Mapping ductal carcinoma in situ in space and time to reveal biomarkers that can predict invasive risk - ABSTRACT Each year, 50.000 women are diagnosed with ductal carcinoma in situ (DCIS) in the US. DCIS is a non-invasive form of breast cancer, which respects the natural tissue barriers of the breast, and is therefore not life- threatening. However, some of these DCIS lesions overrule natural tissue barriers imposed by healthy breast tissue, leading to invasive and metastatic breast cancer (IBC), a life-threatening disease. There is currently no way to predict which DCIS will progress into IBC, causing all women diagnosed with DCIS to be treated with invasive surgery and systemic therapy, causing overtreatment, harmful side effects, and psychological burden, while also using valuable resources in healthcare systems. One of the caveats of clinical DCIS research is the natural progression, which cannot be followed over time, as almost all patients are treated with surgical excision when a lesion is detected. Therefore, the dynamics and natural course of DCIS growth and invasive transformation remain poorly understood. To resolve this lack of dynamic information, PI Scheele at VIB, together with collaborating labs at the Netherlands Cancer Institute, has developed the first living biobank of >100 patient- derived xenograft (PDX) models of DCIS directly derived from human samples through orthotopic injections. This biobank is a full representation of the patient population, and represents all DCIS molecular subtypes, morphological growth patterns, histological features and grade, with high concordance with the original patient samples. PI Scheele’s team also developed a way to analyze DCIS progression using a 3D whole-gland imaging technique to study lesion size, location, and growth pattern in the intact tissues to better assess the growth pattern of the different lesions. Using this approach, the team has shown that 3D growth pattern provides a much better prediction for invasive progression than the classical markers derived from 2D histopathology. The VIB team has identified that all DCIS lesions can be classified into two distinct 3D growth patterns: replacement growth and expansive growth. Only the expansive growth pattern strongly correlates with invasion, whereas replacement lesions never progress into invasive disease. However, the mechanisms by which one growth pattern progresses into invasive breast cancer, whereas the other growth pattern remains indolent remain unknown. Additionally, it is uncertain if these PDX derived growth patterns occur in patient samples, and whether histological analysis will allow to identify these growth patterns. PI Scheele proposes to assess these uncertainties and to further investigate the growth patterns of DCIS by understanding how DCIS cells can overrule the natural barriers imposed by healthy breast tissue. Using the unique cohort of DCIS-PDX models, combined with optical barcoding, 3D imaging and longitudinal (intravital) microscopy, the VIB team aims to build detailed spatio-temporal (xyzt) maps of DCIS growth and progression. From these 4D maps of DCIS growth and progression, the team aims to extract predictive biomarkers, such as morphological features or marker gene expression, that can discriminate low-risk from high-risk DCIS prior to invasion in clinical pathology practice.
