Population-based assessment of leiomyosarcoma (LMS) and cancers in the Li-Fraumeni syndrome (LFS) spectrum: Implications for genetic testing criteria.

Abstract

11570 Background: Data regarding the heritability of rare tumors is limited and may prevent incorporation into genetic testing criteria. This study utilized the Utah Population Database (UPDB) to evaluate cancer risks among LMS cases and their relatives and the prevalence of meeting Chompret criteria (ChC) for LFS genetic testing. Methods: Between 1995-2021, 429 LMS cases with UPDB genealogies were identified from the Utah Cancer Registry. Diagnoses were confirmed, when possible, by pathology reports. Cases were individually age- and sex-matched 1:5 to population controls with similar pedigrees and follow-up (n=2145 controls). Cancers from 1966-2021 were obtained for study subjects and their first- through third-degree relatives. LFS spectrum cancers included breast, soft tissue sarcomas, osteosarcomas, CNS/brain, and adrenocortical. Hazard rate ratio (HRR) estimates of self- and familial relative cancer risks in LMS compared with controls was calculated from a Cox model adjusting for the number of relatives, degree of relatedness, and person-years at risk. Results: A 2.2-fold risk (p<0.001) of a cancer in the LFS spectrum was seen in cases with a non-uterine LMS site (n=323) at any age and a 4.5-fold risk (p<0.001) for developing an LFS cancer at age <50 y. Non-uterine LMS cases had similarly increased risks for developing a non-LFS-spectrum cancer at any age and <50y (Table). Increased risk of LFS or non-LFS-spectrum cancer was not seen in uterine LMS cases (n=106). Although increased cancer risk was not generally observed in relatives of LMS cases compared with control relatives, we observed that non-uterine LMS and their first-degree relatives had an increased risk of colorectal cancer (CRC) (Table). CRC is not an LFS cancer but is known to occur in LFS families. Excluding the LMS diagnosis, non-uterine LMS cases were more likely to meet ChC compared with controls (Table 1). Uterine LMS cases were no more likely to meet ChC than their respective controls. Conclusions: LMS is associated with cancers outside the spectrum, and further studies are needed to determine if LMS is associated with other cancer predisposition genes. Family history should be evaluated broadly and not restricted to ChC. As uterine LMS appears less likely to be associated with genetic predisposition, considering non-uterine and uterine cases separately may be important for future studies of the genetic basis of LMS. Cancer risks in LMS cases and relatives compared with controls. Non-uterine LMS =323, Controls=1615 Uterine LMS =106, Controls=503 Case HRR P FDR HRR P Case HRR P FDR HRR P LFS cancer    2.2     <0.001     1.1    0.68    0.6    0.36    1.3    0.24    LFS cancer <50    4.5      <0.001     1.2    0.72    1.3    0.77    2.1    0.08    Non-LFS cancer    2.2     <0.001     1.2    0.06    1.0    0.98    1.1    0.44    Non-LFS cancer <50    4.1     <0.001     1.2    0.38    1.5    0.52    1.4    0.35    Colorectal    2.7     0.04     1.6     0.02     1.1    0.91    1.4    0.41    HRR=hazard ratio; FDR=first degree relative.