High-dose tamoxifen and sulindac as first-line treatment for desmoid tumors




Desmoid tumors are mesenchymal nonmetastasizing neoplasms. Although rare in the general population, they are a common extracolonic manifestation of familial adenomatous polyposis (FAP). Because of high tumor recurrence rates, surgery has been less than satisfactory in the treatment of desmoid tumors. In the current study, high doses of tamoxifen in combination with sulindac were used to treat severe desmoid tumors to avoid surgery.


Since 1992, 25 patients at Heinrich Heine University (Dusseldorf, Germany) were treated with a combination of tamoxifen and sulindac. In the current study, 17 patients with FAP-associated and 8 patients with sporadic desmoid tumors received 120 mg of tamoxifen and 300 mg of sulindac daily. Every 6 months, the protracted course of desmoid growth was measured by computed tomography and/or magnetic resonance imaging scans. Tumor responses were characterized as progressive disease, stable disease (SD), partial regression (PR), and complete regression (CR).


Of the group of patients who received tamoxifen and sulindac as a primary treatment, all three patients with sporadic desmoid tumors demonstrated cessation of growth, and 10 of the 13 patients with FAP-associated tumors achieved either a PR or CR. In the sporadic desmoid tumor group, eight of nine patients developed tumor recurrences after undergoingsurgery at other institutions. Of these, two patients had SD and two patients had a PR to CR.


The patients with desmoid tumors who were managed conservatively with high-dose tamoxifen and sulindac had the best outcome. Desmoid tumor recurrence after surgery was high and in the FAP-associated tumor group, therapy with tamoxifen and sulindac was found to be less successful. Based on this experience, the authors recommended high-dose tamoxifen and sulindac as the primary treatment for patients with FAP-associated desmoid tumors. However, to our knowledge, the best approach after surgical intervention for patients with sporadic desmoid tumors remains to be determined. Cancer 2004. © 2003 American Cancer Society.

“Desmoid” is derived from the Greek word “desmos,” meaning band-like.1 These tumors are defined as benign fibrous tissue tumors arising in the musculoaponeurotic structures throughout the body. They do not metastasize, but are locally invasive.2 Histologically, mature fibroblasts of uniform size and shape are observed, with mitosis being unusual.

Although rare in the general population with an annual incidence of only 2–4 cases per 1 million population,3 desmoid tumors are a common extracolonic manifestation of familial adenomatous polyposis (FAP). Approximately 2% of all desmoid tumors are associated with FAP, and patients with FAP have a 1000-fold increased risk of developing desmoid tumors, compared with the general population.4 The association between desmoid tumors and FAP first was made in 1923 by Nichols.5 In 1951, Gardner6 reported the familial occurrence of intestinal polyposis, osteomas, fibromas, and epidermal or sebaceous cysts. In 1958, Smith7 proposed the name of Gardner syndrome for this clinical entity. The current view is that Gardner syndrome is a variant of the phenotypic expression of FAP.8 As screening for FAP becomes more efficient, the relative importance of extracolonic manifestations of FAP is increasing12 because surgery has been less than satisfactory in the treatment of desmoid tumors.9–11 Desmoid tumors are reported to be the second most common cause of death in patients with FAP, after colorectal carcinoma.13 The combination of the location and unpredictable nature of FAP-associated desmoid tumors may be responsible for desmoid disease being the leading cause of death among patients with FAP who have undergone prophylactic colectomy.14

Desmoid tumors are poorly understood. Consequently, several medical approaches have been combined with or without surgical resection with mixed results. These include chemotherapy with doxorubicin-based combinations,15 antiestrogen therapy with tamoxifen,16, 17 testolactone (which inhibits steroid aromatase activitiy) and its consequent reduction in estrogen synthesis,18 nonsteroidal antiinflammatory drugs (NSAIDs) such as indomethacin and sulindac,19, 20 radiotherapy,20, 21 and colchicines.22 It is evident from all these medical alternatives that to date there is no established or evidence-based approach for the treatment of this neoplasm. The goal of the current study was to report our experience with high-dose tamoxifen (or the analog, raloxifen) in combination with sulindac as the first line of therapy in patients with sporadic or FAP-associated desmoid tumors. Thirty patients with desmoid tumors represented a heterogenous group. Some patients developed recurrent tumors after (multiple) surgeries and others had fast-growing desmoid tumors occurring in different locations. Patients with very small or slow-growing desmoid tumors did not receive medication, but were observed closely.


Patient Characteristics

FAP-associated desmoid tumors

The current study comprised 115 patients with FAP. A physical examination and regular abdominal ultrasonography are an integral part of patient follow-up at Heinrich Heine University (Dusseldorf, Germany). Patients with FAP were included in the study if they had a clinically evident mass measurable by sonography (extraabdominal desmoid tumor) or if a tentative diagnosis of a mesenteric desmoid tumor by sonography was confirmed with a subsequent computed tomography (CT) scan. Measurable desmoid lesions were identified in 22 patients, corresponding to 19% of all patients with FAP. Most patients entered in the study underwent a colectomy at the study institution and eventually presented with a desmoid tumor. Because these patients were examined regularly, the biologic behavior of the desmoid tumors at the time of the initial diagnosis could well be assessed. Depending on the overall size of the tumor at diagnosis and the interval required for tumor growth, some patients received medication and others did not. Five patients were determined to have benign desmoid tumors and did not receive medication.

Of the 72 female patients with FAP, 18 (25%) developed desmoid tumors. Of the 43 male patients with FAP, 4 (9%) were diagnosed with a desmoid tumor. Therefore, the female-to-male ratio was 3:1. The age at diagnosis of desmoid tumors ranged from 4–49 years and showed a plateau between ages 27–35 years. The mean age was 32 years, with no significant difference noted between males and females (age 33 years in females vs. age 30 years in males). For 17 patients (74%), desmoid tumors were clinically evident a mean of 5.1 years after FAP-related surgery (range, 1–21 years). For 3 patients (13%), the time of surgery and the diagnosis of desmoid tumors were concomitant. For 2 patients (9%), the desmoid tumors developed before the diagnosis of FAP.

Sporadic desmoid tumors

Eight patients presented with a clinically evident and histologically verified desmoid tumor. In general, these eight patients were referred to us for further treatment. FAP was excluded by colonoscopy and negative family history in all cases. The mean age at the time of diagnosis of sporadic desmoids was 43 years, which was significantly higher than the age of patients with FAP-associated desmoid tumors. The youngest patient was age 25 years at the time of diagnosis and the oldest patient was age 64 years. A predominance for the female gender was found (i.e., there was a female-to-male ratio of 3:1).

Site of the Desmoid Tumors

Desmoid tumors occur at single or multiple anatomic sites. Patients with desmoid tumors were classified into four groups according to the anatomic location of the tumors: 1) intraabdominal, for mesenteric/retroperitoneal sites; 2) abdominal, for the abdominal wall; 3) abdominal/intraabdominal when there was a combination of both sites; and 4) extraabdominal, for other body areas. The distribution of treated sites of disease is shown in Table 1.

Table 1. Anatomic Distribution of FAP-Associated Desmoid and Sporadic Tumors
Site of the desmoidsNo. of patients with FAP (%)No. of patients with sporadic tumors (%)
  1. FAP: familial adenomatous polyposis.

 Exclusive15 (68) 
 Exclusive3 (14) 
Both sites  
 Total3 (14) 
 Exclusive0 (0)8 (100)

FAP-associated desmoid tumors

Desmoid tumors were found exclusively within the abdomen in 15 patients (68%), in the abdominal wall only in 3 patients (14%), and at both sites in 3 patients (14%). One patient had an extraabdominal tumor. Multiple desmoid tumors were observed within the abdomen in seven patients.

Sporadic desmoid tumors

In eight patients, all tumors (100 %) were confined to the extraabdominal region. In one patient, multiple desmoid tumors were encountered.

Diagnostic Methods and Follow-Up

Every 4–6 months, depending on the growth rate of the tumor and the patients' symptoms, growth of the desmoid tumors was reassessed by ultrasound (if easily accessible), CT scan, or magnetic resonance imaging scan. Tumors were assessed to determine whether they were still entirely solid or whether they demonstrated central liquefaction as previously described.23 Tumor responses were characterized as progressive disease (PD), stable disease (SD), partial regression (PR), and complete regression (CR). Because all patients had PD at the time of referral (except five patients for whom medical treatment was not indicated), SD was considered a useful response. Increasing grades of response were PR and CR. PD was defined as a 25% increase in the size of at least 1 measurable lesion or the appearance of any new lesions. SD was achieved when neither a 25% reduction in total tumor size nor a 25% increase in the size of 1 or more lesions could be established. A PR was defined as a 25% reduction in total tumor size detected by 2 observations ≥ 4 months apart with no appearance of new lesions and no progression of any existing lesions. The sum of the product of the perpendicular dimensions of all lesions determined the total tumor size. A CR was defined as the clinicoradiologic disappearance of the desmoid tumor(s).

Drug Regimen

We recommend 300 mg of sulindac in combination with either 120 mg of tamoxifen or 120 mg of raloxifen daily. The rationale for raloxifen is the decreased rate of ovarian cysts in premenopausal women. When significant tumor regression occurred and/or SD was observed over a prolonged period of time, the medication was eventually decreased stepwise. In the rare event when a patient did not tolerate the high dosage, a reduction to 60 mg daily was well tolerated and continued for the required duration.

Sulindac inhibits cyclooxygenase (COX), which catalyzes the oxidation of arachidonic acid to prostaglandin precursors.24 It is a long-acting analog of indomethacin.25 Tamoxifen is a nonsteroidal triphenylethylene previously classified as a pure antiestrogen. It competes with estrogen binding sites in target tissues and has been shown to have estrogen-like activity. It is now better classified as a selective estrogen receptor modulator.26


Clinical Course of FAP-Associated Desmoid Tumors Treated with Antiestrogens and NSAIDs after Previous Desmoid Tumor-Related Surgery

Four patients with FAP-associated desmoid tumors were referred to our hospital after being treated with desmoid tumor-related surgical resection (Table 2). All of these patients developed tumor recurrences. Two patients received complete resections with negative surgical margins. In the other two patients, the surgical margins were intralesional. The patients were treated with tamoxifen and sulindac after presenting at our institution. Three patients (75%) had PD. Among these patients, a 28-year-old female was compliant with a low dosage of tamoxifen (i.e., 20 mg) for a period of 6 months, which is usually not long enough to cause a clear response. After receiving subsequent chemotherapy with doxorubicin and cisplatin, cessation of desmoid tumor growth was achieved in all but one (extraabdominal) patient and the patients' symptoms decreased. Only one patient treated for primary disease with tamoxifen and sulindac after partial resection experienced tumor regression.

Table 2. Clinical Course of FAP-Associated Desmoids Treated with Antiestrogens and NSAIDs after Previous Surgical Interventiona
Patient no.Age (yrs)GenderSiteProphylactic colectomybPrevious resectionResponseDrug in mg (mos drug received)Response
  • FAP: familial adenomatous polyposis; NSAIDS: nonsteroidal antiinflammatory drugs; F: female; T: tamoxifen; S: sulindac; PD: progressive disease; IAPA: colectomy with ileo-anal pouch anastomosis.

  • a

    Clinical course with or without radiologic responses.

  • b

    Surgery in years to first appearance of desmoid tumor.

149FAbdominalNoneResection (R0)Recurrence: extraabdominalT120 (36) S 300 (30)PD
223FExtraabdominalNoneMultiple resectionsRecurrences: abdominal, extra- and intraabdominalT 120 (17) S 300 (17)PD
328FIntraabdominalIAPA (−4)Resection (R2)Recurrence: abdominal desmoidT 20 (6) S 300 (6)PD
426FIntraabdominalIAPA (−1)Resection tumor (R2)PDT 90 (24) S 300 (24)PR

FAP-Associated Desmoid Tumors Treated with Antiestrogens in Combination with NSAIDs

Table 3 shows the clinical course of FAP-associated desmoid tumors treated with 300 mg of sulindac in combination with either 120 mg of tamoxifen or 120 mg of raloxifen daily. No additional surgical treatment was performed. In this study group, 10 of 13 patients (77%) demonstrated cessation of tumor growth, a PR, or a CR. Of these 13 patients, 6 (46%) had SD, 4 (31%) achieved a PR or CR for > 6 months, and 3 (13%) had SD for > 30 months. Only one patient had PD during this therapy. However, it must be taken into account that medication was administered for only 6 months, which might have been too short a period to lead to an effect. The clinical outcome of patients receiving medication was compared according to whether previous desmoid tumor-related surgery was performed. As summarized in Table 4, therapy with tamoxifen and sulindac was more beneficial when no previous desmoid surgery had been performed.

Table 3. Clinical Course of FAP-Associated Desmoids Treated with Antiestrogens and NSAIDsa
Patient no.Age (yrs)GenderSitePrevious surgerybDrug in mg (mos drug received)Response
  • FAP: familial adenomatous polyposis, NSAIDs: nonsteroidal antiinflammatory drugs; F: female; IAPA: colectomy with ileoanal pouch anastomosis; T: tamoxifen; S: sulindac; SD: stable disease; PR: partial response; CR: complete regression; IRA: colectomy with ileorectal anastomosis; Ra: raloxifene; M: male; PD: progressive disease; Ra: raloxifen.

  • a

    Clinical course with or without radiologic response.

  • b

    Years to first appearance of desmoid tumor.

515FIntraabdominalIAPA (−2)T 30 (30) S 150 (30)SD
636FIntraabdominal, abdominalIAPA (−3)T 80 (60) S 300 (60)PR after 1 yr, CR after 4 yrs
736FIntraabdominal, abdominalIRA (−1)Ra 80 (10) S 300 (10)SD
834MIntraabdominalIAPA (−5)T 120 (6) S 300 (6)PD
933FIntraabdominal, abdominalIRA (−1)T 120 (60) S 300 (60)SD
1029FIntraabdominalIRA (−21)Ra 120 (10) S 300 (10)SD
1135MIntraabdominalIRA (−2)T 120 (84) S 300 (84)PR
1234FIntraabdominalnoRa 80 (60) S 300 (84)SD
1334FIntraabdominalIAPA (−13)T 120 S 300Lost to follow-up
1440FIntraabdominalcolectomy (−13)T 60 (8) S 300 (8)CR after 4 mo
    hysterectomy (−3)  
1538FIntraabdominalIRAT 120 (> 24) S 300 (> 24)SD
1621MIntraabdominalIRA (−0)T 120 (15)PR
1731FIntraabdominalIAPA (−5)T 30 (15) S 300 (15)Lost to follow-up
Table 4. Clinical Outcome after Treatment with Antiestrogens and NSAIDsa
OutcomeNo. of patients with no surgeryNo. of patients with surgery
  • NSAIDs: nonsteroidal antiinflammatory drugs.

  • a

    Comparison of the clinical outcome between patients with no previous desmoid-related surgery and patients with previous desmoid-related surgery to therapy with antiestrogens and nonsteroidal antiinflammatory drugs.

Static disease60
Partial regression21
Complete regression20
Progressive disease13

Clinical Course of Patients with FAP-Associated Desmoid Tumors without Tamoxifen/NSAID Therapy

Table 5 shows the clinical course of five patients with FAP with small and slow-growing desmoid tumors. No specific therapy was indicated and the patients were followed closely. Two patients with FAP-associated desmoid tumors in the abdominal wall experienced spontaneous CR without tumor recurrence. One patient with a mesenteric desmoid tumor received radiotherapy as a first treatment at a previous institution. A residual tumor measuring 6 cm in greatest dimension was evaluated after cessation of therapy. We opted for follow-up without medication. Eight years later, no desmoid tumor was detectable. One patient developed PD after subtotal excision of a mesenteric desmoid tumor. After 4 years, the tumor arrested spontaneously and the patient had achieved SD at the time of last follow-up. In one patient, multiple mesenteric desmoid tumors were found incidentally during laparatomy, but no specific therapy was recommended. This patient was lost to follow-up.

Table 5. Clinical Course of Desmoid Tumor Patients with FAP and no Tamoxifen/NSAIDs Therapya
Patient no.Age (yrs)GenderSiteProphylactic colectomybTherapyClinical course
  • FAP: familial adenomatous polyposis, NSAIDs: nonsteroidal antiinflammatory drugs; F: female; IAPA: colectomy with ileoanal pouch anastomosis; PD: progressive disease; SD: stable disease; PR: partial regression; CR: complete regression; M: male; Ca: carcinoma.

  • a

    Clinical course with or without radiologic responses.

  • b

    Previous colectomy years to first appearance of desmoid tumors.

1838FIntraabdominalIAPA (−5)R2 resectionPD, after 4 yrs SD
1939FIntraabdominalIAPARadiationPR, after 8 yrs CR
2030FAbdominalIAPA (−2)NoneCR
2129MIntraabdominalColon Ca, segment resectionNoneLost to follow-up
2227FAbdominalIAPA (−2)NoneCR

Clinical Course of Patients with Sporadic Desmoid Tumors Treated with Antiestrogens and NSAIDs

Eight patients with extraabdominal desmoid tumors were followed. Five patients had undergone single or multiple desmoid tumor-related surgical resections before they were referred to our hospital. Four of these 5 patients (80%) developed tumor recurrences. After subsequent drug treatment with 120 mg of tamoxifen (or raloxifen) and 300 mg of sulindac, 1 patient developed PD, 2 patients had SD, and 1 patient achieved a CR. The patient with no tumor recurrence remained disease free, having started prophylactic antiestrogen/NSAID therapy 9 months after undergoing complete surgical resection. Three patients were treated for primary disease with tamoxifen and sulindac at our institution. They had achieved SD at the time of last follow-up (Table 6).

Table 6. Clinical Course of Sporadic Desmoid Patients Treated with Antiestrogens and NSAIDsa
Patient no.Age (yrs)GenderSitePrevious surgery/traumabDesmoid-related surgeryAdditional therapyClinical courseDrug in mg (mos drug received)Response
  • NSAIDs: nonsteroidal antiinflammatory drugs; F: female; PD: progressive disease; T: tamoxifen; S: sulindac; CR: complete regression; To: toremifene; SD: stable disease; M: male; Ra: raloxifene.

  • a

    Clinical course with or without radiologic responses.

  • b

    Previous colectomy years to first appearance of desmoid tumors.

2334FShoulder girdleYes (−3)R2 resectionRadiationPDT 120 (12) S 300 (12)CR
2425FLower extremity, pelvic girdleNoFour resectionsMistelinstallation3 recurrencesTo 240 (24) S 399 (24)PD
2561FShoulder girdleNoNoNo T 120 (43) S 300 (43)SD
2633MThoracicNoNoNo T 120 (34) S 300 (34)SD
2746FLower extremityNoSeveral resectionsNoRecurrencesT 120 (36) S 300 (36)SD after 12 mos
2853FLower extremityNo41 resectionsNo3 recurrencesT 120 (37) S 300 (37)SD
2964MParavertebralYes (−2)NoNoNoT 120 (31) S 300 (31)SD
3034FLower extremityNoR0 resectionNoNo recurrenceRa 120 (5) S 300 (5)Disease free

Side Effects

Of the 20 female patients who were treated with tamoxifen, 8 developed ovarian cysts during therapy. Two of these women experienced spontaneous reduction of tumors. None of these patients required therapeutic intervention other than percutaneous needle aspiration. Dose reduction or administration of raloxifen in place of tamoxifen was beneficial. Five patients suffered from chronic tiredness and fatigue, but were able to comply with their daily obligations. Three patients reported weight gain, which necessitated a dosage reduction.


The basic underlying etiology for desmoid tumor development is a defect in connective tissue formation, although to our knowledge the complete mechanism is not fully understood.27 Surgical trauma can induce desmoid growth and appears to be one of the most important etiologic factors.28, 29 The findings of the current study are consistent with this hypothesis. In the FAP study group, 77% of the patients developed desmoid tumors at a mean of 5 years after undergoing a colectomy. Desmoid tumors developed with similar frequency in patients who received total proctocolectomy compared with those who underwent colectomy with ileorectal and ileoanal anastomosis, or those who underwent a partial colectomy.30 It appears likely that the occurrence of desmoid tumors depends on the surgical trauma per se and this effect does not appear to depend on the particular type of colorectal surgery.

Considering these findings, it seems reasonable that surgical removal should not be the first line treatment of patients with desmoid tumors. Jones et al.17 and Harvey et al.10 proposed that radical surgery should not be performed in patients with FAP and mesenteric desmoid tumors except for symptomatic tumors. The high morbidity associated with resection lengths of small intestine that must be removed to obtain a tumor-free surgical margin and early tumor recurrence have been cited as the primary reasons for these recommendations. Jones et al.17 reported a 60% morbidity rate for 10 patients who underwent resection or attempted resection. In the current series, all four patients with FAP-associated desmoid tumors and four of five patients with sporadic desmoid tumors developed disease recurrences after receiving desmoid tumor-related surgery at other institutions. A bias against surgical control cannot be excluded in the current study because, most likely, only the patients who developed recurrences after desmoid tumor-related surgery were transferred to our hospital for further treatment. Therefore, we cannot generalize the expected recurrence rate for all patients with desmoid tumors treated surgically based solely on the results of the current study. However, a comparison of these results to other data in the literature shows that desmoid tumor recurrence is the rule rather than the exception. For patients with extraabdominal tumors, local recurrence rates are reported to range from 24–77%.31, 32 An overall recurrence rate of 40% with surgical excision alone and a 77% recurrence rate for mesenteric desmoid tumors have been reported.11 These results combined patients with sporadic and those with FAP-associated desmoid tumors. Berk et al.33 observed a recurrence rate of ≤ 85% in patients with FAP-associated mesenteric desmoid tumors after resection. In the current study, when possible, we avoided additional surgical intervention after patients with FAP-associated and sporadic desmoid tumors received prophylactic colectomy. However, despite the risks, there remains a role for surgery in the management of desmoid tumors. In case reports by other authors, surgery was beneficial for patients with desmoid tumors who required life-saving surgery as a result of failure of nonsurgical treatments or the development of complications.34

In the current study, we observed a significant 3.5-fold increased risk for the development of desmoid tumors among females. This finding is in agreement with the data from other authors, who reported a predominance of female cases.8, 17, 34 Clark and Phillips35 reported an increase in the risk for desmoid tumors with the current use of oral contraceptives, as well as a reduction of risk after menopause, pinpointing the role of sexual hormones in the pathogenesis of desmoid tumors. The predominance of cases affecting young women during or after pregnancy is particularly apparent.17, 34–36 However, studies have shown that desmoid tumors in association with pregnancy had a significantly more benign course and a tendency toward smaller, less symptomatic tumors that required treatment less frequently.37 These findings raise two questions: 1) is it the trauma of pregnancy that induces desmoid growth and 2) does the hormonal influence itself ameliorate disease expression? In the current series, none of the female patients were pregnant or had recently given birth. However, we are aware of two female patients with FAP-associated desmoid tumors (one mesenteric and one extraabdominal) who became pregnant. In both women, the desmoid tumor showed SD during pregnancy but grew rapidly after delivery. An increased risk for the female gender is consistent with the observation that desmoid cells may express estrogen receptors and are inhibited in their in vitro proliferation by antiestrogen compounds.38 Estrogen treatment may induce, both in animals and in humans, the formation of desmoid tumors that regress after cessation of drug administration or after addition of progesterones.39 Tamoxifen has antiestrogen activity and has been shown to be partially responsible for the antidesmoid effect.40 However, tamoxifen also has been shown to produce a response in desmoid tumors without estrogen receptors. In the study by Lim et al.,38 estrogen receptors were found to be present in only 33% of the desmoid tumors analyzed (5 of 15 tumors) with equal incidence reported in males and females. It is interesting to note that antiestrogen binding sites were detected more frequently (in 77% of the tumors). Later studies have shown tamoxifen to display other mechanisms of action such as the alteration of metabolic pathways by inhibition (ornithine decarboxylase, cholesterol synthesis, microsomal lipid peroxidation, proteine kinase C).41 Another finding was the correlation between a decreasing concentration of transforming growth factor-β1 and the inhibition of aberrant fibroblasts in desmoid tumors.42

In addition to the avoidance of surgical trauma, our therapeutic regimen reflects the finding that hormonal factors have been implicated in the growth rate of desmoid tumors with estrogen predominating over progesterone.27 In comparison with literature reports with low-dose tamoxifen (30 mg daily), higher dosages of tamoxifen were found to be more effective in inhibiting desmoid tumor growth in the current observational study of a large patient group with long follow-up. This observation is consistent with other reports in literature. For example, Brooks et al.43 described the use of antiestrogens in 20 patients with progressive desmoid disease. In their study, the standard tamoxifen dose was 200 mg per day and clinical responses ranging from SD to CR were observed in 65% of patients. Side effects were infrequent and generally mild. Other literature reports confirmed the success rate and the minimal level of associated toxicity with tamoxifen.44–46 In the current study, a follow-up of ≥ 3 years demonstrated a long-lasting effect with no tumor recurrences observed after the cessation or reduction of therapy.

The use of prostaglandin as a therapeutic agent in patients with desmoid tumors is based on the observation of regression of fibrosarcomas in rats after prostaglandin blockade.47 A role of trauma-mediated stimulation of prostaglandin synthesis and of T-suppressor cell number in desmoid formation has been suggested. Based on these findings, NSAIDs were introduced in the treatment of FAP-related desmoid tumors with some encouraging results,48 including the use of sulindac.9, 49 Sulindac is relatively nontoxic and has been recommended as a combination agent after encouraging results were obtained in larger studies.49 The value of new agents, such as the COX-2 inhibitors or others, remain to be assessed.46 However, the success rate of the combined therapy in the current series leads us to recommend sulindac as the first-line agent in combination with antiestrogens.

Of 13 patients in the current study who received tamoxifen and sulindac as primary treatment for FAP-associated desmoid tumors, 6 had SD and 4 achieved PR to CR for > 6 months. Three of 13 patients were stable for > 30 months. In the group of patients with sporadic desmoid tumors, 3 patients had SD for > 30 months. Because all patients had PD at the time of referral, SD was considered a meaningful response. Based on our experience, a PR to CR after further tumor stabilization is probable.

Eight of nine patients developed tumor recurrences or PD after complete or partial resection of the desmoid tumor. With receipt of the tamoxifen/NSAID therapy, three of the four patients with FAP and one of the four patients with sporadic desmoid tumors had PD. The other four patients demonstrated either no change or a PR to CR. One patient (Patient 23) with a recurrent sporadic desmoid tumor achieved a CR after only 6 months of tamoxifen/NSAID therapy, having undergone previous radiotherapy. A delayed response because of the previous radiotherapy is possible. In the FAP group, we attribute the poor response noted among patients who received previous desmoid surgery to the finding that these tumors were in a state of maximal surgically induced progression. Compared with patients who did not receive previous surgery, patients who received previous surgery had a poor response, even if they received medication because the time period required is prolonged. Because tumor-free surgical margins most likely will not be obtainable without severe morbidity to these patients, we consider high-dose tamoxifen and sulindac to be the first-line treatment even in this setting in which response rates are less satisfactory. For patients with sporadic desmoid tumors, it remains to be determined whether surgery reduces the beneficial effect of the tamoxifen/sulindac therapy. However, in severely ill patients, radiotherapy has led to responses.

Desmoid tumors respresent a heterogenous group of tumors and it is difficult to predict the biologic behavior. Patients presenting with small and slow-growing desmoid tumors were estimated to have a good prognosis. They did not receive medication and our estimate was confirmed by the observation of 1 patient with SD after 4 years, 2 patients with spontanous CR, and 1 patient with CR after radiotherapy during follow-up. Desmoid tumors and their inherent destructive potential are most difficult to predict. To our knowledge, there are no histopathologic or molecular criteria to identify a more “benign” or a more “malignant” tumor. For clinical purposes, we propose a clinical “DES” classification for desmoid tumors that considers the primary dimension (D) in centimeters of the tumor, the expansion (E) of growth defined by the doubling rate in months, and the site (S) of the tumor. This classification has been helpful in our studies of desmoid tumors and could present the basis for further studies enabling “staging” and standardized therapy recommendations for this potentially lethal tumor.