Abstract P2-05-17: Development of a recurrence risk stratification schema to inform treatment decisions in breast cancer patients undergoing mastectomy with 1-3 positive nodes

• Published in: Cancer research (Chicago, Ill.) Vol. 77; no. 4_Supplement; pp. P2 - P2-05-17
• Main Authors: Scanlan, JM, Ellis, ED, Kaplan, HG, Kieper, DA, Morris, AD, Atwood, M
• Format: Journal Article
• Language: English
• Published: 15.02.2017
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.SABCS16-P2-05-17

Abstract Introduction Controversies exist regarding the appropriateness of radiation therapy (XRT) in mastectomy patients who have 1 – 3 positive nodes (LN+). We sought to create a straightforward recurrence and survival prediction model derived from clinical and pathologic parameters common in the community cancer center setting at the time of surgery (SX) that may help determine the patients most likely to benefit from XRT. Subjects In this retrospective observational study, we examined our institution's breast cancer patient registry to identify mastectomy patients with one to three LN+ at the time of diagnosis and tumor pathology results including Her-2 receptor status. Consistent with current treatment standards, Her-2 positive tumors that were not treated with Herceptin were excluded. Methods Breast cancer registry data elements including clinical, pathology, treatment and outcomes data were extracted. Logistic multiple regressions were used to predict loco-regional recurrence (LRR), distant recurrence (DR) and total recurrence (TR) as well as breast cancer mortality (BCa mortality) and all-cause mortality (ACM) based on clinical and pathologic parameters available at SX. The parameters included in the regression model were: tumor receptor status - estrogen (ER), progesterone (PR) and HER-2, the number of nodes examined, presence or absence of lymphovascular invasion (LVI), extension of LVI, number of LN+, node positivity ratio (positive/examined), surgical margins or “tumor on ink”, (SxM); and patient age. Results The application of these filters to our breast cancer registry yielded 935 patients with a mean follow up time of 7 years. Our sample was “modern”: 95% of our patients were diagnosed after 1999, 80% after 2004. Across multiple analyses, the four most consistent risk indicators were PR(-),LVI (+), LN+>1 and SxM (+). In our model, patients with none of the four factors were designated low-risk, those with only one factor were medium-risk, those with 2+ were high-risk (see Table 1). Risk model Validation: All patients N =LRRDRTRBCa Mort.ACMLow Risk2160%3.7%3.7%2.8%8.8%Medium Risk4912.4%9.6%11.2%5.7%12%High Risk2284.4%12.7%16.7%11.8%18.9%3 grp. p-value .01.001.001.001.01 From this complete group, we extracted the relatively small number of patients (N = 137) who, due to a variety of factors, received no chemotherapy or radiation therapy and which showed stronger contrasts between risk groups (see Table 2). Risk model Validation: No chemotherapy or XRT N=LRRDRTRBCa Mort.ACMLow Risk511.9%3.9%3.9%3.9%23.5%Medium Risk708.6%10%18.7%8.6%27.1%High Risk1625%31.3%50%50%68.8%3 grp. p-value .01.001.001.001.01 Discussion We believe this risk group concept is straightforward, feasible and clinically useful in all clinical settings. Our future work seeks to validate this concept in another independent cancer registry. In our analyses we noted that while both PR- and ER- were predictive of patient outcomes in simple correlations, PR- was more predictive in all multivariate equations. Citation Format: Scanlan JM, Ellis ED, Kaplan HG, Kieper DA, Morris AD, Atwood M. Development of a recurrence risk stratification schema to inform treatment decisions in breast cancer patients undergoing mastectomy with 1-3 positive nodes [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-05-17.