UCSF Current Issues in Anatomic Pathology Pleomorphic sarcomas: MFH - PDF document

UCSF Current Issues in Anatomic Pathology Pleomorphic sarcomas: MFH where did you go? 2018 Andrew Horvai MD PhD Clinical Professor, Pathology UCSF Medical Center Mission Bay 1825 4th Street, Room M2354 San Francisco, CA 94158 1

INTRODUCTION The diagnosis of malignant fibrous histiocytoma ( MFH ) was historically accepted as a defined clinicopathologic entity but has largely been eliminated from the literature. This presentation aims to (1) describe the origin of MFH as a pathologic diagnosis (2) provide an overview of how MFH has been gradually reclassified into more specific diagnoses and, most importantly, (3) provide a practical, clinically meaningful, approach to classifying pleomorphic sarcomas The tables and figures corresponding to this text is available in the corresponding PowerPoint handout. THE EMERGENCE OF MFH In the 1960s, Stout and others published several reports of neoplasms referred to as “fibrous xanthoma” and “histiocytoma.” 1,2 In retrospect, these were likely a heterogeneous mix of benign and malignant tumors of various lineages. The first appearance of MFH in the literature was introduced by Ozzello, Stout and Murray based on culture results. 2 The authors proposed that cultured cells from MFH tumors initially demonstrated ameboid movement, suggesting histiocyte lineage, but later formed an adherent layer of bipolar spindled cells, suggesting fibroblast lineage. They concluded that the cells were “facultative fibroblasts” which were cells of macrophage (histiocyte) derivation but which were capable of fibroblast differentiation. The inclusion criteria for MFH grew over the ensuing decades. In 1972, Kempson and Kyriakos included fibroxanthoma in the family of MFH giving rise to the connection between storiform growth and MFH. 3 By the 1980s, MFH became the most common sarcoma in adults and was defined by the presence of a storiform or fascicular pattern, striking nuclear atypia and pleomorphism, necrosis, high mitotic activity including atypical forms, occasional bizarre multinucleated cells and genomic instability. 4 Attempts to sort out the biology of MFH expanded to ultrastructure and immunohistochemistry in the 1970s and 1980s. Theories were proposed for the histotype and histogenesis of MFH. The tumors were purported to contain facultative fibroblasts, as described above, or a biphenotypic malignancy containing both fibroblasts and histiocytes or malignant transformation of a stem cell with dual differentiation into both fibroblasts and histiocytes. 5 The term histiocyte is vague, although probably inescapable given how deeply ingrained it has become in the pathology literature. The current understanding of the monocyte- macrophage lineage differentiation recognizes a variety of cell types with phagocytic and/or antigen presenting properties derived from bone marrow (dendritic cells, 2

macrophages, osteoclasts) and non-marrow derived antigen presenting cells (follicular “dendritic” cells, respectively). 6 Figure 1 demonstrates a summary of the cell types as currently understood. Despite being a more detailed version than shown in the lecture, Figure 1 still overemphasizes the stringency of the classification. Experimental evidence has shown remarkable plasticity of these lineages in vivo . 7 The discussion of all of these pathways is beyond the scope of the presentation, but it is relatively clear that pleomorphic sarcomas virtually never arise from hematopoietic precursors nor express markers of follicular “dendritic” cells. Rather, they are probably derived from pluripotent mesenchymal stem cells with limited differentiation toward recognizable lineages. Figure 1. Summary of cell types recognized as “histiocytes” by pathologists. This is a more detailed version of the figure presented in lecture but still overemphasizes the distinction between lineages. Greater overlap between subtypes likely exists in tissues. 3

Reappraisal of MFH As it became increasingly clear that the “facultative fibroblast” was not a justifiable cell lineage, and immunohistochemistry improved, the diagnosis of MFH grew problematic. Two seminal articles by Fletcher carefully re-evaluated large series of MFH and demonstrated that the majority could be re-classified using careful review of clinical, histologic and immunophenotypic findings. 8, 9 These studies demonstrated not only that the majority of MFH were sarcomas with reproducible differentiation (smooth, skeletal muscle, fat, bone) or distinct clinicopathologic entities (myxofibrosarcoma) but also found clinical relevance in subclassifying MFH. 9 Though Fletcher’s contribution ultimately reclassified most MFH into more specific categories, approximately a third of MFH remained in the “undifferentiated” category. Over the ensuing decades, better characterization of rare sarcomas histologically, immunophenotypically and increasingly using genetic methods, has whittled the “undifferentiated” category further. An informal retrospective review of pleomorphic sarcomas diagnosed at UCSF from 2003 to 2017 demonstrated that only 20% could not be further classified. This latter category in current WHO designation is referred to as undifferentiated pleomorphic sarcoma ( UPS ). 10 UPS, as currently defined, is a diagnosis of exclusion: a malignant mesenchymal tumor usually with high-grade cytomorphology but that does not fit into any other clinicopathologic entity. UPS, therefore, is the standard to which more specific diagnoses are compared diagnostically and clinically. Clinically meaningful classification of MFH The number of diagnostic entities in the pleomorphic sarcoma literature is ever increasing, especially with application of modern molecular methods. However, a practical approach seeks to limit the list only to entities with therapeutic or prognostic implications. To date, targeted therapies for most pleomorphic adult sarcomas remains limited. 11 However, there are prognostic implications to some diagnoses (Table 1). An important consideration when evaluating a pleomorphic malignancy is to consider whether the tumor could represent something other than a sarcoma. Pleomorphic carcinomas and melanomas can mimic sarcomas and these diagnoses can have profound impact on therapy and prognosis. Clues to suggest carcinoma include superficial location, in-situ component, cohesive cells, keratin and especially p63 expression. 12 Melanomas can occasionally show striking pleomorphism, although in my experience this is more common in late recurrences and metastases. Cell nesting, the presence of pigment, immunophenotype (S100, SOX10 etc) and BRAF V600E mutation are clues to melanoma. It should be remembered that some metastatic melanomas can lose melanocytic markers requiring molecular methods for diagnosis. 13 4

Germ cell tumors and hematolymphoid tumors are theoretically in the differential diagnosis. Preliminary data of a tissue microarray study underway at UCSF suggests that UPS virtually never represents a misdiagnosed hematopoietic tumor. Table 1. Summary of prognosis of pleomorphic sarcomas Diagnosis Recurrence Metastasis Mortality Dedifferentiated liposarcoma 51% 15% 26% @ 5 yrs. Myogenic sarcomas ?% 56% 39% @ 20 yrs. Myxofibrosarcoma (grade 1‐3) 31% 15% 23% @ 3 yrs. Atypical fibroxanthoma 8% 0‐3% 0‐2% @ 20 yrs. Pleomorphic dermal sarcoma 28% 5‐10% 0‐2% @ 2 yrs. Pleomorphic hyalinizing 33% 0% 0% @ 4 yrs. angiectatic tumor Myxoinflammatory fibroblastic 51% 2% 0% @ 5 yrs. sarcoma Undifferentiated pleomorphic 13‐42% 31‐35% 37% @ 5 yrs. sarcoma PLEOMORPHIC SARCOMAS WITH DEFINED HISTOTYPE Assuming non-mesenchymal tumors can be excluded, the remaining differential includes a variety of pleomorphic sarcomas. Of those with a defined histotype, only a subset of these has been evaluated for prognostic significance. Dedifferentiated liposarcoma is defined as a biphasic tumor with both lipogenic and non-lipogenic components either synchronous or metachronous. The non-lipogenic component is often highly pleomorphic with bizarre cells and atypical mitoses. Genetic or immunohistochemical (MDM2, CDK4) evidence of chromosome region 12q13-15 is diagnostic. 14 Clinical clues include deep (retroperitoneal, mediastinal, scrotal) location 5