Abstract PR004: Radiogenomics for predicting underestimation of invasiveness in ductal carcinoma in situ (DCIS) diagnosed with vacuum assisted breast biopsy: study rationale and design

Background: Concerns about overdiagnosis and overtreatment of DCIS have led to an interest in de-escalating treatment for DCIS. However, accruing patients to active surveillance trials poses several issues, the most pressing of which is the high risk of synchronous invasion at the biopsy level (~26%). Since DCIS-associated stromal changes and influx of immune cells might be mediators of progression to invasive breast cancer, the present project will investigate whether radiogenomics is able to predict the upstaging in patients preoperatively diagnosed with DCIS with Vacuum-Assisted Breast Biopsy (VABB). Hypothesis: We hypothesize a different immune gene expression signature between pure DCIS (p-DCIS) and DCIS with an invasive component at final surgery (i-DCIS). -We hypothesize to be able to identify these different immune gene expression signatures in VABB tissue samples. -We hypothesize that a subsequent gene expression profiling-derived immunohistochemistry signature in DCIS diagnosed on VABB, integrated with other histopathological characteristics and preoperative imaging - Standard Mammography (SM), Digital Breast Tomosynthesis (DBT), Contrast-Enhanced Spectral Mammography (CESM), and Magnetic Resonance (MR) - will predict with higher accuracy the upstaging at final surgery. We hypothesize that the implementation of a radiogenomic approach will reproduce all this information with ultimate precision. Aims: Aim 1: to investigate the relationship between immune gene expression signature at diagnosis of DCIS at VABB and at final surgery outcomes. -Aim 2: to investigate if the association of imaging plus genomics is better than genomics alone. -Aim 3: to create and validate a gene expression profiling-derived immunohistochemistry signature. -Aim 4: to implement radiomic in DCIS management. Experimental Design: Starting from a cohort of more than 2000 consecutive patients with a diagnosis of DCIS on VABB and subsequently operated at the European Institute of Oncology, training and testing datasets will be created. VABB p-DCIS and i-DCIS will be profiled using a next-generation sequencing gene expression assay, targeting 395 genes associated with tumor-immune systems interactions and performed on RNA extracted from DCIS samples. Gene expression and the subsequently derived immunohistochemistry profiles will be integrated with preoperative imaging (SM, DBT, CESM, MR). Finally, we will perform a radiomic feature analysis by SM, DBT (when available), and CESM-detected DCIS at VABB. Impact on Cancer: Patients with features that place them at low risk may be more likely to forego additional therapy. Clinicians would distinguish patients who are candidates to surveillance only, because the "pure" DCIS has been completely removed, from patients going to active surveillance plus endocrine treatment because the residual lesion is with high probability a DCIS, from patients to be sent to surgery because most certainly the invasive component will be present.

Citation Format: Matteo Lazzeroni, Luca Nicosia, Elena Guerini Rocco, Sara Gandini, Sara Raimondi, Massimo Barberis, Nicola Fusco, Marta Cremonesi, Daniela Origgi, Enrico Cassano, Aliana Guerrieri Gonzaga, Davide Serrano, Debora Macis, Harriet Johansson, Bernardo Bonanni. Radiogenomics for predicting underestimation of invasiveness in ductal carcinoma in situ (DCIS) diagnosed with vacuum assisted breast biopsy: study rationale and design [abstract]. In: Proceedings of the AACR Special Conference on Rethinking DCIS: An Opportunity for Prevention?; 2022 Sep 8-11; Philadelphia, PA. Philadelphia (PA): AACR; Can Prev Res 2022;15(12 Suppl_1): Abstract nr PR004.