MSACL 2025 Abstract

Identifying ‘DESI’red Prognostic Features of Human Ductal Carcinoma In Situ and Invasive Ductal Carcinoma using Spatialomics

Natasha Iaboni (1), Teaghan Kooster (1, 2), Martin Kaufmann (1, 2), Madeleine Carew (1), Amoon Jamzad (1), Abdulhameed Abdulhameed (1, 2), Rachel Rubino (1), Kevin Ren (1, 2), John F. Rudan (1, 2), Parvin Mousavi (1), Sonal Varma (1, 2), Christopher J.B. Nicol (1)
(1) Queen’s University, Kingston, ON, (2) Kingston Health Sciences Centre, Kingston, ON.

Natasha Iaboni, BSc. (Hon) (Presenter) Queen's University >> POSTER (PDF)

Presenter Bio: Natasha is a senior PhD candidate in the Department of Pathology & Molecular Medicine at Queen’s University under the supervision of Dr. Christopher Nicol. She graduated with a Honours Bachelor of Science with a specialization in Life Sciences in 2020. Her research project uses spatial metabolomics to reveal dysregulated metabolomic profiles distinguishing aggressive from non-aggressive breast and colorectal tumours, that may eventually be exploited therapeutically. Longterm, Nastasha’s research aims to advance understanding of breast and colorectal cancers, and improve outcomes for patients diagnosed with these diseases.

Relevant Financial Disclosures (within past 24 months, reported on Apr 22, 2026) No relevant financial relationship(s) to disclose. Conflict mitigation NOT REQUIRED.

INTRODUCTION:
Breast cancer is the leading cause of new cancer cases and second leading contributor to cancer-related deaths among Canadian women. In 2024, ~30,500 Canadian women were diagnosed with breast cancer, and >5500 women died of breast cancer related complications. Between 20-25% of all breast cancer cases diagnosed were classified as ductal carcinoma in situ (DCIS), a form of breast cancer that presents as an abnormal epithelial cell population contained within the breast ducts. While DCIS is non-invasive, it may act as a precursor to invasive ductal carcinoma (IDC) by growing through the ducts and invading surrounding tissue. Women diagnosed with DCIS face challenging treatment decisions due to our current inability to predict the risk of DCIS tumour invasiveness or DCIS recurrence. Due to these knowledge gaps, many women are over treated with aggressive therapies, leading to physical and mental distress, surgical comorbidities and an overall lower quality of life.

Our lab is interested in the peroxisome proliferator-activated receptor γ (PPARγ), a key nuclear transcriptional regulator of lipid and sugar metabolism, and known to stop the growth and spread of breast tumours. Activation of PPARγ signaling via endogenous fatty acids and anti-inflammatory prostaglandins suppresses breast tumour progression via many putative signaling pathways, which are often cell and context specific. PPARγ-dependent protective pathways may include increasing expression of tumour suppressor genes such as BRCA1 or PTEN, decreasing inflammatory pathways by downregulating COX-2, and/or promoting apoptosis-related genes, all leading to decreased growth and spread of breast tumours. Notably, a report of increased PPARγ expression was associated with lower grade DCIS versus invasive breast cancer samples, suggesting PPARγ may normally act to reduce the invasive progression of DCIS. It remains unknown if loss of PPARγ signaling, common in some forms of aggressive breast tumours, plays a role in the transition from DCIS to IDC.

OBJECTIVE:
To address these clinical challenges, the primary objective of this study is to define the metabolomic profiles of human ex vivo DCIS and IDC tumours to enrich the pathological prognostic information available at time of diagnosis.

METHODS:
Formalin fixed and paraffin embedded (FFPE) human ex vivo DCIS/IDC tumour samples (n=8 DCIS, n=16 DCIS/IDC, n=10 IDC, n=34 total)) were obtained from the Kingston Health Sciences Centre. Sections of FFPE tissues were deparaffinized and then assessed using desorption electrospray ionization-mass spectrometry imaging (DESI), over a spectrum of m/z 50-1200 in negative ionization scanning mode, using a spatial resolution of 100um. The DESI-analyzed slides were then stained and annotated by a pathologist for DCIS, IDC, and non-tumour pathological regions. Of the n=34 FFPE samples, n=17 had additional corresponding tissues that were embedded in OCT and DESI assessed to serve as a direct comparison to the analytes detected via FFPE.

The pathologist-guided annotations and DESI 2D heatmaps were then meticulously overlaid with one another using the imaging analysis SlicerMSI software MassVision. Regions of Interest (ROI’s) composed of nine 100X100um pixels were plotted in each pathological zone per slide and merged to form one mean spectra per ROI. A total of n=38,757 ROI’s were selected across 5 pathological zones. Supervised machine learning was conducted along with binary statistical analysis to identify highly significant (p<0.001) ions that had a Log2 fold change >1.5 between DCIS and IDC pathological regions. To complement the metabolomic expression profiles, the samples were assessed using immunohistochemistry (IHC) to obtain an H-Score (intensity x percent positivity) of specified protein expression.

RESULTS:
Notable ions significantly elevated in FFPE IDC versus DCIS tissues included 6 different omega-6 fatty acids such as Linolenic acid (m/z 279.2303), Arachidonic acid (m/z 303.2301) and Docosatetraenoic acid (m/z 331.2600). Omega-6 fatty acids are involved in the production of pro-inflammatory molecules such as pro-tumorigenic prostaglandins (PGs) from arachidonic acid metabolism, which promote invasiveness, progression and immunosuppression. In contrast, we observed the anti-inflammatory PG 15-Deoxy-d-12,14-PGJ2 (m/z 315.11) was significantly increased in FFPE DCIS versus IDC tissue. Interestingly, 15d-PGJ2 is also a natural ligand of PPARγ, and reported to induce apoptosis and decrease oxidative stress via PPARγ activation.

The samples were then stained for PPARγ using IHC to define PPARγ expression within DCIS, IDC and surrounding benign tissues. Our data suggest a significant increase in PPARγ expression both within benign glands and ducts, and in DCIS versus IDC. PPARγ expression was not significantly different when comparing benign glands and ducts to DCIS. Within our cohort, there were no significant differences between intermediate versus high DCIS grades or moderate versus high IDC grades.

CONCLUSION:
This data suggests the significant increased expression of 15d-PGJ2 and PPARγ within DCIS versus IDC samples may aid in revealing the aggressive potential of these tumours and as well as further our understanding of those patients who are at risk for progression. Taken together, our findings also suggest a decrease in PPARγ expression correlates to progression from DCIS to IDC. Validating these expression profiles will allow us to provide pathologists with potential prognostic targets that may be routinely used to help optimize clinical treatment decision-making for DCIS patients.