Article Text
PDF
Abstract
Objectives Uterine carcinosarcoma is a rare, aggressive form of uterine cancer with a high recurrence rate and poor survival at all stages. We sought to evaluate the outcomes of patients treated with chemotherapy versus a combination of chemotherapy and radiation (chemoradiation) to determine survival.
Methods A multicenter retrospective analysis of patients with stage I–IV carcinosarcoma was conducted from January 2000 to December 2017. Inclusion criteria were primary surgical management, defined as hysterectomy ± salpingo-oophorectomy, comprehensive surgical staging and/or tumor debulking, followed by adjuvant chemotherapy or chemoradiation. Differences in the frequencies of stage, cytoreduction status, treatment delays and sites of disease recurrence were identified using Pearson’s χ2 test. Progression-free and overall survival rates were calculated using Kaplan-Meier estimates.
Results Final analysis included 148 patients; 40.5% (n=60) chemotherapy and 59.5% (n=88) chemoradiation. The mean age was 67 years (range 39–89). Stage distribution included 24.3% stage I, 12.2% stage II, 37.2% stage III, and 26.3% stage IV. There was no difference in the frequency of stage (p=0.81), cytoreduction status (p=0.61), treatment delays (p=0.57), or location of recurrence (p=0.97) between cohorts. The most frequent location of recurrence was the abdomen (50.0%). The median progression-free survival favored chemoradiation over chemotherapy (15 vs 11 months, respectively), as did the median overall survival (26 vs 20 months, respectively). Chemoradiation was associated with a statistically significant improvement in 2 year progression-free survival (22.5% vs 13.6%; p=0.006) and 2 year overall survival (50.0% vs 35.6%; p=0.018) compared with chemotherapy alone. On subanalysis of patients receiving chemoradiation, ‘sandwich sequencing’ (chemotherapy–radiation–chemotherapy) was associated with superior overall survival compared with alternate therapy sequences (chemotherapy–radiation and radiation–chemotherapy) (34 months vs 14 months and 14 months, respectively) (p=0.038).
Conclusions Chemoradiation was associated with improvement in both progression-free and overall survival for all stages of carcinosarcoma compared with chemotherapy alone.
- carcinosarcoma
- radiation
Statistics from Altmetric.com
HIGHLIGHTS
Uterine carcinosarcoma is a rare, aggressive, high-grade endometrial carcinoma.
There was a significant improvement in survival with multimodal therapy compared to chemotherapy alone.
Chemoradiation delivered in ‘sandwich sequence’ was associated with improved survival versus alternate sequencing.
Introduction
Uterine cancer is the most common gynecologic malignancy, affecting 61 880 patients in 2019.1 The vast majority of cases are epithelial in origin, most commonly adenocarcinoma. Uterine sarcomas are rare and aggressive malignancies that comprise the remaining 5–6% of uterine malignancies.2 Uterine carcinosarcoma is histologically composed of both distinct cellular components: epithelial and mesenchymal elements. Although traditionally grouped into the sarcoma category, recent evidence suggests a monoclonal origin of carcinosarcoma as an epithelial tumor, with a late divergence and metaplasia of the carcinoma into sarcomatous components.3 These studies demonstrate that the carcinomatous component of the tumor is the primary driver of metastatic disease and recurrence, and therefore these tumors are better classified and managed as high-grade endometrial tumors.4–7 Based on these findings, supported by The Cancer Genome Atlas, carcinosarcoma is now classified as a high-grade endometrial carcinoma.3
Carcinosarcoma is a rare, aggressive form of uterine cancer associated with a high recurrence rate and poor overall survival for all stages. Unlike the more common endometrioid adenocarcinoma, the majority of patients with carcinosarcoma present at advanced stage and more than half will recur despite aggressive surgery and adjuvant therapy.4 5 Over 60% of patients will present with extra-uterine disease at the time of initial diagnosis.8 Along the same accord, the estimated 5 year survival for all stages of carcinosarcoma is poor, ranging from only 33–39%.9 10
Due to the overall rarity of carcinosarcoma, attempts to conduct prospective trials and define treatment regimens have been challenging. At the present time, there is no consensus on the optimal management of carcinosarcoma. Chemotherapy and radiation therapy have been assessed and offer varying degrees of success. Historically, ifosfamide-based regimens have dominated the landscape of systemic therapy, based on phase III prospective trials supporting the use of ifosfamide combinations over single-agent ifosfamide and whole abdominal irradiation.9 11 12 However, due to toxicity concerns with ifosfamide combinations, interest in carboplatin–paclitaxel has evolved, extrapolating data from endometrial carcinoma. A phase II trial demonstrated the activity of this combination in carcinosarcoma, with an overall response rate of 54%.13 Recently, this regimen became accepted as the mainstay of systemic therapy based on the results of the prospective, randomized trial conducted by the Gynecologic Oncology Group (GOG) 261, demonstrating the non-inferiority of carboplatin–paclitaxel versus ifosfamide–paclitaxel in all stages of carcinosarcoma.14
Data on the efficacy of radiation therapy are limited by the fact that the majority of studies investigating radiation have grouped carcinosarcoma into the sarcoma family of uterine tumors, with carcinosarcoma representing only a fraction of the patients. A prospective trial of early stage uterine sarcomas, including carcinosarcoma, reported improvement in local recurrence with the use of adjuvant radiation; however, this did not translate into improvements in survival.15 Furthermore, the existing retrospective data are conflicting, with some studies demonstrating an improvement in recurrence-free survival and others failing to observe a benefit with adjuvant radiation.16–19 This paucity in the data has led to great variation in treatment regimens between institutions, with no clear evidence supporting or refuting the use of radiation.
In this study, we evaluated the outcomes of patients with all stages of carcinosarcoma treated with chemotherapy alone versus those treated with a combination of chemotherapy and external beam radiation therapy ± vaginal brachytherapy (chemoradiation), to determine if there is a survival advantage associated with a particular protocol. Additionally, we aimed to determine if there was a survival advantage associated with a particular sequence of adjuvant chemoradiation.
Methods
From January 2000 to December 2017, a multicenter retrospective analysis of patients with stage I–IV carcinosarcoma was conducted. Participating institutions included SUNY Downstate Medical Center-Health Science University, King’s County Hospital Center, New York Presbyterian-Brooklyn Methodist Hospital, and Good Samaritan Hospital Medical Center. Internal review board approval was obtained at all participating sites. Tumor registries were reviewed to identify all patients with carcinosarcoma who received primary surgical treatment, followed by adjuvant therapy. Inclusion criteria were patients with a diagnosis of uterine carcinosarcoma or malignant mixed mullerian tumor who had undergone primary surgical management, consisting of hysterectomy with or without bilateral salpingo-oophorectomy, comprehensive surgical staging and/or tumor debulking, followed by adjuvant chemotherapy or chemoradiation. Comprehensive surgical staging was defined as pelvic ± para-aortic lymph node dissection ± omentectomy. Tumor debulking was defined as removal of extra-uterine gross tumor from the abdominopelvic cavity. Additional subanalysis was performed on patients undergoing chemoradiation based on the sequence of treatment; chemotherapy followed by radiation then further chemotherapy (sandwich), chemotherapy followed by radiation (chemotherapy–radiation sequence) and radiation followed by chemotherapy (radiation–chemotherapy sequence). Exclusion criteria included patients with a diagnosis of uterine cancer other than carcinosarcoma, patients with incomplete surgical staging, patients receiving neoadjuvant chemotherapy and/or radiation, and death due to non-cancer related causes. Patients receiving vaginal brachytherapy without external beam radiotherapy were also excluded from the final analysis.
Clinical and demographic data were obtained from a review of the tumor registry, operative notes, pathology reports, and both inpatient and outpatient charts. Data regarding the date of diagnosis, surgical procedure, cytoreduction status, type of adjuvant therapy, date of recurrence, site of recurrence, chemotherapy regimen, number of chemotherapy cycles received, type of radiation therapy received, treatment delays, adverse events, and date of death were extracted. Optimal cytoreduction was defined as total residual tumor ≤1 cm in diameter, and suboptimal debulking was defined as tumor amounting to >1 cm of disease. A one-way analysis of variance (ANOVA) test was used to compare differences in mean age between treatment arms. Differences in the frequencies of stage, cytoreduction status, treatment delays, and sites of disease recurrence were identified using Pearson’s χ2 test. Treatment delay was defined as interruption in receipt of planned adjuvant therapy by ≥ 7 days. Progression-free survival was defined as the time of surgery to the time of first recurrence. Diagnosis of recurrence was made by imaging studies or clinical examination in case of vaginal recurrence. Overall survival was defined as time of surgery to time of death. Patients who were alive at the date of last follow-up were censored. Progression-free and overall survival rates were calculated using Kaplan-Meier estimates. Statistical significance was defined as p<0.05. Analysis was performed using Statistical Package for the Social Sciences (SPSS) version 25.0 (International Business Machines Corporation (IBM), Armonk, New York).
Results
We identified 181 patients with uterine carcinosarcoma. A total of 33 patients were excluded because of radiation therapy alone, refusal of adjuvant therapy, receipt of neoadjuvant therapy, incomplete surgical staging, or death due to non-cancer related cause. Final analysis included 148 patients. The mean age was 67 years (range 39–89) and the majority of patients in both cohorts were African-American (76%). The performance score was 0 for 67% of patients. Sixty (40.5%) patients received chemotherapy alone and 88 (59.5%) received a combination of chemoradiation. The combination therapy cohort included 30.7% (n=27) sandwich, 38.6% (n=34) chemotherapy–radiation sequence, and 30.7% (n=27) radiation–chemotherapy sequence. Stage distribution included 24.3% stage I, 12.2% stage II, 37.2% stage III, and 26.3% stage IV. A total of 89% of patients had no residual disease at completion of surgery. There was no difference in the age (p=0.24), race distribution (p=0.97), stage distribution (p=0.81), performance status (p=0.92), or cytoreduction status (p=0.61) between adjuvant therapy regimens (Table 1).
Clinical and pathological characteristics
The majority of patients in both arms received platinum-based chemotherapy (126/148; 85.1%). The most common regimen was carboplatin–paclitaxel (85/148; 57.4%). Other regimens included cisplatin–ifosfamide (36/148; 24.3%), paclitaxel–ifosfamide (15/148; 10.1%), carboplatin–paclitaxel–bevacizumab (5/148; 3.4%), ifosfamide single agent (4/148; 2.8%), and pegylated liposomal doxorubicin (3/148; 2.0%). There was no significant difference in treatment regimens between adjuvant therapy arms (p=0.28). The median number of cycles received was 6 (range 3–10). Of the patients receiving chemoradiation, the majority received external beam radiotherapy to the pelvis with or without extended field (68/88; 77.3%). The remainder received a combination of external beam radiotherapy ± extended field plus vaginal brachytherapy (20/88; 22.7%).
Sixty patients (40.5%) experienced a delay in treatment; 43.3% of the chemotherapy alone cohort and 38.6% of the chemoradiation cohort. The most common reason for treatment delay in both arms was neutropenia (20.9%). Other toxicities leading to treatment delay included thrombocytopenia (6.8%), anemia (4.7%), neurotoxicity (3.4%), and nephrotoxicity (2.0%). Although there was no significant difference in the frequency of treatment delays between regimens, we observed a trend towards greater treatment delay in the radiation–chemotherapy sequencing arm versus sandwich and chemotherapy–radiation sequences (41% vs 33% and 32%, respectively) (p=0.57). There were 119 total recurrences during the study period and 132 individual sites of disease recurrence. The most frequent location of disease recurrence was the abdomen (66/132; 50%), followed by the lungs (29/132; 21.9%), pelvis (24/132; 18.2%), vagina (8/132; 6.1%), and extra-peritoneal distant sites (5/132; 3.8%). The distribution of recurrence did not differ between treatment modalities (p=0.97) (Table 2).
Recurrence site based on adjuvant therapy regimen
The median follow-up for patients receiving chemotherapy alone was 5 years, and 3.4 years for patients receiving chemoradiation. There was a significant improvement in both the progression-free survival and overall survival in patients receiving chemoradiation. The median progression-free survival favored the combination of chemoradiation compared with chemotherapy alone: 15 versus 11 months, respectively (p=0.006). The median overall survival also favored combined adjuvant therapy: 26 versus 20 months (p=0.018) (Figure 1) (Table 3). Chemoradiation also demonstrated superiority over chemotherapy alone in terms of 2 year progression-free survival (22.5% vs 13.6%, respectively) and 2 year overall survival (50.0% vs 35.6%, respectively). On sub-analysis of patients receiving chemoradiation, sandwich sequencing was associated with a significant improvement in overall survival compared with those receiving alternate therapy sequences, 36 months versus 24 months, respectively (p=0.038). This translated into a significant 2 year overall survival benefit in the sandwich cohort (70.0%) compared with both the chemotherapy–radiation sequence (48.4%) and the radiation–chemotherapy sequence (42%). There was a non-significant trend towards improved progression-free survival among patients receiving sandwich (18 months), patients receiving chemotherapy–radiation (14 months), and radiation–chemotherapy (14 months) (p=0.16) (Figure 2) (Table 3). Similarly, we observed a 2 year progression-free survival benefit with the use of sandwich (30.4%) compared with alternate sequencing (chemotherapy–radiation 29.0% and radiation–chemotherapy 7.7%). There was no difference in the stage distribution (p=0.97), frequency of treatment delays (p=0.98), or distribution of recurrence (p=0.73) between chemoradiation sequences.
Kaplan-Meier survival analysis by treatment group. (A) Progression-free survival analysis. (B) Overall survival analysis. C, chemotherapy alone; C-RT, combination chemotherapy and radiation.
Median survival by treatment regimen
Kaplan-Meier survival analysis by sequence of multimodality therapy. (A) Progression-free survival analysis. (B) Overall survival analysis. CR, chemotherapy followed by radiation; CRC, chemotherapy followed by radiation, then further chemotherapy; RC, radiation followed by chemotherapy.
Discussion
We observed an improvement in both progression-free and overall survival with the addition of radiation to chemotherapy in all stages of carcinosarcoma. We observed a significant improvement in overall survival in patients receiving chemoradiation in sandwich sequencing compared with alternate sequences. The majority of data surrounding the use of radiation therapy in carcinosarcoma is based on retrospective reviews which group carcinosarcoma with uterine sarcomas. Collectively, these studies demonstrate a decrease in local recurrence with the use of radiation; however, these studies fail to demonstrate a significant survival benefit.15–17
One prospective trial, in which carcinosarcoma represented almost half of the patient population (n=91), observed a significant decrease in local recurrence (18% vs 36%) in early stage carcinosarcoma treated with radiation, compared with observation.19 However, this trial failed to demonstrate a survival benefit with radiation. Conversely, a large Surveillance, Epidemiology, and End Results (SEER) database analysis observed a 21% reduction in cancer-specific mortality in patients with stage I/II carcinosarcoma receiving adjuvant radiation. This mortality benefit was confined to patients who did not undergo lymphadenectomy, suggesting that the benefit of radiation was limited to unstaged patients likely harboring occult lymphatic metastasis.18 Similarly, another SEER database analysis conducted by Nemani and colleagues failed to demonstrate a survival benefit to radiation therapy regardless of lymphadenectomy.20 However, the superiority of chemotherapy ± radiation over radiation alone was demonstrated by Makker and colleagues, in a retrospective analysis of all stages of carcinosarcoma. Radiation alone was inferior in terms of progression-free and overall survival compared with systemic therapy ± radiation.21
Based on these observations, radiation has been utilized in combination with systemic therapy despite the lack of clear survival benefit. We hypothesize the survival benefit conferred by radiation is likely a result of sterilization of the pelvis and targeting of the retroperitoneum, leading to a delay in recurrence and a subsequent longer survival. In accord with this, we observed a trend towards lower pelvic and vaginal recurrences in the chemoradiation group compared with those patients receiving chemotherapy alone.
At present, there is no consensus on the optimal sequence of chemoradiation. Sandwich sequencing offers several hypothetical benefits compared with alternate therapy sequences. First, the combination of sandwich permits treatment of systemic disease up-front, while controlling micrometastatic disease in the pelvis and retroperitoneum in a timely fashion, and again targeting systemic disease after radiation. Importantly, the sandwich regimen limits toxicity, allowing for maximum therapeutic dosing of both chemotherapy and radiation.22–24 Complete irradiation of a tumor bed prior to the administration of chemotherapy leads to vascular alterations and may impair the delivery of chemotherapeutic agents, and therefore diminish the efficacy of subsequent systemic chemotherapy.22 Additionally, radiation–chemotherapy sequencing has been associated with delays in chemotherapy administration in patients with advanced endometrial cancer.23 Disadvantages to the administration of six cycles of chemotherapy prior to radiation include increased toxicity and more radiation delays, subsequently leading to inferior oncologic outcomes. Consistent with these observations, we experienced a trend towards greater treatment delay in the radiation–chemotherapy sequencing arm versus sandwich and chemotherapy–radiation sequences.
A retrospective review of advanced endometrial cancer observed an improvement in progression-free and overall survival when chemoradiation was delivered in sandwich sequence.23 The sandwich sequence was also explored in a small phase II analysis of carcinosarcoma patients. In this review, cisplatin and ifosfamide or ifosfamide alone was administered in sandwich with radiation. Although there was no comparison group, the authors reported a favorable toxicity profile and survival with the use of sandwich sequencing.22 Extrapolating data from these studies, we have treated patients with carcinosarcoma using the sandwich model. We observed a statistically significant improvement in overall survival with sandwich sequence of therapy, suggesting this adjuvant therapy sequencing is worthy of prospective evaluation.
In the current report, we observed a significant 6 month improvement in overall survival, with a median survival of 26 months in patients receiving dual chemoradiation. This median overall survival far surpasses that of original landmark trials in carcinosarcoma, utilizing chemotherapy alone regimens consisting of ifosfamide with or without paclitaxel or cisplatin.11 19 However, our median survival falls short of that reported by the recent prospective GOG 261. We suspect this is due to the fact that 40% of patients in GOG 261 were stage I disease. In the current report, the majority of patients were stage III/IV disease, skewing our progression-free and overall survival curves towards those patients with advanced disease.
The major limitation to our study is its retrospective nature. Additionally, as this study spans a 17 year time period, there has been variation in the trends of specific chemotherapy regimens over time. Despite this, the majority of patients received platinum-based chemotherapy, and the use of specific chemotherapy regimens was balanced between treatment arms. The strengths of this trial include its relatively large sample size based on the overall rarity of this tumor. Additionally, the present study included only completely surgical staged patients and was well balanced in terms of stage distribution. Despite the inherent limitations, this report demonstrates a 6 month improvement in progression-free survival and overall survival with the addition of radiation to chemotherapy in patients with all stages of carcinosarcoma. In a disease with a dismal prognosis, the authors believe this 6 month improvement represents a clinically significant benefit to patients and warrants further prospective investigation. Additionally, in the era of targeted therapy, future investigations should not only focus on the combination of chemoradiation, but also on molecular tumor characterization to identify novel treatment pathways and new therapeutic options.
Combination adjuvant therapy with chemoradiation was associated with improvement in both progression-free and overall survival in patients with all stages of carcinosarcoma compared with those receiving chemotherapy alone. In patients receiving chemoradiation, sandwich sequencing was associated with improved overall survival when compared with alternate therapy sequencing.
References
Footnotes
Contributors JM: conceptualization, methodology, formal analysis, writing – original draft, visualization. TH: data curation. LS: data curation. VT: data curation. MF: data curation. KE: conceptualization, methodology. CG: conceptualization, writing – review and editing. PS: conceptualization, writing – review and editing. Y-CL: conceptualization, writing – review and editing, supervision. MJK: conceptualization, methodology, supervision, project administration.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. De-identified data are available upon request by contacting the corresponding author. Re-use of the data is not permitted without the written consent of the corresponding author and approval of co-authors.