Patterns of Major Wound Complications Following Multidisciplinary Therapy for Lower Extremity Soft Tissue Sarcoma

Abstract

Background and Objectives:

The purpose of this study was to determine the pattern and timing of major wound complications (MWCs) in patients at our institution who received multimodality treatment for lower extremity soft tissue sarcoma (LE-STS) and to evaluate the impact of MWCs on tumor control and patient outcomes.

Methods:

The medical records of 102 LE-STS patients treated with limb-sparing surgery and radiation therapy were reviewed. MWCs were defined as secondary operations with anesthesia, seroma/hematoma aspiration, admission for IV antibiotics, or persistent deep packing.

Results:

MWCs occurred in 22% of patients, with 45% of events occurring >120 days after resection. On multivariate analysis, preoperative external beam radiation therapy (EBRT) (OR 4.29, 95% CI 1.06–17.40, P = 0.042) and skin graft placement (OR 6.39, 95% CI 1.37–29.84, P = 0.018) were found to be independent predictors of MWCs. MWC occurrence did not predict for chronic toxicity and did not impact tumor control or survival.

Conclusions:

A considerable proportion of MWCs occur >120 days from surgical resection with preoperative EBRT and skin graft placement independent predictors for MWCs. While an additional source of morbidity, MWC occurrence did not impact tumor control, nor did it predict for chronic toxicity.

INTRODUCTION

The goals of treatment of extremity soft tissue sarcoma (STS) are to maximize functional outcome and cancer control while minimizing morbidity. To that end, the National Cancer Institute of Canada (NCIC) trial compared the overall treatment outcomes, including the acute and late toxicity, of pre- versus postoperative external beam radiation therapy (EBRT) for extremity STS [1,2]. While there were no significant differences at 5 years between the two groups in local control (LC), recurrence-free survival, cause-specific survival, or overall survival (OS), there were differences in toxicity. Patients in the preoperative arm had more wound complications at 120 days after resection than patients in the postoperative arm, 35% versus 17% (P = 0.01). A follow-up study reported a higher rate of subcutaneous fibrosis, joint stiffness, and edema in the postoperative group although none of the differences were statistically significant and there was no difference in patient-reported function [3].

Wound complications can adversely affect patient quality of life and have been associated with a higher incidence of chronic complications caused by EBRT [4]. In multiple studies, including the NCIC trial, the lower extremity was associated with a higher risk of wound healing complications [1,5–8]. The definition of a major wound complication (MWC) in the NCIC trial included events which developed up to 120 days after surgery. While it is unclear if any MWCs occurred after 120 days as it was not one of the trial endpoints, chronic radiation skin injury may result in delayed ulcerations and prolonged wound healing [9]. The purpose of this study was to determine the pattern and timing of MWCs in patients at our institution who received multimodality treatment for lower extremity STS. Given the high rate of wound complications in this subgroup of extremity STS reported in the NCIC trial, our hypothesis was that a significant number of MWCs may occur more than 120 days from initial surgical resection. Our goal was to identify risk factors for MWCs in this patient population and evaluate the impact of MWCs on tumor control and patient outcomes. In addition, we sought to determine if there was an association between the timing of the MWC and chronic toxicity including chronic pain and edema.

MATERIALS AND METHODS

Following approval from the institutional review board at the Ohio State University James Cancer Hospital, we identified all patients who received radiation therapy for STS between 1998 and 2013 using electronic treatment records maintained by the Department of Radiation Oncology. Eligibility criteria for this study included: age ≥18 years, STS of the lower extremity defined as extending from the iliac crest to the toes, and definitive limb sparing treatment performed with radiation therapy and surgical resection. Patients with a diagnosis of Ewing’s sarcoma, chondrosarcoma, osteosarcoma, or desmoid tumor were excluded from analysis because of the different treatment approach for these histologic subtypes. Patients treated for recurrent disease or for persistent disease following initial non-oncologic resection were included to evaluate the role of multiple surgical resections on MWCs. Patients treated with radiation therapy alone without surgical resection were excluded. We identified 102 patients who met inclusion criteria for this study.

Data for the cohort was collected through electronic and paper record review. Histologic grade was determined using the French Federation of Cancer Centers Sarcoma Group grading system and tumor size was based on the largest dimension reported on final pathology. MWCs were defined as in the NCIC trial [1]—“secondary operation under general or regional anesthesia, seroma/hematoma aspiration, admission for IV antibiotics, or persistent deep packing.” For our definition, we included wound complications which occurred up to 1 year postoperatively and classified the complications based on the initial intervention. Tumors at the knee joint and above were considered proximal tumors, while tumors below the knee joint were considered distal. Data regarding chronic effects of treatment were also collected. Chronic edema was defined as edema in the treated lower extremity which was noted in the chart as persisting more than 3 months after the final day of treatment (chemotherapy, surgery, or radiation therapy) requiring intervention such as a compression garment or referral to lymphedema clinic. Chronic pain was defined as the regular use of prescription medication for pain related to the primary tumor more than 3 months after the final day of treatment (chemotherapy, surgery, or radiation therapy).

Treatment

For each patient, a full medical history and physical examination was completed along with biopsy of the primary and imaging of the extremity and chest. Each patient was discussed at multidisciplinary tumor board and a definitive treatment plan was agreed upon. Radiation treatment technique was based on the treatment decade and the proximity of the tumor to organs at risk. All radiation therapy planning was CT-based. Typical radiation fields included a 5 cm distal and proximal margin and a 2–3 cm radial margin from the gross tumor with alterations of the margin dependent on the presence of adjacent bones and fasciae. Intensity modulated radiation therapy (IMRT) was utilized in a minority of patients (11/100, 11%) while 2D/3D planning was used in 89/100 (89%) of patients. For patients who received preoperative EBRT, a postoperative boost was delivered for positive margins at the discretion of the radiation oncologist. Intraoperative electron-beam radiation therapy (IOERT) was used at the discretion of the surgeon and radiation oncologist for tumors located adjacent to a critical structure including nerves, vessels, bone, or for close surgical margins adjacent to muscle or fascia.

The waiting period between surgery and EBRT was planned for 4–6 weeks with delays to ensure complete healing prior to the next phase of treatment left to the discretion of the surgeon and radiation oncologist. Patients underwent definitive resection with limb-sparing surgery performed by a surgical or orthopedic oncologist specializing in sarcoma. Plastic surgery assistance with closure and graft/flap placement was left to the discretion of the primary surgeon.

Statistical Analysis

Patient outcomes were analyzed using the Kaplan-Meier method. LC was defined as the time from the final treatment date to the date of disease recurrence in the treatment field based on imaging or biopsy findings. Disease-free survival (DFS) was defined as the time from the final treatment date to the date of disease recurrence either locally or distantly. OS was defined as the time from initial biopsy to date of death.

The median cohort age of 56 years and median tumor size of 8 cm were used to subcategorize patients for statistical analysis. The associations between patients’ clinical covariates and the risk of developing MWCs were first assessed by Fisher’s exact test. The covariates with P < 0.1 were further analyzed by using multivariable logistic regression models to validate the predictability of those risk factors after adjusting for other covariates. The cumulative risk of MWCs was determined using the Nelson-Aalen cumulative hazard estimate. The prognostic effect of MWCs on LC, DFS, and OS was analyzed with the log-rank test. All statistical tests were 2-sided and P < 0.05 was considered statistically significant. All statistical analyses were performed using SAS 9.4 (SAS, Inc.; Cary, NC).

RESULTS

Patient and Tumor Characteristics

A total of 102 patients with STS of the lower extremity were identified and included in this analysis. Characteristics of the patients, tumor, and treatment are summarized in Table I. The majority of patients were male (55/102, 54%) with 95/102 patients (93%) undergoing treatment for a primary diagnosis of STS. Although all patients in this cohort were treated definitively, four patients were found to have metastatic disease prior to the completion of treatment. Tumors were located in the proximal lower extremity in 89/102 patients (87%) and undifferentiated pleomorphic sarcoma/malignant fibrous histiocytoma (UPS/MFH) was the most common histologic subtype (29/102, 28%). The median tumor size was 8 (range 1.5–23) cm with 78/102 patients (76%) diagnosed with high grade tumors.

Table I.

Patient Characteristics

Characteristic	N (%), (N = 102)
Age (y), median (range)	56 (21–86)
Gender, female/male	47 (46)/55 (54)
BMI at surgery ≥30kg/m2
Yes	38 (37)
No	64 (63)
History of diabetes
Yes	16 (16)
No	86 (84)
Any smoking history
Yes	51 (50)
No	51 (50)
Active smoker at diagnosis
Yes	12 (12)
No	90 (88)
Preoperative albumin <3.5 g/dl
Yes	15 (15)
No	30 (29)
Presenting diagnosis
Primary diagnosis	95 (93)
Recurrent disease	7 (7)
Initial non-oncologic resection
Yes	12 (12)
No	90 (88)
Location
Proximal	89 (87)
Distal	13 (13)
Histopathology
UPS/MFH	29 (28)
Liposarcoma	20 (20)
Leiomyosarcoma	15 (15)
Synovial	8 (8)
Other	30 (29)
Tumor size (cm)
≤8	53 (52)
>8	49 (48)
Pathologic grade
Intermediate/high	84 (82)
Low	16 (16)
Indeterminate	2 (2)
Surgical margins
Positive	15 (15)
Negative	87 (85)
Plastic surgery closure
Yes	24 (24)
No	78 (76)
Vascularized tissue flap placed
Yes	22 (22)
No	80 (78)
Skin graft placement
Yes	13 (13)
No	89 (87)
Chemotherapy administered
Yes	40 (39)
No	62 (61)
EBRT Timing
Preoperative	25 (25)
Postoperative	75 (75)
EBRT dose per fraction (Gy)
2	52 (52)
1.8	47 (47)
Unknown	1 (1)
EBRT delivery
2D/3D	89 (89)
IMRT	11 (11)
IOERT boost delivered
Yes	30 (29)
No	72 (71)

BMI, body mass index; UPS/MFH, undifferentiated pleomorphic sarcoma/malignant fibrous histiocytoma; EBRT, external beam radiation therapy; IMRT, intensity modulated radiation therapy; IOERT, intraoperative electron-beam radiation therapy.

Treatment

Seventy-two patients (72/102, 71%) underwent EBRT alone, 28/102 patients (27%) received EBRT plus an IOERT boost, and 2/102 patients (2%) received IOERT alone. One of the patients treated with IOERT alone had recurrent disease following EBRT and resection 2 years prior while the second patient failed to complete EBRT. Twenty-five patients (25/100, 25%) received radiation preoperatively and 75/100 patients (75%) were treated postoperatively. The median preoperative EBRT dose was 50 Gy with a range of 45–50.4 Gy while the median postoperative dose was 60 Gy (range 45–70.2 Gy) with two patients completing a postoperative EBRT boost to a total dose of 64 Gy. The median EBRT dose when IOERT was also delivered was 50 Gy (range 45–59.4 Gy). The fraction size was 2 Gy in 52/100 patients (52%), 1.8 Gy in 47/100 patients (47%), and unknown in one patient. A higher proportion of patients treated with postoperative EBRT received a dose of 1.8 Gy per fraction (43/75, 57%) compared to patients treated with preoperative EBRT (4/25, 16%). The median IOERT dose was 10 Gy (range 10–15 Gy) with a median energy of 6 MeV (range 6–12 MeV) and a median cone size of 9 cm (range 6–12 cm). Chemotherapy was administered to 40/102 patients (39%) with doxorubicin, ifosfamide, and mesna used in the majority of patients (25/40, 63%). Chemotherapy was delivered preoperatively in 13/40 (33%) cases, postoperatively in 17/40 (42%) cases, and both pre- and postoperatively in 10/40 (25%) cases. All patients underwent gross total resection with 15/102 patients (15%) found to have positive microscopic margins. Twelve patients (12/102, 12%) underwent initial non-oncologic resection and were referred to our institution for definitive treatment while 7/102 (7%) patients required re-resection for close or positive microscopic margins. The median time between surgery and EBRT for all patients was 51 days with a median of 63 days (range 49–98 days) for preoperative EBRT and 46 days (range 26–168 days) for postoperative EBRT. Eleven patients had delays between surgery and radiation therapy of >90 days. Of those 11 patients, 6 underwent flap placement and 10 were treated with postoperative EBRT. All delays were secondary to prolonged wound healing. Plastic surgery performed closure in 24/102 patients (24%) with 22/102 (22%) patients requiring flap placement and 13/102 (13%) patients receiving full/split thickness skin grafts for primary closure.

Patient Outcomes/Wound Complications

At a median follow up of 52 months (range 7–150 months), the 5-year LC, DFS, and OS rates were 87% (95% CI 76–94%), 53% (95% CI 42–63%), and 68% (95% CI 57–77%), respectively (Fig. 1A). As summarized in Table II, MWCs occurred in 22/102 patients (22%) with 45% (10/22) of events occurring more than 120 days after surgical resection. Of note, 8/10 (80%) MWCs which occurred >120 days after the date of surgical resection were in patients treated with postoperative EBRT. One of the MWCs occurred in a patient treated with IOERT alone. The majority of complications were either debridement under general anesthesia (7/22, 32%) or admission for IV antibiotics (7/22, 32%). Second surgical interventions requiring general anesthesia accounted for 59% (13/22) of the MWCs. The median time to complication was 88 days (range 10–365 days). In the preoperative EBRT group, 70% (7/10) of the MWCs required a second surgery while 55% (6/11) in the postoperative EBRT group required a second surgery. Two patients (2/11, 18%) in the postoperative EBRT group underwent debridement with regional anesthesia while none of the patients in the preoperative EBRT group were treated with this procedure. Admission for IV antibiotics was approximately 30% in both groups (3/10 in preoperative EBRT and 3/11 in postoperative EBRT). The cumulative risk estimate of MWCs in the pre- and postoperative EBRT groups is shown in Figure 1B. The cumulative risk of a MWC in the pre- and postoperative EBRT groups at 3 months after surgery was 37.6% (95% CI 18.7–75.6%) and 2.7% (95% CI 0.67–10.7%), respectively. At 6 months, the cumulative risk of a MWC was 49.7% (95% CI 26.6–93%) and 8.3% (95% CI 3.7–18.4%) in the pre- and postoperative EBRT groups, respectively.

Fig. 1.

Kaplan-Meier curves for the entire patient cohort including local control (LC), disease-free survival (DFS), and overall survival (OS) (A). Cumulative risk estimate of major wound complications (MWC) in patients treated with preoperative (Pre-op RT) and postoperative external beam radiation therapy (Post-op RT) (B).

Table II.

Major Wound Complications: Type of Complication and Timing

Intervention	Overall, N (% of all patients)	≤120 days of surgery date, N (% of all patients)	>120 days after surgery date, N (% of all patients)	Preoperative EBRT, N (% of preop patients)	Postoperative EBRT, N (% of postop patients)
Total	22 (22)	12 (12)	10 (10)	10 (40)	11 (15)
Secondary operation
Debridement with secondary wound closure	4 (4)	2 (2)	2 (2)	2 (8)	2 (3)
Debridement alone under general anesthesia	7 (7)	6 (6)	1 (1)	5 (20)	2 (3)
Debridement alone under regional anesthesia	2 (2)	0 (0)	2 (2)	0 (0)	2 (3)
Operative drainage	2 (2)	1 (1)	1 (1)	0 (0)	2 (3)
Admission for IV antibiotics	7 (7)	3 (3)	4 (4)	3 (12)	3 (4)

EBRT, external beam radiation therapy.

Univariate analysis was performed to identify predictors for MWCs (Table III). The only patient-related factor found to predict for MWCs was diabetes (8/16, 50% with vs. 14/86, 16% without, P = 0.006). Surgery-related factors predictive for MWCs included closure by plastic surgery (9/24, 38% yes vs. 13/78, 17% no, P = 0.045) and placement of a skin graft (7/13, 54% yes vs. 15/89, 17% no, P = 0.006). The delivery of preoperative EBRT (10/25, 40% preoperative vs. 11/75, 15% postoperative, P = 0.011) and a fractional dose of 2 Gy versus 1.8 Gy (17/52, 33% vs. 4/47, 9%, P = 0.006) were radiation-related statistically significant predictors for MWCs on univariate analysis. On multivariate analysis, preoperative EBRT (OR 4.29, 95% CI 1.06–17.40, P = 0.042) and skin graft placement (OR 6.39, 95% CI 1.37–29.84, P = 0.018) were found to be independent predictors of MWCs (Table IV). The overall rate of chronic pain in the patient cohort was 13% (13/101) with 23% (5/22) of those with MWCs experiencing chronic pain as opposed to 10% (8/79) of patients without MWCs. However, the presence of a MWC did not predict for chronic pain (OR 2.6, 95% CI 0.76–8.90, P = 0.128). Chronic edema occurred in 22% (22/101) of the patients. Although the rate of chronic edema was higher in patients with MWCs 27% (6/22) compared to those without MWCs (20%, 16/79), the presence of a MWC did not predict for chronic edema (OR 1.48, 95% CI 0.5–4.4, P = 0.482). Data on chronic pain and edema was not available for one patient.

Table III.

Univariate Analysis for Predictors of Major Wound Complications

Variable	Major wound complication frequency (%)	P-value
Age ≥56		0.480
Yes	25
No	18
Gender		0.470
Female	26
Male	18
BMI at surgery ≥30kg/m2		0.457
Yes	26
No	19
History of diabetes		0.006
Yes	50
No	16
Any smoking history		0.810
Yes	24
No	20
Active smoker at diagnosis		1
Yes	17
No	22
Preoperative albumin <3.5 g/dl		0.062
Yes	40
No	13
Presenting diagnosis		1
Primary diagnosis	22
Recurrent disease	14
Initial non-oncologic resection		0.454
Yes	8
No	23
Location		1
Proximal	21
Distal	23
Tumor size		1
≤8 cm	21
≥8 cm	22
Pathologic grade		1
Intermediate/High	21
Low	19
Surgical margins		1
Positive	20
Negative	22
Plastic surgery closure		0.045
Yes	38
No	17
Vascularized tissue flap placed		0.559
Yes	27
No	20
Skin graft placement		0.006
Yes	54
No	17
Chemotherapy administered		0.085
Yes	12
No	28
EBRT timing		0.011
Preoperative	40
Postoperative	15
EBRT dose per fraction (Gy)		0.006
2.0	33
1.8	9
EBRT delivery		1
2D/3D conformal	21
IMRT	18
IOERT boost delivered		0.599
Yes	17
No	24

BMI, body mass index; EBRT, external beam radiation therapy; IMRT, intensity modulated radiation therapy; IOERT, intraoperative electron-beam radiation therapy.

Table IV.

Multivariate Analysis for Predictors of Major Wound Complications

Variable	Odds ratio	95% CI	P-value
Diabetes (yes vs. no)	1.82	0.43–7.80	0.418
Plastic surgery closure (yes vs. no)	0.95	0.24–3.70	0.944
Skin graft placement (yes vs. no)	6.39	1.37–29.84	0.018
Chemotherapy administered (yes vs. no)	0.26	0.06–1.13	0.072
EBRT dose per fraction (2 vs. 1.8 Gy)	2.89	0.73–11.46	0.132
EBRT timing (pre- vs. postoperative)	4.29	1.06–17.40	0.042

EBRT, external beam radiation therapy.

The presence of a MWC did not impact local or distant control (Fig. 2A–C). The 5-year LC, DFS, and OS for patients with and without MWCs were 94.1% (95% CI 65–99.1%) and 86.4% (95% CI 73–93.4%), 50.3% (95% CI 27.2–69.5%) and 54.7% (95% CI 41.7–65.9%), and 57.9% (95% CI 26.1–80.1%) and 70.2% (95% CI 57.4–79.8%), respectively, with none of these differences reaching statistical significance. The timing of the MWC also did not impact patient outcomes. The 5-year LC, DFS, and OS for patients with MWCs within 120 days of surgery and those with MWCs more than 120 days from surgery were 100% and 85.7% (95% CI 33.4–97.9%), 49.4% (95% CI 20.1–73.3%) and 50.8% (95% CI 15.7–78.1%), and 50.9% (95% CI 9.5–82.4%) and 60.0% (95% CI 19.5–85.2%), respectively, with none of these differences reaching statistical significance. Seven patients (7/102, 7%) ultimately underwent amputation with a median time to amputation of 18 months from the initial resection. Two of the amputations were secondary to MWCs which persisted as chronic non-healing wounds with one of the patients treated with preoperative EBRT and the other treated postoperatively. One patient presenting with a MWC was ultimately found to have ischemic changes in the treated limb and amputation was performed as the patient was not a candidate for a bypass. This patient received IOERT and was planned for postoperative EBRT, but it was never completed due to the MWC. One of the amputations occurred due to severe wound healing complications following surgical repair of a pathologic fracture in the treated limb 4 years after his initial limb-sparing therapy including postoperative EBRT. The remaining three amputations were secondary to recurrent disease.

Fig. 2.

Local control (A), disease-free survival (B), and overall survival (C) for patients with and without major wound complications (MWC).

DISCUSSION

Wound complications are a major source of morbidity for patients with extremity STS receiving radiation therapy, particularly in the lower extremity [1]. The overall rate of MWCs determined in our study cohort was 22% (22/102), consistent with prior reports [1,4,6]. The rate of MWCs in patients receiving pre- versus postoperative EBRT was 40% (10/25) compared to 15% (11/75) which is consistent with the NCIC trial [1]. However, if only the MWCs which occurred within 120 days of surgery were included, the overall MWC rate is 12% (12/102) with a rate of 32% (8/25) and 4% (3/75) in the pre- and postoperative EBRT groups, respectively. Prior studies have reported a MWC rate of 24–44% for the delivery of preoperative EBRT [1,4,6,7,10–12] and 5–17% for postoperative EBRT [1,4,6,13] although the definition of a wound complication between these studies varied. Of the 22 MWCs found in this study cohort, 12/22 (54%) occurred early (≤120 days from surgery) and 10/22 (45%) were late (>120 days from surgery). For the late MWCs, 80% (8/10) were in patients treated with postoperative EBRT. All patients treated with postoperative EBRT were given time for the incision to fully heal prior to starting treatment. However, in six of the patients with late complications, breakdown of the skin around the wound site occurred during EBRT which failed to heal without intervention. Thus, while the overall incidence of MWCs associated with postoperative EBRT is low, MWCs are most likely to occur >120 days from surgical resection in the setting of postoperative EBRT based on our findings.

After reclassifying the MWCs using the Clavien-Dindo classification [14] of surgical complications, all complications would be either a grade II or III complication with 68% (15/22) of the MWCs classified as grade III. There were 7 (7%, 7/102) grade II complications with three events occurring ≤120 days after surgery and 4 occurring >120 days after surgery. Three of the grade II events occurred with preoperative EBRT, three occurred with postoperative EBRT, and one occurred with IOERT alone. Two grade III-a complications (2%, 2/102) occurred with both events occurring >120 days after surgery and both with postoperative EBRT. The remaining 13 events were grade III-b complications with nine occurring ≤120 days after surgery and four occurring >120 days after surgery. Seven grade III-b complications occurred with preoperative EBRT and six occurred with postoperative EBRT.

On univariate analysis, we found that diabetes, plastic surgery closure, skin graft placement, preoperative EBRT, and a dose of 2 versus 1.8 Gy per fraction were predictors of MWCs with preoperative EBRT and skin graft placement persisting as independent predictors of MWCs. Due to the overall low number of wound healing complications, an analysis to determine which factors predict for early MWCs versus late was not possible. The finding that 2 versus 1.8 Gy per fraction was a predictor of MWCs is likely related to a bias to treat preoperative EBRT patients with 2 Gy per fraction (21/25, 84%) and postoperative EBRT patients with 1.8 Gy per fraction (43/74, 58%). Preoperative EBRT as a predictor for MWCs is consistent with the results of the NCIC trial and a number of retrospective single institution reviews [1,4–6,15]. The increase in MWCs observed in patients with skin graft placement is consistent with results reported by Baldini et al., although in that study vascularized flap and split thickness skin graft closures were combined and analyzed as a single variable [10]. We analyzed vascularized flap and skin graft placement as two separate variables and did not observe a statistically significant increase in MWCs following vascularized flap placement. This result may be explained by the difference in blood supply between the two reconstruction techniques. While tissue flaps have their own blood supply [16], skin grafts are dependent on the recipient site for its blood supply, which may be tenuous in either a pre- or postoperatively radiated surgical bed, and may contribute to impaired wound healing [17].

Plastic surgery closure and the use of vascularized flaps have been hypothesized to reduce the incidence of MWCs in extremity STS. A study performed by Rosenberg et al. demonstrated a non-significant reduction in MWCs when plastic surgery performed closure [11]. No difference in frequency of MWCs was observed with the use of vascularized flaps. In our study, the increased frequency of MWCs in patients with primary closure by plastic surgery may be secondary to selection bias since patients predisposed to wound complications may have been more frequently selected for plastic surgery closure. Other factors associated with an increase in MWCs include both treatment and patient related factors. Preoperative chemotherapy has been associated with adverse wound healing by Chmell et al. [18]; however, in our study, systemic therapy was not associated with an increased risk for MWCs. Diabetes, smoking, and nutrition status have been previously associated with an increased risk for wound complications [6,10,18,19]. We found a significant increase in MWCs in patients with diabetes and a trend for significance in patients with low preoperative albumin.

Overall, this patient population had excellent LC (5-year 87%) with marginal systemic control (5-year DFS 53%), consistent with other studies of extremity STS [20–22]. We hypothesized that patients with MWCs may be subject to treatment delays or incomplete treatment which may adversely affect tumor control. However, we did not observe any difference in LC, DFS, or OS between the patients with and without MWCs. We further evaluated the impact of early (≤120 days from surgery) MWCs versus late (>120 days from surgery) MWCs on tumor control and patient survival. While the analysis is limited by small patient numbers, no difference in outcomes was observed in patients who experienced early versus late MWCs.

Strategies to address the toxicity associated with radiation therapy for extremity STS is an area of active investigation. One solution is to treat a smaller volume of normal tissue. A recently reported phase 2 study by O’Sullivan et al. evaluated if the use of preoperative image-guided IMRT could reduce the wound complication incidence as defined by the NCIC trial [21]. While the wound complication rate (30.5% vs. 43%, P = 0.2) and requirement of secondary operations (33% vs. 43%, P = 0.55) was lower than reported in the NCIC trial, neither metric was statistically significant. RTOG-0630 is a multi-institutional phase II trial which investigated using preoperative EBRT with daily image-guided radiation therapy and reduced planning volumes based on the size and grade of tumor for extremity STS [23]. The primary endpoint of the study was late toxicity including grade ≥2 lymphedema, subcutaneous fibrosis, or joint stiffness at 2 years after completion of EBRT. While the study demonstrated a significant decrease in late toxicity compared to the preoperative arm of the NCIC trial, the incidence of MWCs on the trial was comparable with the frequency of MWCs reported in the NCIC trial. An additional strategy to overcome acute wound complications at the expense of an increase in late complications is the delivery of postoperative EBRT. The VORTEX study has completed accrual in the United Kingdom and is a prospective randomized trial assessing the impact of a reduced postoperative radiotherapy volume on limb function and LC. An additional method to reduce the toxicity associated with radiation therapy is to omit radiation in appropriately selected patients. A nomogram has been developed based on the Memorial Sloan-Kettering Cancer Center prospective sarcoma database that can be used to quantify the risk of local recurrence at 3 and 5 years without adjuvant irradiation [24]. The nomogram includes factors such as age, size, margin status, grade, and histology and was shown to be a better predictor of local recurrence than the AJCC staging system.

Treatment of impaired wound healing following radiation therapy is also an area of active investigation and of great clinical need. Conventional therapy includes local wound care, pentoxifylline, and hyperbaric oxygen therapy in select cases [25]. The use of agents to stimulate wound healing including growth factors, platelet rich plasma, and stem cells are promising, but await human trials [26–28]. With advances in treatment planning and delivery and the development of effective agents to promote healing of irradiated tissue, the morbidity associated with definitive treatment of extremity STS will hopefully continue to be diminished.

This study has several limitations. It is a retrospective cohort study, and, therefore, is subject to selection bias present in all retrospective analyses. While the identification of a second operation or an admission for IV antibiotics was well-documented in the patient chart, documentation regarding persistent packing was less reliable and, therefore, the actual incidence of MWCs may be underestimated. In addition, assessing the impact of early versus late complications on quality of life using the surrogate endpoints of chronic pain and chronic edema certainly has its limitations. Our definitions of chronic edema and chronic pain were based entirely on chart review and documentation. Prospective physician-rated toxicity and patient-reported outcomes would be ideal, but were not possible in this retrospective study. Finally, the study was limited by a small sample size and patients were treated over a long time period during which changes to surgical treatment and radiation therapy of extremity STS occurred.

CONCLUSIONS

In conclusion, about 25% of patients with lower extremity STS treated with multidisciplinary therapy will have major wound complications. About 50% of those complications will occur more than 120 days from surgical resection primarily limited to patients treated with postoperative radiation therapy. Preoperative EBRT was again identified as a risk factor for MWCs. While preoperative EBRT should continue to be favored because of the lower dose and smaller field size, patients are at increased risk for MWCs which may result in additional interventions, although no compromise in tumor control or increased late toxicity is anticipated. The development of new strategies both in the delivery of radiation therapy and in wound healing is imperative to help mitigate the excess morbidity associated with the current treatment of extremity STS.

Abbreviations:

DFS

disease-free survival

EBRT

external beam radiation therapy

IMRT

intensity modulated radiation therapy

IOERT

intraoperative electron-beam radiation therapy

LC

local control

LE-STS

lower extremity soft tissue sarcoma

MWC

major wound complication

NCIC

National Cancer Institute of Canada

OS

overall survival

STS

soft tissue sarcoma