[118] Predictors of Local Recurrence (LR) in Patients with Ductal Carcinoma In Situ (DCIS) Treated by Breast Conserving Therapy (BCT): Value of the Memorial Sloan-Kettering (MSK) Nomogram
Laura C Collins, Ninah Achacoso, Zaineb Sharafali, Reina Haque, Larissa Nekhlyudov, Suzanne W Fletcher, Charles P Quesenberry, Laurel A Habel, Stuart J Schnitt. Beth Israel Deaconess Medical Center, Boston; Harvard Medical School, Boston; Kaiser Permanente, Northern CA, Oakland; Kaiser Permanente, Southern CA, Pasadena; Harvard Vanguard Medical Associates, Boston
Background: Various patient, treatment and pathologic factors have been associated with an increased risk of LR following BCT for DCIS. However, the strength and importance of individual factors has varied and the extent to which combining factors may improve prediction of risk is undetermined. Our aim was to assess, in a large population-based cohort of DCIS patients treated with BCT and known outcome, the value of the recently published MSK nomogram which combines clinical and pathologic features to predict the risk of LR in patients with DCIS treated with BCT.
Design: We identified patients diagnosed with a first unilateral DCIS between 1990-2001 treated with BCT at 3 integrated health plans. Slides from the index DCIS of patients with recurrences (cases=190) and controls (N=305) were reviewed. Regression methods were used to estimate relative risks (RR) of LR associated with clinical and pathologic factors. The MSK DCIS nomogram was then applied to the study population to compare the nomogram-predicted and observed risks for LR at 5 yrs.
Results: Among the 495 patients in the case-control analysis, the only pathologic features associated with increased LR in univariate analysis were larger lesion size (RR=3.0 for ≥20 low power fields of DCIS;95% CI 1.6-5.6) and involved (RR=2.9;95%CI 1.6-5.3), or close (<1mm)(RR=2.4;95%CI 1.6-3.8) margins. The risk estimates provide by the MSK nomogam showed approximately 90% correlation with the observed rates of LR and were stronger than those provided by evaluation of individual features.
| MSK nomogram probability quartiles | 5 yr Observed LR (%) | Nomogram Predicted RR | 95% CI |
| Quartile 1 (2-<8) | 4.8 | 1.0 | Ref |
| Quartile 2 (8-<13) | 8.5 | 1.8 | 1.0-3.4 |
| Quartile 3 (13-<20) | 20.9 | 5.2 | 2.6-10.3 |
| Quartile 4 (20-53) | 33.1 | 9.9 | 4.8-20.4 |