Surgery has been widely used in all types of resectable desmoid tumors. Surgery is the first-line treatment for tumors with a clear margin in most cases (17). It was found that patients in case reports with intra-abdominal desmoids attain a non-recurrent duration varying from <12 months (18) to 30 months (19), excluding those who were lost to follow up and cases that were not mentioned. In addition, the local recurrence rate is 8–31% (20,21) after surgery with a clear margin. A study involving 176 patients receiving a complete excision with a clear margin showed a 10-year local recurrence rate of 58% (22). A meta-analysis containing 1,005 patients receiving surgery alone showed a recurrence of 30% (23). Other reported recurrence rates are even higher. It was reported that local recurrence rates still range between 19–77% with a wide excision (24,25). However, according to a systemic review (26), the local control rate of desmoids with R0 excision from several studies vary from 47–86%, which is not so disappointing.

As for the impact of margin on prognosis, there are different views (16). Several studies (27–29) favored that negative excision did not make a difference in prognosis, however others (16) did not discover such an association. For example, R0 excision did not favor R1 excision in progressive-free survival (PFS) in a univariate analysis (16), but a multivariate proportional hazards regression analysis revealed that state of margin significantly impacted the prognosis (positive vs. negative; P=0.001; relative risk, 2.9) (30). Considering its infiltrate characteristics, it is difficult to distinguish the boundaries of desmoids. Simultaneously, an aggressive excision for negative margin cannot exactly assure a better outcome, which may sacrifice important function or even lives. Hence, the benefits and faults of excision should be carefully considered. In addition, post-operational radiotherapy is supposed to improve the prognosis of those with a positive margin after excision; the 10-year recurrence rates for patients treated with surgery alone and patients treated with combined surgery and radiation therapy were 54% (95% CI; 37–69%) and 31% (95% CI; 15–53%), respectively (P=0.007) (30). Overall, it is risky and difficult to perform complete excision.

Surgery has substantial limitations. To begin with, although wide local excision is the gold-standard treatment for desmoid tumors in the past, no solid evidence has affirmed that surgery yielded superior outcomes over other treatments (16). Even other conservative measures are favored as these measures have lower complication rates. Secondly, scientists have found that in a series of 203 patients with extra-abdominal desmoid tumors treated with surgery over a 35-year period, patients without antecedent interventions had longer disease-free survival and presence of recurrence was the prognostic factor indicating strong recurrence (31). Given that surgery, as a type of trauma, is an etiological factor for desmoid tumors and long-term recurrence rates of surgery are quite high, surgery should be treated with caution. Simultaneously, the boundaries of the tumors are hard to distinguish intraoperatively from scars and connective tissue, causing difficulty in R0 resection (32). Recurrence after surgery not only occurs in situ but also in adjacent areas, as it is unable to preserve connective tissues well (33). Under these conditions, repeated recurrent desmoid tumors would tend to invade surrounding tissues in a concentric way if treated with surgery, which would gradually invade more tissues and important organs or major vessels. Radical surgery is always infeasible in mesenteric desmoid disease as the mortality of surgery is 10–60% and the recurrence rate is 78% (34). Surgery itself subjects patients to post-operational risks in addition to its high recurrence rate (up to 40%) (3).

With multi-institutional studies having demonstrated spontaneous regression rates of 28–50% in recent years (35–37), interventions should be treated discreetly. The effect of the watch-and-wait policy is endorsed by the number of cases. A clinical study demonstrated growth arrest in two thirds of the nonsurgical group (38). In a study (16) involving 27 patients with sporadic desmoid tumors treated with the watch-and-wait policy, only six progressed with a delay to progression ranging from 7.8 to 46.2 months. The remaining 21 patients avoided surgery with an ideal outcome, in which the disease attained a stable course, and even five cases regressed spontaneously. Cho et al (39) reviewed five studies in which desmoids are managed by the watch-and-see policy with varying rates of spontaneous regression from 65.1–96.1%.

The watch-and-wait policy is recommended by NCCN 2018 guidelines where observation applies to selected patients in non-life-threatening situations. However, it is confusing to judge the criterion for ‘selected patients’, and Zhao et al (40) believes that surgery should be performed once the disease has progressed. Others believe that the endpoint of this type of modality depends on the patient's symptoms, medical status and the biological properties of the tumor (41). Briand et al (42) reported a low probability of transiting to other salvages which was only 5.7% at 1 year and 9.6 at 5 years. Hence, it seems that watch and wait is quite safe and may be a promising treatment for first-line therapy, especially considering the high local recurrence rate of other interventions, along with its low rates of dropping out.

TKIs were found to have response rates of 5–20% at 1 year, and the rate of progression-free disease varies between 60–70% (57,58). An uncontrolled, non-randomized, phase III study showed high rates of stable disease up to 50–80% (59,60). Imatinib, sorafenib and sunitinib are common TKI drugs targeted for desmoids, which have been extensively studied (60–66).

Imatinib therapy has a response rate of <10% for solid tumors (41). Imatinib treatment is reported to attain progression-free survival (PFS) rates of 58% at 24 months and 55% at 36 months (59,60). A prospective phase II trial demonstrated that PFS rates were 94% at 2 months, 88% at 4 months and 66% at 1 year. However, it is associated with toxicities such as grade 3/4 neutropenia, rash, fatigue, abdominal pain, diarrhea, myalgias and asthenia (59,60).

Skubitz et al (67) reported a case of multifocal desmoids that was refractory to imatinib but responded well to sunitinib. Based on the aforementioned study, a prospective multicenter phase II study (68) evaluated the efficacy of sunitinib, in which the rates of partial response and stable disease were 26.3 and 42.1%, respectively. Additionally, the 2-year progression-free and overall survival rates were 74.7 and 94.4%, respectively. Side effects that occurred in >5% of patients included neutropenia (33.3%), diarrhea (5.3%) and hand-foot syndrome (5.3%). In the early phase of sunitinib, mesenteric mass bleeding, bowel perforation along with bowel fistula due to bowel mass necrosis were respectively observed in three patients. The author believed sunitinib to be a useful management of non-mesenteric desmoids.

A study elucidated the efficacy of sorafenib and its side effects (69). Sorafenib is administered as the first-line or substitute treatment with a striking clinical benefit comprising 16 of 22 patients. Moreover, no difference was found in radiological benefit (P=0.9). Toxicities include hand-foot syndrome, fatigue, skin rash, trichodynia, hypertension, mild alopecia and diarrhea, which can be controlled with dose reductions and the use of antidiarrhea and antihypertensive drugs. However, another double-blind, phase III trial study comparing the efficacy of sorafenib vs. placebo demonstrated objective response rates [sorafenib vs. placebo: 33% (95% CI, 20–48%); 20% (95% CI, 8–38%)]. In addition, the 2-year PFS rate was 81% (95% CI, 69–96%) in the sorafenib group and 36% (95% CI, 22–57%) in the placebo group (hazard ratio for progression or death, 0.13; 95% CI, 0.05–0.31; P<0.001). The sorafenib group has similar side effects as the previous study. In addition, desmoids harboring the S45F mutation with anti-cytotoxic effects caused by autophagy may respond well to sorafenib and hydroxychloroquine (an autophagy inhibitor), as reported by Braggio et al (70).

The combination of MTX and VBL (MTX/VBL) were studied for desmoid treatment and yielded a promising effect (54,71–74). According to the literature, the response rate for MTX/VBL varied from 31–52% (72,75), but the data is too conservative if more prognosis data are included. A systematic review that evaluated the efficacy of MTX/VBL combination chemotherapy on primary or recurrent desmoids showed a mean response rate of 36% (11–57%) or 85% in patients with stable disease (71). In addition, it was reported that patients with stable disease progression accounted for >35 and >28% attain a PFS duration of a mean of 43.4 months in a phase II study examining the effect of MTX/VBL therapy on children (72). Another similar phase II study (73) focusing on adults showed a 100% rate of clinical benefit with ratios of stable disease and partial response accounting for 60 and 40%, respectively. The percentage of patients with a 10-year PFS interval was up to 67%. As reported, young age is an indicator of bad prognosis, which applies to those treated with chemotherapy according to aforementioned studies.

This combination modality also has some side effects, although the rates of adverse effects are low according to studies on children (54,72,75). Neutropenia (myelosuppression or myelotoxicity) (71,72) was reported to be the main adverse effect. Moreover, anemia, nausea, mild alopecia, vomiting and elevation in hepatic transaminases, which were reversible with interruption of chemotherapy, are all common side effects (54,72,73). The duration of response to MTX/VBL is longer in children, typically ~9 months (76). Due to toxicities of this modality, the combination of vinorelbine and MTX can substitute the previous regimen for decreasing toxicities without sacrificing the response rate (77). Considering its low rates of side effects which are dominantly due to myelotoxicity, which would make it feasible to impose prophylaxis ahead of time, the MTX/VBL modality is a promising and safe modality.

Conventional chemotherapy is generally applied to patients who fail the ‘low-dose’ regimen (78). PLD or LD is a variant of doxorubicin (a type of topoisomerase II inhibitor), which is decorated by liposomes with or without polyethylene glycol for decreasing the cardiotoxicity of doxorubicin (79). Several studies have shown its high rates of response. A study reported that four patients with advanced desmoids responded well to PLD without significant toxicities (80). Another study involving 12 patients demonstrated satisfactory results in which the rate of stable disease was up to 64, and 75% of patients did not attain progression during the entire follow up with a median duration of 14 months (81). A retrospective review of pediatric patients analyzed the effects of LD and an average 4.5% reduction of tumor size and a median PFS time of 29 months, which showed more ideal results (76). In addition, doxorubicin can also be used to assist carboplatin in further reducing the tumor size by intralesional injection, which was reported in a cohort study (82). Additionally, some researchers have discontinued doxorubicin or adjusted the dose of carboplatin to a cumulative of 400–500 mg/m², due to its cardiotoxicity. Although promising, doxorubicin also has some side effects, such as palmar-plantar erythema, mucositis, cardiotoxicity (83) and dose-relevant reduction (81).

Loco-regional chemotherapy is rarely reported in some cases, in which TNF-α and melphalan is perfused into isolated limbs (84). Data from three sarcoma centers of the European Organization for Research and Treatment of Cancer indicated that TNF-α and melphalan could prevent amputation in 88% of patients with advanced disease. Up to 61% showed regression or stabilization of disease during a median follow-up of 84 months and the PFS time was as long as 27 months (85). It acts as a substitute in patients affected by carcinoma in extremities, especially for those who have multifocal desmoids of hand or foot (84). This regimen helps patients avoid routine operation since state of stable disease or slow regression is often observed in patients after therapy (84).

The VAC regimen was firstly reported by Raney (86) to treat fibromatosis in a series of cases including six patients aged 3 months to 7 years with primary or recurrent desmoid tumors. This regimen is an alternative to doxorubicin-based chemotherapy with comparable responses, particularly in pediatric patients. The main side effects of VAC are sterility and carcinogenesis (55,87). The duration from the first administration of the VAC regimen to response is estimated to be at least 12 weeks, according to the Intergroup Rhabdomyosarcoma Study (88). However, others (89) believed that VAC regimens should be given up if no response was observed within a duration of 16–20 weeks following the initial administration. Additionally, most clinical reports regarding the VAC regimen involve children. Prognostic data are not available considering the rare cases and clinical reports, which can be supplemented by more trials.

The present review listed five different regimens that are formed by combinations of different agents and commonly used in desmoid chemotherapy, although other agents such as hydroxyurea (76), γ-secretase inhibitor (90) and pazopanib (91) are also involved in chemotherapy trials. These five combinations are mainstream chemotherapy regimens whose efficacy and side effects have not been comprehensively reviewed in previous literatures.

Percutaneous cryoablation currently delivers room temperature argon gas through a sealed, segmentally insulated probe to cause rapid cooling locally due to the Joule-Thomson effect (100). According to Kurup and Callstrom (101), percutaneous cryoablation is a less-invasive treatment potential to cure desmoid tumors. A study (33) was performed to evaluate the efficacy, safety and advantages of cryoablation for both first-line treatment and remedial treatment of extra-abdominal desmoid tumors. A total of 89% of patients reported improvement and the average change in viable volume was −80% (range, −100 to +10%) with complete response of 36%, partial response of 36% and stable disease of 28% after a 1-year follow-up. The high rate of incomplete ablation, achieving a 90% clinical response rate as well as its less-invasive feature suggested that percutaneous cryoablation is a promising modality to cure extra-abdominal desmoid tumors. Similar results affirmed the efficacy of cryoablation for its high rate of local control (102,103). In addition, cryoablation has been investigated as both first-line and salvage treatment modality (33) and an ideal outcome has affirmed its reliability.

Cryoablation owns several advantages over other therapies. In contrast to a 23% recurrence rate in a 2-year follow-up after R0 excision by imaging, complete A0 ablation extinguishes diseases (16). Moreover, it was reported that the recurrence after cryoablation often occurs in situ. However, recurrence sites after surgery are located in adjacent tissue, complicating further therapy by causing neurovascular injury. This is due to connective tissues or compartments being well-preserved in cryoablation, unlike in surgery (33). Cryoablation has better local control which may substitute surgery or radiotherapy, neither of which alone could attain an ideal local control (47). Finally, nearly all cryoablation is less invasive and patients would return home within one or two weeks, unlike in the surgery group depending on the complexity of excision and recovery state (33).

Percutaneous cryoablation owns several limitations or side effects. Mostly, this type of modality applies to extra-abdominal and abdominal wall desmoid tumors and no reports reported treating intra-abdominal desmoid tumors with this modality. Furthermore, longer term follow up is needed to assess how A0 cryoablation compares to the R0 5-year recurrence rate of almost 40% (47). Injury to skin and nerves are the most common complications caused by cryoablations.

Hormonal therapy typically consists of the use of selective estrogen receptor modulators (tamoxifen, raloxifene and toremifene) and occasionally involves the use of luteinizing hormone-releasing hormone analogues (leuprolide) and other hormones that antagonize estrogen (medroxyprogesterone, testosterone and progesterone).

Tamoxifen, raloxifene and toremifene are all anti-estrogen agents that are used for desmoids. Lackner et al (114) described the efficacy of tamoxifen and toremifene with a response rate of 65% and scarce side effects in treating desmoids. According to a systemic review, the overall response (complete plus partial response) rates and disease stabilization rates are similar in tamoxifen and toremifene (115), which are 58 and 18% vs. 56 and 19%, respectively. Moreover, according to Okuno (116), dosage increases of tamoxifen are not more effective although some clinical trials elevate the dosage of tamoxifen when its effects are not satisfactory. Toremifene is effective in some cases refractory to tamoxifen. Toremifene does not exhibit long-term complications (117–120), however it should be reserved for patients with a history of venous thromboembolism or pulmonary embolism given the risk of thromboembolic events observed following long-term treatment with raloxifene.

Some reports have demonstrated the effects of anti-estrogens combined with other drugs. Hormone-based treatment showed a response in 40–51% of cases (115,121). It was reported that ~50% of patients with mesenteric desmoids respond to anti-estrogens (108). In addition, a systemic review focused on the effects of anti-estrogen with or without other drugs, such as NSAIDs, concluded that overall response (complete response plus partial response) rates and disease stabilization rates are 51% (46/91) and 27% (24/91), respectively in desmoids with FAP, and 48% (24/50) and 38% (19/50), respectively, in sporadic desmoids (115). The author of this review also found a higher rate of complete or partial response in anti-estrogen alone compared with in combination with NSAIDs (P<0.001). In addition, the regimen that combined sulindac (an NSAID) with anti-estrogen agents was tapered after a median duration of medication of 42.4±24.3 months, which could not be influenced by genetic state, sex or a concrete type of anti-estrogen agent applied to desmoids (122).

The most common hormone-based therapy is tamoxifen in combination with or without other drugs. Bauernhofer (123) reported a female patient with FAP suffering from recurrent unresectable intra-abdominal desmoids. Tamoxifen with subsequent goserelin acetate (an analogue of the luteinizing hormone-releasing hormone) attained a PFS duration of up to 17 months, and tumor progression was again inhibited by goserelin acetate combined with IFN-γ. Tamoxifen combined with sulindac have been well investigated in clinical trials. A phase II study within the Children's Oncology Group demonstrated overall response, 2-year PFS and survival rates of 36% (95% CI, 0.23–0.48) and 96%, respectively (124). This study also reported relevant side effects including gastritis, emesis, hemorrhage events, ocular problems, pain in the abdomen or head, embolism-related events, tinnitus, prolonged QTc interval, fever, fatigue, skin breakdown/decubitus, nausea, elevated alanine transaminase/aspartate transaminase ratio, hypomagnesemia, ovarian cysts and dizziness. Another systemic review revealed that the side effects of anti-estrogen occur at a lower rate compared with surgery and chemotherapy (125). Simultaneously, selection of drugs was significantly impacted in cases of side effects of corresponding drugs (122); raloxifene was preferred in female patients in order to reduce the risk of endometrial cancer and ovarian cysts, while male patients were preferably administered with tamoxifen.

NSAIDs used to be the first-line therapy due to their low toxicity (39). Common NSAIDs used for desmoids include meloxicam, indomethacin, sulindac and celecoxib.

Sulindac is always used in combination with tamoxifen. Tsukada et al (126) found that the overall response rate was up to 57% with a mean response time lag of 24 months in assessing the efficacy of sulindac in 14 patients with FAP suffering from recurrent desmoids. Quast et al (122) reported common side effects of sulindac such as high elevation in liver function tests which occurred in 1.5% of patients and venous thrombosis in 2.2% of patients.

Meloxicam, which belongs to the cyclooxygenase-2 (COX-2) inhibitor agent group, has shown its effects in managing extra-abdominal desmoid tumors (127). A retrospective review analyzing the outcomes of primary or recurrent extra-abdominal desmoid tumors treated with meloxicam showed state of partial relief or stability in 65% of patients. The rates of dropping out from this treatment remain at 35% at both 1 or 5 years (39). Meloxicam efficacy was demonstrated by Nishida, who followed 20 patients treated for 3–81 months; over the 44 months of follow up, 95% of patients experienced equal to or better than stable disease (128).

Yang et al (129) first reported a case of patients with aggressive fibromatosis (desmoid tumor) treated with celecoxib (a type of NSAID) under the guidance of genetic testing, in which the patients demonstrated significant regression without any signs and symptoms during a 20-month follow-up. They recommend this modality under the condition that genetic testing showed the presence of p.T41A mutations on the CTNNB1 gene, which could predict whether the patient is sensitive to the COX-2 inhibitor celecoxib or not. A previous study used celecoxib combined with excision, in which complete remission without local recurrence was accomplished within a duration of a 24-month follow-up (130). However, multicenter randomized controlled trials are required in the future.

An early report indicated that indomethacin at an oral dose of 100 mg/d caused resolution of the tumor (131). According to the report, this drug caused complete resolution of desmoids which was partially responsive to radiotherapy in one previous case. Another case showed immediate but short response and resumed resolution for 14 months with large doses of ascorbic acid being supplemented with indomethacin. The remaining two cases received combination indomethacin from the beginning and showed shrinkage of tumors during the entire follow-up.

In addition, other drugs used for the attempted treatment of desmoids showed certain efficacy such as 1,25-(OH)2-vitamin D3 (132,133), IFN-α (134), retinoic acid, tranilast (135), predinisolone (136,137) and ascorbate (138,139). These strategies are not included in conventional therapy but may act as second line therapies. Ferah et al (133) reported a female patient with a huge desmoid mass in the right shoulder region, resistant to radio-chemo-hormonal therapy, that eventually responded to radio-chemo-hormonal therapy. A clinical trial including 13 patients who previously received surgery for desmoids were recruited to examine the efficacy of further management of IFN-α with or without tretinoin (134). The results (seven patients had no evidence of disease and a mean disease-free interval of 22±18 months; in two patients, progressive disease occurred after only 7 and 9 months, respectively, during observation) revealed that this regimen could significantly prolong the disease-free interval. Tranilast was administered to a Japanese male with desmoid tumors on his chest wall who refused to surgery-radio-chemical therapy and the tumor was successfully controlled (135). Predinisolone was reported to induce intra-abdominal desmoids with FAP to regress or even disappear in an unresectable case (137).