Systemic Inflammation Is Associated with Oncological Outcome in Patients with High-Grade Myxofibrosarcoma of the Extremities: A Retrospective Analysis

Myxofibrosarcoma (MFS) is a soft tissue sarcoma (STS) which primarily affects the extremities and girdles, especially in elderly patients [1]. The prognosis of MFS is relatively good compared to other STS, with a risk of distant metastases ranging between 20 and 25% [2] and a disease-specific survival (DSS) rate of 83% [3]. Size and depth of the tumor are generally considered prognostic factors. Recently, we also reported that FNLCC histologic grade may be considered an additional prognostic factor for survival in MFS [4]. On the other hand, the local recurrence (LR) rate is particularly high in this sarcoma (range from 32 to 60%) [2, 5], higher than other STS [6]. Tumor features at magnetic resonance imaging (MRI) can help in identifying risk classes of LR [7-11]. Systemic inflammation in STS was reported to be indicative of aggressive tumor characteristics [12-14]. Nakamura et al. [15] and Szkandera et al. [13] observed that elevated C-reactive protein (CRP) and neutrophil-lymphocyte ratio (NLR) values can be associated with inferior survival rates and an increased risk of LR in patients affected by STS. However, these series were very heterogeneous, in particular regarding histologic subtypes and sites of the tumor. MFS has the highest median macrophage infiltration and T-cell infiltration among STS. For this reason, we aimed to verify whether preoperative systemic inflammatory markers (serum CRP levels and NLR) can help in predicting the DSS and local tumor control in adult patients affected by localized MFS of the extremities.

A total of 397 patients affected by MFS of the limbs (hands and feet excluded) underwent surgical excision at a single institute between 1993 and 2017.

Patients were excluded if they presented with LR or if they had a previous unplanned excision [16]. Patients were also excluded if they presented metastasis at the time of diagnosis and if preoperative MRI images were not available.

In the end, 126 adult patients (>18 years) with primary and localized MFS of the extremities (hands and feet excluded) were included in the study.

All cases were histologically revised and classified according to the 2013 World Health Organization classification of STS [17]. A 3-step system (FNCLCC) was used to assess MFS grades [18]. On preoperative MRI, the presence of a “tail sign” was checked, and myxoid component and gadolinium (Gd) enhancement of the tumor were evaluated and graded with a semiquantitative method, as previously described [9]. All patients had blood tests preoperatively which included CRP and absolute neutrophil and lymphocyte count.

The NLR was defined as the absolute neutrophil count divided by the absolute lymphocyte count. We defined 2 different cut-off values for high versus low CRP and NLR based on mean and median values of our cohort, as there were no standard levels of CRP and NLR in previous series.

The tumor size was assessed on surgical specimens using the larger diameter as a reference, while depths were divided into superficial (above the fascia) and deep (below the fascia), according to preoperative MRI. Margins were evaluated according to the classification of Enneking et al. [19]. Radical and wide margins were considered “adequate.”

The use of radiotherapy (RT) and chemotherapy (ChT) was decided collegially with a multidisciplinary team (orthopedic surgeon, radiotherapist, and oncologist).

Patients’ characteristics were presented by frequencies and percentages for categorical variables, and median and range for continuous variables. The Kaplan-Meier method was used to estimate DSS and LR rates. The DSS interval was defined as the time between surgery and death or last follow-up, whichever came first. Similarly, LR-free survival was defined as the time between surgery and the first LR, respectively, or last follow-up available, whichever came first. Patients who died of other causes were censored. Differences in survival rates were assessed by the log-rank test. Multivariable analysis of DSS and LR were based on cause-specific hazards and therefore carried out by Cox regression models. p values <0.05 were considered significant.

All analyses were completed using the Statistical Package for Social Sciences (SPSS Statistics for Windows, version 22.0, released 2013; IBM Corp., Armonk, NY, USA).

Median age at the time of surgery was 68 years (range 19–92); 60 patients (47.6%) were male, and 66 (52.4%) were female (Table 1). The majority of MFS (86, 68.3%) were localized in the lower limb, 20 (15.9%) in the upper limb, 11 (8.7%) in the pelvic girdle, and 9 (7.1%) in the shoulder girdle.

Thirty-seven tumors (29.4%) were small (<5 cm), 38 (30.2%) between 5 and 10 cm, and 51 (40.5%) were larger than 10 cm. Thirty-nine MFS (31.0%) were superficial, and 87 (69.0%) were deep. Eight (6.3%) MFS were grade 1, 23 (18.3%) grade 2, and 95 (75.4%) grade 3.

Median CRP was 0.4 mg/dL (mean 2.0 mg/dL, range 0.1–26.6 mg/dL). Fifty-two patients (41.3%) had preoperative CRP >0.5 mg/dL, in particular in higher-grade MFS (p = 0.043) and in tumors >10 cm (p < 0.001). Median NLR was 2.8 (mean 3.4, range 0.1–11.2). Forty-four (34.9.9%) had NLR >3.5, mostly observed in MFS >10 cm (p = 0.028). High values of CRP were also correlated with NLR >2.8 (p < 0.001).

At preoperative MRI, a tail sign was identified in 37 (29.4%) cases. Most of the MFS had high Gd enhancement (80.7%, grade 2–3) and high myxoid features (66.7% type 2–3).

Excision was possible in 122 (96.8%) patients, whereas in 4 (3.2%) patients an amputation was necessary. Among excised MFS, 96 (78.7%) had adequate (wide) margins and 26 (21.3%) inadequate margins (marginal in 19 cases and intralesional in 7 cases). Overall, 89 patients (73.0%) received adjuvant or neoadjuvant RT as complementary treatment to surgery. ChT was administered to 25 (20.5%) patients.

At the last follow-up (median 30 months, range 6–175), 91 patients (74.6%) were alive with no evidence of disease, 8 patients (6.6%) died of other causes, and 23 (18.9%) died of the disease. Kaplan-Meier analysis with DSS as a primary endpoint showed an estimated survival of 77.8% (95% CI 70.1–85.5%) at 5 years and 67.4% (95% CI 59.4–75.4%) at 10 years.

A worse DSS was found in patients who had preoperative CRP >0.5 mg/dL (66.5 vs. 85.9% at 5 years, p = 0.002; Fig. 1A). Similarly, a worse DSS was observed regarding preoperative NLR (≤/>3.5; 70.2 vs. 80.3% at 5 years, p < 0.001; Fig. 1B). Grade 1 tumors had the best DSS rates (100%) compared to grade 2 and grade 3 MFS (86.5 and 73.8%, respectively at 5 years, p = 0.037; Fig. 1C). A better DSS was observed in superficial MFS (88.2 vs. 71.0% at 5 years, p = 0.046), while tumors larger than 10 cm MFS had a worse DSS (51.8 vs. 92.5% at 5 years, p < 0.001).

Patients who received ChT had a worse outcome, regardless of the timing (neo/adjuvant; p = 0.028). Locally recurred MFS, considered as a time-dependent covariate, had a worse OS (84.7 vs. 61.8% at 5 years, p = 0.021).

Multivariate analysis was carried out to identify any prognostic factor for DSS: tumor size and grade as well as preoperative CRP values (≤/>0.5 mg/dL) and NLR (≤/>3.5) were confirmed to be prognostic factors in terms of DSS (Table 2). LR was observed in 28 patients (22.2%) after a median of 18 months (range, 4–73).

An LR rate of 35.1% (95% CI 28.8–41.4%) at 5 years and 40.5% (95% CI 33.4–47.6%) at 10 years was found. A tendency toward a worse local control of the disease was observed in patients with preoperative CRP >0.5 mg/dL at short-term follow-up (p = 0.083), whereas no differences were observed among NLR values (p = 0.602; Fig. 2A). A higher LR was observed in patients with high (grade 2–3) Gd enhancement at MRI (39.0 vs. 10.4% at 5 years, p = 0.002) and in those with high myxoid (type 2–3) characteristics (42.2 vs. 21.5% at 5 years, p = 0.012; Fig. 2B, C). Patients who had a tail sign at MRI also had an increased risk of LR (49.3 vs. 24.3% at 5 years, p = 0.023; Fig. 2D). No LR was observed in grade 1 MFS, whereas no differences were observed between grades 2 and 3 (p = 0.807). Margins assessed according to Enneking classification and RT seemed not to affect the risk of LR (p = 0.620 and 0.499, respectively).

Multivariate analysis was carried out to identify any prognostic factor for LR: the presence of a tail sign as well as high Gd enhancement (grade 2–3) at preoperative MRI were confirmed to be prognostic factors in terms of LR (Table 3).

Systemic inflammation in STS was reported to be indicative of aggressive tumor characteristics [12-14], inferior DSS, and an increased risk of LR [13, 15]. However, data regarding the influence of CRP blood levels before initial surgical treatment on clinical outcome of STS patients are sparse. The existing studies investigated the role of preoperative CRP and in heterogeneous cohorts of patients (various histotypes, different sites – extremities and trunk/neck, and primary/recurrent STS) [12-15]. MFS is an STS with the highest median macrophage infiltration and T-cell infiltration [20, 21]. Ogura et al. [22] observed that the average fraction of infiltrated CD8+ T cells was significantly higher in the better prognostic cluster in MFS. Hu et al. [23] reported that a high level of CD4+ T-cell infiltration was associated with significantly better relapse-free survival [24]. To the best of our knowledge, this is the first study in which primary, localized MFS of the extremities are evaluated separately to assess the role of systemic inflammation markers as prognostic factors.

Previous series on STS defined different cut-off values for CRP and NLR. Thus, on the basis of median and mean values of the present series, we set 2 different cut-off values for CRP (0.5 and 2.0 mg/dL) and NLR (2.8 and 3.5). As previously reported for STS, in our cohort systemic inflammation markers were associated with aggressive tumor characteristics [12-14]. The main finding of this series is that elevated pretreatment CRP levels (>0.5 mg/dL) and NLR (>3.5) were associated with decreased DSS rate in 126 adult patients affected by localized MFS of the extremities, both in univariate and multivariate analyses. Our results are concordant with those of Nakamura et al. [15] and Szkandera et al. [13], who observed that elevated CRP values (>1.4 and >0.69 mg/dL, respectively) may be associated with inferior survival rates in patients affected by STS.

This is not surprising because it is known that systemic inflammation can be associated with cancer development and progression [25]. The molecular basis of the relationship between elevated plasma CRP levels and poor clinical outcome in cancer patients has not been fully elucidated, and several possible explanations have been postulated. First, tumor growth can induce tissue inflammation and hence increase the CRP levels [26]. Second, CRP could represent an indicator of an immune response of the host-to-tumor antigens [27]. Third, cancer cells and inflammatory cells of the tumor microenvironment, such as granulocytes, lymphocytes, and macrophages, could increase the production of inflammatory cytokines, which may induce CRP production in cancer patients [26, 28]. In multivariate analysis, we also confirmed our previous findings [4]. DSS was correlated with the histological tumor grade. As far as the tumor grade is concerned, the main difference in outcome was observed between grades 2 and 3 tumors, while less relevant differences were identified between grades 1 and 2 MFS. Additionally, tumors larger than 10 cm had a worse DSS.

Patients who received ChT had a worse outcome in univariate analysis, regardless of the timing (neo/adjuvant). These data lost significance at multivariate analysis. Patients who received ChT were unfavorably selected: the choice being a clinical one performed on an individual basis. This makes any consideration in this sense impossible. The role of ChT in MFS is still debated [29], and this series cannot help in resolving it. Considering that MFS mostly affects the elderly population, it might be of paramount importance to arrange adjuvant treatments such as ChT only for higher-risk patients.

The incidence of LR was definitely higher than that of STS in general (approx. 10%). The LR rate remained stable both in the medium- and long-term follow-up. This emphasizes the need for a strict local control, especially during the first 5 years.

Differently to the reports of Nakamura et al. [15] in STS, we did not find any association between preoperative CRP and NLR values and LR rate. This may possibly be due to the different cohorts that have been analyzed. It is known that STS behavior can be extremely different according to histotype and anatomic region [2]. As we previously reported [4, 9, 11], in multivariate analysis we found an increased risk of LR in MFS with an infiltrative growth pattern (tail sign) and high Gd enhancement at MRI.

There is a limitation to the present work that must be acknowledged, and that is the fact that it is a retrospective study. Nevertheless, we report data from a selected cohort of patients including only adult patients affected by primary, localized MFS of the extremities. Moreover, we lacked data on systemic diseases and different drugs such as angiotensin-converting enzyme inhibitors, angiotensin blockers, and statins, which may be associated with possible higher levels of markers of inflammation. The limited follow-up in some patients will also somewhat affect the results, although the statistical methods used have largely taken this into account. Further studies may be required to confirm these results, and a validation will be performed on a large-scale independent database prospectively.

In conclusion, our single-center experience clearly confirms the potential prognostic utility of preoperative plasma CRP and NLR levels as independent prognostic markers in MFS patients who underwent curative surgical resection. We recommend routine measurement of the CRP level as well as neutrophil and lymphocyte counts, as these tests are familiar to most physicians and are readily accessible. These results suggest that patients with high preoperative CRP and NLR levels, as well as large and high-grade tumors, might be considered as candidates for additional, more aggressive treatment approaches or more stringent follow-up schedules. Higher-risk patients may be suitable for future trials of intensive therapy (e.g., ChT).

The study was approved by the local Ethics Committee (Istituto Ortopedico Rizzoli, Bologna, Italy). All patients gave their written informed consent.

All authors have no conflicts of interest to declare.

The authors did not receive any funding.

A.S., D.M.D., and M.D.P. conceived and designed the analysis. R.Z., P.S., C.G., L.C., and M.F. collected the data. G.B., P.S., M.F., and C.G. contributed data or analysis tools. A.S., M.F., and R.Z. performed the analysis. A.S., D.M.D., M.D.P., and G.B. wrote the paper.