What Are the Results of Resection of Localized Dedifferentiated Liposarcomas in the Extremities?

Abstract

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.

Introduction

Dedifferentiated liposarcoma (DDLPS) is a rare malignancy, accounting for approximately 2% to 4% of soft-tissue sarcomas [11, 14]. DDLPS is seen most frequently in the sixth and seventh decades of life, and male patients predominate. Approximately 90% of DDLPSs are de novo, while 10% occur as a recurrence of an atypical lipomatous tumor (ALT) [4, 23]. DDLPS originates in the retroperitoneum (64% to 75% of patients), extremities (13% to 29% of patients), trunk (7% of patients), and other sites [1, 4]. DDLPS is a biphasic neoplasm that transitions from an ALT or well-differentiated liposarcoma to a sarcoma of variable morphologic appearance and histologic grades [4, 22]. Two studies have shown that immunohistochemistry for p16 combined with cyclin-dependent kinase-4 (CDK4) or mouse double minute 2 (MDM2) increases the sensitivity and specificity of distinguishing DDLPS from other adipocytic neoplasms [9, 22]. Surgical resection is the definitive treatment for resectable DDLPS. Resection with R0 margin is achievable for DDLPS in the limbs, but for obvious anatomic reasons, resection is more challenging for retroperitoneal tumors [7, 13]. Retroperitoneal DDLPS is thus associated with a high proportion of recurrence [7, 13]. Chemotherapy is often used for patients with advanced disease, but only limited data are available regarding the role of chemotherapy in patients with localized DDLPS [7].

The oncologic outcomes for all DDLPS sites are generally poor, with local recurrence occurring in 20% to 52% of patients, distant metastases in 7% to 30% of patients, and disease-related mortality in 28% to 40% [12, 16, 24]. However, in most instances, these reports include patients with retroperitoneal tumors. To the best of our knowledge, only one report assessed the clinical outcomes of primary DDLPS in the extremities [10]. Moreover, no studies have performed an imaging analysis with positron emission tomography (PET)/CT or evaluated the risk factors for local recurrence and distant metastasis in patients with DDLPS. We, therefore, performed an imaging and clinicopathological analysis of primary DDLPS in the extremity.

Specifically, for patients with extremity DDLPS we asked: (1) How frequently do additional primary malignancies with DDLPS occur? (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.

Patients and Methods

To answer our questions, we retrospectively evaluated the medical records of patients with DDLPS in the extremities who underwent surgery in our institution between May 2003 and August 2017. The lower extremity was defined as the groin or gluteal region or distal to these regions. The upper extremity was defined as the shoulder or distal to this region. Inclusion criteria were a pathologically proven diagnosis of DDLPS, no evidence of metastatic disease at presentation, and no previous treatment for DDLPS.

Diagnosis of DDLPS

In all patients, the histologic diagnosis was established according to the WHO’s Classification of Tumours by expert pathologists [2]. Macroscopic appearance was a solid fleshy area of dedifferentiated component next to an ALT component with fatty appearance (Fig. 1A) Microscopically, the ALT component represented proliferation of fat cells including multilobular lipoblasts, and the dedifferentiated component resembled high-grade, undifferentiated pleomorphic sarcoma or myxofibrosarcoma (Fig. 1B). The grade was determined using the Fédération Nationale des Centres de Lutte le Cancer grading system, which is based on tumor differentiation (DDLPS is classified as Grade 3), mitosis count, and tumor necrosis [3].

Fig. 1.

A-D These macroscopic and microscopic images show the findings for DDLPS. (A) A macroscopic appearance was solid fleshy area of dedifferentiated component next to an atypical lipotamous tumor (ALT) component with fatty appearance. (B) A sharp transition is seen between two components: an atypical lipomatous zone and a high-grade non-lipogenic zone that appears as an undifferentiated pleomorphic sarcoma (stain, HE; original magnification, x100). (C) Diffuse expression of CDK4 is seen in both the nuclei of adipocytes and atypical spindle cells in the stroma in the atypical lipomatous zone and in the nuclei of atypical cells in the high-grade non-lipogenic zone (stain, CDK4; original magnification, x100). (D) Diffuse expression of p16 is seen in the nuclei of adipocytes and atypical spindle cells in the stroma in the atypical lipomatous zone, as well as in the nuclei of atypical cells in the high-grade non-lipogenic zone (stain, p16; original magnification, x100). A color image accompanies the online version of this article.

Study Population

During that time, 16 patients were treated for this diagnosis; one was excluded from this study because that patient did not have an MRI, leaving 15 patients (nine men, six women; median [range] age 67 years [42 to 87]) for evaluation. All patients had a minimum of 2 years of follow-up (median [range] 54 months [25 to 136]); 14 of 15 have been seen in the last 5 years (one, who was doing well at the time, was lost after 9 years of follow-up) (Table 1). Three patients were younger than 50 years, four were 50 to 64 years old, and eight patients were older than 64 years. All tumors were located at layers deeper than the fascia. The tumor site was the thigh (n = 12), buttock (n = 1), upper arm (n = 1), and forearm (n = 1). The median (range) follow-up period was 54 months (25 to 136).

Table 1.

Patient characteristics

Characteristics	Category	Number of patients
Gender	Male	9
	Female	6
Age	Median (range)	67 (42–87)
	< 50	3
	50-64	4
	≥ 65	8
Tumor size (cm)	Median (range)	13 (6–35)
	< 15	9
	≥ 15	6
Anatomical location
	Thigh	12
	Buttock	1
	Upper arm	1
	Forearm	1
Histological grade (FNCLCC)
	Grade 1	0
	Grade 2	9
	Grade 3	6

FNCLCC = Fédération Nationale des Centres de Lutte le Cancer.

Histopathology and Imaging

According to the Fédération Nationale des Centres de Lutte le Cancer grading system, the tumor was Grade 2 in nine patients (mitosis count 1 in nine patients; tumor necrosis 0 in five patients and 1 in four patients) and Grade 3 in six patients (mitosis count 2 in three patients and 3 in three patients; tumor necrosis 0 in one patient, 1 in four patients and 2 in one patient) (Table 2).

Table 2.

Fédération Nationale des Centres de Lutte le Cancer grading

Variable	Category	Histologic grade
		Grade 1; total score 2, 3	Grade 2; total score 4, 5	Grade 3; total score 6, 7, 8
Mitosis count	1: 0-9/10 HPF	0	9	0
	2: 10-19/10 HPF	0	0	3
	3: > 19 /10 HPF	0	0	3
Tumor necrosis	0: no necrosis	0	5	1
	1: < 50% tumor necrosis	0	4	4
	2: ≥ 50% tumor necrosis	0	0	1

HPF = high-power field (x400).

We performed immunohistochemical studies of p16 and CDK4 for all tumor specimens. Paraffin sections were cut to a thickness of 4 µm and mounted on glass slides. The primary antibodies used were p16 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and CDK4 (Abcam, Cambridge, MA, USA). The amount of immunoreactivity was semiquantitatively evaluated as negative (0% of cells stained), focally positive (1% to 10% of cells stained), multifocally positive (11% to 50% of cells stained), or diffusely positive (> 50% of cells stained).

In immunohistochemical studies, positive staining was noted for p16 and CDK4 in 87% (13 of 15) and 87% (13 of 15) of the tumors, respectively. The amount of immunoreactivity of p16 was diffusely positive in 13 of 15 patients. The amount of immunoreactivity of CDK4 was focally positive in five patients, multifocally positive in one patient, and diffusely positive in seven patients. The expression of CDK4 was seen in the nuclei of adipocytes and atypical spindle cells in the stroma in the ALT zone (Fig. 1C), as well as in the nuclei of atypical cells in the high-grade non-lipogenic zone. The expression of p16 was also seen in the nuclei of adipocytes and atypical spindle cells in the stroma in the ALT zone and in the nuclei of atypical cells in the high-grade nonlipogenic zone (Fig. 1D). In 11 of 15 patients, MRI demonstrated two components: an ALT component and a dedifferentiated component. The border between these two components was relatively distinct. The ALT component showed signal hyperintensity on T1-weighted (Fig. 2A-B) and T2-weighted sequences (Fig. 2C-D), consistent with a lipomatous tumor. The dedifferentiated component had low-to-intermediate signal intensity on T1-weighted sequences (Fig. 2A-B) and heterogeneous signal hyperintensity on T2-weighted sequences (Fig. 2C-D), resembling tumors such as high-grade undifferentiated pleomorphic sarcoma or myxofibrosarcoma. A gadolinium-enhanced T1-weighted MRI showed heterogeneous enhancement of the tumor after gadolinium injection and fat saturation (Fig. 2E-F). In patients with no biphasic pattern, we could not diagnose them as DDLPS by MRI. These tumors were low intensity on T1 and high intensity on T2 without other specific features. Biopsy were done in all patients. In one patient, adipocytic cells were detected, which was diagnosed as DDLPS. However, adipocytic cells were not detected in three other patients, which were diagnosed as undifferentiated sarcoma or myxofibrosarcoma. No recurrence and two metastases were detected among them.

Fig. 2.

A-H These images show MRI findings in a patient with DDLPS. MRI demonstrates two components: an ALT component and a dedifferentiated component. The border between these two components is relatively distinct. The ALT component shows signal hyperintensity on (A-B) T1-weighted and (C-D) T2-weighted sequences. The dedifferentiated component shows (A-B) low-to-intermediate signal intensity on T1-weighted sequences and (C-D) heterogeneous high-signal intensity on T2-weighted sequences. (E-F) A gadolinium-enhanced T1-weighted MRI shows heterogeneous enhancement of the tumor after gadolinium injection and fat saturation. PET/CT shows a biphasic pattern with a close relationship to MRI findings. (G-H) The dedifferentiated component shows high FDG uptake, while the ALT component displays almost no FDG uptake. Figs A, C, E, G are axial sections, Figs. B, D, and F are sagittal sections, and Fig. H is a coronal section. A color image accompanies the online version of this article

Now that it is covered by insurance, we routinely use PET/CT in patients with DDLPS. Other patients were investigated by CT. Nine patients were evaluated using PET/CT. Any area of increased FDG uptake was evaluated. Tumor metabolic activity was measured using the maximum standardized uptake value (SUV_max), calculated based on the following formula: tissue concentration (MBq/g)/injected dose (MBq)/body weight (g). PET/CT showed a biphasic pattern bearing a close relationship to MRI findings. The dedifferentiated component showed high FDG uptake (median SUV_max, 5.1; range, 1.9 to 22.6), while the ALT component showed almost no FDG uptake (Fig. 2G-H).

Treatment

All patients underwent limb-sparing surgery. The margins of surgical excision were estimated based on the American Joint Committee on Cancer residual tumor classification (R classification) [25]. These surgical margins classified an R0 margin as free of malignancy, an R1 margin as microscopic tumor cells present at the inked border of the specimen, and R2 as grossly positive margins. For patients who developed local recurrence, we resected as much as possible. However, for the small ALT lesion without a dedifferentiated component, we sometimes watched their course and resected it when it became larger because gross detection was sometimes difficult. The surgical margins were estimated as R0 in 10 patients, R1 in four patients, and R2 in one patient. For the dedifferentiated component, R0 margins were achieved in all patients. For the ALT component, R1/R2 margins were achieved in all but five patients because the tumor was close to major nerves or vessels

Radiotherapy was not administered to any patient preoperatively. Typically, we use conventional external beam radiation therapy after surgery when the tumor is close to major nerves or vessels and marginal resection is achieved. Postoperative radiotherapy was administered to five patients to prevent recurrence. The radiation dose was 45-60 Gy with boost radiotherapy. No patients received preoperative or postoperative chemotherapy.

Among the 10 patients with R0 margins, two had local recurrence (Table 3). The local recurrence was identified 9 months after surgery as a tumor with a dedifferentiated component and was again resected with an R0 margin. The patient showed no further recurrence as of the last follow-up examination. Among the five patients with R1/R2 margins, one had local recurrence. None of the four patients who received radiotherapy experienced local recurrence, and the one patient who did not receive radiotherapy had local recurrence. One patient who achieved R1 margins of the ALT component close to major vessels had local recurrence of the ALT component 48 months after surgery. Although re-resection was performed, re-recurrence of a small ALT component was found 3 years later. The patient was too old (90 years) to resect it again. One patient with R0 margin of the ALT component had local recurrence of a small ALT component 29 months after surgery. The patient did not wish to resect it again, so careful follow-up was undertaken. Both patients remained asymptomatic and the tumors grew very slowly without the development of a dedifferentiated component.

Table 3.

Recurrence rate associated with surgical margin and radiotherapy

Surgical margin	Radiotherapy (+) (n = 5)	Radiotherapy (−) (n = 10)
R0 (n = 10)	0 of 1	2 of 9
R1/R2 (n = 5)	0 of 4	1 of 1

Assessment of Study Outcomes

First, we analyzed the presence of additional primary malignancies of DDLPS. CT images of the chest and abdomen were examined for all patients to determine if there was a distant metastasis at the initial presentation or follow-up. We examined CT images of the chest and abdomen and the treated limb for all patients with high-grade sarcomas after surgery looking for a distant metastasis every 3 months to 4 months in the first 1 year to 3 years, then twice a year up to the fifth year and once a year thereafter. MRI was used to evaluate signal intensity and the size and depth of tumors [26]. Tumor size was defined as the maximum diameter measured on MRI. The median (range) maximum tumor diameter was 13 cm (6 to 35).

We also analyzed local recurrence-free survival (LRFS), metastasis-free survival (MFS), and overall survival (OS). LRFS was calculated from the date of surgery to the date of local recurrence or to the date of the last follow-up visit. Those who were dead without recurrence or metastases were censored on the date of their death. MFS was calculated from the date of diagnosis to the date of diagnosis of metastases or the last follow-up visit. OS was calculated from the date of diagnosis to the date of death because of DDLPS or the last follow-up visit. Survival rates were estimated using the Kaplan-Meier method.

The Wilcoxon exact test was used to determine the association of the following variables in terms of survival: age, sex, maximum tumor size, radiotherapy, and surgical margins.

For all analyses, associations were considered significant for a p value < 0.05. Bell Curve for Excel (Social Survey Research Information Co, Ltd, Tokyo, Japan) was used for the analyses.

Results

How Frequently Do Additional Primary Malignancies in Patients with DDLPS Occur?

Seven additional primary malignancies developed in 5 of 15 patients, comprising one additional primary malignancy in four patients and two additional primary malignancies in one patient. These primary malignancies were synchronous with DDLPS in one patient, synchronous or metachronous in one patient, and metachronous in three patients. The patient with a synchronous malignancy was a 71-year-old man with DDLPS in whom PET/CT confirmed pharyngeal accumulation. Biopsy revealed it was a non-Hodgkin lymphoma. The patient with a synchronous or metachronous malignancy was a 67-year-old woman with uterine cancer in whom DDLPS in the thigh was incidentally identified on preoperative CT. Renal cell cancer and lung cancer were identified 0.5 and 2 years, respectively, after the diagnosis of DDLPS. Among the patients with metachronous malignancies, two had a sarcoma (osteosarcoma and solitary fibrous tumor) 10 and 12 years before the diagnosis of DDLPS, and one patient had lung cancer 8 years after the diagnosis of DDLPS.

What Is the Rate of Overall Survival, Metastases, and Local Recurrence in DDLPS?

The 3- and 5-year MFS rates were 86% (95% CI 0.67 to 1.00) and 75% (95% CI 0.49 to 1.00), respectively (Fig. 3). One patient had a single lung metastasis 29 months after surgery. That patient decided against further surgical treatment and died 27 months after the lung metastasis was identified. One patient had metastases in a rib and a single lung at 50 months and 68 months after surgery, respectively; these lesions were resected and the patient remained disease-free as of the last follow-up examination. One patient had intramuscular metastases in the thigh, which developed in another muscle than the original lesion 22 months and 32 months after surgery; these lesions were resected and the patient remained disease-free as of the last follow-up examination. Among all 15 patients, one died of the disease and one had died of lung cancer by the last follow-up examination. The 3- and 5-year OS rates were 100% (95% CI 1.00 to 1.00), and 88% (95% CI 0.65 to 1.00), respectively (Fig. 4). Among the 15 patients, three had local recurrence, and the median (range) time to recurrence was 29 months (9 to 48). The 3- and 5-year LRFS rates were 86% (95% CI 0.67 to 1.00), and 75% (95% CI 0.49 to 1.00), respectively (Fig. 5).

Fig. 3.

This Kaplan-Meier curve shows metastasis-free survival (MFS). The 3- and 5-year MFS rates are 86% and 75%, respectively.

Fig. 4.

This Kaplan-Meier curve shows overall survival (OS). The 3- and 5-year OS rates are 100% and 88%, respectively.

Fig. 5.

This Kaplan-Meier curve shows local recurrence-free survival (LRFS). The 3- and 5-year LRFS rates are 86% and 75%, respectively.

Are There Factors That Are Associated with Metastasis-free Survival and Local Recurrence in DDLPS?

No factors were associated with MFS (Table 4). Three of our 15 patients had distant metastases, but we could not identify specific factors that might predict distant metastases because of the small number of patients. Tumor size greater than 15 cm was the only factor influencing LRFS (p = 0.04) (Table 5). Among the six patients with tumor size greater than or equal to 15 cm, three experienced local recurrence, and the 3- and 5-year LRFS rates were 63% (95% CI 0.21 to 1.00) and 42% (95% CI 0.00 to 0.85), respectively (Fig. 6). On the other hand, none of the nine patients with tumors smaller than 15 cm showed local recurrence. With the numbers we had, we could not show a relationship between recurrence and the tumor’s margin. The proportion of patients with R0 and R1/R2 margins who had recurrence was 20% and 20%, respectively (p = 1.00). With the numbers we had, we could not show a difference in local recurrence in patients who received surgery plus radiotherapy and those who were treated with resection alone. Patients treated with radiotherapy experienced recurrences, and three of 10 patients treated with surgery alone showed recurrence (p = 0.51).

Table 4.

Risk factors of distant metastases

Variable	Number of patients with distant metastases	Number of patients without distant metastases	p value
Age, years
< 65	1	6
≥ 65	2	6	1.00
Gender
Male	2	7
Female	1	5	1.00
Tumor size
< 15 cm	2	7
≥ 15 cm	1	5	1.00
FNCLCC
Grade 2	2	7
Grade 3	1	5	1.00
Margin
R0	3	7
R1, 2	0	5	0.51
Radiotherapy
Yes	0	5
No	3	7	0.51

FNCLCC = Fédération Nationale des Centres de Lutte le Cancer.

Table 5.

Risk factors of local recurrence

Variable	Number of patients with local recurrence	Number of patients without local recurrence	p value
Age, years
< 65	2	5
≥ 65	1	7	0.56
Gender
Male	1	8
Female	2	4	0.52
Tumor size
< 15 cm	0	9
≥ 15 cm	3	3	0.04
FNCLCC
Grade 2	3	6
Grade 3	0	6	0.22
Margin
R0	2	8
R1, 2	1	4	1.00
Radiotherapy
Yes	0	5
No	3	7	0.51

FNCLCC = Fédération Nationale des Centres de Lutte le Cancer.

Fig. 6.

This figure shows local recurrence-free survival (LRFS) in patients with tumors with a maximum diameter of least 15 cm and those with a tumor less than 15 cm in diameter. The 5-year LRFS rates are 42% and 100% in patients with tumors at least 15 cm and less than 15 cm, respectively.

Discussion

DDLPS is a rare sarcoma that occurs more frequently in the retroperitoneum, more so than in the extremity. Local recurrence has been reported to occur in 20% to 52% of patients, distant metastasis has been reported to occur in 7% to 30%, and the disease-related mortality rate has been reported to be 28% to 40% [12, 16, 24]. However, these reports primarily included patients with retroperitoneal tumors. Very few studies have reported on patients with DDLPS in the extremities only. To the best of our knowledge, only one report has assessed clinical outcomes specifically in patients with primary DDLPS in the extremities [10]. Moreover, imaging with MRI and PET/CT and the potential risk factors for local recurrence and distant metastasis have not been widely reported for patients with DDLPS in the extremities. We found that DDLPS had a high proportion of developing other malignancies, and 5-year LRFS, MFS, and OS rates were 75%, 75%, and 88%, respectively. We found no factors associated with an increased risk of distant metastases, and we found only large tumor size was associated with an increased risk of local recurrence, although we were underpowered on this important endpoint for this rare malignancy.

This study had a number of limitations. First, because DDLPS is such a rare malignancy, the sample size was small, at only 15 patients, which limited our ability to identify factors associated with local recurrence. However, we still caution the reader that the fact that some factors we explored were not associated with poorer survival, it is possible that in larger series those factors might be discovered to be important. With only 15 patients, our analysis was grossly underpowered on this important endpoint. Related to this is that there is a high likelihood of confounding; for example, large tumors often are deep tumors or high-grade tumors; with only 15 patients, we could not perform a multivariable analysis to control for the effect of confounding. Another limitation was the fact that wide resection was not performed in all patients because some tumors close to major nerves and vessels received R1/R2 resection; this could well have influenced survival, but our sample size was likely underpowered here, as well. In addition, there may have been selection bias in our application of radiation therapy; we used radiation therapy in patients with surgical margins of R1/R2 when the tumors were close to major nerves or vessels. However, radiation therapy can influence the outcome of local recurrence. Moreover, there are selection bias in which we excluded patients without MRI, though there was only one patient.

We found that on MRI, many of these tumors showed two components: an ALT component and a dedifferentiated component. PET/CT showed a biphasic pattern, similar to MRI findings. Dedifferentiated component showed high FDG uptake (median SUV_max 5.1; range 1.9-22.6). A few reports have focused on the radiologic findings of PET/CET in patients with DDLPS [5, 6, 21]. According to these reports, the dedifferentiated component displayed high FDG uptake (median SUV_max 3.5-5). In this study, positive staining rates for p16 and CDK4 were 87% and 87%, respectively. Recent studies have advised using p16 in combination with other immunohistochemical stains such as CDK4 and MDM2 to increase the sensitivity and specificity of the immunohistochemical panel to distinguish DDLPS from other sarcomas [9, 22]. In two reports, positive staining rates for the combination of p16/CDK4, p16/MDM2, and CDK4/MDM2 were 82% to 83%, 79% to 94%, and 79% to 83%, respectively, in DDLPS [9, 22]. We currently routinely investigate the combination of p16 and CDK4 immunostaining in our institution. DDLPS harbors high-level amplifications of chromosome 12q14-15, which includes the genes for MDM2, CDK4, high-mobility group AT-hook 2, and carboxypeptidase M [19, 20, 23]. MDM2 binds to and negatively regulates p53, which results in inhibited cell cycle arrest and apoptosis, leading to tumor proliferation [21, 23]. CDK4 inhibits p53 activity, leading to increased transcriptional activity of multiple genes, including MDM2. Overexpression of these genes can lead to the development of other malignancies besides DDLPS [19, 20].

In this study, multiple primary malignancies were seen in five of 15 patients, with three primary malignancies occurring in one patient. Jung et al. [8] reported that 8.3% of patients with DDLPS showed a second primary malignancy within 2 years after the diagnosis of DDLPS, and patients younger than 50 years at the time DDLPS was diagnosed had a higher ratio of second primary malignancies than older patients. Although no known genetic cancer syndromes are associated with 12q14-15 amplification, liposarcoma has been infrequently reported in select patients with Li-Fraumeni syndrome and neurofibromatosis [8]. In this study, no familiar syndrome could be identified. Further validation and investigation is needed because this finding may have implications such as closer screening for patients with DDLPS.

Five-year OS and MFS were 88% and 75%, respectively in the present study. To the best of our knowledge, only two reports have assessed OS in patients with DDLPS in the extremities [10, 15] (Table 6). Kito et al. [10] reported a 5-year OS rate of 60% in patients with DDLPS in the extremities. Okada et al. [15] reported a 5-year OS rate of 61% and described that large dedifferentiated area (greater than 8 cm), high MIB-1 labeling index (greater than 30%) of the dedifferentiated area, and lung metastasis at the initial presentation were associated with a poor prognosis.

Table 6.

Comparison of DDLPS treatment results in the extremities

	Number of patients	Local recurrence	Distant metastases	5-year OS rate (%)	Risk factors
Okada et al. [15]	7	1	4	61	Overall survival Large dedifferentiated area (> 8 cm) High MIB-1 labeling index Lung metastasis at the initial presentation
Kito et al. [10]	18	0	2	60
This study	15	3	3	88	Local recurrence Large (> 15 cm) tumors

With the small number of patients in this series, we could not show an association between local recurrence and the tumor’s margin or the use of radiotherapy, although we found that tumors larger than 15 cm appeared to recur locally more often than smaller tumors. Because other factors may influence this observation, larger studies will be needed to confirm a possible association. Although wide excision may be associated with good local control of DDLPS [4], no relationship between local recurrence and surgical margins was identified in this study. Kito et al. [10] reported that although there was no recurrence in patients with marginal margins of the ALT component, wide margins should be the goal because the recurrence rate would presumably increase during the follow-up period. Their theory was based on their previous report of a 23% local recurrence rate with a median time to recurrence of 7.2 years after marginal resection of ALT. The role of radiotherapy in the treatment of extremity soft-tissue sarcomas has been defined by two randomized clinical trials that showed radiotherapy was an option to treat high-risk patients with high-grade tumors that had positive surgical margins, minimizing the risk of local recurrence and preserving limb function [18, 27]. Both reported a reduced risk of local recurrence with brachytherapy or external-beam radiotherapy. In several studies, preoperative radiotherapy combined with complete R0/R1 resection for unifocal, primary DDLPS was associated with improved local recurrence in patients with retroperitoneal DDLPS [1, 4, 17]. However, previous studies have shown no relationship between radiotherapy and local recurrence in patients with DDLPS of the extremities [10, 15]. None of the five patients who received postoperative radiotherapy because marginal resection was performed for an ALT component close to major vessels or nerves showed any recurrence. The utility of radiotherapy for local control of primary DDLPS of the extremities should thus be investigated further. We now have a treatment strategy for DDLPS of the extremities aiming at R0 resection of the whole tumor and we use radiotherapy when the resection proves R1/R2 in tumors close to major nerves and vessels.

In conclusion, the patients with DDLPS in our study had 5-year LRFS, MFS, and OS rates of 75%, 75%, and 88%, respectively. 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 either before, concurrent with, or after this sarcoma is detected. We believe patients should be assessed and monitored carefully for this possibility. 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. Future, larger studies might also reveal the role of radiotherapy and surgical margins for this rare malignancy.