Intravaginal brachytherapy alone for intermediate-risk endometrial cancer
Despite the results of the Gynecologic Oncology Group trial No. 99 (GOG#99), some unanswered questions still remain about the role of adjuvant radiotherapy (RT) for intermediate-risk endometrial cancer. First, can intravaginal brachytherapy (IVRT) alone substitute for external beam RT but without added morbidity? Second, is the high-risk (HR) definition from GOG#99 a useful tool to predict pelvic recurrence specifically? The purpose of this study was to try to answer these questions in a group of patients with Stage IB-IIB endometrial carcinoma treated with high-dose-rate (HDR) IVRT alone.
Between November 1987 and December 2002, 382 patients with Stage IB-IIB endometrial carcinoma were treated with simple hysterectomy followed by HDR-IVRT alone at our institution. Comprehensive surgical staging (CSS), defined as pelvic washings and pelvic/paraaortic lymph node sampling, was performed in 20% of patients. The mean age was 60 years (range, 2992 years). Lymphovascular invasion (LVI) was present in 14% of patients. The median HDR-IVRT dose was 21 Gy (range, 621 Gy), given in three fractions. Complications were assessed in terms of late Radiation Therapy Oncology Group (Grade 3 or worse) toxicity of the GI tract, genitourinary GU tract, and vagina.
Results: With a median follow-up of 48 months, the 5-year vaginal/pelvic control rate was 95% (95% confidence interval [CI], 9398%). On multivariate analysis, a poor vaginal/pelvic control rate correlated with age ≥60 years old (relative risk [RR], 3, 95% CI, 112; p = 0.01), International Federation of Gynecology and Obstetrics (FIGO) Grade 3 (RR, 9, 95% CI, 235; p = 0.03), and LVI (RR, 4, 95% CI, 113; p = 0.051). The depth of myometrial invasion and CSS, however, were not significant. With regard to pelvic control specifically, the presence of GOG#99 HR features did not affect the pelvic control rate. The 5-year rate for HR patients was 96% (95% CI, 90100%) vs. 96% (95% CI, 9499%) for those without HR disease (p = 0.48). Even when the CSS effect was taken into account, the influence of HR features on pelvic control was still not significant (p = 0.51). In contrast, pelvic control was significantly influenced when patients were grouped according to CSS and stage/grade substages. For those with Stage IB Grade 3-IIB and no CSS, the 5-year pelvic control rate was 86% compared with 97% for those with Stage IB Grade 3-IIB and CSS, 97% for Stage IB, Grade 12 without CSS, and 100% for those with Stage IB, Grade 12 and CSS (p = 0.027). The 5-year disease-free survival rate was 93% (95% CI, 9096%). On multivariate analysis, poor disease-free survival correlated with age ≥60 years (RR, 5; 95% CI, 118; p = 0.002), FIGO Grade 3 (RR 5, 95% CI 217; p = 0.013), and LVI (RR 3, 95% CI 18; p = 0.054). Unlike pelvic control, disease-free survival was significantly affected by GOG#99 HR features, with a 5-year rate of 87% (95% CI, 7699%) vs. 94% (95% CI, 9197%) for those without HR features (p = 0.027). The 5-year overall and disease-specific survival rate was 93% and 97%, respectively. The overall 5-year actuarial rate of Grade 3 or worse complications was 1% (95% CI, 02%).
Conclusion: Tumor grade, depth of invasion, and the use of CSS were better predictors of pelvic control than the GOG#99 HR factors. IVRT alone seemed to provide adequate tumor control with very low morbidity. Therefore, it seems prudent to consider it for intermediate-risk patients because of its superior therapeutic ratio compared with that for surgery alone or pelvic RT. Additional follow-up, however, with a larger number of patients is needed, especially for those with LVI.
The Gynecologic Oncology Group trial No. 99 (GOG#99), randomizing patients with Stage IB-IIB endometrial cancer to surgery vs. surgery and pelvic RT, was especially important because it tried to address the role of pelvic RT for patients who had undergone comprehensive surgical staging (CSS). In addition, the authors of GOG#99 divided the patients into high-risk (HR) and low-risk groups. Patients were considered HR if they met a combination of the following criteria: advancing age, Grade 23 tumors, lymphovascular invasion (LVI), or outer-third invasion (2). However, despite the results of the GOG#99, some unanswered questions still remain. First, can intravaginal brachytherapy (IVRT) alone substitute for external beam RT? Second, is the HR definition from GOG#99 a useful tool to predict pelvic recurrence per se or is it a predictor of recurrence in general? Answering those two questions is very critical. Perhaps, rather than abandoning adjuvant RT altogether for intermediate-risk endometrial cancer, it would be useful to determine whether IVRT alone can provide equivalent locoregional control to that with external beam RT but without the added toxicity. Furthermore, if the GOG#99 HR factors are not good predictors of pelvic control, we could not reliably use those factors to determine who needs adjuvant RT.
Several reports have been published in the literature about IVRT, but not all of them have been applicable to everyday practice. Most of the reports included patients with Stage IA, Grade 12 disease, for whom most authors would not recommend adjuvant RT, thus diluting the results. In addition, the use of IVRT in some reports was linked to pelvic and paraaortic lymph node dissection. Yet, it could be argued that lymph node dissection is not needed for all subsets of patients with early-stage endometrial cancer. Therefore, the purpose of this study was to try to answer those questions in a group of patients with Stage IB-IIB endometrial carcinoma treated with simple hysterectomy with or without CSS followed by high-dose-rate (HDR) IVRT alone.
Interpreting the results of the GOG#99 trial represents a major challenge to those interested in the treatment of Stage IB-IIB endometrial cancer. According to some, the lack of a difference in overall survival and the associated morbidity of adjuvant pelvic RT are sufficient reasons to abandon all forms of RT for this group of patients. However, it is essential to realize that the sample size in this trial, especially in light of the high number of intercurrent deaths, was not large enough to answer an overall survival question. To prove that point, one needs only to quote the authors of GOG#99: Approximately half of the deaths were due to causes other than endometrial cancer or treatment (NAT: 19 of 36; RT: 15 of 30). With this number of intercurrent deaths in both arms, even if RT reduces the risk of endometrial cancer-related deaths, the size of this trial is not adequate to reliably detect an overall survival difference. More importantly, the primary endpoint in GOG#99 was not overall survival but progression-free survival, which improved significantly (relative hazard 0.42; p = 0.007) with pelvic RT.
Rather than abandoning RT altogether, perhaps it would be more realistic to try to determine whether IVRT could maintain locoregional control equivalent to that with pelvic RT while providing a better toxicity profile. In the current study, the 5-year rate of vaginal/pelvic recurrence was 3.9%, which compares very favorably with that of GOG#99 (1.6% control rate with RT and 8.9% with surgery alone), keeping in mind that only 20% of patients in the current study underwent CSS and all patients in GOG#99 did so. Opponents of adjuvant RT consistently point to the high rate of complications associated with pelvic RT in both the GOG#99 and the Post Operative Radiation Therapy in Endometrial Carcinoma trials. However, it is important to note that most of these complications were Grade 1 or 2 and, for the most part, were transient. Furthermore, the mechanism for reporting toxicity is slightly different for RT compared with surgery. In the current study, the 5-year actuarial rate of Grade 3 or worse toxicity was only 1%, which also compares very favorably with that in GOG#99. It could be argued that the low rate of significant complications with IVRT reported in this study, as well as in many other reports, is a reflection of the retrospective nature of these studies. That could be true, especially for low-grade complications, but it is unlikely that many Grade 3 or worse complications would be underreported.
To determine which group of patients are at high risk of recurrence, GOG#99 separated the patient population into HR vs. low-risk groups, as determined by advanced age, Grade 2 or 3 tumors, LVI, and outer one-third myometrial invasion. Such grouping is helpful in determining recurrence, in general, but it is not clear whether it is applicable to pelvic recurrence per se. Minimizing pelvic recurrence is the pivotal point in deciding on surgical lymph node staging or pelvic RT. In the current study, the 5-year DFS rate was 87% (95% CI, 7699%) for those with HR features compared with 94% (95% CI, 9197%) for those without (p = 0.027). In contrast, the 5-year rate of pelvic control for patients with HR features was 96% (95% CI, 90100%) compared with 96% (95% CI, 9499%) for those without (p = 0.48). Therefore, it seems that the presence of HR features is a better guide for systemic therapy than deciding whether adjuvant RT should be given.
Perhaps a better guide to determining how best to manage the pelvis would be to categorize patients according to the depth of myometrial invasion and tumor grade. In the Norwegian Radium Hospital trial, 540 patients with Stage IB-IC endometrial cancer who were treated with total abdominal hysterectomy and bilateral salpingo-oophorectomy and postoperative IVRT (60 Gy to the vaginal mucosa) were then randomized to observation (n = 277) or supplemental pelvic RT to 40 Gy (n = 263). A significant reduction in locoregional recurrence was seen with the addition of pelvic RT (1.9% vs. 6.9%, p < 0.01). However, the impact of additional pelvic RT was not the same across the board. When the subset of patients with Stage IB, Grade 12 was evaluated (n = 257) in that trial, the rate of local recurrence either in the vagina/pelvis was 4% (5 of 126) in those treated with IVRT alone compared with 2.3% (3 of 131) for those treated with brachytherapy and external beam RT. The rate of vaginal/pelvic recurrence in the subset of patients with Stage IB Grade 3-IC, however, was 9.3% (13 of 137) for those treated with brachytherapy alone compared with 1.3% (2 of 146) for those treated with brachytherapy and external beam RT (17). The high rate of recurrence in the latter group of patients treated with brachytherapy alone should not be surprising, because lymph node dissection was not performed in that trial.
In the current study, patients with Stage IB Grade 3-IIB, who did undergo CSS also had poor pelvic control compared with those with Stage IB Grade 3-IIB, and CSS or Stage IB, Grade 12 irrespective of surgical staging status (p = 0.027). Thus, a case could be made for IVRT alone for patients with Stage IB, Grade 12, irrespective of surgical staging status, and for patients with Stage IB Grade 3-IIB, who underwent CSS. For patients with Stage IB, Grade 12, one might argue that observation is the best option, and we concur that the option should be presented to patients. However, we should also recognize that this substage represents about 40% of all newly diagnosed endometrial cancers and that the small reduction in recurrence with IVRT is similar to that with adjuvant chemotherapy for node-negative breast cancer patients. For patients with Stage IB Grade 3-IIB and CSS, the data from this study, as well as from others, seem to indicate that IVRT alone is sufficient. Chadha reported on 38 patients with Stage IB Grade 3-IC, who underwent lymph node dissection followed by IVRT. With a median follow-up of 45 months, none of the patients developed pelvic relapse. Similarly, in the report by Horowitz 102 of 164 patients with Stage IB Grade 3-IIB, were treated after lymph node dissection with IVRT alone, and the rate of pelvic recurrence was only 1.9% (2 of 102).
Despite the encouraging results with IVRT alone after CSS in patients with Stage IB Grade 3-IIB, it is important to note that external beam RT has a longer track record with a greater number of patients. Therefore, until we have a real comparison of IVRT alone to pelvic RT in this subset of patients, that shows equivalency, it is important to continue to consider pelvic RT a viable option. Maintaining this option is very crucial, especially for patients with LVI. Several reports, including this one, have demonstrated the significance of LVI as a predictor of recurrence in endometrial cancer. However, the impact of LVI on pelvic control if the patient undergoes CSS needs additional evaluation.