The solitary fibrous tumor of the pleura (SFTP) is a rare neoplasm that accounts for less than 5% of all pleural tumors.1 Its incidence is approximately 2.8 cases out of 100,000 registered hospitalized patients, and it usually occurs in all age groups with a peak incidence in the fourth and sixth decade of life.2,3 Since Klemperer and Rabin4 first described its pathologic characteristics in 1931, more than 900 cases of SFTP were reported in the literature up to 2005.5 According to several different authors, SFTPs have been characterized by dishomogenous taxonomies.6,7 The absence of a clear nomenclature has only added to the confusion about these tumors, where uncertainty exists with regard to their clinical behavior; also, a sound classification with diagnostic and prognostic potential remains far from reach. The development of electron microscopy in the early 1980s and the introduction of immunohistochemistry techniques in the standard pathologic assessment of neoplasms have clarified that the tumor does not originate from the mesothelial layer but from the submesothelial, noncommitted mesenchymal layer.8,9 Thus, the various names used for this disease have slowly become unified, and the disease is now referred to as solitary (or localized) fibrous tumor of the pleura. Although the majority of these neoplasms have benign features (histological and clinical), approximately 10% to 20% of the cases reported in literature are malignant.10,11 Criteria for distinguishing pathologically benign forms from malignant ones have been established by England and colleagues,11,12 and these are now widely accepted. Easily controlled by surgical resection, benign forms have been reported to recur locally14 (this has been hypothesized as related to a malignant transformation). Surgery, as well, remains the standard of treatment for malignant forms. The clinical decision-making process in malignant (m) SFT (mSFT) management is supported, so far, by series of single-case reports.1,12–15 To the best of our knowledge, in fact, the number of mSFTP reported cases is far too small for a substantial analysis of recurrence or survival rates: these, in fact, are inconsistently reported. The result is that no established treatment modality or follow-up plan has been yet agreed to. Also, the prognostic features of mSFTP do not seem to correlate with its histological characteristics and this remains a point to be further clarified.11,16,17
We established a multicenter collaboration among high-volume thoracic surgery centers to collect a consistently homogeneous population (the largest so far in the English literature); we analyzed this population by reviewing its clinical presentation, histopathologic characteristics, treatment modalities, and long-term survival (LTS) results. The outcome of this analysis is reported herein.
PATIENTS AND METHODS
In the period between January 2000 and July 2010, 50 patients were surgically treated for malignant SFTP in the four centers involved in the present analysis (Catholic University of Rome [n = 9], Forlanini Hospital of Rome [n = 17], European Institute of Oncology of Milan [n = 8], and University of Turin n = 16]) and these formed the basis for this retrospective analysis. Before undertaking our data analysis, we obtained the approval of the Catholic University (Promoting Centre) Institutional Review Board for research use of retrospectively collected data (observational) stemming from standard clinical practice. Data related to sex, age, tumor location, signs and symptoms, surgeons’ notes, pathological features, postoperative therapy, recurrence patterns, and long-term follow-up were systematically reviewed and recorded (see Table 1).
In all the patients, the preoperative evaluation included history, physical examination, routine blood tests, standard chest radiographs, and thoracic computed tomography (CT) scan. Fiberoptic bronchoscopy was performed in case of lesions located proximally to major airways. Electrocardiography, spirometry, and arterial gas analysis were routinely performed and noted. Perfusion lung scan was performed only in selected cases as for echocardiography, which was carried out only if a history of cardiovascular disease was present or if a pneumonectomy was anticipated. 18-Fluorodeoxyglucose positron emission tomography/CT was performed in two cases only. At pathological evaluation, tumors were eventually reclassified as malignant when one or more of the criteria reported by England et al.9 were met: (1) mitotic count more than 4 mitoses/10 high-power fields; (2) presence of necrosis; (3) hypercellularity as judged by nuclear crowding and overlapping; and (4) presence of nuclear atypia.
Flow-cytometric analysis was performed on fresh material. Immunostaining for a panel of cell markers including bcl-2 (apart from two cases), CD34 was routinely done in all four centers. Upon pathological re-review, immunohistochemistry was performed on formalin-fixed, paraffin-embedded tissue, using the streptavidinbiotin immunoperoxidase method, in addition to standard hematoxylin and eosin staining (three sections for each case). In fact, the following antibodies were not routinely used: cytocheratin (AE1–AE3) and vimentin.
A centralized pathological revision of the samples was performed by an expert—dedicated in lung pathology—pathologist (PG) to avoid any variation in pathological diagnosis and to achieve a concordance on the histopathological characteristics of the samples reviewed.
Postoperative radiotherapy and/or chemotherapy were performed under the care of referring oncologists: although a certain degree of homogeneity was present within the four centers, this was lost when the centers are compared among themselves. After completion of the study (July 2010), information about the health status of patients was obtained by the clinical records of outpatient clinics.
The t test, the χ2 test and the Fisher’s exact test were used where appropriate. Actuarial patient survival and disease-free survival (DFS) were obtained by the Kaplan-Meier method. Comparisons between survival curves were done with the log-rank test. The Cox multiple regression analysis was subsequently applied and the risk factors for both the mortality and the relapse of the tumor were identified. Differences were considered significant at the level at which the p value was less than 0.05. All statistical analyses were conducted by using SPSS 13.0 for Windows (SPSS, Chicago, IL).
Demographics and clinical and pathological features are summarized in Table 1. There were 26 women and 24 men with age ranging 44 to 87 years (mean ± SD: 66.0 ± 10.4 years). The tumor was right-sided in 29 patients (58%) and left-sided in 18 (36 %); in two cases the neoplasm was located in the mediastinum, and one patient presented with bilateral pleural lesions. At the time of diagnosis, 32 patients (64%) presented symptoms related to pleural fibroma, although no history of asbestos exposure was recorded; when symptoms were present, the predominant ones were dyspnea, chest pain, and coughing. Eight patients (16%) had more than one symptom. Mild-to-severe hypoglycaemia (Doege-Potter’s syndrome) was observed in three cases (6%), and in one patient it was associated with weight loss, fatigue, and dysarthria. Eighteen patients (36%) were totally asymptomatic and pleural fibroma was incidentally found on a chest radiograph or more rarely on a CT scan performed for other reasons.
Chest radiographs were obtained for the entire study group. In all patients but one, a preoperative CT scan was carried out. Usually, the density of the neoplasm was quite low (30–45 UH) and heterogenous to different extents in large tumors because of myxoid degeneration and localized hemorrhage or necrosis (Fig. 1A). Magnetic resonance imaging was obtained in four patients (8%) only.
A moderate-to-large malignant pleural effusion was observed in 12 patients (24%) (bloody fluid except in two cases). Material for cytological examination by transthoracic fine needle aspiration biopsy (FNAB) was obtained in 32 cases (64%). The analysis showed the presence of malignant mesenchymal cells compatible with an mSFTP in 16 cases. This assessment provided a generic diagnosis of mesenchymal tumor with no other specifications in eight cases and failed to provide any result in the remaining eight.
Bronchoscopy was performed in 10 patients (20%). No abnormalities were found in seven cases; an extrinsic compression was present in three, and cytology on transbronchial biopsy confirmed the diagnosis of mSFT in one of them. Overall, a certain diagnosis had been obtained in 17 patients (34%).
Surgical Findings and Perioperative Outcome
All the patients were treated in one of the four institutions as reported above. The operative approach and surgical procedures are summarized in Table 1. Tumors arising from the parietal pleura were more difficult to resect than those from the visceral pleura because they were large, affixed to the chest wall, and required extrapleural resection or extended resection.
When reviewing the surgical behavior in the four participating centers it was clear that the extent of resection was decided on a case-by-case basis when the situation was appraised intraoperatively. In particular, an isolated mass excision was performed in 13 cases (26%) only whereas an en bloc resection of surrounding structures was carried out in all the remaining cases, as reported in Table 1. In six patients, a formal lung resection (lobectomy [n =2] and pneumonectomy [n = 4]) was necessary because of large tumors extending deeply in the lung parenchyma. A chest-wall resection was needed in 10 cases (20%) whereas in two cases the lung resection was extended to the diaphragm and/or the pericardium. In the three cases in which the neoplasm was deemed unresectable or inoperable (extension to adjacent structures/pleural malignant colonization—not previously detected at the preoperative workup) only a biopsy was performed.
Intraoperatively, a polypoid/pedunculated lesion (Fig. 1B) was observed in 11 cases (27%); 31 (73%) presented with sessile tumors. In eight cases the surgical report did not contain any indication about the gross morphology of the tumor and this could not be understood from the final pathology. The visceral pleura was considered the point of origin in 30 patients (60%); and the parietal pleura in 16 (32%). In eight cases (16%) it was not possible to discriminate the exact point of origin of the disease because it extended largely into the pulmonary parenchyma and infiltrated the chest wall. Margins were intraoperatively confirmed free of tumor in all those cases (especially in the sessile morphology group) in which a doubt regarding the completeness of resection was present: a complete resection was achieved, as confirmed by final pathology, in all but four cases (92%) with three (6%) gross residual tumor (R2) situations (2 pleural carcinosis and 1 infiltration of the aortic arch) and one microscopic (R1), in a case where the neoplastic mass measured 25 cm circa in its maximum diameter and largely infiltrated the chest wall.
Postoperative Results and Adjuvant Therapy
The postoperative recovery was uneventful in 37 patients (74%); complications occurred in 13 patients (26%) and these included: atrial fibrillation (n = 3), pneumonia (n = 3), bleeding (n = 3), which in one case required a reintervention, pulmonary embolism (n = 2), persistent air leak (n = 1), and wound infection (n = 1). Two patients (4%) died in the perioperative period. One patient (a 78-year-old man) with no prior history of coronary artery disease died suddenly on postoperative day 3 because of a myocardial infarction after a left extended pneumonectomy with pericardial reconstruction for a very large tumor (24 cm of maximum diameter). The second patient (a 77-year-old woman) died on postoperative day 9 because of pneumonia complicated by massive pulmonary embolism after a right upper lobectomy with combined chest-wall resection.
Fifteen patients (30%) underwent adjuvant treatment (radiotherapy in 11, chemotherapy in 3, and combined radiochemotherapy in 1 case). Clearly, the adjuvant treatment was substantially not homogeneously administrated in the cohort of patients analyzed in this report in which the clinical decision-making process was established by each single center.
The tumors measured from 30 × 18 × 22 to 320 × 250 × 170 mm (mean maximum axis =12.8 ± 6.5 cm; range, 2–32 cm) and the largest one weighed 2130 g (Fig. 1C). Grossly, the majority of the tumors were well circumscribed, lobulated in appearance, and firm in consistency; the cut surface was vaguely nodular storiform, occasionally microcystic and myxoid, and white-gray. On cut sections, tumors showed a whorled appearance, sometimes alternating with some myxoid areas.
Microscopically, the typical pathological findings of mSFTs were confirmed in all the specimens; in particular, a patternless pattern characterized by a proliferation of uniform elongated spindle cells with a various amount of connective tissue was detected in most of the cases. Sometimes, specifically fibrous or hypercellular patterns, along with mixed bundles of tissue with marked amounts of blood vessels were seen in different sampled areas of the tumor (Fig. 2A). The diagnosis of malignancy was based on the four criteria by England et al., which we evaluated and eventually matched one by one, also arbitrarily extracting a risk-score (0–4) (Table 1) thereafter to be correlated with prognosis.
The mitotic count was also evaluated in all patients and this ranged from 1 to 36 mitosis for 10 high-power fields (mean 10.02 ± 8.2). The proliferation index ki-67% was calculated in 38 of the 50 patients (mean ki-67% = 14.8% ± 11%, range, 1%–60%).
The immunohistochemical analysis showed in all tumors a positivity for vimentin and CD-34 (Fig. 2B and C), although no expression of low molecular weight cytokeratin and of desmin was detected. A clear positivity for bcl-2 was found to be present in almost the entire cohort of patients (96%).
The follow-up was completed in 100% of the cases at the time of writing this report (median 52.5 months, range, 1–194). Censoring found 40 patients alive and 10 dead. The overall LTS (Kaplan–Meier) was 81.1% (95% CI: ± 10.2%) at 5 years and 66.9% (95% CI: ± 18.6%) at 10 years (Fig. 3A).
The most significant prognostic factor in our series proved to be the completeness of resection. In fact, patients with an R+ disease did not reach the 5-year survival, with a striking difference with R0 ones (overall: 87.1%, p < 0.001) (Fig. 4A). Interestingly, no differences in LTS was found when attempting a stratification of patients according to (1) the size of the tumor, (2) its pleural pattern (tumor arising from visceral versus parietal pleura), or (3) the pedunculated versus sessile morphology (Table 2). No evidence of different survival functions was found, similarly, stratifying for age and sex. Moreover, as stated above, we calculated a score (0–4) based on England’s criteria (Table 1); no significant differences were found when comparing the four classes with a trend for better survival in scores 1 and 2 (100% alive at 5 years) if compared with scores 3 and 4 (90% and 79.6% alive at 5 years, respectively) (Table 2).
LTS was significantly worse for those patients with malignant pleural effusions if compared with those without (5-year LTS, 85.5% versus 69.4%; p = 0.04; (Fig. 4B). No further statistically significant difference was found, considering the other characteristics of the population, and this is very probably because of the relatively small number of observations.
As reported in Table 3, the Cox multiple regression analysis confirmed that the incomplete resection represents a risk factor for earlier death. In particular, it was estimated that patients with R+ resections had a risk of death 39.02 times higher than patients in whom an R0-resection was achieved (95% CI: 4.04–380.36; p = 0.002). The presence of malignant pleural effusion had a heavily negative impact on survival with an estimated HR of 3.44 (95% CI: 0.98–12.05) and a p value almost statistically significant (p = 0.053). No other additional risk factors were evidenced.
During the follow-up period, 15 patients (30%) experienced a recurrence of the tumor (mean time to recurrence: 34 months, range: 2–128). The recurrence pattern was localized (intrathoracic) in six patients (40%) and diffuse (metastatic) in nine (60%). Interestingly, in the entire population, no significant differences in the survival functions were observed between patients who experienced a recurrence versus those who did not (p = 0.371). In four patients with localized recurrences a re-resection was performed (in one case 3 further iterative resections were needed in a 4-year-long period). In all the other cases (2 with local relapse and 9 with metastases) chemotherapy was administered (in 2 combined with radiation therapy). The chemotherapeutic protocol was not uniform: most patients underwent single-agent treatment (doxorubin) or a combination of doxorubicin and ifosfamide; the clinical response was substantially unsatisfactory in almost all cases.
For the entire population, the 5- and 10-year estimated DFS rates were 72.1% (CI 95%: ± 15.6%) and 60.5% (CI 95%: ± 20.0), respectively. (Fig. 3B). The visual inspection of the disease-free curves suggested four factors potentially associated with the recurrence rate (Fig. 4C and D). In particular, the log-rank test confirmed that the patients who had an mSFTP involving the chest wall showed a lower rate of 5-year DFS (no chest-wall invasion 79.5% versus chest-wall invasion 50%; p = 0.003); similarly, the analysis of survival demonstrated a significantly worse DFS in patients with malignant pleural effusion (no pleural effusion 78.9% versus pleural effusion 55.6%; p = 0.023). Differently, no further pathological or immunohistochemical features were associated with different DFS.
The Cox multivariable regression analysis confirmed that the presence of chest-wall invasion and the malignant pleural effusion both increased the risk of recurrence (HR: 3.48, 95% CI: 1.1%–11.0%, p = 0.038; and HR: 4.34, 95% CI: 1.5%–12.6%; p = 0.002, respectively) (Table 3).
SFTPs are uncommon neoplasms (with just over 900 cases reported in the English literature in decades),5,6 but a pattern of augmented incidence could be detected in the recent years. Although the majority of these neoplasms are benign, approximately 10% to 20%10,11 show a more aggressive biological behavior, with pathological findings of invasiveness, appearance of distant metastases, and local relapse after resection. Sung et al.5 reported that the incidence of malignant SFTP varies from 7% to 60% in the English literature, a variation attributed to slight differences in the pathologic criteria differences used in each institution.
SFTs with malignant morphology and behavior are generally larger (mean tumor size of 12.8 cm in our series) if compared with benign ones,18,19 significantly so in some studies,20,21 though very large tumors can be also benign. Some authors suggest that larger tumors are more likely to be malignant because they have had time to undergo genetic changes.22
Furthermore, the symptoms, if any, are more frequently associated with malignancy.5 Intrathoracic symptoms are usually nonspecific and include cough, chest pain, and dyspnea. Possible causes may be: (1) the irritation of adjacent structures by invasion or peritumoral adhesion, (2) the paracrine action of unknown factors secreted by the mSFTP, or (3) the presence of pleural effusion (which is generally more frequent in those malignant tumors). Indeed, according to England and colleagues,11 pleural effusion affects 16% of the patients, with a higher prevalence (32%) among those with mSFTP (8% in those with benign tumors). In the present series, we could detect a malignant pleural effusion in 12 patients (24%) and, very interestingly, these patients had a significantly worse prognosis in terms of LTS (69.4% versus 85.5%, p = 0.04) and DFS (55.6% versus 78.9%, p = 0.023) if compared with those without any sign of pleural effusion. This finding probably reflects a more advanced stage of the disease in which the presence of pleural fluid can be considered the expression of a pleural direct invasion or dissemination (M+ disease).
There is also evidence that the hypoglycemia has a paraneoplastic syndromic appearance and is related to the production of insulin-like growth factor II by the tumor cells; in fact, it normally resolves after the removal of the tumor itself.5,23 The incidence of hypoglycemia varies from 2% to 4%3,8,14 and topped at 6% in our series (possibly because of the greater representation of malignant forms).
Although the chest CT scanning remains the test of choice, the radiological differentiation between the benign or malignant nature of the neoplasm remains difficult. However, the following findings could suggest malignancy: a diameter exceeding 10 cm, central necrosis, and ipsilateral pleural effusion, according to criteria described in the work by Ferretti et al.24 Nuclear magnetic resonance may provide some help, showing the fibrous character of the lesion25 and by providing a topographic and differential diagnosis and differentiating the tumor from surrounding structures. Although FDG/positron emission tomography-CT has been demonstrated to be useful in selected mSFTP cohorts because it provided some differential diagnostic elements to distinguish benign and malignant variants of the tumor,18,26 further evidence is needed for this methodology to be routinely inserted in the mSFTP clinical workup guidelines. Even the cytological assessment of these neoplasms in the preoperative setting is invariably unsatisfactory because the accuracy of the CT-guided FNAB procedure ranges in levels below 50%.18,20,26,27 We share the opinion that FNAB should be reserved for those cases in which the clinical condition contraindicates surgery (for functional and/or oncological reasons).
Surgery is the first step treatment of choice for mSFTPs and the completeness of the resection represents its primary objective.11,12 An extended resection through standard thoracotomy should be performed to warrant completeness of the resection. Video-assisted thoracic surgery procedures have been sometimes used (in 10% only of our population) because these approaches cannot guarantee the completeness of the resection, especially in large and invasive/malignant tumors. However, in those cases in which the pedunculated morphology is accompanied by the absence of pleural dissemination and/or chest-wall clear infiltration, we believe that video-assisted thoracic surgery can be reasonably used. We suggest intraoperative checking of the surgical margins to confirm the radicality of the excision when this approach is used. In the present series, pulmonary wedge excision was the most common surgical procedure, whereas formal anatomic lung parenchymal resections alone were performed in only six patients (12%). Chest-wall resection alone was required in three patients (6%) whereas in seven patients (14%) it was combined with additional lung parenchymal resection. The operative approach and extent of surgical excision is clearly dictated by the mSFTP’s size, location, and attitude toward surrounding structures.
The final diagnosis of malignancy in the cases reported in this review was established on the basis of histological criteria suggested by England et al.11 These criteria are currently widely accepted and have been employed in the most recent surgical series,12–15,27 and their usefulness is also suggested by the American Registry of Pathology.28 In our series we did not observe the pleomorphic variant as described by Perrot et al.,6 in which isolated pleomorphism can be detected without any other sign of malignancy (England score 1) and that is generally considered as an intermediate form, borderline with frank malignancy.
Regarding the long-term outcome after surgical resection, we may assume that LTS after resection of mSFTPs is possible because it ranges from 46% to 100% in the main surgical series reported until now (77.6% in the present series; Table 4).
In agreement with other authors,5,11,15,18,20,27 our data demonstrated an LTS is dramatically better (84.9% versus 0%) when a complete resection is achieved (Fig. 3A). Univariate analysis shows that patients with invasion of neighboring structures and with malignant pleural effusion have a worse DFS; furthermore, the multivariate analysis confirmed the invasion of neighboring structures as the only independent factor impacting significantly on the recurrence rate of the disease (Table 3). Regarding the tumor size, differently from some studies,11,17,21 we could not confirm the tumor size as a factor significantly impacting survival, in agreement with other authors.18,29 Similarly, we found that prognosis was not affected by the site (visceral or parietal pleura) of origin or by the gross morphologic features of the tumor (sessile versus peduncolated) (Table 2). In this setting, Perrot and colleagues6 classified SFTP on the basis of a combination of gross morphologic features and pathologic subtype (benign pedunculated, benign sessile, malignant pedunculated, and malignant sessile). The authors reported significant differences in the recurrences and survivals of these subgroups and in particular, the recurrences were reported to occur in 63% of those patients with sessile mSFTPs and in 14% of patients with peduncolated mSFTPs. Of all the patients enrolled in the present study, morphopathologic information could be obtained in 42 cases: 31 were sessile mSFTPs and 11 pedunculated. Local recurrence occurred in seven patients (21.6%) and four patients (36.4%). Although our data are not completely in line with those from Perrot et al.,6 we should also consider that: (1) our sample is composed basically by a surgical series, (2) all patients were surgically treated in high-quality and high-volume surgical centers with a very high resection completeness rate (< 92%). The significant differences of the analyzed series could justify this observation. We are of the opinion that the Perrot morphopathological classification, although useful, should be further validated regarding its prognostic potential in large prospective series. On a different note, when data from our cases are matched with England’s criteria, we could detect a linear correlation with DFS shortening according to the increase of the score; in fact mSFTPs with scores 3 to 4 were associated with a DFS rate of 78.8% and 66.4% whereas all cases of mSFTPs with scores 0 to 1 were 100% free from recurrence at the moment of the follow-up.
Therefore, our results, although based on a relatively small cohort of observations, allow us to confirm that England’s criteria may have a prognostic value. Adjuvant treatment did not induce, in our series and in agreement with others,12 any significant benefit to the overall survival of operated patients. On the contrary, we could detect a trend showing a negative impact. Nevertheless, we believe that this observation is clinically irrelevant because of the great heterogeneity of criteria that led, in each center, to the indication of adjuvant treatment. Finally, we could confirm the fact that SFTP, despite their tendency to grow locally, have a malignant behavior that justifies a full long-term follow-up. In the surgical R0 population, the recurrence rate (local plus distant) ranges from 14% to 86%5,12,27,30 (30% in our series, 15 cases of which 9 with evidence of distant metastasis). This behavior also justifies an aggressive surgical strategy where redo resection should be performed in all cases in which it is reasonably possible. Our long-term results on the four redo surgery cases, in line with others,15,31 support this attitude.
Limitations and Strengths
This is a retrospective analysis and, as such, has significant limitations in solving the uncertainties regarding the SFTPs. However, the relatively large number of observed cases, the homogeneity of the cohort, the long-term follow-up information (10 years), and the comprehensive analysis of the prognostic factors we have performed support the conclusion that radical surgery and redo surgery, if performed in selected high-volume centers, are able to achieve a more-than-satisfactory disease control in a biological setting, in which so far, alternative treatments (radio and chemotherapy) have failed to provide significant benefits.
Surgical resection, when technically and medically feasible, is recommended for the treatment of localized mSFTP, and long-term survivals could be expected. Additional prognosis stratifiers must be further investigated and validated.
1. Lu C, Ji Y, Shan F, Guo W, Ding J, Ge D. Solitary fibrous tumor of the pleura: an analysis of 13 cases. World J Surg. 2008;32:1663–1668
2. Komatsu T, Shoumura Y, Tomii K, Ishihara K, Imai Y, Takahashi Y. A case of solitary fibrous tumor of the pleura with reference to its treatment and so-called ambiguous characteristics. Tanaffos. 2005;4:57–59
3. Okike N, Bernatz PE, Woolner LB. Localized mesothelioma of the pleura: benign and malignant variants. J Thorac Cardiovasc Surg. 1978;75:363–372
4. Klemperer P, Rabin LB. Primary neoplasms of the pleura: a report of five cases. Arch Pathol. 1931;11:385–412
5. Sung SH, Chang JW, Kim J, Lee KS, Han J, Park SI. Solitary fibrous tumors of the pleura: surgical outcome and clinical course. Ann Thorac Surg. 2005;79:303–307
6. de Perrot M, Fischer S, Bründler MA, Sekine Y, Keshavjee S. Solitary fibrous tumors of the pleura. Ann Thorac Surg. 2002;74:285–293
7. Chan JKC. Solitary fibrous tumour—everywhere, and a diagnosis in vogue. Histopathology. 1997;31:568–576
8. Hernandez FJ, Hernandez BB. Localized fibrous tumors of the pleura: a light and electron microscopic study. Cancer. 1974;34:1667–1674
9. Al-Azzi M, Thurlow NP, Corrin B. Pleural mesothelioma of connective tissue type, localized fibrous tumor of the pleura, and reactive submesothelial hyperplasia: an immunohistochemical comparison. J Pathol. 1989;158:41–44
10. Robinson LA. Solitary fibrous tumors of the pleura. Cancer Control. 2006;13:264–269
11. England DM, Hochholzer L, McCarthy MJ. Localized benign and malignant fibrous tumours of the pleura. A clinicopathologic review of 223 cases. Am J Surg Pathol. 1989;13:640–658
12. Cardillo G, Facciolo F, Cavazzana AO, Capece G, Gasparri R, Martelli M. Localized (solitary) fibrous tumours of the pleura: an analysis of 55 patients. Ann Thorac Surg. 2000;70:1808–1812
13. Suter M, Gebhard S, Boumghar M, Peloponisios N, Genton CY. Localized fibrous tumours of the pleura: 15 new cases and review of the literature. Eur J Cardiothorac Surg. 1998;14:453–459
14. de Perrot M, Kurt AM, Robert JH, Borisch B, Spiliopoulos A. Clinical behavior of solitary fibrous tumors of the pleura. Ann Thorac Surg. 1999;67:1456–1459
15. Rena O, Filosso PL, Papalia E, et al. Solitary fibrous tumour of the pleura: surgical treatment. Eur J Cardiothorac Surg. 2001;19:185–189
16. Robinson LA. Solitary fibrous tumors of the pleura. Cancer Control. 2006;13:264–269
17. Carretta A, Bandiera A, Melloni G, et al. Solitary fibrous tumors of the pleura: Immunohistochemical analysis and evaluation of prognostic factors after surgical treatment. J Surg Oncol. 2006;94:40–44
18. Cardillo G, Carbone L, Carleo F, et al. Solitary fibrous tumors of the pleura: an analysis of 110 patients treated in a single institution. Ann Thorac Surg. 2009;88:1632–1637
19. Chang YL, Lee YC, Wu CT. Thoracic solitary fibrous tumor: clinical and pathological diversity. Lung Cancer. 1999;23:53–60
20. Harrison-Phipps KM, Nichols FC, Schleck CD, et al. Solitary fibrous tumors of the pleura: results of surgical treatment and long-term prognosis. J Thorac Cardiovasc Surg. 2009;138:19–25
21. Gold JS, Antonescu CR, Hajdu C, et al. Clinicopathologic correlates of solitary fibrous tumors. Cancer. 2002;94:1057–1068
22. Miettinen MM, el-Rifai W, Sarlomo-Rikala M, Andersson LC, Knuutila S. Tumor size-related DNA copy number changes occur in solitary fibrous tumors but not in hemangiopericytomas. Mod Pathol. 1997;10:1194–1200
23. Filosso PL, Asioli S, Ruffini E, et al. Radical resection of a giant, invasive and symptomatic malignant Solitary Fibrous Tumour (SFT) of the pleura. Lung Cancer. 2009;64:117–120
24. Ferretti GR, Chiles C, Choplin RH, Coulomb M. Localized benign fibrous tumors of the pleura. AJR Am J Roentgenol. 1997;169:683–686
25. Versluis PJ, Lamers RJ. Localized pleural fibroma: radiological features. Eur J Radiol. 1994;18:124–125
26. Kohler M, Clarenbach CF, Kestenholz P, et al. Diagnosis, treatment and long-term outcome of solitary fibrous tumours of the pleura. Eur J Cardiothorac Surg. 2007;32:403–408
27. Magdeleinat P, Alifano M, Petino A, et al. Solitary fibrous tumors of the pleura: clinical characteristics, surgical treatment and outcome. Eur J Cardiothorac Surg. 2002;21:1087–1093
28. Battifora H, McCaughey WT Tumors of the Serosal Membranes. 1995 Washington, DC, WA American Registry of Pathology, Armed Forces Institute of Pathology:100–11
29. Liu CC, Wang HW, Li FY, et al. Solitary fibrous tumors of the pleura: clinicopathological characteristics, immunohistochemical profiles, and surgical outcomes with long-term follow-up. Thorac Cardiovasc Surg. 2008;56:291–297
30. Santos RS, Haddad R, Lima CE, et al. Patterns of recurrence and long-term survival after curative resection of localized fibrous tumors of the pleura. Clin Lung Cancer. 2005;7:197–201
31. Park CK, Lee DH, Park JY, Park SH, Kwon KY. Multiple recurrent malignant solitary fibrous tumors: long-term follow-up of 24 years. Ann Thorac Surg. 2011;91:1285–1288
Solitary fibrous tumor; CD34; Malignant solitary fibrous tumor; Localized fibrous tumor