Endometrial carcinoma represents the most common gynecological malignancy in the United States. In 2004, an estimated 34,000 new cases occurred Standard therapy for endometrial carcinoma is total abdominal hysterectomy, bilateral salphingo-oophorectomy with lymph node sampling. Factors that influence the selection of adjuvant treatment include age, depth of invasion, tumor grade, lymphovascular space invasion, histologic type, and lymph node status. Three randomized studies in patients who had intermediate-risk or high-risk features have demonstrated that postoperative radiotherapy reduces the risk of pelvic recurrence but with no clear benefit to overall survival. This risk reduction may be a result of the effective salvage methods for patients who experienced recurrence.

The purpose of this retrospective study is to analyze the outcome of patients who had isolated vaginal recurrences treated with radiotherapy at Washington University School of Medicine. We reviewed the records of 50 patients treated at our institution between 1967 and 2003 for an isolated vaginal recurrence of endometrial carcinoma. Initial treatment for endometrial carcinoma was definitive surgery in 49 patients and definitive radiotherapy in 1 patient. The median time from initial diagnosis of endometrial carcinoma to recurrence was 25 months (range, 4–179 months). Three patients (6%) received external-beam radiotherapy alone, 8 patients (16%) received brachytherapy only, and 39 patients (78%) received combined external-beam radiation therapy and brachytherapy. Median dose of radiation to the recurrence was 60 Gy (range, 16–85 Gy). Overall survival was calculated by the Kaplan-Meier method. Endpoints were measured from the date of diagnosis of the vaginal recurrence. Median follow-up of survivors after recurrence was 53 months (range, 8–159 months).

Results: The 5-year and 10-year disease-free and overall survivals were 68% and 55%, and 53% and 40%, respectively. On multivariate analysis, age, Grade 1 or 2 vs. Grade 3 tumor , and size of recurrence were significant predictors of overall survival. All patients who had Grade 3 disease were dead by 3.6 years from the time of recurrence. Five patients experienced a Grade 3 or 4 complication.

Discussion
Survival rates after vaginal relapses after salvage radiotherapy have been reported in the literature to range between 25% and 68% . A previous report from our institution concluded that patients who had extravaginal disease were unable to be salvaged with radiotherapy alone and are excluded from this analysis. Similar results have been reported elsewhere. The majority of recurrences regardless of initial site occur during the first 3 years from the time of original treatment.

Several factors have been previously reported to correlate with overall survival. These factors include size of recurrent tumor, grade, dose of radiotherapy, and time to recurrence. High tumor grade has been reported to correlate with poor outcome . Hartin a review of 26 patients treated with definitive radiotherapy for isolated vaginal relapses, found that moderate to poor differentiation correlated with poor overall survival. Jhingran e reported in a similar cohort that patients who had Grade 3 lesions had poorer overall survival compared with patients who had Grade 1 or 2 lesions, as was seen in our study. In a surgicopathologic staging study conducted by the Gynecologic Oncology Group of 895 patients who had clinical Stage I or II endometrial carcinoma, Grade 3 tumors was the single most important predictor of recurrence. In our analysis, initial grade of the tumor was significantly predictive of overall survival; no patients survived beyond 3.6 years with Grade 3 tumors. Patients who had Grade 2 tumors had an outcome similar to patients who had Grade 1 tumors. This observation is consistent with what has been reported by others

Time to local recurrence has also been reported to be a significant predictor of overall survival, with a better prognosis for late recurrences . Most studies have reported recurrences within the first 3 years after initial diagnosis. Sears  reported actuarial survival rates of 40% and 70% for patients who relapsed less than 1 year and more than 1 year, respectively, from time of diagnosis. We did not find time to recurrence to be predictive of overall survival.

Local control of tumor has also been reported to be related to size of recurrence . Greven   reported size to be the most important prognostic factor for local control. Wylie et al. (9) found that disease bulk greater than 2 cm resulted in worse local control, but had no influence on overall survival. In our study, overall survival was significantly worse for patients who had disease greater than 2 cm; however, local control was independent of size.

Several groups have reported the influence of dose to the recurrence on overall survival and/or local control. Curran et al.  reported that patients who received 60 Gy or more had a significantly improved overall survival and pelvic-control results. Wylie also reported a trend toward improved local control at doses greater than 80 Gy (p = 0.07). Jhingran also found that doses of 80 Gy or more had higher local control rates. The median dose of radiation in our study was 60 Gy. No statistically significant difference occurred in local control or overall survival for patients with respect to dose in our series

Five-year disease-free survival in our study was 68%, with local control of disease in 37 of 50 patients. This outcome is similar to what has been reported by Jhingran. They found local control at 5 years to be 75%. We found no factors that were prognostic of local control in our study. This observation may be a result of the few pelvic or vaginal failures that were seen.

Two recent reported randomized studies have addressed the role of adjuvant radiotherapy in intermediate-risk patients . A Phase III study by GOG-99 randomized patients to receive pelvic EBRT or observation after surgery. Overall survival estimates at 48 months for patients in the observation arm was 86% vs. 92% in the radiotherapy arm . However, a statistically significant difference occurred in cumulative incidence of recurrence: 2% for radiotherapy arm vs. 12% for the observation arm (p = 0.007). Similar results were previously reported by the PORTEC study group: a 5-year locoregional recurrence rate of 14% in the no-adjuvant-treatment group vs. 4% in the pelvic-radiotherapy group   Despite the higher locoregional control rate, no survival difference was reported. The absence of a survival difference in these studies is likely related to the rate of successful salvage after relapse. In a follow-up report, the PORTEC study group analyzed the rates of local control and survival after relapse (3). The majority of recurrences occurred in the vaginal vault and could be salvaged by radiotherapy, surgery, or both, with the 3-year actuarial survival rate after vaginal relapse at 73%. In our analysis, 29 patients that would have met initial criteria for GOG-99, and at least 9 of those patients would have criteria for high intermediate-risk disease. Initial adjuvant radiotherapy for these patients may possibly have prevented subsequent local recurrence. Our current treatment policy at Washington University is to deliver adjuvant radiotherapy to patients who have high intermediate-risk endometrial cancer

.Conclusions: Patients treated with radiotherapy for an isolated vaginal recurrence can be cured in over 50% the cases. Radiotherapy is well tolerated, with a low risk of complications. Factors predictive of overall survival include tumor grade, patient age at recurrence, and tumor size.

Recurrent endometrial cancer after surgery alone: results of salvage radiotherapy
Jereczek-Fossa B, Badzio A, Jassem JInternational Journal of Radiation Oncology*Biology*Physics September 2000 (Vol. 48, Issue 2, Pages 405-413)

Cancer of the endometrium continues to be the most common gynecologic tumor worldwide. Surgery alone in low-risk patents, or surgery followed by irradiation in high-risk patients, remains the standard management. In the latter category, postoperative radiotherapy reduces the risk of local recurrence to about 5%   as compared with the risk of up to 27% in nonirradiated patients  On the other hand, irradiation is associated with numerous side effects and its effect on survival is unclear. The lack of prospective controlled studies evaluating the role of postoperative radiotherapy in relation to known prognostic factors has led to significant differences in indications for its use between centers . Only recently, two randomized studies confirmed better local control but not a survival benefit with postoperative irradiation   Irrespective of the practice in a particular institution, a number of previously nonirradiated patients are referred to radiotherapy departments for the irradiation of postoperative local recurrence. In view of controversies over the use of immediate postoperative irradiation, it is of particular importance to define the efficacy of radiotherapy applied at the time of relapse. Despite a large body of literature on the management of endometrial cancer, the issue of salvage radiotherapy has been addressed only occasionally . Our aim was to determine the efficacy of irradiation applied for local recurrence in previously not irradiated patients and to analyze the impact of patient- and treatment-related factors on survival. Additionally, we tested the usefulness of staging classification used in primary vaginal carcinoma for evaluation of the impact of recurrent tumor extension on treatment outcome.Postoperative irradiation of endometrial cancer patients decreases the risk of local recurrence but is associated with a number of long-term sequelae. In a proportion of patients, no immediate postoperative radiotherapy is applied and this treatment is introduced only at relapse. The aim of our study was to assess the long-term results of salvage radiotherapy in previously nonirradiated endometrial cancer patients who developed local recurrence, and to evaluate the impact of patient- and treatment-related factors on treatment efficacy.

Methods and Materials: We performed a detailed retrospective analysis of 73 endometrial cancer patients given radiotherapy for local recurrence after the initial surgery only. The mean age at diagnosis of the recurrence was 63 years (range, 39–78 years). Median time to recurrence was 11 months (range, 1–19 months). All recurrences were staged with the use of Perez modification of the International Federation of Gynecology and Obstetrics (FIGO) staging system for primary vaginal carcinoma. There were five (7%) Stage I patients, 43 (59%) Stage II patients, and 25 (34%) Stage III patients. Forty-four patients (60%) received both external beam irradiation (EBRT) and endovaginal brachytherapy (BRT), 17 (23%) received only BRT, and 12 (17%) received only EBRT. The mean total physical radiation dose was 75.9 Gy (range, 8–130 Gy), and the mean normalized total dose (NTD) calculated on the base of the linear-quadratic model was 86.6 Gy (range, 8.5–171.9 Gy). Median follow-up for alive patients was 8.8 years (range, 3–21 years). The impact of patient-, tumor-, and therapy-related factors on the treatment outcome was evaluated with the use of uni- and multivariate analyses.

Results: Three- and 5-year overall survival rates were 33% and 25%, respectively. In the univariate analysis, lower stage of recurrent disease (p < 0.0005), combined EBRT and BRT (p = 0.027), higher total radiation dose (p = 0.031), and higher NTD (p = 0.006) were significantly correlated with better survival. In the multivariate analysis, only stage of recurrent disease (p < 0.005) and high total dose (p = 0.047) were independently correlated with better survival. Lower FIGO stage of recurrence (p = 0.023) and higher total dose (p = 0.005) were also independently correlated with longer time to progression, whereas higher radiotherapy dose was the only factor correlated with better local control (p = 0.029).

Discussion  

Up to 27% of nonirradiated patients develop local recurrence after surgery for endometrial carcinoma (2, 4, 8). Most frequently, the failure is initially diagnosed in the vagina although at careful examination it is frequently accompanied by other tumor locations. Four mechanisms have been proposed as a cause of the vaginal failure in endometrial carcinoma patients: 1) seeding and implantation of tumor cells at the time of surgery; 2) unrecognized microscopic residual disease remaining in the vaginal cuff following surgery; 3) retrograde lymphatic dissemination; and 4) venous spread. The local recurrence is probably caused by a combination of these factors (31).

Radiotherapy, along with surgery, is the cornerstone of treatment of postoperative local recurrence. Our analysis showed that success of radiotherapy for recurrence depends on the early diagnosis of relapse (early FIGO stage) and radicality of irradiation.

The FIGO stage reflects the tumor extent. Most treatable are stage I failures, i.e., those located in the vagina. Unfortunately, isolated vaginal recurrences are rare; usually they are an indication of an underlying deep pelvic disease, which carries a poor prognosis. In our series only 7% of all patients had isolated vaginal mucosa involvement, whereas in 34% of cases the disease was extended to the pelvic wall. Progressing endometrial carcinoma was the cause of death in 16 of 29 stage I–IIa patients (55%) and in 22 of 25 stage III patients (88%). We found that the patients with lower FIGO stage at the diagnosis of recurrence responded better to radiotherapy, had higher probability of local control, longer time to progression, and longer survival. Thus, our study confirms the predictive value of modified FIGO classification in the local recurrences of endometrial cancer. This system has been found to be a good predictor of treatment outcome in other reports. In the study of Curran et   actuarial 3-year pelvic control rates were as follows: stage I—100%, stage IIa—53%, stage IIb—35%, and stage III—0%. In the series of Sears 5-year local control rates were 77%, 51%, and 17%, and 5-year disease-specific survival rates for stages I, II, and III: 86%, 38%, and 13%, respectively. Some authors also found the tumor size to be predictive of local control  Greven  observed an 86% local control rate for recurrences not larger then 2 cm and only 20% for larger tumors. The retrospective assessment of our material precluded precise definition of tumor size in each case; therefore, we could not analyze this factor. Other authors reported also the correlation between treatment results and initial stage of disease (18) or location of recurrence in the vagina (16, 19, 20). The salvage rate is higher for recurrence in the lower third part of the vagina; this location is also correlated with lower risk of distant metastatic disease (20). In our series no significant difference in survival was found for proximal and distal tumors; however, this factor could be tested in 29 patients only (Stage I and IIa), which makes the interpretation of the results difficult.

As in many other series, the majority of recurrences were diagnosed within the first 2 years after surgery. Therefore we recommend frequent follow-up examinations at least in that period. A trend toward worse treatment outcome in symptomatic, as compared to asymptomatic patients suggests the need for more intensive schedule of follow-up examinations. This issue has been a subject of recent studies. Berchuck  concluded that due to low recurrence rate of FIGO Stage I/II endometrial cancer and the paucity of effective second-line treatment, surveillance Pap smears and chest radiographs appear to have little impact on survival. Until more effective therapies are available for recurrent disease, these authors recommend, in the absence of symptoms, biannual visits for 5 years without Pap smears or chest radiographs. Based on a similar study, Shumsky et al.(33) proposed no routine follow-up for low-risk patients and a tailored schedule of follow-up for high-risk patients. Such a risk-specific follow-up allows using the health care resources more efficiently. Burke (34) recommended the specialist follow-up during the first 3 years and a family physician follow-up thereafter if a patient is disease-free. These less intensive approaches should, however, be tested within prospective randomized trials as has been the case for breast cancer (35, 36). Another common problem, observed also in our series, is omission of radiation oncologist consultation for gynecologic cancer patients operated in nononcologic centers. As a consequence, some high-risk patients are referred for irradiation only at the time of relapse.

The data on predictive value of time to recurrence are contradictory. In some studies , as in ours, disease-free interval had no impact on treatment outcome. Other authors, however, observed better treatment outcome if recurrences occurred after a longer disease-free interval

Predictive value of age has also not been clearly established. Similar to our results, Curran  did not find age to be correlated with prognosis, whereas Sears  showed a superior outcome for older women treated with radiotherapy for local recurrence. In contrast, higher age was demonstrated to have an unfavorable impact on the risk of pelvic and extrapelvic failures . As in the study of Kuten et al.(18), we did not find any influence of histology on treatment results. Vavra  observed significantly lower survival rates for papillary endometrial carcinoma, but this result was due to a strong tendency to distant dissemination associated with this subtype (Vavra’s report included both local and distant recurrences).

The important prognostic factor revealed by our analysis was the total radiation dose. The correlation between radiation dose and treatment results was earlier suggested in endometrial carcinoma patients applied definitive  or preoperative (45,   radiotherapy. Earlier studies on the endometrial cancer cell cultures showed the radiation sensitivity of endometrial carcinoma in vitro. However, no data on the radiosensitivity of recurrent endometrial cancer are available in the literature. Clinical tumor response was seen in the majority of our patients, similarly to the observation of Kuten .

The combination of EBRT and BRT seems to be the best treatment approach for postoperative local recurrence. Such management allows one to administer higher doses both to the vaginal tumor and to the extravaginal disease. In our series the mean dose in patients who received EBRT and BRT was higher (89.9 Gy), as compared to those treated with EBRT or BRT only (59.9 Gy and 45.8 Gy, respectively). Additionally, in patients given BRT only, treatment volume was, by definition, much smaller than in those given EBRT. We must admit that some of our patients were clearly undertreated. By analogy with primary vaginal cancer, BRT alone should be reserved to patients with tumors limited to the vaginal mucosa (Stage I) (48), whereas in our series such a treatment was also administered in some Stage II and III cases. Sears et al.(16) observed better local control in patients given EBRT followed by BRT than in those with the opposite sequence. The importance of high radiation dose has also been reported by Curran . Better survival and local control were seen in patients who received total radiation dose of at least 60 Gy . In our analysis total physical radiotherapy dose was an independent prognostic factor both for time to progression and overall survival. We demonstrated the lack of independent correlation between NTD and treatment outcome, although such a correlation was found in a univariate analysis. Analysis based on the NTD model may be obscured by some radiobiological uncertainties in the clinical application of the linear-quadratic model. For example, ?/? ratio for endometrial carcinoma has not yet been clearly defined. Another limitation of dose–outcome analysis is a retrospective character of all reports on recurrent endometrial cancer. Patients might have been selectively allocated to different treatment schedules and doses.

Similar to other reports , in our series pelvic progression was the predominant site of failure. Thus, the optimization of local treatment seems to be essential. Higher radiation doses, debulking surgery before or after irradiation, as well as innovative approaches including hyperthermia and intraoperative irradiation should be further investigated. In highly selected patients who have exhausted other treatment modalities, pelvic exenteration may by taken into consideration (49). Opposite to the first discouraging studies of Brunschwig in 1961 (50), Morris et al.(49) achieved with this procedure a 45% 5-year disease-free survival.

The small number of patients receiving concomitant progestational therapy precluded analysis of this factor in our series. Curran et al.(15) did not find any additive advantage of this treatment over radiotherapy alone in 34 patients treated for local recurrence. Several recent phase II studies evaluated the value of salvage cytotoxic and hormonal therapies in patients with recurrent or advanced endometrial cancer (51–54). Encouraging results of these trials have still to be verified within prospective randomized studies, since relatively high response rates were only occasionally translated into long-term survival.

The current management of local recurrences after previous surgery is still of limited efficacy. Successful salvage therapy can only be sought in patients with limited vaginal relapses. Overall 5-year survival rates measured from the time of recurrence ranged from 20% to 50% In our series the cure was only achieved in less than one-third of cases and the median survival after the diagnosis of recurrence was merely 20 months. This is comparable with the median survival since the diagnosis of local recurrence (18 months) reported in endometrial cancer patients who received postoperative irradiation at our institution . Surprisingly, Ackerman  who observed the relatively good therapeutic results in radiotherapy-naive patients with pelvic recurrence (5-year survival rate of 44%), suggested no need of postoperative irradiation in moderate-risk patients who underwent extensive surgery. However, the results coming from small studies (Ackerman’s series included 32 patients with postoperative pelvic recurrence) should be interpreted with caution. On the other hand, the frequency of distant dissemination after local failure demonstrated in other series points to the role of local control in endometrial cancer (56).

Conclusions: The efficacy of salvage radiotherapy in endometrial cancer patients with local failure after previous surgery is limited. Factors determining treatment outcome include advancement of the tumor at relapse and radiotherapy dose.