Determination of endometrial cancer molecular subtypes using a whole exome-sequencing based single-method approach

In modern oncology, molecular diagnostics are becoming increasingly important, particularly when it comes to individualized therapy decisions (Malone et al. 2020; Prasad et al. 2016). The incorporation of molecular markers into therapy stratification is intended to improve the treatment response and clinical outcome and to increase the safety of oncological therapy by reducing therapy-related toxicity (Murali et al. 2018; Urick and Bell 2019; Arend et al. 2018). The TCGA molecular subtypes of endometrial cancer, namely POLE, MSI, CN low, and CN high, and analogue classifications, in particular the ProMisE algorithm, are firmly established in clinical routine and are recommended for consideration in treatment decisions by national and international guidelines (Emons et al. 2023; Oaknin et al. 2022; Concin et al. 2021). Patients with POLE EC, typically displaying high-grade endometroid carcinomas, exhibit an excellent prognosis. Considering this excellent prognosis, overtreatment by means of adjuvant chemotherapy and/or radiation can be avoided. In contrast, the CN high subgroup comprises EC with predominantly serous-like histology, exhibit a very poor prognosis necessitating an extended therapy approach. The prognostic value of these retrospectively obtained four molecular subtypes was validated in a large prospective EC cohort (León-Castillo et al. 2020). From a clinical point of view, it is critically important to identify these different subtypes to avoid undertreatment in high-risk patients and overtreated in low-risk patients. To achieve this in clinical routine, a suitable diagnostic tool is crucial. Such a diagnostic tool is required to provide (i) a precise classification with high prognostic values and, (ii) should be methodologically feasible. The TCGA original publication is based on an extremely complex methodology comprising genomic, transcriptomic, and proteomic data (Cancer Genome Atlas Research Network et al. 2013). Talhouk et al. reported molecular surrogate markers obtained from immunohistochemistry and DNA sequencing to precisely identify four EC molecular subtypes with analogue prognostic values to the TCGA subtypes (ProMisE; (Talhouk et al. 2015; León-Castillo et al. 2020). In the present study, we investigated the feasibility of determining EC molecular subtypes in analogy to both, the TCGA- and molecular surrogates classification applying a WES-based single method approach that has been recently published (Mustea et al. 2023). Determination of EC molecular subtypes in analogy to TCGA-classification applying the single method approach has not yet been able to address high-risk patients with sufficient certainty. This could be due to methodological reasons. The DNA quality extracted from FFPE tissues in the EC-cohort was very heterogeneous. This might also be reflected by the comparatively high rate of MSI/dMMR cases (Hwang et al. 2017; Pécriaux et al. 2021). DNA quality analysis exhibited a very high proportion of short nucleic acid fragments, which may have led to the supposed detection of MSI/dMMR. Consistently, an increased number of MSI/dMMR tumors reduces the group of high-risk carcinomas. Comparable conclusions can be extrapolated regarding the CN status. Here, poor DNA quality may have also led to a decreased detection of CN high tumors. In this regard, we further investigated whether there is a correlation between MSI/dMMR status determined by MH Guide analysis and selected quality parameters. We found a strong correlation between MSI status and the fraction of mapped and on-target reads, the fraction of short fragments, and the amount of adapter contamination per sample (data not shown). EC samples were defined as MSI/dMMR if they had any of the following characteristics: Percentage of mapped reads < 92%, percentage of on-target reads < 66%, percentage of short fragments > 93%, or adapter contamination > 6.34%. All four parameters demonstrated a strong correlation with the age of the samples. All samples collected from 2011 onwards were within the non-critical range of parameter values. The default thresholds at which MH Guide triggers quality warnings for the fraction of mapped and on-target reads or short fragments are far more stringent than the critical values for these parameters determined by MSI definitions. It remains speculative whether the analysis would have been different with higher DNA quality. However, this has to be clarified conclusively in further investigations. The surrogate marker-based stratification by DNA sequencing, however, reliably identified low- and high-risk patients in the EC-cohort. The surrogate marker-based classification differs from the TCGA-analogue classification mainly with respect to the determination of the CN high/low or p53abn/NSMP group. Our data imply that based on p53 mutation status, at least in our cohort and in the presence of poor DNA quality, a stratification into high- or low-risk patients can be more reliably determined. Of note, TP53 mutations were identified in all four subgroups, but did not show prognostic values in POLE and MSI/dMMR EC. This is clinically highly relevant and testing for all four subtypes should always be performed in routine clinical practice. If due to the complexity of the diagnostic procedure, testing for POLE is omitted, a p53abn but undetected POLE positive EC may be considered as high-risk with corresponding overtreatment.

Our data strongly support the inclusion of molecular subtyping in the diagnostic regimen of EC. Further, conduction of interventional trials performing stratification based on molecular subtypes is crucial and currently ongoing (e.g. RAINBO trial; NCT05255653; (RAINBO Research Consortium 2022). Our data further indicate, that identification of molecular subtypes in EC with subsequent risk stratification is feasible based on molecular surrogates obtained by a single method approach provided by MH Guide. The implementation of such a diagnostic tool in routine clinical practice is crucial to ensure safer and more effective treatment of EC patients. Due to the growing understanding of tumor biology, molecular markers will become increasingly important in the future (Arend et al. 2018; Stope et al. 2016; Vermij et al. 2020). In this context, new genetic markers can be easily integrated into the WES-based method allowing rapid translation of data from research into clinical practice. Furthermore, tumor genetic characterization provides the potential to identify targeted therapies based on tumor genetic alterations. This is therapeutically of great value especially in advanced disease stages after passing through different lines of therapy.