Treatment of Recurrent or Metastatic Uterine Adenosarcoma

Purpose This study retrospectively evaluated overall survival (OS) by treatment of recurrent or metastatic uterine adenosarcoma including surgery, radiation, chemotherapy, and hormonal therapy and evaluated OS and progression-free survival (PFS) after 1st line systemic chemotherapy. Methods 78 patients with recurrent or metastatic adenosarcoma comprised the study population. The Kaplan-Meier method was used to estimate OS and PFS. The log-rank test was performed to test the difference in survival between groups. Results Median OS from diagnosis of recurrent or metastatic disease was 1.8 yrs. OS was influenced by pathology on recurrence, p=0.035. Median OS differed by surgery for 1st recurrence 26.3 months versus 15.1 months. OS was not influenced by chemotherapy, p=0.58, palliative radiation, p=0.58, or hormonal therapy, p=0.15. The response rate (CR + PR) per RECIST 1.1 for chemotherapy was 31.2% for doxorubicin-based regimens and 14.3% for gemcitabine/docetaxel. OS since 1st line chemotherapy was not significantly different among chemotherapy regimens. However, the median PFS was superior for doxorubicin/ifosfamide (15.4 months) compared to gemcitabine/docetaxel (5.0 months), platinum-based regimens (5.7 mo), or other doxorubicin-based regimens (6.5 months). Conclusion These results suggest that surgery is an important treatment modality for recurrent or metastatic uterine adenosarcoma, and the most effective chemotherapeutics are doxorubicin/ifosfamide and gemcitabine/docetaxel.


Introduction
Uterine adenosarcoma is an extremely rare subtype of uterine sarcoma, which represents only 5.5 to 9% of all uterine sarcomas [1,2]. Uterine adenosarcoma was rst described by Dr. Phillip B. Clement and Dr. Robert E. Scully in 1974 [3].
is tumor is composed of a malignant mesenchymal component and a benign epithelial component [4,5]. is biphasic cellular di erentiation is characteristic of adenosarcomas and required for the diagnosis of this tumor. Adenosarcomas can arise from the uterus but have also been noted to arise from the ovaries, vagina, cervix, and pelvis usually in the setting of prior endometriosis [6][7][8][9].
ere is a speci c FIGO uterine adenosarcoma staging system which divides between stage Ia, Ib, and Ic by the presence and extent of myometrial invasion [10]. e majority of adenosarcoma patients (70 to 80%) will present with stage I disease [11][12][13]. Despite this, many patients, even with stage I disease, will develop recurrent or metastatic disease. Survival varies with the presence and extent of myometrial invasion and sarcomatous overgrowth [12][13][14][15].
Unfortunately, given the rarity of adenosarcoma, there are limited data to guide treatment decisions in the recurrent or metastatic setting.
ere is no standard treatment for recurrent or metastatic disease, though surgery is often preformed [12]. A recent review suggests that uterine adenosarcomas can respond to doxorubicin/ifosfamide and gemcitabine/docetaxel chemotherapy [16]. Furthermore, the role of hormonal therapy in recurrent or metastatic disease is limited to case reports and case series. e purpose of this study was to examine the role of surgery, radiation, and hormonal therapy in the recurrent or metastatic setting. Furthermore, this is the rst study to report response rates, overall survival, and progression-free survival of adenosarcoma patients treated with chemotherapy for recurrent or metastatic disease.

Patients and Methods
With approval by the MD Anderson Cancer Center Institutional Review Board, a search was conducted of the Institutional Tumor Registry. Seventy-eight patients treated at the MD Anderson Cancer Center from August 1982 to December 2014 with recurrent or metastatic uterine or extrauterine adenosarcoma were identi ed. e diagnosis of uterine adenosarcoma was con rmed by MD Anderson sarcoma or gynecologic pathologists. Demographic, clinicopathologic, and treatment characteristics were abstracted from the patients' medical records.
en, a deidenti ed database was constructed of all adenosarcoma patients. e stage was assigned using the International Federation of Gynecology and Obstetrics 2009 uterine adenosarcoma staging system [10].
Patients' demographic and clinical characteristics were analyzed, with categorical variables summarized in frequency tables while continuous variables summarized using mean (±S.D.) and median (range). e product limit method of Kaplan and Meier was used to estimate overall survival (OS) and progression-free survival (PFS) [17]. OS was determined from the date of rst recurrent or metastatic disease diagnosis to the date of death or date of last contact. OS since the start of chemotherapy for recurrent or metastatic disease was determined from the date of initiation of 1st line chemotherapy for recurrent or metastatic disease to the date of death or date of last contact. Progression-free survival was determined from the date of initiation of chemotherapy for recurrent or metastatic disease to the date of progression or death. Complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD) were determined by RECIST 1.1 [18]. e log-rank test was performed to test the di erence in survival between groups [19]. Regression analyses of survival data utilized the Cox proportional hazards model [20].

Patient Characteristics.
Seventy-eight patients with recurrent or metastatic uterine or extrauterine adenosarcoma who received treatment at the MD Anderson Cancer Center were identi ed from a retrospective review of the Institutional Tumor Registry and were included in this study. e median follow-up time since recurrent or metastatic disease was 8.2 years. e demographic and clinical characteristics for all patients included in this study are summarized in Table 1. e median age at diagnosis was 55 years (range 27 to 79 years). e majority of patients were Caucasian (80.8%). e most common presenting symptom was abnormal uterine bleeding (44.9%), and the second most common presenting symptom was pelvic pain (16.7%).
irty-six patients had stage II-IV primary tumor (46.2%). e primary tumor location was the uterine corpus in the majority of patients (74.4%). At last follow-up, 24 patients were alive, and 54 have died. Local recurrences occurred within the abdomen and pelvis in seventy-two patients, and sixteen patients developed distant metastasis. Patients who had local and distant recurrences commonly developed the local recurrence rst. Sites of distant disease included the lung (14 pts), bone (4 pts), liver (3 pts), brain (1 pt), and subcutaneous tissue (4 pts).

Treatment Characteristics for Recurrent or Metastatic
Disease. Treatment on recurrence varied greatly and was in uenced by the location and extent of recurrence. Table 2 describes the treatments that patients received on 1st and 2nd recurrence including surgery, palliative radiation, chemotherapy, and hormonal therapy. First-line chemotherapy for recurrent or metastatic disease was given to 59 patients, at 1st recurrence in 40 pts, at 2nd recurrence in 10 pts, at 3rd recurrence in 2 pts, and at 4th recurrence in 2 pts. ree additional patients had residual disease after primary treatment, and upon progression, they received 1st line chemotherapy (Table 2). One patient had 7 prior surgeries and hormonal therapy prior to receiving chemotherapy, and another patient received chemotherapy for 1st recurrence, though documentation regarding the speci c chemotherapy regimen was lacking. e speci c chemotherapy regimens utilized greatly di ered as well. Table 2 lists the speci c chemotherapy regimens that patients received for 1st, 2nd, and 3rd line chemotherapy, as well as the number of patients that received surgery, radiation therapy, or hormonal therapy either before or after their chemotherapy as part of their treatment for recurrent or metastatic disease.

Pathology on Recurrence.
On recurrence, 58 patients had a biopsy available for review: high-grade sarcoma in 36 patients and mixed epithelial and mesenchymal components in 22 patients. Median overall survival di ered by the pathology of recurrence. Patients with high-grade sarcoma on recurrence had a median overall survival of 17.6 months versus 33.5 months in patients with mixed histology, p � 0.035, HR � 0.47 (95% CI 0.23-0.96) (Figure 1(a) and Table 3).

Surgery for Recurrent or Metastatic Disease.
Forty-ve patients underwent surgery for recurrence or metastatic disease.
Forty-one patients underwent resection of recurrence within the abdomen and pelvis. Twenty-ve patients underwent one resection, eight patients had two resections, four patients had three resections, three patients had four resections, and one patient had nine resections. Four patients underwent thoracotomies for resection of metastatic disease to the lung. Median overall survival was improved in those patients who underwent resection for recurrent or 2 Sarcoma metastatic disease, for 1st recurrence 26.3 months versus 15.1 months, p � 0.54 (Figure 1(b) and Table 3).

Chemotherapy for Recurrent or Metastatic Disease
3.5.1. Response Rates. Of the fty-nine patients that received 1st line chemotherapy for recurrent or metastatic disease, 9 received chemotherapy after surgical resection to obtain no evidence of disease, so response rate per RECIST 1.1 could not  (Table 3). e majority of the responses with doxorubicin-based regimens were seen in patients receiving the combination of doxorubicin and ifosfamide (4/5 pts) or doxorubicin and dacarbazine (1/5 pts). e rate of CR + PR + stable disease (SD) was 87.5% for doxorubicin-based regimens and only 42.9% for gemcitabine/docetaxel. Only ve patients with a platinum-based regimen were evaluable for RECIST 1.1, 80% had SD, and the one patient with a partial response, also, received ifosfamide.
Patients were analyzed for other factors which may have in uenced the e ectiveness of their chemotherapeutic regimens, such as chemotherapy dosing, # of chemotherapy cycles, presence of sarcomatous overgrowth (SO), point in their disease course when they received chemotherapy, and time to recurrence prior to initiation of chemotherapy, a potential indicator of the aggressiveness of their disease. Patients did not signi cantly di er in terms of chemotherapy dosing, # of chemotherapy cycles, presence of SO, or point in their disease course when they received chemotherapy. e median time to recurrence prior to initiation of 1st line chemotherapy was 10.1 months for all doxorubicin-based regimens, 12.9 months for AI, 9.6 months for other doxorubicin-based regimens, 8.1 months for platinum-based regimens, and 21.7 months for gemcitabine/docetaxel, indicating a possible physician bias for treating patients with doxorubicin-based regimens in patients with quicker relapses, so more aggressive disease.

Hormonal
erapy for Recurrent or Metastatic Disease. Twenty-eight patients received hormonal therapy at some point during their treatment course. Seven patients received more than 1 line of hormonal therapy. Initial hormonal therapies included GnRH agonists (leuprolide, 9 pts), progesterones (megestrol acetate, 7 pts), SERMs (tamoxifen 3 pts and raloxifene 1 pt), and aromatase inhibitors (anastrozole 3 pts and letrozole 1 pt). e medial overall survival for patients with recurrent disease who received hormonal therapy was 34.7 months compared to 17.6 months, a trend toward improved outcomes that was not statistically signi cant, p � 0.15.
Out of these twenty-eight patients, there were four that derived several years of bene t from hormonal therapy. Two patients treated with leuprolide had disease control for >2 years. eir tumors were not stained for the estrogen receptor (ER) or progesterone receptor (PR). A third patient was placed on leuprolide after surgical resection with the development of progression after two years. is patient was then placed on megestrol acetate without response and then on anastrozole with a complete response lasting for eight years. is patient developed a 2nd malignancy. She was taken o anastrozole at the time of surgery for her cholangiocarcinoma. Shortly after completing adjuvant chemotherapy for her cholangiocarcinoma, she developed recurrence of her adenosarcoma, biopsy proven. She was placed back on anastrozole with response lasting for another ve years. More recently, she developed progression and is now on systemic chemotherapy with trabectedin. is patient's tumor stained for ER 80% and PR 40%. e mesenchymal portion of her initial tumor was described as endometrial stromal sarcoma, though she did have sarcomatous overgrowth noted as well. A fourth elderly patient with locally advanced unresectable disease was treated with leuprolide and carboplatin for seven cycles leading to a partial response. Her leuprolide was continued after chemotherapy resulting in a complete response after ve years with resolution of her pelvic mass and remaining subcentimeter pelvic lymphadenopathy. She continued leuprolide for 14 years, at which point her leuprolide was stopped. She then developed radiographic recurrence within the abdomen and was restarted on leuprolide with stable

Discussion
In this study, we report the largest single-institution experience with recurrent or metastatic uterine and extrauterine adenosarcoma. ere is no standard treatment for patients with local recurrence or metastatic uterine adenosarcoma. Treatment options include surgery, radiation, chemotherapy, and hormonal therapy. is study is the rst to show that the pathology on recurrence in uences outcomes.
Speci cally, those patients that present with a recurrence that is purely high-grade sarcoma have signi cantly worse outcomes than those patients who present with a recurrent tumor with mixed epithelial and mesenchymal components. is may represent selective clonal evolution of these tumors such that patients with a recurrence of pure high-grade sarcoma have a more clinically aggressive course. e majority of uterine adenosarcomas recur locally, suggesting that resection of a local recurrence may improve overall survival and time to next recurrence. Previous reports have indicated a bene t for secondary cytoreduction of recurrent adenosarcoma in terms of overall survival and 6 Sarcoma time to next recurrence [12,21]. Our study shows an improvement in overall survival for those patients that underwent surgery for 1st recurrence, though this was not statistically signi cant. ere is a selection bias in this result, in that patients able to undergo surgery likely had better functional status at the time of surgery, less medical comorbidities, and recurrence amenable to surgical resection. However, given the improvement in overall survival, it may be worth considering surgical resection of a recurrence, for those patients amenable to surgery. is is the rst retrospective report to examine the use of systemic chemotherapy in a large population of recurrent uterine adenosarcoma. Case reports and case series have described responses in adenosarcoma with the use of doxorubicin-based regimens [21][22][23][24][25][26], gemcitabine/docetaxel [21,27], trabectedin [28], or platinum-based regimens [21].
is report shows that active agents in uterine adenosarcoma are doxorubicin/ifosfamide (AI), doxorubicin/dacarbazine (ADIC), and gemcitabine/docetaxel. e most e ective agents for adenosarcoma in terms of response per RECIST 1.1 were the combination of doxorubicin and ifosfamide. All three are reasonable chemotherapeutic choices for recurrent or metastatic uterine adenosarcoma. e dosing of chemotherapy in this retrospective study varied greatly; however, the patients that received gemcitabine/docetaxel received 675 to 900 mg/m 2 of gemcitabine and 75 to 100 mg/m 2 of docetaxel; the patients that received doxorubicin-based regimens received 60 to 75 mg/m 2 of doxorubicin, 7.5 to 10 gm/m 2 of ifosfamide, and 750 to 1000 mg/m 2 of dacarbazine. is is standard sarcoma chemotherapy dosing. e OS after 1st line chemotherapy for recurrent or metastatic disease was not statistically di erent between AI, gem/doc, or platinum-based regimens. However, OS may be a ected by subsequent therapies such as surgery, hormonal therapy, or further chemotherapy. Additionally, there was a trend toward worse survival for patients that received other doxorubicin-based regimens excluding AI, suggesting that ifosfamide is required to obtain the higher response rate and clinical bene t. However, whether this e ective is limited to ifosfamide alone or a result of synergy  between doxorubicin/ifosfamide, as noted in soft tissue sarcomas, is undetermined [29]. PFS may be a better measure of bene t from systemic chemotherapy. Patients who received AI had the longest PFS (15.4 months), with a trend toward statistical signicance when compared to patients who received doxorubicin alone or in combination with a 2nd agent or platinumbased regimens. If the goal of therapy is to produce reduction in tumor size prior to surgical resection of recurrent disease, AI chemotherapy may have advantages over gemcitabine/ docetaxel. If patients are older with poor function status and multiple medical comorbidities, they are not a doxorubicin/ ifosfamide candidate; then gemcitabine/docetaxel may be the preferred regimen. Cisplatin-and carboplatin-based regimens appear to be the least e ective in uterine adenosarcomas and should not be recommended for treatment of recurrent or metastatic disease. Further study is required to evaluate the role of trabectedin in the treatment of uterine adenosarcomas.
It should be noted that there was likely a physician bias in choosing treatment with doxorubicin-based regimens over gemcitabine/docetaxel. ere was a shorter median time to recurrence prior to the start of 1st line chemotherapy with AI (12.9 months) or other doxorubicin-based regimens (9.6 months) compared to gemcitabine/docetaxel (21.7 months), indicating that physicians were more likely to treat with doxorubicin-based regimens than with gemcitabine/docetaxel for patients that had quicker relapses, or more aggressive disease. is suggests a bene t of doxorubicin/ifosfamide over gemcitabine/docetaxel in the treatment of uterine adenosarcomas, despite similar median OS, as patients with more aggressive disease would be expected to have worse survival.
Evidence for the use of hormonal therapy in uterine adenosarcoma is even more limited than evidence for the use of chemotherapy. Case reports or case series have noted responses to hormonal therapy lasting 10 months to 7 years [12,[30][31][32][33]. Agents used include GnRH agonists, progesterones, selective estrogen receptor modulators, or aromatase inhibitors. Responses have been occasionally correlated with the presence of estrogen receptor (ER) and progesterone receptor (PR) staining. Loss of ER and PR expressions has been associated with sarcomatous overgrowth [34]. Furthermore, loss of response to hormonal therapy was associated with reduced ER/PR expression in one case report [33]. is suggests ER and PR as possible predictors of response to hormonal therapy in uterine adenosarcomas, though this has not been studied in a systematic manner. In this study, the majority of patients (86%) did not receive bene t from hormonal therapy, though there were 4 patients that derived bene t in terms of stable disease and improved survival from leuprolide or anastrozole for 2 to 15 years, suggesting that select patients may have a bene t from hormonally targeted therapy. Unfortunately, we are currently unable to determine which patients will receive such a large bene t from hormonal therapy.
is study is limited by its retrospective nature and small sample size, though this is the largest single-institution recurrent or metastatic uterine or extrauterine adenosarcoma series to date. Overall, uterine adenosarcoma is a rare disease with limited evidence-based data to determine treatment recommendations. Treatment of recurrence or metastatic disease can consist of surgery, chemotherapy, preferable with doxorubicin/ifosfamide or gemcitabine/docetaxel, or hormonal therapy in select patients.

Conflicts of Interest
ere are no con icts of interest to report.