Abstract 122P
Background
Cribriform (cbf) morphology has been shown in carcinomas from various organs, including prostate and breast, to be associated with worse patient outcomes and unique molecular features. We sought to examine the multimodal data from the landmark Cancer Genome Atlas (TCGA) study to clarify molecular and prognostic factors associated with cbf morphology in colorectal adenocarcinomas.
Methods
We retrospectively analyzed the available TCGA CRC cohort (n=594). Clinicopathological, digital slides, and molecular data were extracted. Cases with insufficient clinical data, no diagnostic slides, or poor slide quality were excluded. TCGA coded data, pathology reports, and digital diagnostic slides were reassessed by one gastrointestinal pathologist and a resident. Cbf morphology was graded using five categories based on prostate cribriform studies: (0) no-cribriform, (1) loose cribriform, (2) small cribriforming glands, (3) diffuse large cribriform, and (4) intraductal carcinoma of the prostate-like. Cases were grouped into cbf 'low' (0-1) and cbf 'high' (2-4) groups, which were correlated with clinicopathological, histological and molecular features. Immune infiltration was characterized by cell type fraction. An adjusted α = 0.05 for statistical significance was used for all tests.
Results
573 CRC cases were included. Cbf 'high' CRCs (159/573) were associated with more left-sided tumors (p=0.049), metastatic diseases (p=0.0001), less MSI-H (p=0.003), and different aneuploidy and fraction genome altered (p=0.003-0.008). Cbf 'high' vs 'low' tumors did not show a significant difference in grading distribution. T and N staging were not associated with cbf morphology. By cell fraction, cbf ‘high’ CRCs showed an immune infiltrate with increased regulatory T cells (p<0.0001) but decreased memory B cells (p=0.046), M2 macrophages (p<0.0001), neutrophils (p=0.018), resting dendritic cells (p=0.046), compared with cbf ‘low’ CRCs. Cbf 'high' tumor morphology was associated with worse overall survival outcomes by log-rank testing (p=0.037).
Conclusions
Cribriform morphology in colorectal adenocarcinoma TCGA cases was associated with poorer outcomes and distinct transcriptomic immune profiling.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Q-H. Trinh.
Funding
Fonds de recherche Santé Québec.
Disclosure
All authors have declared no conflicts of interest.
Resources from the same session
5P - Clinicopathologic features and genomic profiling of occult breast cancer
Presenter: Liansha Tang
Session: Poster Display
Resources:
Abstract
6P - Tumor cell-released autophagosomes (TRAPs) promote lung metastasis through inducing PD-L1 high expression of pulmonary vascular endothelial cells (PVECs) in breast cancer
Presenter: Xuru Wang
Session: Poster Display
Resources:
Abstract
7P - Tumor cell-released autophagosomes (TRAPs) promote breast cancer lung metastasis by modulating neutrophil extracellular traps formation
Presenter: Xiaohe Zhou
Session: Poster Display
Resources:
Abstract
9P - Clinicopathological features and prognosis of mucinous breast cancer: A retrospective analysis of 358 patients in Vietnam
Presenter: Hoai Hoang
Session: Poster Display
Resources:
Abstract
10P - Comparison of 28-gene and 70-gene panel in risk-prediction of Chinese women with early-stage HR-positive and HER2-negative breast cancer
Presenter: Lei Lei
Session: Poster Display
Resources:
Abstract
11P - Multimodal analysis of methylation and fragmentomic profiles in plasma cell-free DNA for differentiation of benign and malignant breast tumors
Presenter: Hanh Nguyen
Session: Poster Display
Resources:
Abstract
12P - Plasma cell-free mRNA profiles enable early detection of breast cancer
Presenter: Chi Nguyen
Session: Poster Display
Resources:
Abstract
13P - Relationship of distress and quality of life with gut microbiome composition in newly diagnosed breast cancer patients: A prospective, observational study
Presenter: Chi-Chan Lee
Session: Poster Display
Resources:
Abstract
14P - Classification of molecular subtypes of breast cancer in whole-slide histopathological images using a novel deep learning algorithm
Presenter: Hyung Suk Kim
Session: Poster Display
Resources:
Abstract
15P - The regulation of pregnenolone in breast cancer
Presenter: Hyeon-Gu Kang
Session: Poster Display
Resources:
Abstract