Pediatric aggressive fibromatosis

A retrospective analysis of 13 patients and review of the literature




Aggressive fibromatosis (AF) is a soft tissue tumor and is rare in childhood, with high potential for local invasiveness and recurrence. General recommendations for the clinical management of pediatric patients with AF remain undetermined.


The authors retrospectively analyzed 13 children with AF who were diagnosed from 1987 until 2004 in the Erasmus MC-Sophia Children's Hospital, and a review of the pediatric literature was conducted.


Two patients received preoperative chemotherapy with combined vincristine, actinomycin-D, and cyclophosphamide (VAC). All 13 patients underwent surgery. Three of six patients who underwent incomplete resection received adjuvant treatment, two patients received radiotherapy, and one patient received chemotherapy (VAC). The median follow-up was 3.9 years (range, 0.6–14.0 years). Three patients developed recurrent AF, including two recurrences after patients underwent incomplete resection without adjuvant treatment. Secondary resection was performed, which was incomplete in one patient who subsequently received chemotherapy (VAC). At the time of the current report, all 13 patients were in complete remission. Ten pediatric AF studies, including the current study, with a total of 187 patients were reviewed. Incomplete resection was the most important determinant for disease recurrence; in the authors' opinion, the role of adjuvant therapy needs to be studied further.


Primary surgery with negative surgical margins was found to be the most successful primary treatment modality for children with AF. Positive margins after surgery indicated a high risk for disease recurrence. Multicenter, prospective (randomized) trials will be necessary to clarify the role of adjuvant treatment for patients with pediatric AF. Cancer 2005. © 2005 American Cancer Society.

Aggressive fibromatosis (AF) (i.e., desmoid tumor) is a soft tissue tumor and is rare in childhood. AF arises principally from the connective tissue of muscle and the overlying fascia (aponeurosis) and consists of elongated, fibroblast-like cells. Although AF is a nonmetastasizing tumor with benign histologic features, it has significant potential for local invasiveness and recurrence.1, 2 The overall incidence of AF in childhood is estimated at 2–4 new diagnoses per 1 million per year.3 Childhood AF has an age distribution peak at approximately 8 years (range, 0–19 years) with a slight male predominance.4–12 The pathogenesis of AF most likely is multifactorial; and genetic predisposition,13–16 endocrine factors,17, 18 and trauma5, 6, 8 seem to play an important role.3 The incidence of AF is remarkably higher in families with familial AF, familial adenomatous polyposis (FAP), and Gardner syndrome.19, 20 The typical clinical presentation of AF is a painless, slowly growing, deep-seated mass. Predilection sites are the shoulder, chest wall and back, thigh, and head/neck.4–12 Primary surgery is the most common treatment in adults and children with AF unless there is a risk of significant mutilation and/or functional impairment.5, 6, 12, 20–25 To our knowledge, the role of adjuvant radiotherapy and/or chemotherapy in childhood AF has not been established to date. A disease recurrence rate of approximately 50% has been reported in children with AF.4–12 However, mortality in AF is rare. To our knowledge, only small series of patients with pediatric AF have been described to date, reflecting the rarity of this disease.4–12 Therefore, general recommendations for the clinical management of AF are lacking. We conducted a retrospective analysis of all patients with childhood AF who were treated in the Sophia Children's Hospital over the last 17 years and reviewed the literature to define general principles for the management of AF in childhood.



Between January 1987 to January 2004, all pediatric patients (age < 18 years) with histologically proven AF who were diagnosed in our department were evaluated. Clinical, epidemiologic, and follow-up data were collected from the medical records and data base. Outcomes were analyzed as of July 1, 2004. Adverse events were disease recurrence, development of secondary malignancy, and death.


The diagnosis of AF is based on histology, and it arises principally from the connective tissue of muscle and the overlying fascia (aponeurosis). Characteristically, AF is circumscribed poorly and infiltrates the surrounding tissue, which usually is striated musculature. In AF, proliferation consists of elongated, fibroblast-like cells of uniform appearance surrounded by and separated from one another by abundant collagen and with little or no cell-to-cell contact. The cells lack hyperchromasia or atypia, and the mitotic rate is variable. The spindle muscle cells exhibit strong immunohistochemical staining with vimentin, whereas smooth muscle actin (SMA) and muscle-specific actin staining varies. There are rare instances of tumors that also stain with desmin and S-100.1, 2 All histologic diagnoses were revised in the Department of Clinical Pathology, Erasmus MC-Sophia Children's Hospital, Rotterdam, The Netherlands.


All chemotherapy given in this study was comprised of a combination of vincristine, actinomycin-D, and cyclophosphamide (VAC). Preoperative VAC chemotherapy in this study was administered before 2001 according to the following 6-week schedule: vincristine at a dose of 2 mg/m2 by intravenous (i.v.) infusion every 2 weeks, actinomycin-D at a dose of 15 μg/kg per day by i.v. infusion for 5 days every 6 weeks, and oral cyclophosphamide at a dose of 200 mg/m2 daily during the first course for 7 days and during every course thereafter, for 5 days every 6 weeks. Postoperative VAC chemotherapy in this study was administered after 2001 in 3-week courses of vincristine at a dose of 1.5 mg/m2 (maximum dose, 2.0 mg) by i.v. bolus once every 3 weeks, actinomycin-D at a dose of 45 μg/kg (maximum, 2 mg) by i.v. bolus once every 3 weeks, and cyclophosphamide at a dose of 1000 mg/m2 by i.v. infusion every 3 weeks.

Literature Review

A review of published pediatric case series of AF was performed using PubMed data bases (National Library of Medicine, Washington DC) from 1986 until January 2004. The keywords “aggressive fibromatosis” and “desmoid tumor” limited to children ages 0–18 years were used. In addition, we cross-referenced references from all identified studies. Inclusion criteria for analysis were that the study reported only children age < 19 years with AF (i.e., desmoid tumor) and reported a series of ≥ 4 patients. If there were multiple reports from the same institution, then the overlap of time frames excluded the least extended report.

Tumor Surgical Margin Status and Response

Multifocal disease was defined as two or more localizations of AF that involved one anatomic region of the body.2 In the current study, response to chemotherapy was assessed using ultrasound and/or magnetic resonance imaging (MRI). Reviewing all series, we attempted to categorize the response to therapy of all available patients in four groups. Response stated as “complete (clinical) response”, defined as the absence of residual tumor on imaging studies, was categorized as a complete response (CR). All responses that were reported as a reduction ≥ 25% in tumor size, a “partial” response, or a “good” response were categorized as a partial response (PR). Patients who responded with a reduction < 25% in tumor size were categorized as having stable disease (SD). Progression of tumor during treatment, disease recurrence within 1 month after the end of treatment, or responses stated as “without response” were categorized as progressive disease (PD). Disease recurrence was defined as “progression of tumor,” “progressive disease,” or “recurrence” ≥ 1 month after the end of treatment. In the current study, all surgical tumor margins were defined as microscopically negative (SN), positive (SP), or uncertain (SU). Reviewing all series, we attempted to divide patients in these groups. Surgical resection margins that were reported as “wide” or “complete” were categorized as SN, whereas surgical resection margins that were reported as “intralesional,” “incomplete,” “partial,” or “subtotal” were categorized as SP. Surgical resection margins that were reported as “close (1–2 mm),” “indeterminate,” or “marginal” were categorized as SU and were included only if the disease recurrence status was reported.

Statistical Analysis

Because many different retrospective studies were pooled for the current analyses, the current study results may have been influenced by biases inherent in each study. Therefore, only descriptive statistics were used to compare variables and differences in outcome between groups of patients. Analyses were performed using the SPSS statistical software package (version 11.5; SPSS Inc., Chicago, IL).



From 1987 until 2004, 13 children in our hospital (6 boys and 7 girls) were diagnosed with AF (Table 1). The clinical presentation in all 13 patients was a painless, slowly growing mass. The median patient age at diagnosis was 4.1 years (range, 0.8–12.4 years). Sites of involvement were the head and neck (five patients), trunk (four patients), and extremities (four patients). Patient 1 had 2 separate lesions located on his back that were classified as multifocal disease (Table 1). None of the patients had a family history of AF, adenomatous polyposis coli (FAP), or Gardner syndrome, and none had suffered a recent trauma. Immunohistochemical analysis showed positive staining with vimentin in 100% of patients, SMA in 78% of patients, actin in 30% of patients, S-100 in 8% of patients, and desmin in 0% of patients (Table 2). In addition, two of the tumors diagnosed in children were tested for estrogen receptor (ER) and progesterone receptor expression. Cytogenetic analysis of the tumor was performed in only three patients (Table 2).

Table 1. Clinical Characteristics of the 13 Children with Aggressive Fibromatosis in Rotterdam
Patient no.Age at Dx (yrs)GenderTumor locationSize (cm)TreatmentRecurrenceTime to recurrence (mos)Treatment for recurrenceOutcomeFollow-up (mos)
  • Dx: diagnosis; Preop: preoperative; S: surgery; Postop: postoperative; MD: multifocal disease; SN: surgery with positive margins; CCR: continuous complete remission; VAC: combined vincristine, actinomycin-D, and cyclophosphamide; SP: surgery with positive margins; Gy: grays (radiotherapy); SU: surgery with unspecified margins.

  • a

    These patients were lost to follow-up; at the time of last follow-up, continuous complete remission was confirmed.

11.3MaleTrunk (MD)<5NoneSNNoneNo  CCR15
28.1MaleTrunk>5NoneSNNoneNo  CCR101
30.8FemaleUpper extremity>5NoneSNNoneNo  CCR132
44.4MaleLower extremity<5NoneSNNoneNo  CCR68
53.1FemaleMandible>5NoneSNNoneNo  CCR36
62.1FemaleMandible>51 × VACSNNoneNo  CCR58
79.2MaleTrunk<5NoneSPNoneNo  CCR15a
812.4FemaleLower extremity>5NoneSPNoneYes11SP, 8 × VACCCR26
109.5FemaleTrunk<5NoneSP59.4 GyNo  CCR82
114.1FemaleNeck<52 × VACSP57.6 GyNo  CCR47
125.7MaleMandible<5NoneSP6 × VACNo  CCR7
131.3FemaleUpper extremity<5NoneSUNoneYes4SNCCR19a
Table 2. Diagnostic Features in the 13 Children with Aggressive Fibromatosis in Rotterdam
Patient no.ImmunohistochemistryaHormone receptorsKaryotype
  • SMA: smooth muscle actin; ER: estrogen receptor: PR: progesterone receptor; ND: not done; Neg: negative; Pos: positive.

  • a

    None of the 13 pediatric aggressive fibromatosis tumors were evaluated immunohistochemically with β catenin.

2NDPosPosNegWeak posNDNDNormal 46,XY
8NDPosWeak posNegNegNDNDND
9NDWeak posWeak posNegNegNDNDND
11NDPosWeak posNegNegNegNegND
12NDPosWeak posNegNegNDNDNormal 46,XY
13NDWeak posNegNegNegNegWeak posND
Total no.49101113223
No. positive (%)4 (100)7 (78)3 (30)0 (0)1 (8)0 (0)0 (0) 


Two patients received preoperative chemotherapy (VAC) with the objective of decreasing tumor volume (Table 1). Patient 6 received one course of preoperative VAC and responded with SD. Patient 11 received 2 courses of preoperative VAC, with an adjustment of the second course; cyclophosphamide was given at a dose of 1000 mg/m2 by i.v. infusion for 2 days. The patient responded with PD.

Surgical resection was performed in all 13 patients (Table 1). This resulted in complete resection (SN margins) in six patients and incomplete resection (SP margins) in six patients. In one patient, the surgical margins were not reported (SU). Three patients with SP margins received adjuvant treatment, two patients received radiotherapy, and one patient received chemotherapy (VAC), which resulted in a CR in all three patients (Table 1). The median follow-up was 3.9 years (range, 0.6–14.0 years). Three patients developed local recurrences after 4 months, 11 months, and 12 months, respectively (Table 1). Two patients with recurrent disease had SP margins at the time of initial diagnosis, and the third patient had SU margins. None had been treated with adjuvant therapy. All three patients underwent surgical resection, which was incomplete in one patient. This patient was treated subsequently with adjuvant chemotherapy (VAC) and responded with a CR. At the time of last follow-up, all 12 patients were in continuous complete remission.


Follow-up regarding toxicity revealed limited range of motion in the primary area in two individuals (Patients 3 and 8) (Table 1). Two of four individuals (Patients 5 and 9) who underwent a large resection of the mandible with direct reconstructive surgery using a rib transplantation developed a very mild asymmetry of the face. Of the two patients who received radiotherapy, one individual (Patient 10) developed mild scoliosis after irradiation of the thoracic-cervical region, and the other individual (Patient 11) developed a subclinical hypothyroidism (i.e., an elevated thyroid-stimulating hormone level with normal free-T4 values) 2 years after irradiation to the cervical region. At the time of last follow-up, no patient had developed a secondary malignancy.

Literature Review

Between 1986 and 2004, 12 pediatric AF case series, including the current study, were identified. Two studies, by Merchant et al.26 and Raney et al.27, reported patients from the same institution as Rao et al.8 and Spiegel et al.,12 respectively, with overlapping time frames. Those investigators26, 27 reported on small series of patients who were treated with radiotherapy and chemotherapy, respectively, and both reports were excluded from this review. The other 10 studies, along with the current study, reported a total of 187 patients (Table 3).4–12 In 64 of the reviewed patients, including the current series, the site of involvement and the age at diagnosis were specified.3, 10, 12 The children with AF of the head and neck had a median age of 3.6 years at the time of diagnosis (range, 0.2–9.9 years), whereas the children with AF of the trunk or limb had a median age of 7.8 years (range, birth–15.7 years) (Fig. 1). None of the pediatric AF studies reported patients with familial history of AF or FAP, and only two patients with Gardner syndrome were reported.6 In 3 studies of pediatric AF, 20% of patients (n = 108 patients) had a history of local trauma before they developed AF.5, 6, 8

Table 3. Overview of Reports Concerning Pediatric Aggressive Fibromatosis
ReferenceNo. of patientsMedian age (yrs) (range)M:F ratioSite of involvement (no)MD (no.)Primary tumorTreatment (adjuvant)Follow-up (yrs) (range)No. of recurrences (%)No. with multiple recurrencesNo. of deaths
  • M:F ratio: male: female ratio; H/N: head and neck; MD: multifocal disease. S: surgery; ± with or without; CT: chemotherapy; RT: radiotherapy; US: unspecified.

  • a

    In three patients, the site of involvement was not specified.

Ayala et al., 19865257 (0–15)15:1041382YesS (± CT)4 (0–25)6 (24)00
Scougall et al., 19871088 (1–14)6:24400YesS (± CT/RT)5 (0–15)3 (38)20
Rao et al., 19878209 (2–18)12:86771YesS (± CT/RT)6 (0–17)9 (45)62
Faulkner et al., 1995663US (0–19)30:33841115aYesS (± CT/RT)6 (US)42 (67)200
Spiegel et al., 199912187 (0–15)12:611251YesS (± CT)8 (1–22)15 (83)61
Godzinski et al., 20037218 (0–17)12:931440YesS (± CT/RT)4 (0–10)9 (43)00
Current study, 2005134 (0–12)6:74451YesS (± CT/RT)3 (0–14)3 (23)00
Atahan et al., 19894410 (3–15)3:13100NoRT (± CT)3 (0–3)1 (25)00
Skapek et al., 1998111012 (6–18)6:43610NoCT1 (0–3)3 (30)10
Reich et al., 19999511 (7–17)5:01401NoCT2 (0–6)1 (20)00
Total no. (%)187 (100)8 (0–18)107:80 (57:43)37 (20)106 (57)41 (22)11 (6)a  4 (0–25)92 (49)34 (19)3 (2)
Figure 1.

This box-whisker plot illustrates patient age at the time of diagnosis in relation to the localization of pediatric aggressive fibromatosis. The upper and lower margins of the box represent the 75th and 25th percentiles, respectively. The midline of the box represents the 50th percentile (median). The upper and lower margins of the whiskers represent the 97.5th and 2.5th percentiles, respectively. The group with head/neck localization (n = 18 patients) included 1 patient who had a value greater than the 97.5th percentile (O).

Seven of 10 pediatric AF studies, including the current investigation, reported treatment of the primary tumor, and all patients generally underwent initial surgery (n = 168 patients) (Table 3).5–8, 10, 12 The other 3 series reported treatment of recurrent tumors, including 2 series in which patients initially were treated with chemotherapy (vinblastine [VBL] and methotrexate [MTX]; n = 15),9, 11 and a third study in which radiotherapy was administered (n = 4 patients).4 The disease recurrence rate in the reviewed children with primary AF was approximately 50% (Table 3). The majority of disease recurrences (89%), including those occurring among the patients in the current study, were observed within 3 years, and nearly all recurrences (97%) were observed within 6 years, although recurrences after 10 years have been reported.6, 8, 10, 12 All reported disease recurrences were local or regional with a pattern consistent with infiltrative growth. Three deaths were reported that were caused by invasive tumor destruction of vital organs; all three patients had head and neck lesions.8, 12 Including the current analysis, 85 patients underwent primary surgery and had information available regarding surgical margin status and disease recurrence (Table 4).5, 6, 8, 10, 12 The recurrence rate among patients with SN margins was 16%, compared with 67% among patients with SP margins. In patients with SP margins, 74% of those who did not receive additional therapy developed recurrent disease, compared with 40% of patients who received adjuvant treatment. Adjuvant treatment consisted of chemotherapy (eight patients) or radiotherapy (two patients). In total, 32 of 187 pediatric patients were treated with chemotherapy only at the time of initial diagnosis (n = 6 patients), with chemotherapy at the time of disease recurrence (n = 17 patients), or with preoperative chemotherapy (n = 9 patients), and information regarding response was available (Table 5).5, 6, 9, 11, 12 A combination of VBL and MTX was the most commonly reported regimen. Among the patients who received chemotherapy only, the response was a CR in 25% of patients, a PR in 34% of patients, SD in 25% of patients, and PD in 9% of patients, and response was not reported in 6% of patients. Two patients died: one of progressive disease and the other of cardiotoxicity after treatment with doxorubicin.5 The overall recurrence rate after treatment with chemotherapy only (n = 21 patients) was 24%.

Table 4. Outcomes in Patients with Positive versus Negative Surgical Margins after Primary Surgery for Pediatric Aggressive Fibromatosis
StudyTotal no. of patientsOutcome after surgery (no.)Adjuvant treatment in patients with SP (no.)
  • SP: positive surgical margins; SN: negative surgical margins; RT: radiotherapy; CT: chemotherapy; DO: doxorubicin; DA: dacarbazine; AMSA: amsacrine; V: vincristine; C: cyclophosphamide; F: 5-fluorouracil; MTX: methotrexate; MH: methylhydrazine; A: actinomycin-D; VBL: vinblastine.

  • a

    These patients also received preoperative chemotherapy with vincristine, actinomycin-D, and cyclophosphamide.

Ayala et al., 198658532- 1DO/DA
Faulkner et al., 1995627141310- 3AMSA; V/C/F/MTX/MH; V/A/C
Rao et al., 198781771010- - 
Scougall et al., 1987107433- - 
Spiegel et al., 1999121411310- 3VBL/MTX (2 patients); V/A
Current study126632a58–59 Gy1V/A/C
Total no. of patients853748382 8 
No. of recurrences (%)38 (45)6 (16)32 (67)28 (74)0 (0) 4 (50) 
Table 5. Outcome after Treatment of Pediatric Aggressive Fibromatosis with Chemotherapy Only
Reference/CTTotal no.Response (no. of patients)Duration of treatment (mos)Recurrence (no.)
  • CT: chemotherapy; CCR: clinical complete remission; PR: partial response; SD: stable disease; PD: progressive disease; US: unspecified; VBL: vinblastine; MTX: methotrexate; NA: not applicable; DO: doxorubicin; DA: dacarbazine; V: vincristine; A: actinomycin-D; C: cyclophosphamide; IF: ifosfamide; ET: etoposide.

  • a

    These patients received treatment with preoperative chemotherapy, and their response was evaluated at surgery.

  • b

    The median duration of treatment was 10 months.

Primary tumor        
 Skapek et al., 199811        
 Ayala et al., 19865        
  DO/DA/V/A/C220000US, 24–290
  DO/DA110000US, 1–50
  US100010PD (dead of disease)NA
 Subtotal641010 0
 Reich et al., 19999        
 Skapek et al., 199811        
  VBL/MTX821500CCR, 12–18; PR, 12; SD, 2–35b 
 Spiegel et al., 199912        
  DO100100US (2 courses, still on CT)NA
 Ayala et al., 19865        
  V/A/C/DO1000019 (dead of toxicity)NA
 Subtotal1743712 5
 Current study        
  V/A/C200110SD, 2; PD, 2NAa
 Faulkner et al., 19956        
  V/A/C101000US (6 courses)NAa
  IF/ET101000US (6 courses)NAa
  V/A then VBL101000US (1 course then 9 courses)NAa
 Ayala et al., 19865        
  US303000US, PD (dead of disease)NAa
 Reich et al., 19999        
 Skapek et al., 199811        
  VBL/MTX1032500CCR, 12–18; PR, 12; SD, 2–35b3
 Spiegel et al., 199912        
  DO100100US (still on CT)0
 Faulkner et al., 19956        
  V/A then VBL101a000US (1 course then 9 courses)NAa
  IF/ET101a000US (6 courses)NAa
  V/A/C101a000US (5 courses)NAa
 Current study        
  V/A/C2001a1a0SD, 2; PD, 1NAa
 Ayala et al., 19865        
  DO/DA/V/A/C220000US, 24–290
  V/A/C/DO1000019 (dead of toxicity)NA
  DO/DA211a000CCR, US, 1–5; PR, USNAa
  US103a010US, PD (dead of disease)NA,a NA
Total no. (%)328 (25)11 (34)8 (25)3 (9)2 (6) 5/21 (24)

The median follow-up was 4 years (range, 0–25 years). Information regarding the toxicity of treatment was available in 4 pediatric AF case series, including our own (n = 128 patients).5, 6, 11, 12 In our own series and in one other series,12 limited range of motion of the primary area was reported as the most frequent late complication (42%). Severe short-term toxicity of treatment was reported in three patients; two patients died of cardiotoxicity after treatment with doxorubicin,5 and one patient died of severe radiation-induced dermatitis with chronic ulcers.6 During this short median follow-up, 1 secondary malignancy was reported: a papillary carcinoma of the thyroid gland, which developed 11 years after radiotherapy.12


No general determinants for the treatment management of pediatric AF are available based on studies in large cohorts of patients. We reviewed all pediatric AF literature to discover the behavior of this type of tumor in children.

A review of the pediatric AF case series, including our own, indicated that children with AF of the head and neck were younger at the time of diagnosis than children with AF at other sites (Fig. 1).5, 10, 12 This difference in age distribution may be influenced by referral and selection bias; however, it may reflect the site distribution in different age groups of children with AF. The diagnosis of AF is based on histology, and it has been suggested that the pathogenesis of AF is multifactorial. Local physical trauma before developing AF was reported in 20% of our reviewed pediatric patients.5, 6, 8 Apparent chromosomal aberrations and nonrandom X-chromosome inactivation in adult and pediatric AF suggests true neoplastic characteristics.13, 28, 29 This finding was supported by 8 patients with pediatric AF reported in 1 study, 5 of whom (63%) had an abnormal karyotype (2 at the time of initial diagnosis and 3 at the time of disease recurrence), with trisomy 8 (n = 4 patients) and trisomy 20 (n = 1 patient) being the only recurrent features.13 Sporadic cases of adult AF contain a somatic mutation in either the adenomatous polyposis coli (APC) gene (21%), identified on chromosome 5q22 and associated with FAP, or in β-catenin gene and protein expression (52%).14–16 Further findings suggest that insulin-like growth factor binding protein 6 (IGFBP-6) is down-regulated directly by the β-catenin/TCF complex in adult AF and implies a role for the IGF axis in the proliferation of AF.30 A recent review reported a high prevalence of desmoid tumor in 126 of 880 adult patients with FAP (14.3%) who had proven AF gene mutation,19 and an extremely high prevalence (38%) was reported for patients with Gardner syndrome.20 In pediatric AF studies, no patient with a history of familial AF or FAP and only two patients with Gardner syndrome were reported.6 Routine karyotyping has a relatively limited value, and to our knowledge the significance of the APC and β-catenin genes in the pathogenesis of childhood AF and their value for differentiating fibroblastic tumors has yet to be established. In adults, a correlation between the tumor growth rate and the level of endogenous estrogen has been suggested in female patients because of high amounts of ER in their tumor tissue.3 It has been suggested that the presence of antiestrogen binding sites distinct from ER play a role in treatment with antiestrogens in adult AF.18 In two studies, including the current analysis, four tested children with AF did not express ER.17 The role of ER expression and antiestrogen binding sites in the pathogenesis of childhood AF and options for treatment have yet to be established.

There remains a lack of general recommendations for the clinical management of patients with pediatric AF. Although spontaneous regression has been observed in sporadic cases,2, 21 surgery generally is the primary treatment modality in adults and children with AF.5, 6, 12, 2–25 Faulkner et al.6 identified surgical margin status as the only significant prognostic factor for disease recurrence in patients with pediatric AF (n = 24 patients). The recurrence-free probability at 3 years in their study was 0.15 for patients with positive surgical margins and 0.70 for patients with negative surgical margins. The current review of all well documented pediatric cases confirmed these findings (i.e., 16% of patients who had negative [SN] margins after primary surgery developed disease recurrences compared with 67% of patients who had positive [SP] margins) (Table 4). The high risk for disease recurrence among these patients indicates that the role of adjuvant treatment in patients with SP margins needs further exploration. In adults, the standard approach for SP margins is adjuvant radiotherapy, which has achieved a high local control rate of approximately 80% and is considered beneficial regardless of surgical margin status.22, 25, 31–33 In pediatric patients with AF, the high doses of radiotherapy (55–60 grays [Gy]) necessary for tumor control harbors a large risk for growth problems and the development of secondary malignancies.4, 6, 12, 22, 26, 32 One pediatric AF study reported 11 children with partially excised or recurrent lesions who received radiotherapy and who had at least 3 years of follow-up.6 Four of those 11 children (36%) developed recurrent disease, including 2 of 5 patients who received radiation doses > 50 Gy. In contrast, in another pediatric AF study, which was from the same institution reported by Rao et al.8 and included overlapping patients, 11 of 13 children (85%) developed recurrent disease after irradiation, including 6 of 8 children who received doses ≥ 50 Gy.26 To our knowledge, the role of radiotherapy in childhood AF as adjuvant treatment in patients with SP margins has not been established to date and will require further study in a prospective, randomized study. The use of chemotherapeutic and other systemic agents may be a reasonable alternative to avoid radiotherapy in the growing child, although chemotherapy also carries a risk for potentially adverse side effects, such as second malignancies, fertility problems, and cardiotoxicity. A recent review concerning mainly adult patients with AF reported a median overall response rate of 50% (range, 17–100%) with combination chemotherapy (doxorubicin, actinomycin-D, MTX, and vinca alkaloids) in 16 single-arm studies.18 According to our current review of all patients with pediatric AF who were treated with chemotherapy only, 25% of patients achieved a CR, and 34% of patients achieved a PR in 34, based mainly on a combination of MTX and vinblastine (VBL), or VAC (Table 5).5, 6, 9, 11, 12 The outcomes of 46 pediatric patients with primary positive margins (SP) suggest that there is an advantage for patients who received adjuvant chemotherapy (n = 8 patients) compared with patients who did not receive adjuvant treatment (n = 38 patients), with a decrease in the disease recurrence rate from 74% to 50% (Table 4). However, the numbers of patients were small and were derived from different series. To our knowledge, the role of chemotherapy in childhood AF has not been established to date and should be studied further. Recently, a collaborative study of MTX/VBL chemotherapy for children with AF has been initiated.11 Based on the experience of others, it has been suggested that the response of pediatric AF to chemotherapy is slow, and it has been suggested that treatment should be continued for prolonged periods from 12 months to 18 months.3, 9, 11 The chronic and prolonged course that many of these children with AF endure as a result of these slow growing lesions suggests that the use of noncytotoxic drugs, such as antiestrogens, nonsteroidal antiinflammatory drugs (NSAIDs), imatinib mesylate, interferon-α, and retinoic acid, for part of their treatment is reasonable.18, 34–37 Treatment with a well tolerated combination of antiestrogens and NSAIDs was reported to result in tumor reduction and SD in adult and pediatric patients with sporadic AF.36, 37

Primary surgery with negative margins is the treatment choice for children with AF. In patients with unresectable tumors, the use of chemotherapy and/or noncytotoxic drugs in children with AF may be a reasonable alternative. Positive margins after surgery indicate a high risk for disease recurrence. Multicenter prospective (randomized) trials will be necessary to clarify the role of adjuvant treatment for patients with pediatric AF.