• malignant fibrous histiocytoma;
  • paraneoplastic syndrome;
  • C-reactive protein;
  • tumor marker;
  • prognosis


  1. Top of page
  2. Abstract

Paraneoplastic syndromes (PNSs) associated with mesenchymal tumors are uncommon. Previous reports sporadically described inflammatory PNSs with elevated serum C-reactive protein (CRP) levels and leukocytosis in patients with inflammatory malignant fibrous histiocytoma (MFH) of soft tissue; however, the relationship between other subtypes of MFH and PNS has not been extensively investigated. Forty-six patients with primary MFH of soft tissues who underwent radical surgery were retrospectively analyzed. These patients were divided into 2 groups according to preoperative serum CRP level: normal (<1.0 mg/dl) and elevated (≥1.0 mg/dl). The correlation between serum CRP level and several clinicopathologic factors was analyzed. Correlation between preoperative serum CRP level and metastasis-free and overall survival was also investigated by univariate and multivariate analyses. Elevated preoperative serum CRP levels were found in 65% of patients with a mean of 3.7 mg/dl. There were statistically significant relationships regarding tumor size, depth, histologic subtypes, grade, stage and metastatic rate among normal and elevated CRP groups (p < 0.001, p < 0.02, p < 0.005, p < 0.001, p < 0.001 and p < 0.05, respectively). When the tumor was removed, the elevated CRP level subsided into the normal range and other abnormal laboratory findings diminished in all cases. In 11/14 relapsed cases that showed elevated CRP preoperatively, the serum CRP level re-elevated with tumor relapse. The normal CRP group showed significantly more favorable prognosis than the elevated CRP group in metastasis-free and overall survival on univariate analysis (p < 0.02, p < 0.05, respectively). Patients with MFH frequently present with an inflammatory PNS, such as elevated serum CRP level, which can be a useful marker of disease activity and a valuable prognostic indicator. © 2002 Wiley-Liss, Inc.

Malignant fibrous histiocytoma (MFH) is one of the most common soft tissue sarcomas in late adult life.1 MFH manifests variable histologic subtypes, including pleomorphic-storiform, myxoid, giant cell and inflammatory types.2 Inflammatory MFHs have been sporadically reported to demonstrate systemic inflammatory symptoms, including fever, leukocytosis with neutrophilia or eosinophilia, accelerated erythrocyte sedimentation rate and elevated serum C-reactive protein (CRP) levels.3, 4 These paraneoplastic syndromes (PNSs) disappear after surgical resection and recur when the tumor relapses. However, such unusual manifestations have rarely been documented in other histologic subtypes of MFH.5

In our study, the frequency of PNS in MFH was determined and serum CRP levels were evaluated continually as a representative marker of PNS. These data were then correlated with several clinicopathologic factors and prognosis.


  1. Top of page
  2. Abstract

Between March 1990 and August 2001, 80 patients with MFH of soft tissues were treated in the Departments of Orthopedic Surgery at Osaka University Medical School, Osaka Medical Center for Cancer and Cardiovascular Diseases and Osaka National Hospital. We obtained informed consent from each patient. Patients with recurrent disease or distant metastasis at diagnosis and those without laboratory data were excluded. Data on 46 primary MFHs were eligible for retrospective analysis. The diagnosis of all cases was based on light microscopic examination with a hematoxylin and eosin stain according to previously described criteria and confirmed immunohistochemically using specific differentiation markers.1 Tumor cells were negative for markers of specific differentiation, such as α-smooth muscle actin, desmin, S-100, factor VIII and epithelial membrane antigen (EMA).

There were 22 men and 24 women, ranging in age from 37–86 years (median 61 years) at diagnosis. Histologic subtypes were 37 pleomorphic-storiform, 8 myxoid and 1 giant cell. Four patients had low-grade, 8 intermediate and 34 high-grade histology. Histologic grade was determined by cellularity, extent of necrosis and number of mitoses.6, 7 Two patients were in stage I, 7 in stage II and 37 in stage III. Clinical staging was determined based on the revised recommendations of the American Joint Committee on Cancer.8 All patients underwent wide local excision of primary tumors. Twenty-nine patients underwent adjuvant chemotherapy. Adjuvant chemotherapy was given to patients with high-grade sarcomas excluding elderly patients over 70 years. Chemotherapy consisted mainly of doxorubicin with several combinations of cisplatinum, ifosfamide, vincristine and dacarbazine. Adjuvant radiotherapy was given to 8 patients postoperatively. The radiation dose ranged from 45–66 Gy (median 52 Gy). The follow-up period for survivors ranged from 11–106 months (median 30 months). Serum CRP levels were measured mainly using the Hitachi (Tokyo, Japan) 7250 autoanalyzer as part of a routine clinical examination.

Although the normal serum CRP level was <0.3 mg/dl, levels up to 1.0 mg/dl were regarded as minor or insignificant elevations found occasionally in apparently healthy individuals.9 We therefore defined cut-off values above 1.0 mg/dl as abnormal. We defined 2 groups: normal CRP level (<1.0mg/dl) and elevated CRP level (≥1.0 mg/dl). Clinicopathologic analysis was performed on factors including age, gender, size, depth, histologic grade, histologic subtypes, stage, local recurrence and metastasis.

Statistical associations of these factors were evaluated using the χ2 test. The duration of overall or metastasis-free survival was defined as the interval between the date of diagnosis and that of death or metastasis, respectively. Survival curves were constructed using the Kaplan-Meier method.10 The log-rank test was used to compare the survival of the 2 groups of patients.

Multivariate analyses based on the Cox proportional hazards model were performed to identify the most significant factors related to metastasis-free survival and overall survival. Statistical analysis was performed using the JMP statistical software package (version 3.1; SAS, Cary, NC). p < 0.05 was considered significant.


  1. Top of page
  2. Abstract

Pretreatment frequency of PNS (serum CRP levels and other laboratory findings)

Table I shows pretreatment serum CRP levels in relation to several physiologic parameters. Elevated serum CRP level (≥1.0mg/dl) was seen in 30/46 patients (65%). The mean value was 3.7 mg/dl (range 1.0–15.8 mg/dl). In 18/30 cases (60%) showing elevated serum CRP, other abnormal laboratory findings were recorded, such as leukocytosis, anemia, thrombocytosis, elevated fibrinogen, hypoalbuminemia and hepatic dysfunction. The frequency of manifestations such as leukocytosis, thrombocytosis, hypoalbuminemia and hepatic dysfunction was significantly higher in the elevated CRP group than in the normal CRP group (p < 0.01, p < 0.05, p < 0.02, p < 0.05, respectively). Furthermore, in the normal CRP group, only 1 case presented anemia.

Table I. Association between Serum CRP Levels and Other Laboratory Findings1
Laboratory findingsElevated CRP group (≥1.0 mg/dl, 30 cases)Normal CRP group (<1.0 mg/dl, 16 cases)p
  1. N.S., not significant; WBC, white blood cell; Hb, hemoglobin; Plt, platelet; GOT, glutamic-oxaloacetic transaminase; GPT, glutamic pyruvic transaminase.

Leukocytosis (WBC ≧ 10,000/mm3)11/30 (36%)0/16 (0%)<0.01
Anemia (Hb ≦ 10 g/dl)7/30 (23%)1/16 (6%)N.S.
Thrombocytosis (Plt ≧ 400,000/mm3)8/30 (26%)0/16 (0%)<0.05
Elevation of fibrinogen (fibrinogen ≧ 450 mg/dl)10/14 (71%)0/1 (0%)N.S.
Hypoalbuminemia (albumin ≦ 3.5 g/dl)9/24 (37%)0/13 (0%)<0.02
Hepatic dysfunction (GOT, GPT ≧ 40 U/I)7/24 (29%)0/14 (0%)<0.05
18/30 (60%)1/16 (6%)<0.001

CRP as a tumor marker

In the elevated CRP group, 14 patients had local recurrence and/or distant metastasis. Ten of 12 metastatic cases (83%) and 4/7 locally recurrent cases (57%) showed re-elevation of serum CRP. In total, 11 of the relapsed cases (78%) demonstrated re-elevation of serum CRP levels. In the normal CRP group, local recurrence or distant metastasis was found in 5 patients. Serum concentrations of CRP were measured again when the recurrent tumors were discovered. CRP levels in this group were normal in all cases after tumor relapse (Table II).

Table II. Re-elevation of Serum CRP Levels at Tumor Relapse
Preoperative CRP groupNumber of patients with re-elevation of CRP
Local recurrenceMetastasisTotal
Elevated CRP group4/7 (57%)10/12 (83%)11/14 (78%)
Normal CRP group0/40/10/5

Relationship between CRP levels and clinicopathologic variables

The correlation between clinicopathologic features and pretreatment serum CRP level is shown in Table III. There are significant correlations between CRP and tumor size, depth, histologic subtype, histologic grade, American Joint Committee on Cancer stage and metastasis (p < 0.001, p < 0.02, p < 0.005, p < 0.001, p < 0.05 and p < 0.05, respectively). Other factors, such as age, gender and local recurrence, had no significant relationship to pretreatment serum CRP levels.

Table III. Correlation between Pretreatment Serum CRP Levels and Clinicopathologic Features1
Patients characteristicsElevated CRP group (≥1.0 mg/dl)Normal CRP group (<1.0 mg/dl)p
  • 1

    AJCC, American Joint Committee on Cancer; N.S., not significant.

Age (years)
Size (cm)
Histologic subtype
 Myxoid/giant cell2/06/1
Histologic grade
AJCC stage
Local recurrence

Prognostic value of serum CRP before treatment

As shown in Table III, a significantly higher probability of metastasis was seen in patients with elevated serum CRP level (≥1.0 mg/dl) compared to those with normal serum CRP level (<1.0 mg/dl) (p < 0.05). The difference between the 2 metastasis-free survival curves was statistically significant (p < 0.02) (Fig. 1). Univariate analysis confirmed that size, grade, stage and CRP had prognostic significance in terms of metastasis-free survival (p < 0.05, p < 0.05, p < 0.02 and p < 0.02, respectively) (Table IV). When subjected to multivariate analysis, grade remained significantly associated. Univariate analysis showed that the following factors were significantly related to overall survival: CRP (p < 0.05) and stage (p < 0.05). In the multivariate analysis, CRP lost prognostic significance. Figure 2 shows the overall survival curves according to preoperative serum CRP levels.

thumbnail image

Figure 1. Distant metastasis according to preoperative serum CRP levels: ≥1 mg/dl (n = 30) or <1 mg/dl (n = 16), p < 0.02.

Download figure to PowerPoint

Table IV. Univariate Analysis of Metastasis-Free and Overall Survival
Clinicopathologic factorsUnivariate p
Size (≧10 vs. < 10 cm)0.03240.0572
Depth (deep vs. superficial)0.29970.3931
CRP (≧1.0 vs. <1.0 mg/dl)0.01070.0327
Grade (high vs. intermediate and low)0.04280.0639
Stage (III vs. II and I)0.01640.0416
thumbnail image

Figure 2. Overall survival based on preoperative serum CRP levels: <1 mg/dl (n = 16) or ≥1 mg/dl (n = 30), p < 0.05.

Download figure to PowerPoint


  1. Top of page
  2. Abstract

Neoplasms can produce a variety of remote effects on the host; these are referred to as PNSs.11 PNS was first described by Viallet.12 The following criteria are generally used for PNS: (i) the tumor produces symptoms and signs at a distance from itself or its metastasis, (ii) the course of the PNS runs parallel to the disease course and (iii) paraneoplastic symptoms and signs subside soon after tumor removal. To date, several cases regarding soft tissue sarcomas with PNS have been reported. The majority of the previously reported cases are coupled to inflammatory MFH characterized by the microscopic appearance of foamy cells and marked infiltration of inflammatory cells admixed with a variety of atypical cells.3, 4 This inflammatory MFH commonly arises in the retroperitoneum and may be associated with systemic symptoms including fever and leukocytosis.1, 2 However, PNSs in other soft tissue sarcomas, including other subtypes of MFH, are rarely reported and the frequency remains unknown.13–15 Here, we have determined that PNS with elevated serum CRP levels frequently occurs in MFH patients (65%, 30/46), particularly in those with pleomorphic-storiform type (93%, 28/30); our series did not include inflammatory MFH.

In advanced malignant neoplasms, CRP may be elevated as a metabolic abnormality. In our series, we analyzed serum CRP levels continually from first presentation to clarify whether CRP might be useful as a tumor marker.

In MFH patients with elevated serum CRP at first presentation, serum CRP levels decreased when wide local excision or adjuvant chemotherapy was performed effectively. Following normalization of the serum CRP level after appropriate treatment, re-elevation occurred on tumor recurrence or metastasis in nearly 80% of relapsed cases. In many relapsed cases, elevation of CRP preceded the detection of recurrent disease (data not shown). These results suggest that serum CRP is a valuable tumor marker for MFH patients with elevated serum CRP at first presentation. Therefore, it is necessary to measure serum CRP in MFH patients continually for early detection of recurrent diseases.

As for soft tissue sarcomas other than MFH, most show normal CRP levels below 0.3 mg/dl before treatment without PNS, though we have seen inflammatory PNS occur in leiomyosarcoma cases (data not shown). Since the pathologic entity of MFH was introduced and subsequently popularized in the 1960s and 1970s, it has become widely regarded as the most common soft tissue sarcoma in adults. Several studies have shown considerable doubt about MFH as a pathologic entity and have suggested that MFH might represent a common morphologic manifestation of poorly differentiated sarcomas in general.2 A molecular study has revealed that MFH and leiomyosarcomas share very similar comparative genomic hybridization imbalance profiles, suggesting that MFH is a morphologic modulation in the tumorigenesis of other sarcomas, particularly leiomyosarcoma.16 In addition, Hisaoka et al.17 reported a dedifferentiated liposarcoma with an inflammatory MFH-like component presenting with a leukemoid reaction. Taken together, these results suggest that PNS in MFH might be a manifestation of advanced sarcoma tumorigenesis. The frequency of PNS in a large number of various soft tissue sarcomas, however, needs to be determined.

The mechanism responsible for PNS in MFH is not fully understood. Release of cytokines or growth factors from the tumor, with distant effects on the hematopoietic system, has been postulated. Melhem et al.18 reported the immunohistochemical detection of various cytokines, IL-6, IL-7, IL-8, IFN-γ and keratinocyte growth factor in 2 inflammatory MFHs presenting with leukemoid reaction. They also serologically detected elevated serum levels of IL-6 and tumor necrosis factor in both tumors. Takahashi et al.19 speculated that a colony-stimulating factor, presumably produced by tumor cells, might be responsible for the leukemoid reaction in 1 case of inflammatory MFH. Thus, previous reports suggest that several cytokines play a role in the local and systemic inflammatory effects of MFH.

Serum CRP concentration reflects IL-6 activity. IL-6 is a pleiotropic cytokine acting on several cell lineages. At the hepatocyte level, it stimulates the synthesis of acute-phase proteins, such as CRP.20 Serum IL-6 levels were measured in 7 MFH patients with elevated serum CRP and all showed high levels, which diminished after surgery and re-elevated following tumor recurrence (data not shown). Hamada et al.5 immunohistochemically demonstrated that IL-6 localized to MFH cells. Several reports suggest that MFH itself might produce IL-6, accounting for the local and systemic inflammatory effects.19, 21 We also established 2 human MFH cell lines derived from patients with inflammatory PNS, which showed overexpression of IL-6 (data not shown). These findings suggest that IL-6 may be a tumor-associated factor involved in the pathogenesis of tumor progression and its clinical manifestations.

In our series, it is of particular interest that tumors associated with elevated CRP were larger than those with normal CRP. Moreover, the pleomorphic-storiform type predominated in the elevated CRP group over the myxoid type. These results suggest that cytokines might be responsible for not only the proliferation of tumor cells but also the maintenance of their morphology, as reflected by histologic subtype.

Although the number of MFH cases here is relatively small and follow-up was short, univariate analysis for survival data indicates that elevated CRP might result in poorer prognosis. Le Doussal et al.22 analyzed 216 patients with localized primary MFH and showed that histologic grade and tumor size were significantly related to metastasis-free survival in those with the poorest prognosis. These results are consistent with our own, supporting our finding that elevation of serum CRP levels, which appears to be IL-6–related, is associated with an increased risk of distant metastasis and overall survival.

In conclusion, inflammatory PNS with elevated serum CRP is a relatively frequent manifestation of MFH. Serum concentrations of CRP may be an objective biomarker of disease activity and of prognostic significance in MFH.


  1. Top of page
  2. Abstract