Leiomyosarcoma is a malignancy with an aggressive course and limited treatment options. At present, prognosis of LMS is predicted using traditional clinicopathologic features15
and the current chemotherapeutic armamentarium for these tumors, when uncontrolled, is based predominantly on doxorubicin and related therapies and results in only marginal improvements in survival. Although molecular biomarkers do not yet inform prognostication and treatment of LMS in the clinical setting, gene expression microarrays have already been employed to detect signatures of probable metastasis in these tumors.10,18
Additionally, macrophage infiltration18
and the CSF1 response signature7
have been identified by our group as poor prognostic factors in LMS. An improved understanding of the pathogenesis of LMS molecular subtypes could help further inform prognosis and shed light on chemotherapeutic vulnerabilities of LMS. Given the improved prognosis of Group 1 LMS when compared to other LMS subtypes, reliable distinction of Group 1 from both other LMS subtypes and other soft tissue sarcomas is critical. Our studies here are intended to determine whether 5 IHC markers shown previously to be useful for the distinction of Type I LMS (with good prognosis) could be useful in a clinical setting.
Our findings suggest that the novel markers ACTG2, CASQ2, CFL2, MYLK, and SLMAP are valuable tools in the diagnosis of Group 1 LMS, a subset of LMS cases that we previously defined by gene expression profiling.1
The presence of strong Group 1 marker expression in 25% of LMS in the current study coincides with previously reported finding that Group 1 comprises ~25% of LMS.1
When used in concert, these markers were also useful at distinguishing Group 1 LMS from a variety of other sarcomas. Individually, however, the antibody stains had variable reactivity profiles and not infrequently stained other non-LMS sarcomas. These findings suggest that while a panel of stains is a powerful tool for the diagnosis of Group 1 LMS, single markers are of only limited utility depending upon the differential diagnosis.
One frequently encountered diagnostic problem for which individual markers may be of value is in the distinction of LMS from de-differentiated liposarcoma. While diagnosis is straightforward in morphologically classic cases, differentiating LMS from liposarcoma may be difficult when the former entraps significant amounts of fat or the latter shows areas of divert differentiation. Expression SLMAP was highly specific (100%) for LMS when the differential was limited to LMS vs. liposarcoma. MYLK performed similarly, showing 96% specificity for LMS when applied in this focused differential.
The presence of strong Group 1 marker expression in significant numbers of ESS and GIST is intriguing. Smooth muscle differentiation is well-documented in both tumors. In the case of ESS, admixed areas of benign smooth muscle are relatively common and are not thought to influence prognosis so long as cytologic atypia is absent; the presence of any significant atypia in areas of smooth muscle differentiation excludes the diagnosis of ESS and places the lesion under consideration for a diagnosis of LMS.20
Group 1 marker staining in bland smooth muscle components of ESS is not surprising, given that a subset of benign smooth muscle tumors react with these immunostains, and should not represent a source of additional diagnostic difficulty in the distinction of ESS and LMS.
Group 1 marker staining was also noted in some GISTs. A subset of these GISTs has been shown to demonstrate histologic features of smooth muscle differentiation with desmin and actin positivity. While initial work suggested distinct differences in gene copy number between LMS and GIST,19
recent molecular studies have demonstrated common regions of 13q and 11q imbalance in GIST and LMS.23
Further study is warranted to evaluate the significance of the morphologic, immunohistochemical, and molecular overlap of some GISTs with LMS.
Although the Group 1 subtype includes a greater proportion of histologically well-differentiated LMS than do the other LMS clusters, the prognostic significance of Group 1 was independent of grade1
and the absence of strong Group 1 staining in the majority of leiomyomas provides argument against the supposition that the Group 1 subtype constitutes a better-differentiated form of LMS. In addition, these clinically well-behaving tumors paradoxically demonstrate increased levels of genomic complexity, with shared affected genes, when compared to other LMS.1
Prior work has also shown that the most frequently lost region in Group 1 LMS is 2.5 MB on 1p36.32, which includes the PRDM16
gene. The loss of PRDM16
promotes differentiation of brown fat precursors into skeletal muscle and results in elevated expression of genes involved in smooth, cardiac, and skeletal muscle (including ACTG2, MYLK, CFL2, SLMAP, and CASQ2).1,21
The heterogeneity of muscle genes expressed in Group 1 tumors may help to explain the frequent overlap of individual Group 1 immunostains with other sarcoma subtypes.
Despite the increased genetic complexity of Group 1 LMS, patients with these tumors experience a survival benefit that is independent of histologic grade1
suggesting that factors beyond degree of differentiation influence prognosis. The finding that the Group 1 phenotype connotes improved prognosis irrespective of differentiation is of particular interest in cases of UPS showing strong Group 1 marker expression by immunohistochemistry, particularly given the finding that a comparable percentage of UPS show high coordinated Group 1 gene expression on hierarchical clustering of cDNA microarrays. These markers may define a small population of Group 1 LMS-derived UPS with improved prognosis and could suggest future therapeutic targets. This is especially relevant given that myogenic differentiation usually confers worse prognosis with increased risk of metastasis and decreased overall survival in UPS. 6,8
Although the rarity of Group 1 marker-positive UPS argues against the possibility that large numbers of Group 1 LMS “hide” within the UPS category, reappraisal of UPS with Group 1 features may provide prognostically useful information.