 mm were 0.9 (95% CI, 0.66-1.17), 1.2 (95% confidence interval [CI], 0.88-1.56), 1.5 (95% CI, 1.05-2.16), and 2.1 (95% CI, 1.15-3.86) for delays of > 12 to < or = 20, > 20 to < or = 52, > 52 to < or = 104, and > 104 < or = 156 weeks, respectively (p(trend) < or = 0.0001), compared with delays of > 4 to < or = 12 weeks. Similarly, odds ratios for lymph node metastasis were 1.0 (95% CI, 0.67-1.42), 1.2 (95% CI, 0.84-1.69), 2.2 (95% CI, 1.48-3.15), and 3.2 (95% CI, 1.84-5.55) for the same time intervals (p(trend) = 0.0033). CONCLUSIONS: The authors' findings suggest that delays to diagnosis of asymptomatic breast carcinoma of 6 to 12 months are associated with progression of breast carcinoma as measured by increasing risk of lymph node metastases and larger tumor size. A policy of early recall rather than biopsy for low suspicion mammographic abnormalities may introduce delays of this magnitude. The tendency to more expediently investigate women with high-suspicion, worse prognosis screens (suspicion bias) obscures whether delays shorter than 20 weeks also worsen prognostic indicators. Suspicion bias should be considered when interpreting the effect of delay on prognosis.