Impact of Microinvasion as a Predictor of Local Recurrence in Ductal Carcinoma In Situ Treated With Breast Conserving Therapy

Abstract

The significance of microinvasion (MI) (invasive cancer ≤1mm) on the risk of local recurrence (LR) in women with ductal carcinoma in situ (DCIS) is unclear. We report a population-based analysis of women with DCIS with and without MI treated with breast-conserving surgery (BCS) alone or with radiotherapy (RT), and evaluated the impact of MI on the risks of any LR (DCIS or invasive) and invasive LR Cohort includes all women diagnosed with pure DCIS or DCIS with MI in Ontario from 1994-2003 that were treated with BCS +/- RT and had undergone expert pathology review. Treatments and outcomes were assessed through administrative databases and validated by chart review. Local recurrence-free survival (LRFS) was calculated using the Kaplan-Meier approach with differences compared using the log-rank test. The impact of MI on the development of LR was assessed using Cox proportional hazards model adjusted for propensity score to account for systematic differences in women treated with and without RT. Population cohort includes 2,988 women with DCIS treated by BCS (N=2,721 with pure DCIS, N= 267 DCIS with MI). Median follow-up was 13 years. Median age at diagnosis was 58 years. RT was given in 58% of cases with MI and 51% of cases with pure DCIS (p=.03). RT boost was given in 16.6% of women with MI and 14.4% of those with pure DCIS (p=.21). Women with MI were more likely to have high nuclear grade (p<.001), and larger tumour size (p<.001) versus pure DCIS. LR developed in 59 (22.1%) cases with MI and 530 (19.6%) cases of pure DCIS. Among women treated with BCS alone, the 15 year LRFS rate for DCIS with MI versus pure DCIS was 66% v 75% (p=.03) for any LR and 83% v 86% (p=.30) for invasive LR. Among women treated with BCS and RT, the 15 year LRFS rate for DCIS with MI versus pure DCIS was 82% v 82% (p=.74) for any LR and 88% v 88% (p=.84) for invasive LR. However, on multivariable analyses adjusted for age, propensity score and year of diagnosis, the presence of MI was not associated with an increased risk of any LR (HR=1.05, 95% CI: .80-1.38) or invasive LR (HR=1.08, 95% CI: .76-1.54). There was no significant interaction between the presence of MI and provision of RT for any LR (p=.15) or invasive LR (p=.67). Factors associated with any LR include RT (HR=.60, 95% CI: .51–.71), high nuclear grade (HR=1.21, 95% CI 1–1.45) and multifocality (HR=1.35, 95% CI: 1.12-1.64). Factors associated with invasive LR include RT (HR=.76, 95% CI: .61–.94) and multifocality (HR=1.40, 95% CI: 1.09-1.79). The presence of MI with DCIS is not associated with an increased risk of any LR or invasive LR after BCS with or without RT. The impact of number of foci of microinvasion on the risk of LR is in progress.