A Simple Risk Scoring System for Predicting Recurrence in Women with Locally Advanced Breast Cancer (LABC) Treated with Neoadjuvant Chemotherapy (NACT)

Introduction: Breast cancer commonly presents in locally advanced stage (LABC) in developing countries, for which NACT followed by surgery and radiotherapy is the standard of care. There is a need for a simple tool to risk categorise patients in the clinic, so that treatment intensification can be offered to women with high risk of recurrence.

Materials and Methods: Data of prospectively maintained database of LABC (between January 2007 - December 2012), who received NACT followed by surgery, radiotherapy and endocrine therapy was retrospectively analysed for clinico-pathological factors associated with disease recurrences. A recurrence risk scoring model was developed on the basis of regression coefficient of identified independent risk factors.

Results: In the data set of 206 patients, the median follow-up was 48 months (range: 6-156 months) and mean and median disease-free survival (DFS) were 87.41 and 85 months. The 1, 5, 10 years DFS was 95%, 54% and 41%. The independent risk factors (on modified p value <0.40) for recurrence were Tumour stage, Nodes stage, grade, age groups, pathologic complete response, intrinsic subtype, and type of surgery. Risk score prepared by regression coefficient (β), was in the range of 1-8 with median score of 5. ROC curve showed that area under ROC Curve of the score was 71.8% (95% CI: 64.8%-78.8%, p<0.001). To detect recurrences, a risk score ≥3 had at least 93.1% sensitivity and 31.9% specificity whereas score ≥4 had at least 73.5% sensitivity and 59.6% specificity. Based on cluster analysis, score 1-4 was identified as low risk whereas 5-6 as moderate risk group and ≥7 identified as high-risk group and their mean/median disease free survival time were 107.86/ NR, 66.99/30 months and 58.34/20 months respectively.

Conclusions: The significant difference in DFS among three risk groups, indicates goodness of the fit of our risk score model. The risk scoring model developed by us is simple, easy to use in clinic and can be used for selecting high risk patients who benefit from treatment intensification.