Combination therapy with sorafenib and celecoxib for pediatric patients with desmoid tumor

To the editor

Desmoid tumors, also known as aggressive fibromatosis, are local tumors of mesenchymal origin that may cause significant morbidity due to their infiltrative nature.^1,2 Risk factors for the development of desmoid tumors include familial adenomatous polyposis, pregnancy, blunt trauma, previous surgery, and oral contraceptive medications.³ Desmoid tumors can develop in any anatomic location, but they are most commonly found in the extremities, trunk, abdominal wall, and intra-abdominal cavity. The behavior and clinical presentation of these tumors depends on their size, location, and degree of infiltration. While many patients are asymptomatic, the mass effect of the tumor may lead to pain, discomfort, paresthesia, neuropathy, and decreased joint mobility.^2,4

There is no established standardized treatment of desmoid tumor. The decision to treat is based on clinical presentation, location, and behavior.^2,4 Options may include observation, surgery, radiation therapy, chemotherapy, nonsteroidal anti-inflammatory drugs (NSAIDs), or hormonal medications.⁵ Systemic therapy should be considered in a symptomatic patient with an unresectable tumor, when surgical resection would cause significant morbidity, or in the presence of progressive disease.^5,6

Immunohistochemical analyses of desmoid tumor specimens demonstrate consistent overexpression of platelet-derived growth factor receptors (PDGFR) and cyclooxygenase-2 (COX-2).⁷ This finding suggests potential benefit from combining a tyrosine kinase inhibitor (TKI) to target the PDGFR pathway and an NSAID to target COX-2.⁷ In a retrospective review of adult patients with desmoid tumors treated with sorafenib, an oral TKI, 25% of patients had a partial response and 70% had stable disease at a median of 6 months of treatment.⁸ A prospective, randomized, double-blind, placebo-controlled study designed to evaluate the safety and efficacy of sorafenib 400 mg daily in the treatment of adult desmoid tumor patients found an estimated 2-year progression-free survival of 81% in the treatment group compared to 36% in the placebo group.⁹ Additionally, treatment with celecoxib, a COX-2 inhibitor, led to regression of tumor size and reported symptoms in a subset of patients.^10,11 In two adults with desmoid tumors, combination therapy with sorafenib and celecoxib resulted in significant regression of tumor size and symptomatology.¹² One of the patients had progressed while on sorafenib monotherapy, suggesting possible enhanced activity with both agents.¹² As sorafenib and celecoxib act by different mechanisms and lack overlapping toxicities, they are an attractive combination therapy for pediatric desmoid patients. In the following case series, we present two pediatric patients with symptomatic desmoid tumors with favorable response to therapy with sorafenib and celecoxib.

Case 1

A 16-year-old African American male presented to the emergency department with abdominal pain, constipation, urinary frequency and left leg paresthesia. He had associated fatigue, anorexia, and weight loss. Computerized tomography (CT) scan revealed a 13 cm pelvic mass in the sacral plexus and bilateral hydronephrosis. Laboratory evaluation was notable for a mildly elevated creatinine. Ultrasound guided biopsy led to a diagnosis of desmoid tumor. A beta-catenin mutation was not present.

He underwent debulking surgery to relieve ureter compression and subsequently failed therapy with vinblastine/methotrexate and celecoxib.^10,11,13 He required bilateral nephrostomy tubes for ureteral obstruction secondary to increased tumor size, and he was trialed on tamoxifen.¹⁴ He overall tolerated tamoxifen well, though he did develop a deep venous thrombosis necessitating treatment with rivaroxaban. After nearly a year of treatment on tamoxifen, his tumor started to increase in size, which prompted a switch to hydroxyurea.¹⁵ Nearly four months after switching therapy to hydroxyurea, interval imaging revealed continued increase in tumor size. Sorafenib 200 mg/m2 daily in combination with celecoxib 200 mg BID was initiated. Pediatric dosing of sorafenib was based on the results of an early phase clinical trial; patients had tolerated a dose of 200 mg/m2 PO BID,¹⁶ and this dose was cut in half based on a similar reduction in sorafenib dosing for adult patients with desmoid tumor (400 mg daily) as compared to renal cell carcinoma (800 mg daily).⁸

Imaging obtained after 10 months of therapy revealed a decrease in the size of his tumor from 21 cm to 11 cm (Figure 1). In addition, he had a decrease in T2 signal intensity on magnetic resonance imaging (MRI), which has been used as a marker of decreased cellularity (Figure 1).^8,9 He has tolerated the medication well without any adverse effects. Of note, his current sorafenib dose is 600 mg daily based on his body surface area. It was proposed to decrease the dose of sorafenib to 400 mg daily as per adult dosing, but the patient elected to continue on his dose of 600 mg per day as he was not experiencing any side effects.

Figure 1.

MR imaging of desmoid tumors in 2 consecutive pediatric patients treated with sorafenib and celecoxib. A) Coronal T2/HASTE image demonstrates a 21 cm pelvic mass with heterogeneous internal T2 signal abnormality. B) Interval scan after 10 months reveals decrease in size of pelvic mass (11 cm) and nearly homogeneous, low T2 signal compared to the prior examination. C) Coronal T2/STIR image demonstrates symptomatic, T2 hyperintense mass in the right forearm, measuring 2.2 cm. D) Scan after 14 months of sorafenib and 13 months of celecoxib (initiation of celecoxib delayed secondary to insurance issues) shows stable size of right forearm mass and marked decrease in T2 signal abnormality. White dotted lines show tumor margin in each panel.

Case 2

A 7-year-old male presented with an eight-month history of intermittent and progressive right elbow pain, a prominence at the proximal right ulna, and limitations in flexion and pronation of his right elbow. There was no known trauma. MRI revealed a mass arising from the proximal ulna with areas containing increased T2 signal. Excisional biopsy was consistent with desmoid tumor. Over the next several months, he had two additional surgeries due to the development of a total of three masses on MRI and worsening pain. His desmoid tumor was positive for the S45F beta-catenin mutation.

Following his third surgery, he was placed on trials of sulindac, hydroxyurea, and imatinib.^15,17–19 Worsening contractures of his digits prompted debulking of his dominant mass and flexor tenolysis. His dominant mass recurred following surgery, and he was placed on sorafenib 200 mg/m2 PO daily in combination with celecoxib 100 mg PO BID. After 14 months of therapy, imaging showed stabilization in tumor size with a markedly decreased T2 signal in the dominant forearm mass (Figure 1). Clinically, the patient reports substantially decreased pain and improved right upper extremity range of motion and functionality. The mass feels subjectively softer under the skin. He has tolerated the medication regimen well without any side effects.

In summary, the combination of sorafenib and celecoxib was safe and effective in two pediatric patients with refractory desmoid tumor. These two patients are the only patients that we have treated with this combination, but we consider this an attractive therapeutic approach for patients with refractory disease. Factors that might prompt consideration for the combination of sorafenib and celecoxib include pain symptoms that could benefit from celecoxib, parental desire to avoid central line access and intravenous chemotherapeutic agents, progression on sorafenib monotherapy, and the non-overlapping side effect profiles of sorafenib and celecoxib.

The treatment of desmoid tumors is rapidly evolving, and there are a number of agents with reported or proven efficacy. In addition, multiple studies support observation alone without any treatment in relatively unaffected patients.^9,20,21 The fact that many desmoid tumors fail to progress with observation alone calls into question the reported efficacy of agents in retrospective case series; efficacy in small, uncontrolled studies may merely reflect the natural history of the disease. This possibility seems unlikely in our cases considering that the tumors were refractory to previous therapies. However, it is possible that the favorable response in our two patients was due to either sorafenib (or celecoxib) alone rather than the combination. A clinical trial would be required to conclusively demonstrate enhanced efficacy of sorafenib plus celecoxib compared to sorafenib alone. However, the rarity of this disease and its natural history, the desire to discover novel agents with efficacy as monotherapy, and the possible lack of appropriate funding can make the conduct of a clinical trial in patients with desmoid tumors very challenging.

Future studies in patients with refractory desmoid tumor could benefit from combination therapy in addition to serial testing of novel single agents. We consider that such studies should include the mutational profile and location of the tumor, change in patient-reported pain/function metrics, and MRI describing T2 signal intensity in lieu of tumor size. In the meantime, we report that sorafenib 200 mg/m2 daily in combination with celecoxib may be an effective, well-tolerated option in a subset of children with refractory, symptomatic desmoid tumors.