Use of Mitotic Activity and the Size of Any Dedifferentiated Component for Risk Assessment in MDM2-Amplified Liposarcoma.

Context.—: The characteristic molecular signature for both atypical lipomatous tumor/well-differentiated liposarcoma and dedifferentiated liposarcoma is amplified sequences derived from chromosome 12q13-15, including MDM2 proto-oncogene (MDM2). As the progression of atypical lipomatous tumor/well-differentiated liposarcoma to the more aggressive dedifferentiated liposarcoma has the potential to adversely affect patient outcomes, the extent of the latter component might be important to evaluate.

Objective.—: To investigate the correlation between clinicopathologic characteristics, including tumor size, modified Fédération Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grade, molecular data, and outcomes in 123 surgically resected MDM2-amplified liposarcomas.

Design.—: Pathology reports and clinical records were reviewed. A log-rank test was used to compare the survival trends, and univariate logistic regression was performed to identify variables associated with adverse events (distant metastasis and/or death), from which the P value was derived to construct a multivariate regression model.

Results.—: In univariate analysis, the largest single dimension of the dedifferentiated component, the percentage of cells with gain of chromosome 12, mitotic count, and the presence of modified FNCLLC grade 3 were associated with adverse events. In multivariate analysis, the largest single dimension of the dedifferentiated component (odds ratio: 1.169; 95% CI: 1.053, 1.299; P = .003), and a higher mitotic count (odds ratio: 1.133; 95% CI: 1.037, 1.237; P = .006) were correlated with adverse events. There was no statistically significant association between current local recurrence status, overall largest tumor dimension, overall tumor volume, MDM2 copy number, or MDM2 to chromosome 12 centromere probe ratio and adverse outcomes.

Conclusions.—: Staging dedifferentiated liposarcoma based on the size of the dedifferentiated component better predicts the outcome.