[1586] Predicting 5-Year Recurrence of Ductal Carcinoma In Situ (DCIS) Following Initial Breast Conserving Surgery and Radiation Therapy

Christopher P Kragel, Shi Wei. University of Alabama at Birmingham, Birmingham, AL

Background: A large number of previous studies have identified various clinicopathologic factors significantly associated with local recurrence (LR) for DCIS. However, the LR rate has been greatly reduced since breast-conserving surgery (BCS) with radiation therapy (RT) became a standard treatment option in women with localized DCIS. Yet, there is still a benefit to identifying patients at lower risk for LR in order to avoid overtreatment. In this study, we sought to establish factors significantly associated with LR and to examine the utility of the recently published Memorial Sloan-Kettering (MSK) nomogram in predicting LR rates in patients with DCIS after BCS and RT.
Design: The Surgical Pathology files of the authors' institution were searched to identify patients with DCIS who underwent BCS and subsequent RT from 2003-2007. Those with coexisting invasive carcinoma or receiving mastectomies following BCS with positive margins were excluded. The odds ratios (OR) were calculated, and Fisher Exact Probability or t test was utilized for statistical analysis, when appropriate. The clinicopathologic factors populated the MSK DCIS nomogram.
Results: A total of 31 cases meeting the inclusion criteria were identified in the study period, of which only 4 patients (13%) had LR within 5 years, 3 with DCIS and 1 with invasive carcinoma. While the presence of necrosis in DCIS showed the highest odds ratio for LR, univariate analysis failed to identify any of the clinicopathologic factors examined significantly associated with LR. Further, there was no clear association between the MSK nomogram-predicted and observed 5-year LR rates. The mean probability of LR was 3.75% in the recurrence group and 2.81% in the non-recurrence group (p=0.16).

Univariate Analysis of Clinicopathologic Factors
VariablesOR95% CIp-value
No. of re-excisions (1 vs. 2&3)1.1[0.1, 11.8]1.0
Family history of cancer0.4[0.03, 3.9]0.4
Endocrine therapy2.1[0.2, 22.5]0.6
ER+0.4[0.04, 3.0]0.3
PR+2.4[0.2, 25.9]0.5
Nuclear grade (3 vs. 1&2)2.9[0.3, 24.3]0.3
Necrosis5.1[0.5, 55.9]0.2
Close margins (2 mm or less)2.6[0.2, 33.2]0.5

Conclusions: As management of localized DCIS has become more standardized to include BCS and RT, it may become more difficult to reliably predict 5-year LR risk. Large-scaled studies may be needed to accurately evaluate the DCIS recurrence risk in patients who received standardized therapy to assist clinical decision making in pursuit of individualized medicine.
Category: Informatics

Monday, March 4, 2013 1:00 PM

Poster Session II # 54, Monday Afternoon


Close Window