BACKGROUND: Dedifferentiated liposarcoma (DDLPS) is a rare malignancy that transitions from an atypical lipomatous tumor to a sarcoma with a variable morphologic appearance. The behavior of this tumor in the retroperitoneum is aggressive, but the behavior of DDLPS in the extremities is less well-defined because it is rare. Few reports have assessed the imaging features and clinical outcomes of primary DDLPS in the extremities. QUESTIONS/PURPOSES: In patients with primary DDLPS of the extremity, we asked the following questions: (1) How frequently do additional primary malignancies occur in patients with DDLPS? (2) What is the rate of overall survival, metastases, and local recurrence in DDLPS? (3) What factors are associated with metastasis-free survival and local recurrence in DDLPS? METHODS: We defined DDLPS as a biphasic neoplasm that transitions from an atypical lipomatous tumor (ALT) to a sarcoma of variable morphologic appearance and histologic grades. We retrospectively evaluated the medical records of patients with DDLPS of the extremities who underwent surgery in our institution between 2003 and 2017. During that time, 16 patients were treated for this diagnosis; one was excluded from this study because the patient did not have an MRI, leaving 15 patients (nine men, six women; their median [range] age was 67 years [42 to 87]) for evaluation. All had a minimum of 2 years follow-up (median [range] 54 months [25 to 136]); 14 of 15 have been seen in the last 5 years (one patient, who was doing well at the time, was lost after 9 years of follow-up). In 11 patients, MRI demonstrated two components: an ALT component with high intensity on both T1-weighed and T2-weighted sequences and a dedifferentiated component low-to-intermediate intensity on T1-weighed and heterogeneous hyperintensity on T2-weighted sequence. Nine patients were evaluated using 2-deoxy-2-18F-fluoro-D-glucose positron emission tomography (FDG-PET) combined with CT (PET/CT). PET/CT showed a biphasic pattern with a close relationship to MRI findings. The dedifferentiated component presented with high FDG uptake (median [range] maximum standardized uptake value 5.1 [1.9 to 22.6]), while the atypical lipomatous tumor component showed almost no FDG uptake. In all patients, immunohistochemical studies of p16 and cyclin-dependent kinase-4 (CDK4) were investigated. Positive staining for both p16 and CDK4 were seen in 13 of 15 patients.We retrospectively evaluated the electronic medical records of all patients in our institution for the presence of additional primary malignancies, local recurrence-free survival, metastasis-free survival, and overall survival. The survival rate was estimated using the Kaplan-Meier method. The Wilcoxon exact test was used to determine the prognostic importance of the following survival variables: age, sex, maximum tumor size, radiotherapy, and surgical margin. RESULTS: Seven additional primary malignancies developed in five of 15 patients (two lung cancers, two sarcomas, one renal cell cancer, one uterine cancer, and one non-Hodgkin lymphoma). The 3- and 5-year metastasis-free survival rates were 86% (95% CI 0.67 to 1.00) and 75% (95% CI 0.49 to 1.00), respectively. With the numbers available, we found no factors associated with metastasis-free survival. The 3- and 5-year overall survival rates were 100% (95% CI 1.00 to 1.00) and 88% (95% CI 0.65 to 1.00), respectively. Three of 15 patients had local recurrence. The 3- and 5-year local recurrence-free survival rates were 86% (95% CI 0.67 to 1.00) and 75% (95% CI 0.49 to 1.00), respectively. Large (> 15 cm) tumors were more likely to have a local recurrence (p = 0.04). CONCLUSIONS: In this small series, we found that the extremities are a favorable site for DDLPS compared with the retroperitoneum, although we did not directly compare the two sites. This rare tumor has a relatively high likelihood of being associated with other malignancies. We believe patients should be assessed and monitored carefully for this possibility. In the future, larger studies are needed to better define predictors of local recurrence, although the tumor's size may be associated with a greater propensity for local recurrence. LEVEL OF EVIDENCE: Level II, prognostic study.