Complications of combined modality treatment of primary lower extremity soft-tissue sarcomas




Correlations between various patient, tumor, and treatment characteristics and complications in patients undergoing combined modality treatment for primary lower extremity soft-tissue sarcomas were investigated.


Using the M. D. Anderson Radiation Oncology database, the records of the subset of patients treated with combined radiation and limb-sparing surgery for primary lower extremity soft-tissue sarcomas were retrospectively reviewed from the years 1960 to 2003.


In all, 412 patients were identified. With a median follow-up of 9.3 years, there were a total of 113 (27%) acute wound complications and 41 (13% at 20 years) chronic radiation-related limb complications. Preoperative radiation and tumor sizes >5 cm were associated with an increased risk of acute wound complications (34% preoperative vs. 16% postoperative, P < .001; and 31% >5 cm vs. 17% ≤5 cm, P = .005). At 20 years the radiation-related complication rate was higher in patients with a groin or thigh tumor location (16% vs. 4% other; P = .008), prior acute wound complications (20% vs. 10% no surgical complication), and a radiation dose ≥60 grays (Gy) (18% vs. 9% for dose < 60 Gy; P = .04). Five fractures occurred, resulting in a crude overall fracture rate of 1.2%.


Patients treated with preoperative radiation for larger tumors are more likely to have acute surgical wound complications. Acute wound complications followed by postoperative radiation are associated with chronic radiation-related limb problems, as are higher radiation dose and proximal tumor location. The fracture rate is so low that prophylactic fixation is not warranted. Cancer 2006. © 2006 American Cancer Society.

External beam radiation therapy (EBRT) can be combined with limb-salvage surgery for the treatment of soft-tissue sarcomas. Previous studies have investigated the relation between EBRT and local complications. Postoperative wound complications include wound dehiscence, wound necrosis, persistent drainage, infection, and seroma formation,1, 2 whereas chronic radiation-related complications include fibrosis, edema, neurologic injury, osteitis, impaired physeal growth, sarcomatous change, and fractures.3–6 These studies have concluded that preoperative EBRT allows the use of smaller field sizes and lower doses,7 but is associated with more acute wound complications,1, 2, 8 whereas postoperative radiation requires the use of larger fields and higher doses and can be associated with chronic radiation-related problems.9 Another important finding is that a complication of any kind after treatment for an upper extremity sarcoma is distinctly uncommon.2, 10, 11

What to our knowledge has not been well characterized to date is the correlation between specific surgical interventions (e.g., vascular or plastic surgery reconstructions) and subsequent acute wound complications and between radiation therapy technique (e.g., whole-bone irradiation or high dose) and subsequent chronic radiation-related complications. To better characterize these complex correlations, we performed this retrospective review and limited our analysis to patients with tumors of the lower extremities only.

One particular area of interest was the incidence of radiation-related fractures and any associated risk factors. Fractures after radiation therapy have been investigated but the results and identified risk factors have varied somewhat. Lin et al.5 found that periosteal stripping, especially when combined with adjuvant chemotherapy and female gender, resulted in an increased fracture risk. Helmstedter et al.3 also found periosteal stripping to be a risk factor for fracture, as were anterior thigh compartment involvement and positive tumor margins. Although Holt et al.4 did not find periosteal stripping to be a risk factor, they did find female sex, age older than 55 years, and high-dose radiation to be significant risk factors. Both Lin et al.5 and Helmstedter et al.3 recommended considering prophylactic fixation in select, high-risk patients. An objective of this study was to determine the risk factors for fracture as well as the role of prophylactic fixation.


This retrospective study was approved by our Institutional Review Board with a waiver of informed consent. Our Department of Radiation Oncology sarcoma database was reviewed to find all patients treated with combined EBRT and limb-salvage surgery for primary soft-tissue sarcomas of the lower extremity between 1960 and 2003 who underwent definitive surgery and radiation therapy at our institution. This cohort represents a subset of the 1588 total patients in the Radiation Oncology sarcoma database, which in turn is a subset of the larger M. D. Anderson Cancer Center sarcoma database. Lower extremity was defined as involving the groin, thigh, knee, leg, ankle, and foot regions. Pelvis and iliac fossa tumors were excluded. Exclusion criteria included amputations, recurrent tumors, and definitive surgery or radiation performed outside of our institution.

Specific details extracted from the medical records included patient age, sex, use of adjuvant chemotherapy, the location of tumors, whether there was exposure of the bone at the time of surgery and whether the periosteum was formally stripped, whether a plastic surgeon was involved in wound closure, vascular reconstruction status, the use of prophylactic fixation, and the timing and nature of any complications. Patients who were treated with EBRT to the groin, thigh, or knee had their radiation records specifically reviewed to determine whether none, a partial circumference, or the entire circumference of the femur received radiation to a dose ≥50 grays (Gy).

Complications were defined as either a postoperative wound complication (< 3 months after surgery) or a chronic radiation-related complication. Complications were graded in our database as “mild” if a complication was noted but required no intervention or alteration in care, “moderate” if a complication required an intervention or alteration in patient care but did not require a surgical procedure, and “severe” if a complication required surgical intervention. The “moderate” and “severe” complications corresponded to major wound complications as defined in a previous randomized, prospective study.2

Actuarial data for local control and disease-free, distant metastasis-free, disease-specific, and complication-free survival curves were calculated with the Kaplan-Meier method,12 and tests of significance were based on the log-rank statistic. Multivariate analysis was performed with the proportional hazards model, with the log linear relative hazards function of Cox.12


Patient, Tumor, and Treatment Characteristics

From 1960 to 2003, 412 patients with primary soft-tissue sarcoma of the lower extremity received combined modality treatment with limb-salvage surgery and EBRT at our institution. There were 206 male and 206 female patients. The median age was 49 years (range, 8–92 years). The median follow-up on the 241 surviving patients was 9.3 years (range, 1.2–31 years). The median tumor size was 8 cm (range, 1.2–30 cm). Among the tumors, 304 measured >5 cm in greatest dimension and 107 measured ≤5 cm. The most common histologic subtype was malignant fibrous histiocytoma (175 patients [42%]), followed by liposarcoma (90 patients [22%]), synovial sarcoma (52 patients [13%]), unclassified sarcoma (38 patients [9%]) leiomyosarcoma (14 patients [3%]), neurogenic sarcoma (14 patients [3%]), fibrosarcoma (6 patients [1%]), and other sarcomas (23 patients [6%]). Tumor grade was as follows: low in 17 patients (4%), intermediate in 119 patients (29%), and high in 276 patients (67%). The site of tumor involvement was as follows: the thigh in 263 patients, the leg in 59 patients, the knee in 48 patients, inguinal in 23 patients, ankle in 9 patients, and foot in 10 patients.

The goal of the definitive surgical procedure was a wide resection (defined as dissection through normal tissue). Despite this policy, final resection margins were positive or uncertain in 63 patients (15%) and negative in 349 patients (85%). Specific surgical procedures included bone exposure in 95 patients (23%), periosteal stripping in 70 patients (17%), vascular reconstruction in 13 patients (3%), and plastic surgery reconstruction in 84 patients (20%). The exact types of plastic surgery reconstructive procedures according to the timing of radiation therapy delivery are shown in Table 1. Plastic surgery reconstruction was more frequent in patients treated preoperatively (P = .04).

Table 1. Plastic Surgery Reconstruction According to the Timing of Radiation Therapy
TypePreoperative XRT (n = 269)Postoperative XRT (n = 143)P
  1. XRT indicates X-ray therapy.

Skin graft33 
Rotational flap269 
Free flap132 
Graft plus rotational flap125 
Graft plus free flap92 
Total63 (23%)21 (15%).04

Radiation technique varied with location, but the field of radiation generally included a 1-cm to 3-cm radial and 5-cm to 7-cm longitudinal margin around the tumor or tumor bed. For patients treated postoperatively, a shrinking field technique was used after 50 Gy and again after 60 Gy if additional boosting was deemed necessary. Two hundred sixty-nine patients (65%) received preoperative EBRT with a median dose of 50 Gy (range, 44–70 Gy) and 143 patients (35%) received postoperative EBRT with a median dose of 60 Gy (range, 50–72 Gy). Of the 334 patients with groin, thigh, or knee tumors, 160 patients received radiation to the entire circumference of the femur and 174 received radiation to none or partial circumference of the femur. The interval of time between surgery and radiation was 39 days (range, 13–104 days) in patients treated preoperatively compared with 33 days (range, 6–233 days) in patients treated postoperatively.

Adjuvant chemotherapy was administered to 169 patients (41%) and was administered concomitantly in 46 patients (11%).

Patient Outcome and Complications

The overall and disease-free survival rates at 10 years were 62% and 62%, respectively. The distant metastasis-free survival rates were 71% at 5 years and 67% at 10 years. Overall local control rates were 89% at 5 years and 88% at 10 years.

The postoperative wound complication rate was 27% (113 of the 412 patients). The most common wound complication was dehiscence or breakdown, followed by infection, and seroma (Table 2). Fifteen complications were mild, 63 were moderate, and 35 were severe. On multivariate analysis, the use of preoperative radiation was associated with an elevated rate of acute wound complications (34% vs. 16% after postoperative radiation; P < .001) (Fig. 1). The rate of acute wound complications was found to be higher in patients treated for tumors measuring >5 cm (31% vs. 17% if tumors measured ≤5 cm; P = .035). Bone exposure, periosteal stripping, vascular reconstruction, and patient age were not found to be significant factors for acute postoperative wound complications. For the 84 patients who underwent plastic surgery reconstruction, the acute wound complication rate was 34% compared with 26% for the 328 patients who did not undergo a plastic surgery reconstruction (P = .1). When this analysis was limited to the patients treated with preoperative radiation only, the 63 patients who underwent plastic surgery reconstruction had an acute wound complication rate of 38% compared with 32% of the 206 patients who did not undergo a plastic surgery reconstruction (P = .2). Tumor location within the lower extremity also was not found to be related to subsequent wound complications (Table 3).

Figure 1.

Kaplan–Meier curves showing the rates of acute wound complications according to the timing of radiation therapy (XRT). Pre-Op indicates preoperative; Post-Op, postoperative.

Table 2. Postoperative Wound Complications
Wound complicationNo. (N = 113 total)
Persistent drainage3
Skin graft breakdown2
Wound edema1
Table 3. Percent Complications According to Tumor Location
LocationNo.Acute complicationsPChronic complicationsP
Thigh26328 16 
Knee4825 8 
Leg5922 2 
Ankle911 0 
Foot1040 0 
Distal12624 4 

The crude chronic radiation-related complication rate was 10% (41 of the 412 patients) (Table 4). The actuarial radiation-related complication–free survival rate was 87% at 20 years. Thirteen complications were graded as mild, 17 as moderate, and 11 as severe. On multivariate analysis, proximal tumor location, prior surgical complication, and a radiation dose ≥60 Gy were found to be significant factors for the development of a chronic radiation-related complication. For patients with an acute wound complication, there was a 20% incidence of a subsequent radiation-related complication versus 10% if no surgical complication occurred (P = .003). After doses ≥60 Gy, there was an 18% incidence of a radiation-related complication versus 9% if doses were < 60 Gy (P = .003) (Fig. 2). When dose was treated as a continuous variable, it remained significantly associated with chronic radiation-related complications. Although not statistically significant, there was a trend between the timing of radiation and chronic complications (P = .079) (Fig. 3). For patients with proximal tumors (groin and thigh), the incidence of a chronic wound complication was 16% versus 4% for patients with distal tumors (P = .002) (Table 3). Concomitant chemotherapy use, adjuvant chemotherapy use, plastic surgery reconstruction, vascular reconstruction, tumor size, and patient age did not appear to correlate with chronic complications.

Figure 2.

Kaplan–Meier curves showing the rates of chronic radiation-related complications according to the dose of radiation. Gy indicates grays.

Figure 3.

Kaplan–Meier curves showing the rates of chronic radiation-related complications according to the timing of radiation therapy (XRT). Post-Op indicates postoperative; Pre-Op, preoperative.

Table 4. Chronic Radiation-Related Complications
Chronic complicationNo. (N = 41 total)
Soft tissue necrosis6
Neurologic injury2

To further examine the relation between the timing of radiation and the development of subsequent complications, a stratified analysis was performed. The 23 patients who developed a postoperative surgical wound complication and who then received postoperative radiation had a 42% incidence of long-term radiation-related problems versus only 13% in those 120 patients without a surgical complication (P < .001). This relation was not found in patients treated with preoperative radiation (15% long-term radiation-related complication rate in 90 patients with a surgical complication vs. 7% in the 179 patients without a complication; P = .1). It is noteworthy that, in patients who received their radiation therapy postoperatively, the median time to the initiation of radiation therapy for those patients who had experienced an acute postoperative wound complication was 60 days compared with only 29 days in patients who had not experienced an acute wound complication (P < .001).

Five fractures occurred for a crude overall fracture rate of 1.2% (5 of the 412 patients). The median time to fracture was 5.2 years (range, 1.7–25 years). The median age at the time of fracture was 47 years (range, 34–71 years). The fracture sites included the femoral shaft in 3 patients, the femoral neck in 1 patient, and the tibial tubercle in 1 patient. All fractures were within the radiation field for patients with groin, thigh, or knee tumors (5 of 334 patients [1.5%]) and were sustained with minimal or no trauma. The small number of fractures did not allow multivariate analysis. Using crude incidence and limiting the analysis to the 334 patients with groin, thigh, or knee tumors, radiation to the entire circumference of bone (3% vs. 0%; P = .03) and surgical exposure of bone (4.5% vs. 0.7%; P = .02) were significant for the development of a fracture. The incidence of a fracture was found to be higher in the 144 patients receiving adjuvant chemotherapy (3% vs. 0.5%; P = .05), but there was also an imbalance in the number of patients having had exposure of bone during the surgical procedure (38 of the 144 patients [26%] who underwent chemotherapy also had bone exposed compared with 28 of the 190 patients [15%] who did not receive chemotherapy; P = .01). The actuarial fracture rate at 10 years in patients who had the whole circumference of bone irradiated and had the bone exposed at the time of surgery for a groin, thigh, or knee primary tumor was 11% (95% confidence interval, 5–30%). The actuarial curve is shown in Figure 4.

Figure 4.

Kaplan–Meier curve showing the rate of fracture in patients having had whole bone circumference radiation and surgical exposure of bone. There were a total of 44 patients with 3 events (fractures).


The current study confirms the findings of a recent randomized trial comparing preoperative with postoperative radiation for patients with soft-tissue sarcomas of the extremities.2 Preoperative radiation therapy is associated with acute wound complications, and higher radiation dose, as used in the postoperative setting, is associated with chronic radiation-related complications. In addition, there appears to be a correlation between acute wound complications and larger tumor size, in which the wounds and fields of radiation are larger and the surgical closures more complex.

Similar to the results seen in other studies, we did not find that wound closure or coverage by a plastic surgeon lowered the rate of wound complications.2, 10 The use of plastic surgery reconstruction has steadily increased at our institution since 1985, when the first complex plastic surgery reconstruction was performed for patients with primary soft-tissue sarcomas of the lower extremities. Between 1985 and 1999 there were 44 reconstructive procedures performed in the 216 patients (20%) compared with the years 2000 to 2002, during which 34 of the 83 patients underwent a reconstructive procedure (41%; P < 0.001). At the current time, plastic surgery reconstruction is used in patients with anticipated wound or soft-tissue coverage issues, thereby self-selecting a patient group with a higher likelihood of a wound complication.

There was also an independent association noted between proximal tumor location and chronic radiation-related complications. Reasons for an increased incidence of chronic complications might be related to the volume of tissue irradiated for patients with groin and thigh tumors that is independent of tumor size. Given the large relative size of the groin and thigh, even small tumors that were radiated via small fields may actually have had large volumes of tissue receiving high doses. Given the retrospective nature of this study and a lack of three-dimensional (3D) planning in the majority of patients, it was impossible to study this question further, but it will be important for future radiation-related complication studies to account for volumes of tissue receiving the prescribed radiation dose.

With regard to fractures, the overall fracture rate was found to be quite low after combined modality treatment of soft-tissue sarcomas of the lower extremity and prophylactic fixation does not appear to be routinely necessary. Patients who undergo surgical exposure of the bone and receive radiation to the entire bone circumference might warrant close observation, but even in this situation the rate of fracture is only reported to be 11% at 20 years.

Our findings confirm those of previous studies that demonstrate that most lower extremity fractures occurred within proximal locations.3–5 In this subset of patients, we were not able to confirm the significance of formal periosteal stripping, but it is possible that the periosteum may have been stripped but not well documented in the surgical report.3, 5 Attempts have been made to quantify the extent of stripping,3 but we found this impossible to do in a retrospective review because the size of periosteal stripping was recorded only infrequently. Periosteal stripping has been shown to decrease the blood supply to the femoral cortex normally supplied by the periosteum.13 In addition, the cambium layer of the periosteum contains progenitor cells that contribute to bone healing.14 The loss of this layer may impair the femur's ability to heal from repetitive microtrauma. The combination of decreased blood supply and impaired bone healing, even after simple bone exposure, most likely increases the chance of fracture.

We did not find radiation dose, adjuvant chemother apy, patient age, or gender to be significant risk factors for fracture, although the number of patients with fractures was so small that some differences may not have been apparent. Although other studies have found these variables to be significant risk factors for the development of a fracture, given the small number of events, it is very difficult to account for other confounding characteristics and the issue remains open to debate. Although the diminished bone density that frequently occurs with female sex, advancing age, and chemotherapy would intuitively appear to contribute to a higher fracture risk, we were only able to convincingly show the significance of whole bone irradiation and bone exposure.

Another interesting finding was the association between acute wound complications and subsequent chronic radiation-related complications. It was previously believed that acute wound complications completely resolve, but the current analysis suggests that some of these complications do in fact develop into chronic problems, particularly in patients treated with postoperative radiation. Although there was a delay in the initiation of postoperative radiation in those patients with an acute wound complication, it is possible that the wound was still healed incompletely and this predisposed the patient to develop a chronic problem.

In summary, the current analysis confirms that preoperative radiation is associated with a higher rate of acute wound complications and that a higher radiation dose, as used in the postoperative setting, is associated with a higher rate of chronic radiation-related complications. Patients with a postoperative wound complication who subsequently receive postoperative radiation have a much higher incidence of chronic radiation-related complications, which lends support for the use of preoperative radiation to minimize the occurrence of chronic complications. When used preoperatively, radiation results in manageable acute wound complications that can be effectively treated with medical and, only occasionally, surgical intervention. The issue of which characteristics actually predispose to fracture in radiated patients remains open, but it would seem that routine prophylactic fixation is not necessary in approximately 90% of even “high-risk” patients.

In addition, the results of the current study appear to suggest that the volume of tissue being manipulated either surgically or radiotherapeutically is related to complications. In the preoperative setting, the size of the tumor and, in the postoperative setting, the proximal thigh location most likely serve as surrogates for volume of tissue being treated. If this is true, then conformal radiation techniques might be useful in the preoperative treatment of large tumors as well as in the postoperative treatment of groin and thigh tumors, in which even small fields of radiation appear to have resulted in some chronic radiation-related complications. Intensity modulated radiation therapy, or at least careful 3D treatment planning, might be useful in these patients for limiting the volume of tissue unnecessarily receiving the prescribed radiation dose. Whether daily image-guided radiotherapy techniques with cone beam computed tomography or kilo-voltage portal imaging could help achieve the needed treatment accuracy needs to be studied. Whether these techniques will result in decreased complications while not increasing the local failure rate will only be known through further investigation.