Introduction

Myxofibrosarcoma (MFS) was first described in 1977 and presents a distinct entity of malignant soft tissue tumours (STS) since reclassification in 2002 by the World Health Organization (WHO). MFS is a rather rare entity representing about 5% of STS. It is mostly found in adults, aged between 50 and 70 years old, with a slight male predominance.

MFS is known for histological characteristics as the myxoid stroma, curvilinear vessels and pleomorphism1,2,3,4.

Typically, MFS appears at limbs and the limb girdles. Other localizations such as trunk, hands, feet, head or neck are quite rare, as well as abdominal or retroperitoneal appearance.

Primarily, tumour formation occurs in the dermal and subcutaneous tissues. Subsequently, MFS is known for its infiltrative growth, penetrating the underlying fascia and skeletal muscle1,2,4,5,6.

MFS often presents a specific pattern of focal infiltration, the so called “Tail sign”. It describes a specific pattern of infiltrative growth, showing a curvilinear shaped tumour spread from the main tumour mass, usually along the fascial plane, probably caused by the lack of a pseudocapsule7,8 Therefore, today’s diagnostic standard is magnetic resonance imaging, as the most sensitive and precise method of imaging.

State-of-the-art treatment for localized disease consists of radical resection combined with (neo-)adjuvant radiotherapy. Yet, the importance of chemotherapy is not clarified3,4,9.

This entity of STS is known for its unfortunate local outcome. Among STS, MFS show highest rates of local recurrence (LR), with a range from 15 to 60% after 5 years1,2,3,4,6,10,11. Some authors have postulated that grading has no influence on rates of LR. In contrast, more recent studies confirmed a correlation between tumour grading an LR-rates. Risk of LR are estimated around 48% for low- grade tumours, up to 62% for high-grade MFS1,12,13.

Progress of tumour grade is seen in up to 38% of patients with LR, which is linked to higher metastatic potential and is described both, in superficial as well as in deep lesions9,14.

However, metastatic potential of MFS is comparatively low in general, but increases with higher tumour grade2,15. Huang et al. showed that distant metastasis (DM) was found in up 50% of the cases in patients with tumour progression. In contrast, without histologic progression, up to 17% of DM occurred8,14,16,17.

The incidence of distant metastases reported in the literature ranges from 9.5 to 34%6,9,11,12,18. Accordingly, MFS shows overall survival rates between 61% and 83%, indicating a relatively favorable prognosis9,19,20,21. Even though majority of patients with STS benefit through radical or wide tumour margin assessment and local control can be achieved, in patients with MFS there is still uncertainty about the optimal width of resection. Therefore it is still in question if standard margin assessment is adequate for patients with MFS11,22,23. Some authors attribute this uncertainty to the infiltrative growth pattern of MFS, which makes it challenging for surgeons to visually determine the true expansion of the tumour12,15.

Kandel et al. presented a clinical practice guideline, saying that there is “no available evidence-based data addressed how to adequately assess margins24.

Patients and methods

A total of 135 patients with histologically verified MFS were included in this study, treated at our institution between December 1999 and October 2023. Some patients were referred after initial treatment elsewhere. All patients had a minimum follow-up of one year. The population consisted of 74 male and 61 female patients, aged between 16 and 92 years. All patients were followed up at our department every 3 months during the first three years, every 6 months in year 4 and 5, and in 12-month intervals after that. Postoperative surveillance consisted of clinical examination, computed tomography of the thorax and abdomen and local magnetic resonance imaging. Complete clinical pathological data were available for 135 patients (i.e., the final study population) and were collected from medical reports retrospectively (sex, age, tumour site, histology, tumour size, tumour depth, resection margins, tumour stage, laboratory parameters, neoadjuvant and adjuvant radiotherapy, adjuvant chemotherapy, local recurrence). Baseline patient characteristics are shown in Table 1. Histopathological diagnoses were done according to the current WHO classification for soft tissue and bone tumours and reconfirmed by an experienced pathologist specialised in STS at our hospital5. Characteristics of the tumour were assessed according to preoperative imaging and histopathological analysis of the specimen. Tumour size was defined using the largest diameter as a reference. Tumour depth was classified as either superficial (above the muscle fascia) or deep (below the muscle fascia). Resection margins were classified according to both the Enneking classification and the Union internationale contre le cancer (UICC) system25,26. Tumour staging followed the criteria of the American Joint Committee on Cancer (AJCC)2, and grading (G1–G3) was based on the French Federation of Cancer Centres Sarcoma Group (FNCLCC) system (29).

The study was approved by local ethics committee of the Medical University of Vienna, Austria. All methods were performed in accordance with relevant guidelines and regulations, under informed consent from all participants.

Assessment of tumour size was performed at surgical resection specimen using the largest diameter as a reference. Further depth of tumour-localization was differentiated between superficial (above the muscle fascia) and deep (below the muscle fascia), according to preoperative imaging and histopathological analysis.

Table 1 Baseline characteristics of the study population.
Full size table

Statistical analysis

All statistical analyses and graphical visualization were performed using IBM SPSS Statistics, Version 29, SPSS Inc, Chicago, IL, USA. Continuous data was summarized using mean values, medians and ranges, and categorical data by absolute frequencies and percentages. Correlations were calculated by the Spearmann´s correlation coefficient. Follow-up time was defined as the time from index surgery to death or last known alive.

Analysis of resection margins were evaluated and performed by an experienced specialist at our clinic.

We defined the day of histopathological confirmed diagnosis as baseline and considered death from any cause as endpoint (overall survival), which in this study could carefully be considered as equal of death of disease. Survival probabilities were calculated with the Kaplan-Meier product limit estimator. To compare the survival functions of two or more patient groups, we applied the log-rank test. Uni- and multivariable Cox proportional hazards regression models were fitted to evaluate the association between baseline variables and survival.

Multivariable analysis of LR were based on cause- specific hazards and therefore carried out by Cox regression models. P values < 0.05 were considered significant.

The Kaplan Meier method was used to estimate overall sarcoma specific survival (OS) and LR-free survival.

Results

Median age of the population group was 66.4 years, IQR 21.9 years and the median follow-up time 94.57, IQR 82.5 months.

Twenty-six MFS were superficial (19.3%), 108 tumours were deep (80%) and in one case the information was missing.

Within our population, 111 MFS were localized at the extremities (82.2%), 80 (59.3%) were found in the lower limb, 27 (20%) in the upper limb, 4 (3%) in the shoulder girdle, 23 (17%) in at the trunk and one at the neck (0.7%). In 128 cases (94.8%) tumour resection could be performed, in 7 cases (5.2%) amputation had to be performed.

LR was found in 14 out of 135 patients, (10.37%). Distant metastasis occurred in 31 cases (22.9%).

Table 2 Characteristics of local recurrences (LR) and resection margins (RM).
Full size table

Median overall survival was 106.5 months, IQR 29.6 months. Median LR-free survival was 109.17 months, IQR 33.73 months. Evidence of death was found of 30 patients (22%).

Taking further analysis on the prior performed tumour resection, 4 cases of LR (7.3%) were found after wide tumour (55 cases) resection after Enneking’s classification and after marginal resection according to Enneking, 6.7% of LR were observed (4 of 60 cases). Intralesional resection was performed in 20 cases and 6 LR (30%) occurred.

Analysing resection margins according to UICC R-classification 7 out of 14 LR were observed after resection with R0 > 1 mm status, which presents 7.3% in 96 cases. R1-dir margin status was found in 22 cases with 6 LR (27.3%).

There was no LR in the 14 patients with R1 < 1 mm status and one LR in three cases with R2a status (33.3%), which means that macroscopically tumour was left. Characteristics of LR and the distribution regarding resection margins can be seen in Table 2.

Univariate statistical risk analysis of these findings confirmed that there was no significant impact of the width of the performed tumour resection on local control, with exception of R1-dir status (p = 0.029 h 3.41, 95% CI 1.14–10.25) and R2a status (p = 0.037 h 10.16, 95% CI 1.15–90.03) Thus, if macroscopically or microscopically tumour is left behind, there is an increased risk for LR.

In multivariate analysis R1-dir status remains significant after adjustment for adjuvant therapy (radiation, as well as chemotherapy) (p = 0.036 h 3.67, 95% CI 1.09–12.40). Moreover, a significant impact of R2a status could be detected (p = 0.028 h 12.412, 95% CI 1.32-117.04).

Furthermore, multivariate risk analysis showed significant results for R1-dir status as well as R2a status, after adjustment to tumour size and tumour depth and high grading (p = 0.026, HR 3.91, 95% CI 1.18–12.99), (p = 0.031, HR 11.74, 95% CI 1.26-109.79).

Results can be seen in Table 3 below.

Table 3 Multivariable risk analysis on local control.
Full size table

The first LR occurred at a median of 16.4 months, IQR 19.7 months, after definitive surgery. Among patients with LR, the median number of recurrences was 2, IQR 0.5, and most patients experienced more than one LR.

Further, there was no statistically significant impact on local control for patients who had underwent tumour resection before (p = 0.45, HR 1.53, 95%, CI 0.51–4.62).

Table 4 Frequencies of radiation therapy administration; frequencies of local reoccurrence (LR).
Full size table

There was no significant correlation of LR and survival (p = 0.245, HR 1.66, 95% CI 0.71–3.88) whereas significance regarding metastasis and survival could be shown (p < 0.001, HR 3.54, 95% CI 1.74–7.21). Survival functions can be seen in Figs. 1 and 2. Tumour size > 5 cm, with a hazard ratio of 8.45 (p = 0.036, 95% CI 1.15–62.05) showed an adverse impact on overall survival.

Tumour localisation (superficial/deep) could not be proven as a significant factor (p = 0.835, HR 1.10, 95% CI 0.45–2.69). Furthermore, there was no statistically significant impact of adjuvant chemotherapy on LR (p = 0.919 h 0.96, 95% CI 0.47–1.99) as well as adjuvant radiation on LR (p = 0.945, HR 1.04, 95% CI 0.39–2.72). Administration of radiotherapy and frequencies of local reoccurrence can be seen in Table 4.

Overall, within our population 104 patients (77%) received adjuvant or neoadjuvant RT as complementary treatment to surgery. Chemotherapy was administered to 44 (32.6%) patients (median age 56.5 years, IQR 19.86 years), mostly in patients with grade 3 MFS and in deeply located, sized > 5 cm.

Fig. 1
Fig. 1
Full size image

Survival function according to local recurrence (LR).

Fig. 2
Fig. 2
Full size image

Survival function according to distant metastasis (DM).

Discussion

Myxofibrosarcoma represents one of the most common STSs and is mostly found in elderly patients. Furthermore, it is known for its unfortunate local outcome.

There is still uncertainty about the optimal width of resection. It is still in question if standard margin assessment is adequate for patients with MFS and whether width of resection margin is essential for local control22,23,24,25.

In the present study we could show that there is no statistically significance of the impact of the width of resection margins on local recurrence, except for the R1-dir status (p = 0.029) and R2a status (p = 0.037). Microscopically and macroscopically tumour residual was shown to have an unfortunate effect on rates of local reoccurrence. In contrast to previous published studies we distinguished width of performed complete tumour resection25,26. These results support our clinical suggestion that LR occurs in patients with MFS, but width of resection margins does not seem to be crucial if the tumour is removed in total. This is an important information for practical approach to tumour resection.

In summary MFS is known as an entity of STS with distinct behavioural characteristics. The infiltrative nature of MFS, along with characteristic findings such as the tail sign, poses a challenge to accurately determining the tumour’s extent during surgery. With our recent findings, surgical margins as they are classified typically for STS are not predictive for the LR of MFS.

Accordingly, we can postulate that the tumour must be removed totally, resection width is not mandatory. If there is a rest part of tumour left, there is a significant increase of risk of LR.

Therefore, we suggest resection with histologically negative margins for best patient’s outcome. While achieving adequate resection margins is a fundamental surgical objective, our findings highlight that the width of the margin may not be the critical factor in ensuring local control in MFS. It also undermarks the importance of recent and high-quality imaging to identify true tumour expansion and therefore plan adequate resection.

In contrast to previous published studies we could not prove a significant impact of high grade sarcoma or radiotherapy on LR, even though, the literature shows variability regarding the influence of radiotherapy2,12,25,26.

Furthermore, while local recurrence did not influence overall survival, effective local control is essential, as recurrent MFS often shows higher grade and greater metastatic potential2,15.

We are aware of the limitations of the present study (retrospective design, sample size prevalent cases), the strength of our report is the homogeneity of our study population (myxofibrosarcomas from a single centre only).

Furthermore, we plan on retrospectively analyse our databank and histopathological resee all MFH (malign fibrous histiocytoma) probes and therefore be able to perform on a larger study population. Moreover, effectiveness of different tissue qualities as defence of tumour growth and infiltration should be exanimated.

Recently there are ongoing explorations on using indocyanine green fluorescence (IGF) imaging for intraoperative evaluation of STS surgical margins. Given the present difficulties for precise tumour resection in MFS because of its infiltrative growth and knowing our findings, we are optimistic that IGF imaging will set benefits in local control.