188P - Spatial transcriptomic analysis of tumor tissue in ovarian cancer patients treated with neoadjuvant chemotherapy

Background

Ovarian cancer (OC) is one of the most common gynecological cancers, with the worst prognosis and the highest mortality rate out of gynecological diseases. Despite a good response to the first line of standard neoadjuvant chemotherapy (NAC), the relapses in patients with sporadic ovarian cancer are detected within a short period of time in 70% of cases. The most relevant task is to identify the key features of tumor tissue of non-responders and to find out how these features can affect the development and outcome of OC.

Methods

We applied 10x Visium technology for spatial transcriptomic analysis of FFPE samples with ovarian cancer tissue. Eight patients, treated with neoadjuvant chemotherapy were included: 5 had unfavorable outcomes (local or distant recurrence or metastasis) and 3 did not experienced progression during 2 years after the treatment. All patients had initial partial response after the completion of NAC course assessed by the CRS score system. Sequencing of 10x Visium libraries was performed using NextSeq2000 platform (Illumina). The Spaceranger pipeline was used for the raw sequencing data processing, namely, the Fastq files generation, the quality control of the raw reads, mapping of the reads and counting of the reads mapped to the individual genes. The filtered expression matrixes were analyzed via the Seurat package in R. Filtered data was normalized via SCTransform, merged and additionally re-normalized with SCTransform. The Harmony R package was used for batch effect reduction. The batch-corrected data were used for non-linear dimension reduction and the SNN clustering.

Results

Bioinformatics analysis allowed us to reveal 16 clusters in combined 8 samples. Patients who experienced unfavorable outcomes had clusters with significantly more pronounced expression of genes, belonging to processes of angiogenesis, extracellular matrix remodeling, invasion and immune activation. They include collagens, matrix metalloproteases and other matrix proteins as well as immunoglobulins.

Conclusions

For the first time we performed spatial transcriptomic analysis of NAC-treated patients with two distinct outcomes, favorable and unfavorable.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Larionova Irina.

Funding

Russian Science Foundation, Grant 21-75-10021.

Disclosure

All authors have declared no conflicts of interest.