Endometrial cancer (EC) remains the leading female genital tract malignancy in industrialized countries.1 In Taiwan, the incidence rose tremendously in the past 30 years from less than 100 cases per year to over 1000 per year. More importantly, according to the data from the Department of Health in 2010, it became the first ranking of gynecologic cancer in Taiwan.2 The incidence of EC increased 21% from 2008 to 2012 in the United States.3
Endometrial cancer is characterized by early diagnosis and long-term survival for most patients,3,4 although the mortality-to-incidence ratio is higher in developing countries than the developed world (5-year overall survival of 67% vs 82%).4 Generally, the prognosis of recurrent EC is poor except for isolated vaginal relapse.3–8 However, survival after isolated vaginal recurrence can be limited in those who have received postoperative radiotherapy (RT).5,7 The 5-year survival of extravaginal recurrences confined to the pelvis ranged from 8% to 29%.4,5,8 Recurrences associated with distant metastasis remain a challenge. Surgical resection/debulking of recurrent EC has been advocated in selected clinical scenario.9–11 Other factors influencing outcomes after treatment failure, such as time to recurrence, performance status, grade, serous/clear cell histology, salvage treatment including surgery, and residual tumor after cytoreductive salvage surgery, were controversial.5,8–13
Our aim was to investigate the failure pattern, outcomes, and prognostic factors after the failure of primary treatment in EC.
In our previous study, we retrospectively reviewed the clinical records of patients with histologically proven EC treated at the Chang Gung Memorial Hospital (CGMH), Taoyuan, Taiwan, between 2000 and 2010.14 The study was approved by the medical ethics committee at the CGMH (#101-2187B). Patients with fertility-sparing treatment, a diagnosis of carcinosarcomas, with synchronous endometrial and ovarian cancers, and primary treatment performed outside the CGMH were excluded. The clinical and pathological data were obtained from our previous study with additional details extracted from medical records and cancer registry. Treatment failure was defined as relapse after complete removal of all cancerous lesions or persistent/progressive disease despite treatment. Patient status follow-up was updated to August 31, 2013.
Primary Treatment and Adjuvant Therapy After Primary Surgery
Surgery was the main method of primary treatment that consisted of hysterectomy, bilateral salpingo-oophorectomy, pelvic lymph node (LN) dissection, and para-aortic LN dissection. Washing cytology was performed at first while entering the abdomen. The type of hysterectomy included simple, extended, or radical hysterectomy; it depended on the extent of cervical involvement. Patients with a high risk of relapse received adjuvant therapy, whereas those with intermediate risk were at the physician’s discretion. Patients with occult cervical stromal invasion treated with simple hysterectomy were given postoperative adjuvant RT, including external beam RT with or without intravaginal brachytherapy as vaginal stump boost. Moreover, adjuvant chemotherapy was recommended for positive cytology, adnexal metastasis, or tumors of nonendometrioid histology (clear cell, papillary serous, or undifferentiated carcinoma). Adjuvant chemotherapy consisted of cisplatin of 60 mg/m2 or carboplatin AUC5-6 in combination with paclitaxel of 175 mg/m2 or Adriamycin/epirubicin 50 to 60 mg/m2 3-weekly for a total of 4 to 6 courses. Patients with LN metastases received adjuvant chemotherapy, concurrent chemoradiation (CCRT) (cisplatin of 60 mg/m2 or carboplatin AUC5-6 in combination with paclitaxel of 175 mg/m2 or Adriamycin/epirubicin 50–60 mg/m2 3-weekly or cisplatin 40 mg/m2 weekly for 4–6 cycles, chemotherapy starting simultaneously with RT and continued after RT completed up to 6 cycles if tolerable), or sequential/sandwich chemotherapy and RT. Radiotherapy for patients with para-aortic node metastases included extended field up to L1 or T12.10
The patterns of recurrence were classified into vagina only, pelvis, distant alone, or pelvic plus distant site(s); distant lymphatic recurrences include the inguinal, para-aortic, or supraclavicular nodes, whereas other distant recurrences include the peritoneum, lung, bone, liver, or other organs. All the patients with recurrence were confirmed by histologic and/or clinical findings. Vaginal or pelvic recurrence was detected by pelvic examination and imaging (computed tomography [CT] or magnetic resonance imaging). Distant metastasis was diagnosed from physical examination, confirmed by imaging methods and image-guided biopsy/aspiration cytology. Positron emission tomography (PET) (before 2006) or integrated PET/CT was performed for treatment plan.
Survival of those who progressed or had persistent disease during primary treatment was calculated from the date of primary diagnosis. Time to failure (TTF) was defined as the period between diagnosis and occurrence of relapse. Survival after recurrence (SAR) was measured from the time of documented recurrence to the time of last follow-up or death. Analysis was performed using SPSS version 17.0. The SAR curves were plotted using the Kaplan-Meier method and compared with log-rank test. Prognostic variables included age at diagnosis and relapse, 2009 International Federation of Gynecology and Obstetrics (FIGO) stage at diagnosis, grade of differentiation, histology subtype, peritoneal cytology, cervical extension, adnexal metastasis, pelvic node metastasis, para-aortic node metastasis, tumor estrogen receptor (ER) and progesterone receptor (PR), serum levels of CA125 at diagnosis or relapse, postoperative adjuvant therapy, and site(s)/pattern of recurrence. Based on the information obtained from univariate analysis, a multivariate Cox proportional hazards regression model was used to identify the significant predictors of SAR. A 2-sided P value less than 0.05 was used to indicate statistical significance in all analyses.
A total of 923 patients with the diagnosis of EC were treated between 2000 and 2010 in CGMH. Primary treatment failure was seen in 109 patients. The median age at treatment failure was 59.0 years (range, 36.0–92.7 years). Among the 109 patients, 18 patients with progressive or persistent diseases had a median survival of 6 months from primary diagnosis (range, 1–23 months). Ninety-one patients with recurrent disease were further analyzed. The initial clinicopathological characteristics of the 91 patients are shown in Table 1. The initial FIGO stages were 27 (29.7%) of stage I, 11 (12.1%) of stage II, 34 (37.3%) of stage III, and 19 (20.9%) of stage IV. The median time to recurrence was 13.3 months (range, 3.2–97.2 months), and 82.4% recurred within 3 years. The median TTF at vagina (9.0 months; range, 4.2–51.4 months) was shorter than those in the pelvis or distant sites (13.9 months; range, 3.2–97.2 months).
The recurrent patterns were 8 (8.8%) in the vagina, 3 (3.3%) in the pelvis, 58 (63.7%) in distant, and 22 (24.2%) in both pelvic and distant site(s). The distant sites involved included the lung (29 of 80, 36.3%), para-aortic LN (22 of 80, 27.5%), peritoneum (20 of 80, 25.0%), bone (12 of 80, 15.0%), supraclavicular LN (9 of 80, 11.3%), and liver (9 of 80, 11.3%). Eighty-one of the 91 patients with recurrent EC received salvage treatment (Table 2). Isolated vaginal recurrences (n = 8) were treated with RT alone if PET or PET/CT confirmed. Pelvic lesions with/without distant metastasis were evaluated by PET or PET/CT to determine if surgery was necessary and treated accordingly in combination with chemotherapy, RT, or in combination. Cases not amenable for curative intent treatment were treated with chemotherapy with/without hormone therapy (n = 17), hormone therapy (n = 1), or supportive care only (n = 10).
The median follow-up after recurrence of survivors was 60.5 months (range, 10.6–121.7 months). The median SAR was 20.3 months (range, 1.9–121.7 months) with a 5-year SAR rate of 32.4%. By univariate analysis, the 5-year SAR was better in patients with initial FIGO stage I disease, type I disease, grade I to II, without cervical invasion, without adnexal involvement, negative peritoneal cytology, and negative LN involvement. Age at diagnosis, depth of myometrial invasion, and lymphovascular invasion were not related to prognosis. Not surprisingly, those who received adjuvant chemotherapy or CCRT yet recurred had worse SAR than those treated with surgery alone or adjuvant RT (Table 1). Age at recurrence was unrelated to SAR (P = 0.818).
Table 2 summarizes TTF, pattern of failure, and salvage therapy in relation to the outcome by univariate analysis. The TTF was unrelated to SAR in the 91 patients (P = 0.074) (Table 2). Including cases with extravaginal recurrences (n = 82), the median TTF was significantly longer in survivors than that in the nonsurvivors (19.7 vs 12.2 months, P = 0.025) (Supplemental Digital Content, http://links.lww.com/IGC/A217). Patients with isolated vaginal recurrence had a better prognosis than extravaginal recurrences (5-year SAR of 83.3% vs 27.7%, P = 0.006), and patients with solitary recurrent site had a better prognosis than multiple recurrent sites (5-year SAR of 45.2% vs 20.4%, P < 0.001). The 5-year SAR rates in patients with solitary vaginal, nodal/liver, or pulmonary/bony recurrence were 83.3%, 50.5%, and 24.2%, respectively (P = 0.062). No 5-year survivors were seen in those receiving chemotherapy or hormone therapy alone, or no salvage treatment, whereas patients treated with salvage surgery plus CCRT had a 5-year SAR of 100% (P < 0.001). Failure at previous RT sites was associated with poorer SAR (0% vs 36.7%, P = 0.005) in comparison with those who had not received adjuvant RT or did not fail in the previous RT sites.
Tumor hormone receptors (ER and PR) and serum CA125 levels at initial treatment and relapse in relating to prognosis are summarized in Table 3. Both positive in ER and PR of the primary tumor had prolonged 5-year SAR as compared with those negative in either ER or PR (P = 0.007). Among the 29 patients with initial ER positive, 17 who had sustained ER positive in recurrent tumor had significantly prolonged 5-year SAR as compared with those whose recurrent tumor turned ER negative (P = 0.019). Interestingly, of the 22 patients with initial ER negative, 5 had an ER-positive recurrent tumor, of which the 5-year SAR was 75%. In contrast, of the 18 patients with initial PR-negative tumor, no one had a PR-positive recurrent tumor. The patients with positive hormone receptors in both initial tumors and recurrent tumors had significantly better SAR rates after recurrence. Eighteen of the 55 patients (32.7%) with positive ER/PR received hormonal therapy (1 used hormone therapy alone, 3 as adjuvant with either chemotherapy (n = 2) or surgery plus chemotherapy (n = 1), 1 as maintenance after remission, and 13 after salvage treatment failed). There was no significant difference of 5-year SAR between hormonal therapy users and those without in patients with positive ER or PR. Patients with serum CA125 level greater than 35 U/mL at recurrence had poorer SAR (P = 0.007), whereas CA125 level at initial diagnosis was unrelated to SAR.
By multivariate analysis, initial stage II to IV (hazards ratio [HR], 3.41; 1.53–7.60; P = 0.003), type II histology (HR, 2.50; 1.28–4.90; P = 0.008), positive peritoneal cytology (HR, 2.23; 1.07–4.68; P = 0.033), and recurrence at multiple sites (HR, 2.51; 1.30–4.84; P = 0.006) were significantly associated with poor SAR. Failure at previous RT sites, grade, and ER/PR were not independent variables, and CA125 level at recurrence was marginally significant (P = 0.066) (Table 4). Under stratified analysis, CA125 level greater than 35 U/mL was associated with decreased SAR in cases with solitary metastasis (P = 0.002). In alternative model, cervical invasion can be selected as independent poor prognostic factor with cervical invasion replacing stage (Supplemental Digital Content, http://links.lww.com/IGC/A217). Cervical invasion with primary simple hysterectomy had more pelvic/pelvic-distant recurrences (44.8%) than those with primary radical hysterectomy (15.4%), despite adjuvant RT was used (simple hysterectomy, 62%; radical hysterectomy, 38.4%). It is noteworthy that the 5-year SAR was better in those with initial cervical invasion who received radical hysterectomy (n = 13, 38.5%) at primary treatment than those received simple hysterectomy (n = 29, 0%; P = 0.034) (Supplemental Digital Content, http://links.lww.com/IGC/A217).
On the other hand, recurrence at multiple sites does not equal to unresectable diseases; 6 cases with recurrence at multiple sites obtained long-term disease free after salvage surgery and postoperative adjuvant therapy (Table 5, patients 1–6; Fig. 1). Six patients (patients 5–10) with 2 to 3 recurrences have long-term survival after multimodality salvage therapy, and 3 remained diseases free at their last follow-up (Table 5). These 10 patients with multiple sites or multiple relapses underwent multimodality therapy when PET scans showed feasibility of curative intent salvage therapy and attained SAR greater than 5 years. These 10 cases belonged to nonserous histology.
Prognosis of persistent/progressive EC despite treatment is extremely poor, all 18 such patients in this study died shortly after primary diagnosis (1–23 months). Novel agents have to be added in this group of patients. Similar to the literature,3–8 isolated vaginal recurrence has excellent prognosis (5-year SAR of 83.3%); therefore adjuvant vaginal cuff brachytherapy or external beam RT to pelvis to intermediated risk group may be spared for type I disease. Sartori et al8 identified time to recurrence (<24 months) and adjuvant RT at primary treatment as significant poor prognostic factors. Otsuka et al12 also found that the longer time to recurrence was an independent predictor of prolonged SAR. In our study, TTF was unrelated to SAR in the whole series, because the median TTF for isolated vaginal recurrence (who had excellent SAR) was 9.0 months (Supplemental Digital Content, http://links.lww.com/IGC/A217). In cases with extravaginal recurrences, longer TTF was a significant predictor of longer SAR.
Mountzios et al15 found relapsed within the RT field to be an independent poor prognostic factor. Nagel et al13 reported a trend toward shorter SAR for papillary serous/clear cell than endometrioid carcinoma (P = 0.08). In this study, type II histology was a significant poor prognostic factor by multivariate analysis. Type II histology was associated with more multiple metastatic sites at recurrence (73.3% in type II vs 44.3% in type I; P = 0.013). Failure at previous RT sites (P = 0.005) was significant by univariate analysis but was not independent after controlling confounding variables in our study. Among 10 patients with in-field failure, 9 patients (90%) had multiple organ metastases at recurrence. For those with in-field failure or type II histology, a PET/CT scan to identify multiple metastases is crucial because phase I/II systemic trials would be appropriate.
Positive ER/PR status was associated with better SAR in univariate analyses but not independent by multivariate analysis. Eighteen of the 55 patients (32.7%) with positive ER/PR received hormonal therapy (1 used hormone therapy alone, 13 after salvage treatment failed, 3 as adjuvant with either chemotherapy (n = 2) or surgery plus chemotherapy (n = 1), and 1 as maintenance after remission). The ER/PR positivity was associated with favorable patterns of failure (more vaginal only or pelvic, and solitary) (Supplemental Digital Content, http://links.lww.com/IGC/A217). Of the 22 patients with initial ER negative, 5 had an ER-positive recurrent tumor, of which the 5-year SAR was 75%. Other studies have also found that the status of hormone receptor may change between the primary and recurrent tumors.16,17 CA125 level greater than 35 U/mL at recurrence was related to poor outcome in univariate analysis but not in the multivariate analysis.
Initial stage II to IV and positive cytology were independent predictors of poor outcome. It is understandable that those with initial stage III to IV or positive cytology were treated with surgery and adjuvant chemotherapy and/or RT. They usually relapsed in distant and disseminated areas (96.2%) or previously irradiated field. In our study, all patients with malignant cytology relapsed in distant/pelvic-distant sites without local-regional recurrence. Those with positive cytology and received adjuvant chemotherapy yet recurred had poor 5-year SAR (0%) and should also be candidate for new drug trials (Supplemental Digital Content, http://links.lww.com/IGC/A217). Response to second-line chemotherapy is limited.18 Cervical invasion was associated with poor SAR in univariate analyses but not independent by multivariate analysis. The fact that the 5-year SAR (0%) was extremely poor for those with cervical invasion but received simple hysterectomy at primary treatment (n = 29) prompts a need for better assessment of cervical extension before primary surgery (Supplemental Digital Content, http://links.lww.com/IGC/A217). Cohn et al19 reported a better disease-free survival of patients treated with radical hysterectomy than extrafascial hysterectomy for surgical stage II EC. Orezzoli et al20 also found radical hysterectomy to be an independent predictor of better prognosis.
Patients with multiple recurrent sites or multiple relapses will often be treated palliatively with systemic chemotherapy or hormone therapy. However, long-term survival can be achieved in some nonserous cases (Table 5). Residual disease after salvage cytoreduction was found to be a significant variable for progression-free and disease-specific survival. Scarabelli et al9 reported 32% survival at 80 months in those without residual tumors. Bristow et al11 reported a 54% survival at 39.8 months in those achieved complete salvage cytoreduction. The role of PET in EC is controversial in primary staging.21 Chao et al22 reported that the clinical impact was positive in 73.1% for posttherapy surveillance and in 57.1% after salvage therapy. A recently published meta-analysis estimated a treatment plan change rate of 22% to 35% by PET or PET/CT.23 In this series, most of the treatment plans were based on the PET or PET/CT findings, which achieved a 5-year SAR of 56.6% in patients with extravaginal recurrences treated with surgery ± RT ± chemotherapy and a 5-year SAR of 83.3% in patients with isolated vaginal recurrence and received RT alone.
In conclusion, initial stage II to IV, histologic type 2, positive cytology, and recurrence at multiple sites were significant poor prognostic factors. The prognosis of persistent/progressive EC patients, who should be candidates for novel therapy, despite treatment is extremely poor. Salvage therapy with curative intent remains a viable option for nonserous cases with resectable or irradiatable multiple recurrences and solitary distant nodal/liver metastasis even after rerecurrence. Chemotherapy alone will not offer long-term SAR. Our results are important for selection and stratification of patients for clinical trials of systemic therapy alone.
1. GLOBOCAN 2008, Cancer Incidence and Mortality Worldwide: International Agency for Research on Cancer. Available at: http://globocan.iarc.fr. Accessed August 20, 2013.
2. Cancer Registry Annual Report, 2010 Taiwan: Ministry of Health and Welfare, Executive Yuan 2012. Available at: http://http://www.hpa.gov.tw
/BHPNet/English/Index.aspx. Accessed August 20, 2013.
3. Sorosky JI. Endometrial cancer. Obstet Gynecol. 2012; 120: 383–397.
4. Tangjitgamol S, Anderson BO, See HT, et al. Management of endometrial cancer in Asia: consensus statement from the Asian Oncology Summit 2009. Lancet Oncol. 2009; 10: 1119–1127.
5. van Wijk FH, van der Burg ME, Burger CW, et al. Management of recurrent endometrioid endometrial carcinoma: an overview. Int J Gynecol Cancer. 2009; 19: 314–320.
6. Huh WK, Straughn JM Jr, Mariani A, et al. Salvage of isolated vaginal recurrences in women with surgical stage I endometrial cancer: a multiinstitutional experience. Int J Gynecol Cancer. 2007; 17: 886–889.
7. Jhingran A, Burke TW, Eifel PJ. Definitive radiotherapy for patients with isolated vaginal recurrence of endometrial carcinoma after hysterectomy. Int J Radiat Oncol Biol Phys. 2003; 56: 1366–1372.
8. Sartori E, Laface B, Gadducci A, et al. Factors influencing survival in endometrial cancer relapsing patients: a Cooperation Task Force (CTF) study. Int J Gynecol Cancer. 2003; 13: 458–465.
9. Scarabelli C, Campagnutta E, Giorda G, et al. Maximal cytoreductive surgery as a reasonable therapeutic alternative for recurrent endometrial carcinoma. Gynecol Oncol. 1998; 70: 90–93.
10. Awtrey CS, Cadungog MG, Leitao MM, et al. Surgical resection of recurrent endometrial carcinoma. Gynecol Oncol. 2006; 102: 480–488.
11. Bristow RE, Santillan A, Zahurak ML, et al. Salvage cytoreductive surgery for recurrent endometrial cancer. Gynecol Oncol. 2006; 103: 281–287.
12. Otsuka I, Uno M, Wakabayashi A, et al. Predictive factors for prolonged survival in recurrent endometrial carcinoma: implications for follow-up protocol. Gynecol Oncol. 2010; 119: 506–510.
13. Nagel C, Moron K, Elwell S, et al. Survival outcomes after recurrence in endometrioid compared to papillary serous/clear cell carcinoma of the uterus. Gynecol Oncol. 2012; 127: S28.
14. Chao A, Tang YH, Lai CH, et al. Potential of an age-stratified CA125 cut-off value to improve the prognostic classification of patients with endometrial cancer. Gynecol Oncol. 2013; 129: 500–504.
15. Mountzios G, Bamias A, Voulgaris Z, et al. Prognostic factors in patients treated with taxane-based chemotherapy for recurrent or metastatic endometrial cancer: proposal for a new prognostic model. Gynecol Oncol. 2008; 108: 130–135.
16. Vandenput I, Trovik J, Leunen K, et al. Evolution in endometrial cancer: evidence from an immunohistochemical study. Int J Gynecol Cancer. 2011; 21: 316–322.
17. Soslow RA, Wethington SL, Cesari M, et al. Clinicopathologic analysis of matched primary and recurrent endometrial carcinoma. Am J Surg Pathol. 2012; 36: 1771–1781.
18. Moxley KM, McMeekin DS. Endometrial carcinoma: a review of chemotherapy, drug resistance, and the search for new agents. Oncologist. 2010; 15: 1026–1033.
19. Cohn DE, Woeste EM, Cacchio S, et al. Clinical and pathologic correlates in surgical stage II endometrial carcinoma. Obstet Gynecol. 2007; 109: 1062–1067.
20. Orezzoli JP, Sioletic S, Olawaiye A, et al. Stage II endometrioid adenocarcinoma of the endometrium: clinical implications of cervical stromal invasion. Gynecol Oncol. 2009; 113: 316–323.
21. Lai CH, Yen TC, Chang TC. Positron emission tomography imaging for gynecologic malignancy. Curr Opin Obstet Gynecol. 2007; 19: 37–41.
22. Chao A, Chang TC, Ng KK, et al. 18F-FDG PET in the management of endometrial cancer. Eur J Nucl Med Mol Imaging. 2006; 33: 36–44.
23. Kadkhodayan S, Shahriari S, Treglia G, et al. Accuracy of 18-F-FDG PET imaging in the follow up of endometrial cancer patients: systematic review and meta-analysis of the literature. Gynecol Oncol. 2013; 128: 397–404.
Endometrial cancer; Recurrence; Prognostic factors; Salvage
Supplemental Digital Content
© 2014 by the International Gynecologic Cancer Society and the European Society of Gynaecological Oncology.