Abstract 809P
Background
Uterine leiomyosarcomas (LMS) is a highly aggressive but rare uterine tumor. However, it is almost impossible to distinguish LMS from the most common benign uterine leiomyomas (LM) through pre-operative diagnosis, leading to poor prognosis of LMS patients. Thus, it is clinically important to identify molecular differences between LMS from LM, which would not only advance our understanding on tumorigenesis of LMS but also lay a foundation for developing effective early detection strategy.
Methods
We performed whole-exome sequencing from 78 LM and 10 LMS treatment-naïve tumor samples along with their matched normal samples as well as RNA-sequencing on 4 LM and 10 LMS samples. We employed a comprehensive bioinformatics analysis based on WES and RNAseq data to identify differential molecular features between the two diseases.
Results
We find that TP53 loss-of-function mutations are exclusively observed in LMS (5 out of 10 LMS samples, 0 out of 78 LM samples; Fisher’s exact test, p = 5.8x10-5), whereas MED12, a cervical cancer driver gene, is significantly mutated in LM group (1 out of 10 LMS samples, 37 out of 78 LM samples; Fisher’s exact test, p = 3.9×10-3). LMS tumors harbor a higher tumor mutation burden (p = 2.2×10-6, two-tailed Wilcoxon rank-sum test). Compared with LM, mutational Signature 5, 7, 10 show relatively higher contributions in LMS (Signature 5, Age, p = 2.0×10-2; Signature 7, Ultraviolet, p = 2.1x10-3; Signature 10, POLE, p=3.5×10-3; two-tailed Wilcoxon rank sum test). LMS tumors also show much significantly higher fractions of loss of heterogeneity (LOH, p=1.2×10-5, two-tailed Wilcoxon rank sum test), suggesting the genome-wide instability. We also noted the well-known cancer hallmark genes are consistently upregulated in LMS, including EGFR, BRAF, PIK3CA, PIK3CB, JAK1/2 and MTOR (|log2FC| > 1, FDR < 0.05). Consistent with the Signature 7 signal, UV_response_DN pathway is significantly upregulated in LMS.
Conclusions
Using unbiased genome-wide molecular profiling data, we identified the mutation and gene expression features that distinguish LMS from LM. LMS appears to be more driven by TP53 mutations, accompanied with genome-wide instability and cancer-hallmark pathway dysregulation.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
H. Yang.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.