Background

Desmoid tumors arise from connective tissue of muscles, the fascia, or the aponeurosis and may occur at multiple anatomic sites [1]. Most often they occur in areas of prior trauma and scars and in Gardners Syndrome in the abdominal cavity. Clinically, they present as a slow growing mass with little associated pain or tenderness, and are often thought to be a hematoma or hernia. Grossly the tumor presents as a dense, poorly vascularized, hard, rubbery, grayish-white mass. Histologically, they are characterized by dense collagenous material with intertwining bundles of spindle cells lacking epithelial components [2, 3]. Most often they are confined locally, but may show infiltration of adjacent structures. Unlike fibrosarcomas, with which they may be confused, desmoid tumors have low mitotic rates [3, 4]. Because they rarely metastasize, the local infiltration and compression of surrounding structures results in a high local recurrence rate particularly when found in areas where complete surgical resection is not possible. Sixty percent of desmoids are at an extra-abdominal location, and of these up to 25% are located in the abdomen wall. Another 15% of desmoids arise in the abdominal cavity [5]. Most of the latter are in patients with a familial adenomatous polyposis (FAP), and when the two occur together the term Gardners Syndrome is often used. This is a special and separate entity of desmoid tumors, and the only form understood in terms of development and molecular pathogenesis. In these desmoids, there is abnormal expression of c-cis leading to heightened production of platelet-derived growth factor [2]. These findings have not been found in non-FAP related desmoids.

Extra-abdominal desmoid tumors in the rectus abdominis muscles in postpartum have been described in individual case reports, but no series such as reported here and in this location, has been forthcoming. While the association of pregnancy and the development of desmoid tumors is well known, the cause and the nature of this relationship is unclear. Typically, these tumors occur in young, gravid women, and as reported here, during the first year after childbirth. The association between pregnancy and the postpartum state and the development of desmoids has been a source of speculation for many years, but no definitive answers have been forthcoming. The expression of estrogen receptors has been examined, and has been reported to be uniformly negative although there are occasional reports of responses to tamoxifen [6, 7]. The mechanism involved in these responses is likewise unclear.

With the increased understanding of the molecular factors that lead to other similar fibrotic tumors such as gastrointestinal stromal tumors (GIST) and dermatofibrosarcoma protuberans (DFSP), desmoid becomes a more interesting area of investigation. We suspect that others have observed the syndrome, namely, a desmoid tumor of the right rectus abdominus muscle in postpartum women, and this clinical picture may lead us to new molecular studies and therapy for this form of desmoid tumors.

Methods

We reviewed the clinical records and pathology of all patients with desmoid tumors seen at University of Colorado Cancer Center during the years 1999–2007. Among these were four postpartum women with identical anatomic location of the tumor, the right rectus abdominus muscle. These were selected for further review after removal of personal identifying data and approval by the Institutional Review Board of our institution in accordance with the ethical standards of the 2004 Declaration of Helsinki.

Clinical characteristics

A summary of the clinical characteristics of the four cases is shown in Table 1, following the individual case descriptions. They ranged in age from 32 to 39 with a mean of 36.5 years. Three were Caucasian and one was of Middle Eastern origin. None had prior history of a desmoid or trauma to the right upper abdomen, family history of FAP or Gardners Syndrome. The time from delivery to discovery of the desmoid tumor varied from 1 to 13 months, with an average of 9 months. The major clinical finding in all was a painless mass located in the right upper abdomen. All were confirmed by biopsy, and pathologic examination showed identical features in all four cases characteristic of desmoid tumor. An example from case 3 is shown in Fig. 1. The tumors varied in size from 2 to 16 cm in maximum diameter. Various imaging studies were done prior to surgery with CT scans being the most frequent and helpful, a typical example of a CT scan, from case 3, is shown in Fig. 2. All were resected surgically without complications, and have been followed by examination and CT scans for at least 2 years with a median follow-up of 2.5 years. None have had further pregnancies, and none local or systemic recurrence.

Table 1 Summary of the postpartum woman cases with abdominal wall desmoid tumor
Fig. 1
figure 1

Microscopic appearance of desmoid tumor resected from case 3. H&E stained section showing spindle cells dispersed hyalinized collagen. No nuclear pleomorphism or mitotic figures are present ×100

Fig. 2
figure 2

Abdominal CT scan from case 3 demonstrating the characteristic features of those desmoids in the right rectus abdominis muscle (arrow)

Case 1

A 36-year-old G1P1 female noted a firm mass in the right upper abdomen a year after delivering twins by cesarean section. Her history was remarkable for successful in vitro fertilization. A preoperative CT scan showed a vague area of fullness in her right upper rectus abdominis. Subsequently, a low-grade 2-cm desmoid tumor was resected with evidence of positive margins on pathology. She did not receive adjuvant therapy, and 3-years post-surgery has no evidence of tumor recurrence on repeat CT scans.

Case 2

A 39-year-old G3P3 female recognized a mass in her right mid upper abdomen 4 weeks postpartum from her third cesarean section. The mass was initially thought to be a hernia, but at surgery was found to be 2-cm desmoid tumor of the right rectus abdominus that was completely resected. Her postoperative course was uneventful, and at her last follow-up 2 years after surgery, has no evidence of recurrence by exam and CT scan. The patient had a uterine myomectomy for fibroids 2 years before, and three prior uneventful pregnancies. She also had fertility treatment for her second and third pregnancies.

Case 3

A 32-year-old G1P1 female reported sustained pressure in her RUQ 1 year after uncomplicated vaginal delivery. The pressure was present over a 3-month period, and then evolved into pain. An ultrasound and subsequent CT scan (Fig. 2) revealed a well-demarcated mass in her right upper abdominal wall. It was thought to be a hematoma derived from occupational heavy lifting. The pain and pressure continued and a core needle biopsy with CT guidance demonstrated a desmoid tumor. She was initially treated with non-steroidal anti-inflammatory drugs (NSAIDs) and tamoxifen for 8 months with no improvement. Imatinib was started without benefit and 3 months later a 16 × 10.5 × 5.8-cm desmoid tumor was excised from the right rectus abdominus with clear margins. A significant abdominal wall defect resulting from the resection was repaired with mesh. Four years after surgery, the patient had no recurrence.

Case 4

A 39-year-old G1P1 female was evaluated for a RUQ mass and found 5.5 × 4-cm desmoid tumor that was resected with clear margins 1-year postpartum. Other than a routine cesarean section in 2005, her history was remarkable for uterine myomectomy due to uterine fibroids, a phylloides tumor of the breast, and papillary carcinoma of the thyroid. The latter was diagnosed after the desmoid, and was treated with a thyroidectomy and radioiodine-131 (RAI).

At follow-up 3 years after surgery, she has no evidence of recurrence of the desmoid tumor by exam or CT scan.

Discussion

There is a clear link between the development of desmoid tumors, pregnancy, and the postpartum state. This association has been noted many times in the past, mainly in single case reports in the literature [810]. What is most interesting about most of these cases, and those reported here is that the desmoids arising in this situation are almost always in the abdominal wall, although in a few cases the tumor arose elsewhere in the body [1114] The pregnant or postpartum state with all the attendant physical, hormonal, and immunologic changes appear to be an important factor, but need some unexplained local effect on the abdominal wall to lead in desmoid formation. Up to 30% of nonpregnant desmoid cases have a history of local trauma, but this is very uncommon in those associated with pregnancy such as those reported here. We have speculated, however, that occult trauma may occur in pregnancy as the result of stretching and/or tearing of the aponeurosis in the abdominal muscles or perhaps associated with fetal movement or kicking.

Numerous studies of desmoid tumors have examined growth rates, hormonal influences (particularly estrogen uptake), and the impact from previous trauma. Extra-abdominal desmoids may express low levels of estrogen receptors (ER), which may account for the occasional response to antiestrogen therapy, but this is uncommon and unlikely to explain their pregnancy-associated development [7].

Surgical trauma has been implicated in the etiology of desmoid tumors; one of the cases reported in the literature is a desmoid tumor arising in a cesarean section scar during pregnancy [5]. Three of our patients had a history of low-transverse cesarean section, but none developed their desmoid tumor, in or close to, their cesarean section scar. We could not find other common features among the patients reported here, except for an older mean age of 39 and prior oral contraceptive use. We speculate that both of these may have lead to more lax abdominal musculature, and the increased likelihood of trauma or tearing of muscles in the abdominal wall.

As with our cases, surgery remains the mainstay of treatment in all patients with extra-abdominal desmoid tumors. There have been no recurrences with a 2-year follow-up in our cases despite one patient having positive surgical margins. The literature presents conflicting data concerning the importance of complete resection. Some authors report that recurrence is independent of negative margins [15, 16] whereas, others demonstrate higher local recurrence rates after close or positive margins, and recommend aggressive resection. The largest review is that of Leithner et al. who examined 12 studies of extra-abdominal desmoid tumors reported between 1969 and 2003. They noted that that 27–30% of the patients with positive surgical margins did not have tumor recurrence as in the cases presented here [15]. Despite this, others have recommended intra-operative microscopic tumor evaluation as a standard procedure to verify that surgical margins are negative; otherwise a more extensive resection should be performed when possible to decrease the chance of recurrence. The question therefore still remains open. It is our feeling that with minimal microscopic involvement, most patients can be safely and adequately followed with modern imaging techniques, particularly when more surgery may result in significant cosmetic or functional defects. Responses of primary and recurrent desmoids tumors to NSAIDs, anti-estrogens, chemotherapy, and radiation have been reported [7]. Their use prior to surgery has not been established. Likewise their role as adjuvant therapy have not been determined. Indeed it is not clear which of these modalities has the greatest and most consistent benefit in recurrent or non-resectable desmoids. This determination will likely come only from new molecular insights into desmoid formation leading to targeted therapies. A number of genes and molecular pathways have been identified as playing some role in desmoids (Table 2), but no clear common molecular pathways have been identified such as those that have been described in GIST and DFSP that can targeted as yet [17, 18]. Henrich et al. examined desmoids from 19 patients for alterations in a number of genes with mixed findings, but no single mutant gene or pathway identified. Further they treated all 19 patients with imatinib; three of these had partial responses and four more had stable disease. They concluded that imatinib is an active agent in desmoids, but the responses were quite low and the mechanism of action unclear. Further studies are clearly indicated and are underway in our laboratory and elsewhere [19]. Despite the common clinical presentation seen in the cases reported here, it is likely that we will find that all desmoids are not alike and may have different molecular basis and require different therapies. We are obviously intrigued by the association of pregnancy and pregnancy-associated trauma. Our studies will concentrate on the genes and receptors involved in muscle and fascia repair and the stimulators of these that may be associated with pregnancy.

Table 2 Genes and receptors that have been found to be associated with extra-abdominal desmoid tumors