Primary appendicular soft‐tissue sarcoma resection: What tumour parameters affect wound closure planning?

Abstract

Wound closure after wide, local excision of an appendicular soft‐tissue sarcoma (STS) can be challenging. This study evaluates the value of magnetic resonance imaging (MRI)‐based tumour parameters in planning wound closure. A total of 71 patients with a primary limb‐based STS, excluding vascular or osseous involvement, excluding the shoulder or hand and hip or foot were included. Axial MRI images were used to measure the circumferences and cross‐sectional areas of the limb, bone, and tumour. The tumour diameter, length, and depth at the level of maximal tumour dimension were measured to identify the tumour's relative contribution to the planning of optimal wound closure management through primary closure (PC) or reconstructive surgery (RS). Eighteen patients required planned wound RS. Wound complications occurred in 14% overall. Tumours, which were closed by PC, were of significantly greater depth, shorter radial diameter, and shorter tumour circumference relative to those closed by RS. On multivariate analysis, tumour depth was the greatest contributory factor in predicting type of wound closure. A quantitative analysis of MRI‐based tumour parameters demonstrates tumour depth as the most predictive factor in planning for the type of wound closure and may prove beneficial in providing greater insight into planned wound management of sarcoma resection.

1. INTRODUCTION

Soft‐tissue sarcomas (STSs) are heterogeneous mesenchymal neoplasms, which account for 1% of adult malignancies. They typically present with a mass effect or metastatic disease. Routine preoperative cross‐sectional imaging is mandatory and aids in defining the appropriate surgical resection zone.1, 2 With current surgical techniques, the treatment of appendicular sarcomas involves limb‐sparing surgery in over 90% of cases.3

The increasing rate of limb salvage has increased the demands on the surgeon for appropriate wound closure. Options, in this regard, include primary wound closure (PC) or planned wound reconstructive surgery (RS). To date, determining the optimal method of wound closure remains subjective but clearly depends on the wound size and location, skin quality, exposed vital structures, vascularity, anaesthetic considerations, and need for adjuvant therapy. Smaller lesions can often be primarily closed, while larger or more complicated lesions typically require RS in the form of skin grafting, myocutaneous flap coverage, or delayed primary wound closure. A more aggressive resection in irradiated soft tissues, combined with primary reconstruction, should be considered in cases with multiple risk factors.4 Flap reconstruction is not associated with increased postoperative complications following extremity sarcoma resection and may mitigate the effects of some risk factors in selected patients.5

It could, however, be proposed that preoperative imaging can be used to not only determine the resection margin but also the likelihood of requiring more complex reconstructive wound closure. Yet, to date, despite routine cross‐sectional imaging being available for most patients, no research has examined the quantitative relationship between tumour parameters and the chosen surgical approach to STS wound closure, the implications of which include more directed patient counselling, earlier involvement of plastic and reconstructive surgical teams, requirement for more intensive postoperative nursing, length of hospital stay, and timing of postoperative radiotherapy.

Therefore, this study aims to assess the relative contributions of magnetic resonance imaging (MRI)‐based tumour parameters in planning wound closure during surgical resection.

2. METHODS

2.1. Setting

All patients were managed at a tertiary referral centre. A weekly multidisciplinary team meeting is held where all confirmed new, suspected, or recurrent STS cases are discussed to formulate a management plan. Neoadjuvant or adjuvant radiotherapy and/or chemotherapy is offered to patients where appropriate. Where skin loss and RS are anticipated, the plastic surgery team is consulted during the index procedure, and delayed reconstruction is planned. We defined a wound complication as a wound dehiscence of greater than 1 cm² or a wound infection as defined by the modified ASEPSIS score.

2.2. Surgical technique

Skin resection margins were determined by bimanual palpation of the lesion to gauge deliverance of the tumour into the wound. The surgical approach involved a longitudinal incision, including the biopsy tract with an ellipse of skin where necessary, all performed by the senior author. Resections were made by marginal excision or wide excision, depending on tumour extension, as guided by preoperative histopathology biopsy and MRI. In either case, where the wound was closed primarily with nylon sutures, two drains were used, and sutures were removed at least 2 weeks after the index procedure. In cases involving RS, negative pressure wound therapy was provided in the interval period until histological margins were confirmed (mean 7 days, range 5‐10 days). Definitive wound closure was subsequently performed. All patients had a 2‐ and 6‐week postoperative appointment where wound healing was checked.

2.3. Patients

Patients with primary appendicular STSs affecting the upper or lower limb presenting to our service between June 2007 and June 2014 were included regardless of their stage or grade. Of the 94 patients presenting with appendicular STS during this period, 71 were deemed eligible for inclusion (75%). Tumours extending proximal to the lesser trochanter or distal to malleoli and tumours proximal to the lesser tubercle or distal to the carpus were excluded. Tumours requiring bone, joint, or vascular reconstruction; revision resections; and tumours where the oncological approach was complicated by the biopsy technique were excluded. Patients who had previously had operative interventions or visible skin trauma in the region of the operative site were excluded, as were those with lymphoedema or steroid usage exceeding 1 year.

2.4. Measurements

Patient‐related risk factors for wound breakdown were recorded prospectively. These included age, gender, diabetes, smoking, concurrent steroid use, preoperative radiotherapy, and body mass index (BMI). Age was categorised into groups of younger than 30, 30 to 50, 50 to 70, and older than 70 years. Smoking was categorised into never, less than 5, 5 to 15, and greater than 15 pack‐years. BMI was categorised into less than 18.5, 18.5 to 25, 25 to 30, and greater than 30 kg/m². Postoperative radiotherapy and any wound complications were recorded at the 6‐week follow‐up appointment.

OsiriX digital software (Version 3.9, Pixmeo) was used by the primary investigator for the analysis of preoperative MRI scans. The examiners were blinded to the wound closure technique and treatment outcomes. The tumour measurement approach was agreed upon by the investigators prior to MRI evaluation, including optimal resection margins, anatomical structures to be sacrificed, and known surgical approaches, and was taken on two separate occasions.

The axial image displaying maximal tumour size was identified in each case. This image was subsequently used for analysis. A polygon region of interest was selected by drawing a series of dots along the resection margin of choice, dictated by the preoperative tumour type, grade, and stage. The programme subsequently yielded a line of best fit and calculated the cross‐sectional area (CSA, cm²). CSA measurements were then taken of the limb and bone(s) at the maximal level. Linear tumour depth was taken from the skin surface to the desired resection margin (cm), along the line of desired surgical approach.

2.5. Statistical analysis

Statistical analysis was performed using SPSS (IBM, New York) version 22.0. Cohen's kappa was used for discrete variables in evaluating inter‐ and intra‐observer correlation. Categorical variables associated with PC and RS were initially explored using chi‐squared tests or Fisher's exact tests where cell counts were ≤5. Tests for normality of distribution of continuous variables were undertaken, and independent sample t tests and Mann‐Whitney U tests were used to compare parametric and non‐parametric data, respectively. P‐values of <.05 were considered statistically significant. Multivariable‐adjusted logistic regression models were built to explore the factors independently associated with the type of surgical approach. All patient‐related factors and tumour‐related factors were considered for inclusion in multivariable models if P < .2 on univariate regression analysis.

Ethical approval was attained from the local ethics committee.

3. RESULTS

The types of appendicular STS treated in this case series are outlined in Table 1. Most STS cases were treated with PC (71.8%, n = 51). The mean (±SD) age of cases was 53.0 (±19.5) years. There was no significant difference in the proportion of STS cases between males and females (47.9% vs 52.1%).

Table 1.

Frequency of type of soft‐tissue sarcoma in the study group (n = 71)

	n	%
Myxoid liposarcoma	14	19.7
Leiomyosarcoma	13	18.3
Pleomorphic undifferentiated sarcoma	11	15.5
Synovial sarcoma	8	11.3
Atypical lipomatous tumour	8	11.3
Alveolar sarcoma	5	7.0
Fibromyxoid sarcoma	5	7.0
Other	7	9.9

The demographic and clinical features of the study group are shown by wound closure approach in Table 2. Wound complications occurred in 14% of cases (n = 10). There was a significant difference in the prevalence of wound complications occurring among those who underwent RS (30%, n = 6) compared with those who underwent PC (8%, n = 4; P = .025). The mean age at wound closure was significantly lower among those who underwent PC (49.6 ± 20.1 years) compared with RS (61.8 ± 14.8 years) (P = .011).

Table 2.

Patient‐related factors in the study group (n = 71)

		Primary closure		Reconst. surgery
		n	%	n	%
Gender	Male	21	41	13	65
	Female	30	59	7	35
Limb	Upper	9	18	5	25
	Lower	42	82	15	75
Smoker	No	45	88	15	75
	Yes	6	12	5	25
Diabetes	No	49	96	20	100
	Yes	2	4	0	0
Steroid use	No	49	96	20	100
	Yes	2	4	0	0
Pre‐op. radiotherapy	No	43	90	17	94
	Yes	5	10	1	6
Post‐op. radiotherapy	No	25	52	6	35
	Yes	23	48	11	65
Wound complications	No	47	92	14	70a
	Yes	4	8	6	30

^aχ²: P < .05.

The tumour‐related factors are shown for both groups in Table 3. Tumours that were closed by PC were of significantly greater depth, shorter radial diameter, and shorter tumour circumference relative to those closed by RS.

Table 3.

Tumour‐related factors among the study group by surgical approach (n = 71)

		Primary closure	Reconst. Surgery	P
Limb circumference (cm)	Median (range)	45.3 (21.0‐67.2)	39.0 (23.8‐92.0)	.45
Bone circumference (cm)	Median (range)	9.0 (4.8‐22.5)	9.7 (4.6‐35.0)	.76
Tumour circumference (cm)	Median (range)	15.7 (4.8‐32.0)	20.1 (10.3‐81.5)	.045
Radial diameter (cm)	Median (range)	5.0 (2.0‐11.0)	6.0 (4.0‐24.0)	.018
Limb CSA (cm2)	Median (range)	158.8 (31.9‐328.3)	107.0 (43.4‐621.0)	.08
Bone CSA (cm2)	Median (range)	5.9 (1.8‐35.1)	6.4 (2.9‐36.2)	.63
Tumour CSA (cm2)	Median (range)	19.1 (1.6‐97.3)	25.3 (7.9‐60.5)	.08
Tumour length (cm)	Median (range)	8.7 (2.2‐17.9)	8.7 (3.2‐32.7)	.61
Tumour depth (cm)	Median (range)	0.9 (0.2‐5.6)	0.2 (0.1‐1.8)	<.001

Note: P were values computed by Mann‐Whitney U tests.

Abbreviation: CSA, cross‐sectional area.

Intra‐observer and inter‐observer correlations were 0.85 and 0.86, respectively.

Radiotherapy details were available for 65 patients. Of these, six patients (9%) received preoperative radiotherapy, and 34 patients (52%) had postoperative radiotherapy. Wound complications were more prevalent among those who had postoperative radiotherapy (n = 8; 24%) compared with those who did not (n = 2; 7%), but this difference was not statistically significant.

On multivariate‐adjusted analysis, tumour depth was the most statistically significant factor determining wound closure. Tumour depth was associated with significantly lower odds of closure by RS relative to PC (adjusted odds ratio [OR] 0.44, 95% confidence interval 0.01‐0.60; P = .02). No other patient or tumour‐related factors remained significantly associated with the type of wound closure approach after controlling for confounding variables (Table 4).

Table 4.

Patient‐ and tumour‐related factors associated with wound closure by reconstructive surgery (multivariable adjusted)

		Crude OR	95% CI	Adj. OR	95% CI	P
Gender	Male	1.0		1.0
	Female	0.38	0.13‐1.10	0.49	0.08‐2.93	.43
Age (per year)		1.04	1.00‐1.08	1.01	0.95‐1.08	.68
Radial diameter (per cm)		1.27	1.02‐1.57	1.20	0.68‐2.13	.53
Tumour depth (per cm)		0.04	0.01‐0.26	0.44	0.01‐0.60	.02
Tumour length (per cm)		1.09	0.98‐1.21	1.17	0.88‐1.56	.28
Tumour circumference (per cm)		1.07	1.00‐1.14	1.00	0.83‐1.20	.99

Abbreviations: CI, confidence interval; OR, odds ratio.

4. DISCUSSION

Planning the surgical resection of STS requires careful anatomical and pathological consideration. Preoperative cross‐sectional imaging is used to determine resection margins and anatomical considerations to optimise the surgical plan.6, 7, 8 MRI has played a pivotal role as the preoperative image modality of choice and has been integral in promoting a shift from limb amputation towards limb salvage surgery.

The focus of this study was to determine whether the preoperative MRI could be used as an aid in the decision‐making of wound closure. We found, on multivariate analysis, that tumour depth was the most important predictor of the wound closure technique required. Specifically, we found that, for every additional 1 cm in tumour depth, the odds of requiring closure by RS decreased by 56% (OR 0.44).

Importantly, given the importance of reproducibility in delineating soft tissue and measuring our parameters,9 we achieved acceptable inter‐ and intra‐observer correlation with OsiriX (freely downloadable software: http://www.osirix-viewer.com/Downloads.html for Apple products).

The spectrum of histopathology displayed by sarcomatous tumours and corresponding resection margins is accounted for with these findings. A standard surgical oncological approach was taken for all tumours in this series. Positive margins are consistently associated with adverse survival‐related outcomes.9 Histological and molecular assay analysis of peri‐tumour soft tissue has shown tumour cells beyond 1 cm and up to a maximum of 4 cm.10, 11 Histological grade on preoperative biopsy would determine a marginal (eg, atypical lipomatous tumour) or wide, local (eg, myxoid liposarcomatous tumour) resection (Figure 1). A 1 cm resection margin is widely quoted as an acceptable target for a safe excision.12, 13, 14, 15 Increasing the circumference of the resection as desired on the preoperative MRI scan thus influences the decision to plan for wound reconstruction (Figure 1).

Figure 1.

Case illustration: A, conservative (marginal) resection margin; B, aggressive (wide, local) resection margin as defined by preoperative magnetic resonance imaging. CSA, cross‐sectional area

Predictive measurements are dependent on a close correlation between the tumour as seen on MRI and the intra‐operative pathology. Disparities may arise between a preoperative two‐dimensional image and intra‐operative management of a three‐dimensional tissue defect as experienced by any skilled operator whose planning is based on preoperative imaging. Postoperative swelling may also vary. Previous computer‐aided tissue reconstruction research has cited constant values for skin elasticity and thickness.16 Skin resection diameter is reliant on bimanual palpation by the operator, which is influenced by the CSA and depth of the tumour. The arc of resectable skin on MRI was reflected by the width of the tumour and represented a large SD in both groups and was therefore excluded.

Systemic risk factors for poor‐quality skin and wound healing, such as diabetes, smoking, obesity, and steroids, are well recognised and evade any quantitative skin assessment. One patient was excluded from this cohort because of prior prolonged steroid use. Reported wound complications for preoperative radiotherapy are 35% compared with 17% for post‐operative radiotherapy.16 This difference is less marked when the upper limb is treated. Radiotherapy demonstrated wound‐healing complications in this series of patients in both groups, particularly in RS patients. This may not have been a causative factor but indicated that the lesions were more complex. While preoperative radiotherapy may encourage the surgeon to use RS,16 this was not demonstrated in this study.

With regard to limitations, the small number of cases makes sub‐group analysis difficult. Comparisons between PC and RS must be interpreted with caution as this is not a randomised study. However, while the specific methodology of this study cannot determine the specific depth at which PC is safe, we would propose that a tumour depth exceeding 1 cm would be considered appropriate to tolerate PC. The unpredictable effect of radiotherapy on wound breakdown has been mentioned. STS tumours with osseous or neurovascular involvement accounted for 14% in a series of STS described by Ravaud et al.17 A total of 9.9% required vascular resection in a series by Schwarzbach.18 STS of the hand accounted for 4% of all STS,19 with STS occurring in the foot accounting for 3% of lower‐limb cases.20 The eligible tumours accounted for 75% of all tumours at this unit, which is similar to international rates. Recurrent tumours have unpredictable tissue vascularity, thus making them inappropriate for this formula. In addition, any insult inflicted to the skin, including undermining, stretching, and radiotherapy, risks diminishing the skins oxygenation,21 which can result in wound dehiscence and necrosis after STS resection.22

In the absence of digital images or software, manual measurement of the factors described is difficult, particularly with small or irregularly shaped lesions. Distance measurement is available on most software programmes, but best‐fit circumference measurements may not be available. However, an adequate version of the software used for this study is freely available, as are many other software programmes.

A prospective study with a larger cohort and with evaluation of all relevant preoperative factors would prove beneficial to providing greater insight into planned wound management for these patients.

5. CONCLUSION

Preoperative cross‐sectional MRI, particularly for tumour depth, is most helpful in determining the optimal wound closure technique.

Cawley DT, Barrett P, Moran B, et al. Primary appendicular soft‐tissue sarcoma resection: What tumour parameters affect wound closure planning? Int Wound J. 2019;16:1553–1558. 10.1111/iwj.13251