5340 - Quantitative imaging and characterization of collagen patterns in high grade serous ovarian carcinoma (HGSOC)

Background

HGSOC is known to demonstrate diverse molecular heterogeneity. It is unclear whether the tumor microenvironments such as the stromal components in the extracellular matrix (ECM) also exist heterogeneity.

Methods

As a proof-of concept study, characterization of collagen patterns was performed on 60 unstained formalin-fixed paraffin embedded (FFPE) HGSOC samples (three sections from each of the 20 patients). Each section with 5 micron thickness and a minimum tumour surface area of 40 mm² were scanned by using a multiphoton imaging system (Genesis 200, HistoIndex) following deparaffinization. Parameters including Collagen Area Ratio (CAR), Collagen Fiber Density (CFD), Collagen Reticulation Index (CRI), Collagen Fiber Number (CFN), Collagen Fiber Thickness (CFT), and Collagen Fiber Length (CFL) were analysed. Unsupervised hierarchical clustering analysis was performed.

Results

Unsupervised hierarchical clustering of the collagen parameters revealed four distinct patterns in HGSOC samples. G1 tumors consisted of long and thick collagen fibers with high CFT and CFL. G2 tumors were low in all the collagen parameters, suggesting a relatively “clean” stromal without collagen deposition. G3 tumors consisted of dense collagen fibers with high CRI suggestive of extensive cross alignment among the fibers. G4 tumors were high in CFN and low in CRI, CFT and CFL, suggesting a stroma loaded with high amount of thin and short collagen fibers without cross alignment.The collagen patterns were not exclusive for specific organ sites (ovary, fallopian tube, other metastatic sites) except that the collagen pattern from the omental metastases were mainly G2 (10/18; 55.6%) and G4 (6/18; 33.3%).

Conclusions

The stromal component in the ECM of HGSOC can be successfully quantitated by the multi-photon imaging technology on unstained FFPE sections. The collagen component exhibited significant heterogeneity in terms of the number, thickness, length, and reticulation patterns. The contribution of different collagen patterns to clinical outcomes and the correlation with known molecular subtypes in HGSOC warrants further investigation in larger cohorts.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

National University of Singapore.

Funding

National Medical Research Council of Singapore.

Disclosure

All authors have declared no conflicts of interest.