Molecular and phenotypic characteristics influencing the degree of cytoreduction in high-grade serous ovarian carcinomas

Abstract

Background: High-grade serous ovarian carcinoma (HGSOC) is the deadliest ovarian cancer subtype, and survival relates to initial cytoreductive surgical treatment. The existing tools for surgical outcome prediction remain inadequate for anticipating the outcomes of the complex relationship between tumour biology, clinical phenotypes, co-morbidity and surgical skills. In this genotype-phenotype association study, we combine phenotypic markers with targeted DNA sequencing to discover novel biomarkers to guide the surgical management of primary HGSOC.

Methods: Primary tumour tissue samples (n = 97) and matched blood from a phenotypically well-characterised treatment-naïve HGSOC patient cohort were analysed by targeted massive parallel DNA sequencing (next generation sequencing [NGS]) of a panel of 360 cancer-related genes. Association analyses were performed on phenotypic traits related to complete cytoreductive surgery, while logistic regression analysis was applied for the predictive model.

Results: The positive influence of complete cytoreductive surgery (R0) on overall survival was confirmed (p = 0.003). Before surgery, low volumes of ascitic fluid, lower CA125 levels, higher platelet counts and relatively lower clinical stage at diagnosis were all indicators, alone and combined, for complete cytoreduction (R0). Mutations in either the chromatin remodelling SWI_SNF (p = 0.036) pathway or the histone H3K4 methylation pathway (p = 0.034) correlated with R0. The R0 group also demonstrated higher tumour mutational burden levels (p = 0.028). A predictive model was developed by combining two phenotypes and the mutational status of five genes and one genetic pathway, enabling the prediction of surgical outcomes in 87.6% of the cases in this cohort.

Conclusion: Inclusion of molecular biomarkers adds value to the pre-operative stratification of HGSOC patients. A potential preoperative risk stratification model combining phenotypic traits and single-gene mutational status is suggested, but the set-up needs to be validated in larger cohorts.

Keywords: cancer genetics; clinical observations; gynaecological oncology; mutations; risk model; surgery.

FIGURE 1

Flow chart depicting the selection from the biobank of patients for inclusion. The inclusion criteria included a confirmed HGSOC diagnosis and the fact that the patients underwent standardised treatment defined as primary cytoreductive surgery and postoperative chemotherapy. The FIGO 2014 classification was used for cancer staging. OS: overall survival; R0: no residual tumour tissue after primary cytoreductive surgery; R ≠ 0: any residual tumour tissue (irrespective of size) after primary cytoreductive surgery; FIGO: the International Federation of Gynaecology and Obstetrics classification of malignant ovarian tumours from 2014.

FIGURE 2

An oncoplot mutation list of all mutations detected by targeted DNA sequencing (360 gene panels) of ovarian cancer patients (n = 97). Genes are altered in 93 (95%) of 97 samples. Genes are listed on the left. Bars on the right indicate the prevalence of each mutation among the tumours analysed. Mutations are colour‐coded based on the type of mutation detected. Patient IDs are given below the columns, and each column represents one tumour/patient. At the top, the tumour mutational burden (TMB) is demonstrated. TMB was calculated by adding all missense, insertions/deletions and frameshift variants within the tumour sample and dividing by the total size of the panel. In the stacked bar plot, the bars are coloured according to the type of mutation discovered: C > T; red, C > G; dark blue, C > A; light blue, T > A; green, T > C; yellow, T > G orange (see also Figure S1). The different TP53 mutations are presented in Figure S2.

FIGURE 3

A bar graph illustrating the mutational status and pathway analysis results, according to surgical outcomes. Along the x‐axis, the genetic pathways are listed above the stipulated horizontal line, while the mutated genes are listed below. In the figure, the blue bars indicate genetic changes in the tumours of patients with complete cytoreductive surgery, while the red bars show which mutations can be found in the tumours of patients with residual tumours after surgery. Percentages of mutations are demonstrated along the x‐axis.

FIGURE 5

Cox survival curves of the HGSOC cohort. The patients are analysed separately based on whether they belong to the R0 or the R ≠ 0 cohorts. The curves show overall survival (OS) for patients with (the three upper plots) or without (the three lower plots) complete cytoreductive surgery. In all plots, the blue lines demonstrate survival curves for patients who do not harbour the specific mutational status while the red curves illustrate the OS for patients whose tumour carry a mutation in the genes: (A) Other DNA repair genes, (B) HRD genes, or (C) part of the chromatin remodelling pathway and/or histone methylation pathway. Below the survival curves a table is inserted showing the number of patients harbouring the specific mutational status in each cohort.

FIGURE 4

Kaplan–Meier survival analysis of the total HGSOC cohort stratified according to surgical outcome (R0 vs R ≠ 0). While the cumulative survival of the HGSOC cohort is depicted along the y‐axis, the duration of the overall survival in months is demonstrated along the x‐axis. Patients with complete surgical removal of tumour tissues (R0) are shown in blue, while the red line shows the patients with incomplete tumour tissue removal (R ≠ 0).