Allelic imbalance analysis of uterine carcinosarcoma: A comparison of two components of the neoplasm

Aditya Deshpande, The University of Texas School of Public Health


Introduction: Uterine carcinosarcoma (UCS), also known as Malignant Mixed Mullerian Tumor, is a neoplasm of the female genital tract that shows histological features of both carcinoma and sarcoma. Some studies have indicated that UCS arises from sarcomatous differentiation of high-grade carcinoma while others have confirmed a bi-clonal nature of the tumor. Given these differences, further investigation in the origin of this tumor with newer approaches and technologies is warranted. In this study we determined the allelic imbalance (AI) profiles of the two components of UCS and compared the AI of events that were shared by them. We further draw a comparison between these events and de novo carcinoma. Methods: Samples were obtained from 10 patients diagnosed with UCS and were preserved in formalin fixed paraffin embedded (FFPE) blocks. The two components (carcinoma & sarcoma) were identified and micro-dissected, and whole-genome DNA micro-arrays were applied to the isolated DNA. We then applied a sensitive computational method that combines haplotype information with SNP array data to identify somatic segmental copy number variations and copy-neutral loss of heterozygosity (cnLOH). After we obtained the AI events we filtered small events (< 2 MB) to rule out germline duplications. We then obtained events that are common to both components by using a criterion of 80% reciprocal overlap. All such events were classified as a gain, loss or cnLOH based on the log R ratio and B allele frequency deviations. We used a crude tool to analyze difference in AI profile between these events in order to assess developmental directionality. We compared these events with de novo uterine carcinoma using the same workflow. After combining these results with developmental directionality analysis, we classified the putative origin of cancer for individual cases. Results: Overall 31% of the events were highly concordant between the two components but this proportion varied among different cases (from 18% to 100% among the 9 cases). Four out of nine cases showed higher number of events that were on average larger than the events in carcinoma components. Three cases had equal number of events. Gains in 2p, 3q, 8p and 8q were seen in 5 out of 9 cases. Losses were seen in 15q, 17p (6 out of 9 cases); 16p, 17q, 22q (5 out of 9 cases). Out of these events, 8q gain, 15q loss, 22q loss and 17p loss have been implicated in this cancer type by previous studies. When compared to de novo carcinoma, 3q gain, 8q gain, 15q loss, 17p loss and 22q loss were seen in major proportions of both groups. 8q gain has been implicated in endometrial carcinoma and is seen very commonly in both components of carcinosarcoma. Based on this we inferred about the origin of tumor in 3 cases based on these analyses. Conclusions: Our analysis showed that the sarcoma component shared various portions of the AI profile with the carcinoma component. The majority of cases have a monoclonal origin. For some cases we could conjecture the origin of the tumor. The majority of the cases showed AI profile similar to de novo carcinoma especially with 8q gain, corroborating the similarity of the two cancers.

Subject Area


Recommended Citation

Deshpande, Aditya, "Allelic imbalance analysis of uterine carcinosarcoma: A comparison of two components of the neoplasm" (2016). Texas Medical Center Dissertations (via ProQuest). AAI10131750.