Immunohistochemical evaluation of CREM in CREB-rearranged mesenchymal tumors and their mimics

Fusion genes between the FET (EWSR1/FUS) and CREB (CREB1, ATF1, CREM) families characterize many tumor types, including mesenchymal entities. Herein, we tested the diagnostic utility of CREM (C-terminus) immunohistochemistry using 51 CREB-rearranged mesenchymal tumors and 159 tumors of 14 mimicking entities. Staining was considered positive if nuclear staining of moderate or strong intensity was observed in at least 10% of the tumor cells. Among the 51 CREB-rearranged tumors, CREM was at least focally positive in 39 tumors (76.5%), diffusely (≥ 50%) positive in 31 tumors (60.8%), and diffuse strong staining was observed in 23 tumors (45.1%). Diffuse strong reactivity was observed in nearly all angiomatoid fibrous histiocytoma (11 of 12 tumors), intracranial FET::CREB mesenchymal neoplasms (2 of 2 tumors), and unclassifiable sarcomas (2 of 2 tumors, including 1 case with SMARCA2::CREM). However, the positivity was more limited in clear cell sarcoma (15 of 20 tumors), keratin-positive malignant FET::CREB tumors in the abdomen (5 of 8 tumors), and gastrointestinal neuroectodermal tumors (4 of 7 tumors). Two separately evaluated CRTC1-rearranged tumors (CRTC1::TRIM11 and CRTC1::SS18) demonstrated diffuse strong positivity. Among the 159 tumors in the comparison cohort, CREM was at least focally positive in 33 (20.8%) tumors, diffusely positive in 19 tumors (11.9%), and diffusely and strongly positive in 5 tumors (3.1%). CREM demonstrated moderate sensitivity and specificity for the diagnosis of CREB-rearranged mesenchymal tumors as a whole cohort, and its overall utility in predicting CREB fusion is limited. However, CREM staining may be useful in a few specific contexts, such as when angiomatoid fibrous histiocytoma is suspected.