![]() | Only 14 pages are availabe for public view |
Abstract Lipomatous neoplasms are among the most frequently encountered soft tissue neoplasms. Liposarcoma is the most common soft tissue sarcoma. It includes four subtypes: Atypical Lipomatous Tumor/ Well Differentiated Liposarcoma (ALT/WDLPS), Dedifferentiated Liposarcoma (DDLPS), Myxoid Liposarcoma (MLPS) and Pleomorphic Liposarcoma (PLPS). ALT/WDLPS and DDLPS are characterized by MDM2 gene amplification which can be detected by immunohistochemical staining. MLPS shows a distinctive gene fusion between DDIT3 and either FUS or EWSR1. Recently, cancer testis antigens -mainly CTAG1B- were found to be expressed in MLPS. Therefore, MDM2 and CTAG1B immunostaining and DDIT3 and EWSR1 gene rearrangements were proposed as diagnostic markers for distinguishing the different types of liposarcoma from their histological mimics. The present study aimed at assessing the utility of these markers by immunohistochemistry or FISH in the diagnosis of different types of liposarcoma. To achieve, this aim 60 cases including different types of liposarcoma and their mimics were chosen. Cases showing adipocytic component (lipoma, ALT/WDLPS, DDLPS or MLPS with lipomatous pattern) as well as spindle cell/pleomorphic sarcoma were stained for MDM2. Cases showing myxoid areas with or without adipocytic areas were stained for CTAG1B and were examined for DDIT3 gene rearrangement by FISH. Only selected cases were examined for EWSR1 gene rearrangement by FISH. To assess the diagnostic role of MDM2 in distinguishing ALT/WDLPS and DDLPS from their mimics, different parameters of MDM2 immunostaining (intensity of staining and percentage of positively stained nuclei) were assessed separately. In addition, the percentage of positively stained nuclei was assessed using three different methods; eyeballing the slides at medium power, examining hotspots at 10 different high-power fields semi quantitatively and obtaining photomicrographs of these hotspots to be subjected to image analysis. Finally, combined scores including intensity of staining and percentage of positively stained nuclei were calculated using each method. Murine Double Minute 2 (MDM2) immunostaining significantly distinguished ALT/WDLPS from lipoma by the staining intensity, percentage of positively stained nuclei as assessed by eyeballing and hotspot semi-quantitative methods with different cut-off values (11.5% and 21.5% respectively) and the combined score of all methods. The highest sensitivity and specificity (100% and 87.5% respectively) were obtained by the strong staining intensity and combined score 6 using semi-quantitative and image analysis methods. Murine Double Minute 2 (MDM2) immunostaining also significantly distinguished DDLPS from its mimics using the staining intensity, percentage of positively stained nuclei (using all methods) and the combined score (using all methods). However, 100% sensitivity and 100% specificity were obtained by the percentage of positively stained nuclei using the eyeballing and interactive image analysis methods (at cut -off values of 20% and 19.77% respectively) and combined score using either eyeballing or interactive image analysis methods with different cut-off values (score 4 and score 5 respectively). Summary 140 In an attempt to use MDM2 in distinguishing between the different types of liposarcoma, the different parameters of MDM2 immunostaining showed a significant difference among ALT/WDLPS, DDLPS and MLPS with adipocytic pattern. However, the combined score as assessed by the interactive image analysis method showed 100% sensitivity and 100% specificity in excluding a diagnosis of MLPS with adipocytic pattern at a cut-off value of 5. However, most of the parameters of MDM2 immunostaining did not show any significant difference between ALT/WDLPS and DDLPS. However, only the combined score when assessed by the interactive counting by image analysis method showed a significant difference at value 6 but with a low specificity (54.55%). To assess the diagnostic role of CTAG1B in distinguishing MLPS from its mimics, the percentage of cells showing positive staining as well as the intensity of staining were both evaluated, and a combined score was obtained for each case. Moreover, the cellular localization of the staining (nuclear versus cytoplasmic) was recorded. CTAG1B immunostaining showed significantly higher intensity, higher percentage of nuclear and cytoplasmic staining, higher overall density irrespective of cellular localization and higher combined scores in MLPS than its mimics. Although nuclear staining was more frequently observed than cytoplasmic staining in MLPS (85.7% and 64.28% respectively), the highest diagnostic accuracy was obtained by percentage of cytoplasmic staining exceeding 1.5% (89.6%) with a sensitivity of 85.7% and a specificity of 93.3%. However, although the other parameters were more sensitive than the percentage of cytoplasmic staining (92.86%), they were all less specific (80%) with less diagnostic accuracy (86.2%). Using FISH to assess the diagnostic role of DDIT3 gene rearrangement to distinguish MLPS from its mimics, DDIT3 gene rearrangement was detected in all cases of MLPS with a mean percentage of nuclei showing split signals of 57.93%. Only one case of DDLPS with myxoid change showed split signals in 50% of the examined nuclei. Therefore, DDIT3 rearrangement could significantly differentiate between MLPS and its mimics with a sensitivity of 100% and specificity of 92.86%. EWSR1 gene rearrangement was not detected in any of the tested five cases of MLPS but was detected in a case of extraskeletal myxoid chondrosarcoma. This difference was not statistically significant in distinguishing MLPS from its mimics. Comparing the results of CTAG1B immunostaining and FISH for DDIT3 gene rearrangement, the present study showed that both are diagnostically useful in distinguishing MLPS from its mimics, with DDIT3 rearrangement being more sensitive than both CTAG1B cytoplasmic staining and combined score (100%, 85.71% and 92.86% respectively) but less specific than cytoplasmic staining (92.86% and 93.3% respectively). However, this difference in sensitivity and specificity between DDIT3 rearrangement and cytoplasmic staining of CTAG1B was not significant. Finally, the present study showed that although using both CTAG1B and DDIT3 gene testing by FISH was more specific, yet DDIT3 gene rearrangement alone remains the most sensitive test for distinguishing MLPS from its mimics. |