Ductal carcinoma in situ (DCIS) is a non-invasive form of breast cancer. The lesions are typically surgically removed. However, there is a risk of recurrence as another DCIS or as a more dangerous invasive cancer. If the DCIS is large, full mastectomy is used, and the chance of recurrence is extremely low, so no further treatment is necessary. Based on large randomized prospective clinical trials, patients who get lumpectomy face a 20-30% chance of recurrence, which would be invasive in about half of the cases. Post-surgery radiation cuts this risk in half, but the absolute number of patients who avoid recurrence is only in the 10% range, and there is no evidence that radiation decreases the rate of breast cancer-associated mortality. Given the modest chance of benefit, convenience issues, monetary treatment costs, and potential short- and long-term health-related issues associated with radiation, it is important to accurately identify patients who need this treatment. In fact, a recet NIH expert panel identified the greatest current need in DCIS research to be an accurate risk assessment tool. Unfortunately, currently available clinicopathologic risk factors like patient age tumor size, tumor grade, and margin size are not very accurate. This study proposes development of a multi-marker personalized test that accurately stratifies DCIS patients into risk of invasive recurrence groups to guide treatment choice. The test will be a compilation of multiple previously validated molecular markers-assayed by immunohistochemistry (IHC) in the tumor tissue-and established clinicopathologic risk factors. Robust, reproducible assays for eight molecular markers will be developed prior to grant funding. Phase I is both a validation and further development study. Eight IHC assays will be run at a CLIA reference lab on archival tissue samples from ~375 DCIS patients with known long-term outcome. Pre-defined published single-marker and multiple-marker molecular and clinicopathologic risk models first will be tested, comparing the outcomes of patients in each risk group (e.g., low, intermediate, and high). Then, model building will be used across all available data to produce a single optimized risk model that divides patients into discrete risk groups. The primary goal is to accurately defin a low-risk group with a long- term local invasive recurrence rate low enough to justify foregoing radiation treatment. Phase II is a blinded pivotal clinical study of ~700 archival lumpectomy-only DCIS patients derived from registries at three major sites. The performance of the optimized model locked down in Phase I will be assessed in this independent set of patients. All scoring will be done without knowledge of patient outcome, and all data and analyses will be handled by an independent data management firm. Secondary analyses will include assessment of a high-risk group who may need more aggressive treatment, DCIS recurrence risk, as well as risk in lumpectomy plus radiation patients.
The prognostic test proposed for development will help guide the aggressiveness of treatment in patients diagnosed with ductal carcinoma in situ (DCIS), a pre-invasive (Stage 0) form of breast cancer. This personalized prognostic tool would spare many low-risk patients from unnecessarily aggressive treatment and ensure that high-risk patients receive sufficient therapy, improving outcome and quality of life.