Introduction

Endometrial cancer is the most common gynecological tumor in developed countries, with increasing prevalence [1, 2]. Clinical and pathological risk factors are taken to triage patients and estimate their prognosis. Lymph node involvement (pN+), lymphovascular space involvement (LSVI), grading, age of the patients, and T classification are internationally accepted risk factors for treatment decisions [3,4,5].

Low-risk disease is defined as myometrial infiltration <50% (FIGO IA) and G1–2. Intermediate-risk patients include those with IAG3 and IB (MI, myometrial infiltration >50%) G1 or G2. According to the GOG-99 study [6], high-intermediate-risk was defined if patients are ≥70 years of age with any one of the risk factors tumor grade 2 or 3, MI > 50%, or LVSI. Patients ≥50 years with any two of the abovementioned risk factors (RF) or patients of any age with all three RF were defined as high-intermediate risk. More advanced stages are defined as high-risk disease.

Clinicians’ experience that some of the patients with the low-risk tumors show relapses and rapid progression with distant metastases while most of them have a good prognosis without any adjuvant treatment challenges the conventional classification of type I (“good”) for estrogen-dependent tumors and the type II (“bad”) tumors [7,8,9].

Recently, integrated genomic characterization created a new classification with (1) POLE ultra-mutated, (2) microsatellite instability hypermutated, (3) copy number low, (4) and copy number high tumors with distinct risks for progression. This might be a future prospect for better patient selection for adjuvant radiation and/or chemotherapy.

Some aspects of clinical and pathological risk factors need to be discussed.

Lymph node status

Thirty years ago, a randomized trial by the GOG [10] underlined the importance of lymph node involvement for the prognosis of patients. 15 and 16% of patients with clinical stage I, II presented with pelvic and paraaortic metastases, respectively. The latter patients had a poor survival of less than 30% after 5 years. Two decades later, the ASTEC trial and Italian trial [4, 11, 12] could not show any benefit in terms of 5‑year overall survival from lymphadenectomy. The use of LNE is declining in stage I patients [13, 14]. In contrast, a Japanese study including mostly patients with intermediate- and high-risk parameters demonstrated a survival benefit of pelvic and paraaortic lymphadenectomy compared with only pelvic lymphadenectomy for all subgroups of included patients—low-, intermediate- and high-risk groups [15]. For advanced stages II and III, lymphadenectomy remains a part of surgical treatment. As for cervical cancer, FDG-PET-CT demonstrated a disappointing sensitivity of 65% for detection of pelvic and paraaortic nodes and cannot be recommended for lymph node staging in endometrial cancer [16].

Sentinel node concept in endometrial cancer

Although not validated in randomized studies, there is an increasing use of the sentinel node technique (SLN) [17]. From radiation oncologists’ point of view, results of ultrastaging after SLN raise new questions. Plante et al. [18] demonstrated that after sentinel procedure, about 50% of 500 patients had macrometastases, but 13 and 36% showed micrometastases and isolated tumor cells within the lymph nodes, respectively. These 50% of the lymph node findings (micrometastases and isolated tumor cells) have never, or at least only accidentally, been seen by conventional histology [19]. The prognostic value of these “new” findings based on ultrastaging and its clinical implication is unclear. Plante and co-authors [18] claim that micrometastases and isolated tumor cells imply a better prognosis. This should be evaluated in a trial randomizing patients with micrometastases and isolated tumor cells to either adjuvant treatment or follow-up.

Grading and stage

For endometrial cancer, most parameters for the treatment decision are given by the pathologists. In the PORTEC-1 study, a substantial shift from grading 2 to grading 1 could be demonstrated after central review of the specimen. Years later again, a central pathology review by expert gynecopathologists changed histological type, grade, or other items in 43% of women with high-risk endometrial cancer [20,21,22]. This raises the question of how reliable our treatment decisions are on the basis of pathology reports, even outside clinical trials.

Role of adjuvant external beam radiation

The results of four randomized trials are summarized in a Cochrane analysis. In stage I patients, external beam radiation decreases the pelvic as well as the vaginal recurrence rate significantly, without impact on overall survival ([23]; Table 1). The abovementioned studies were underpowered with regard to overall survival (OS), because OS was not the endpoint. However, patients with grade 3 tumors and >50% myometrial invasion achieved an improved survival after EBRT [6, 24, 25], although higher rates and grade of gastrointestinal and genitourinary toxicity were documented after EBRT.

Table 1 Randomized trials on external beam radiation (EBRT) in FIGO (classification bevor 2009) stage I, II endometrial cancer
Full size table

Role of adjuvant brachytherapy

In the PORTEC 2 trial, vaginal brachytherapy versus EBRT were compared in patients with high-intermediate-risk endometrial cancer. EBRT significantly reduced pelvic and vaginal relapses, brachytherapy reduced vaginal relapses only. No difference was seen in OS. Given the short treatment time and mild GI toxicity, vaginal brachytherapy has been the standard of care in patients with stage IA G3, IB G1–2 without additional risk factors. Long-term quality of life did not differ between VBT and EBRT ([27, 28]; Table 1).

Some argue against any adjuvant treatment because salvage rates seem to be acceptable in EC patients. Vance et al. [29] evaluated a total of 156 recurrent women with early stage endometrial cancer after hysterectomy with or without adjuvant radiation. The (immediate) adjuvant radiation group had significantly better 5‑year disease-specific survival compared with patients after salvage treatment (95% vs 77%, P < 0.001) and 5‑year OS (79% vs 72%, P = 0.005). Salvage treatment caused more toxicity than adjuvant radiation with higher rates of ≥grade 2 GI and ≥grade 2 GU toxicity (23% versus 54% and 3% versus 44%, respectively) in favor of immediate adjuvant treatment.

Role of adjuvant chemotherapy

Against the background of the PORTEC 1, GOG-99, and Norwegian trials which could not demonstrate a survival benefit in patients after radiation, one idea of the new millennium was to replace radiation by chemotherapy. Three randomized trials were performed—the Japanese phase III [30] trial of pelvic radiotherapy versus cisplatin-based chemotherapy in patients with intermediate and high-risk endometrial cancer, the Italian trial evaluation of adjuvant chemotherapy versus radiotherapy in high-risk endometrial carcinoma [31], and the GOG-122 [32] randomized phase III trial which evaluated the whole-abdominal irradiation versus doxorubicin and cisplatin chemotherapy in locally advanced endometrial carcinoma.

The Japanese trial [30] included patients with endometrioid adenocarcinoma with deeper than 50% myometrial invasion. Among 385 patients, 193 patients received pelvic radiotherapy with 45–50 Gy with an outdated APPA technique. 192 patients received cyclophosphamide, doxorubicin, and cisplatin every 4 weeks for at least three cycles. The 5‑year PFS rates and the 5‑year OS rates did not show statistically significant differences between chemotherapy and radiation. Interestingly, after chemotherapy, there were slightly more extrapelvic recurrences. Chemotherapy was more toxic, with ≥4.8% grade 3 toxicity versus 1.6% in the radiation arm. No treatment-related deaths occurred.

For the Italian trial [31], 345 patients in FIGO stage (before 2009) IcG3, IIG3 with myometrial invasion >50%, and FIGO stage III were randomized to either EBRT with 45 Gy or to adjuvant chemotherapy with cisplatin (50 mg/m2), doxorubicin (45 mg/m2), and cyclophosphamide (600 mg/m2) every 28 days for five cycles. Radiotherapy delayed local relapses and chemotherapy delayed distant metastases. Nevertheless, the 3‑, 5‑, and 7‑year progression-free survival rates were 69, 63, and 56% for radiotherapy and 68, 63, and 60% for chemotherapy, respectively (n.s.).

GOG-122 [32] is the only chemotherapy versus radiation study which showed significantly improved results in favor of chemotherapy. It has been one of the most misinterpreted studies in gynecologic oncology. There is a need to highlight the background: The study used an experimental whole-abdomen radiation (n = 202) versus chemotherapy (doxorubicin 60 mg/m2 and cisplatin 50 mg/m2 every 3 weeks for seven cycles, followed by one cycle of cisplatin; n = 194) in women with stage III or IV endometrial carcinoma having a maximum of 2 cm of residual disease after surgery. That implies that the GOG-122 study is not a study for adjuvant, but for locally advanced disease and/or the palliative situation. Neither the used chemotherapy nor the disturbing concept of whole-abdominal radiation with 30 Gy in 20 fractions and a 15 Gy boost to the pelvis and/or paraaortic nodes has been proven for this indication. As a result, adherence to radiation was 84%, and only 63% of the patients received all cycles of chemotherapy. 69 out of 191 patients experienced grade 4 hematologic toxicity during chemotherapy. There were 5 deaths during treatment. In summary, there is no conclusive evidence that chemotherapy should replace radiation in advanced EC patients. Table 2 gives an overview.

Table 2 Studies on chemotherapy versus radiation in endometrial cancer
Full size table

Role of adjuvant chemoradiation

As a result of the studies discussed above, it has been shown that chemotherapy is able to reduce (or delay) distant metastases and radiation is effective to decrease local (vaginal and/or pelvic) failures. The consequence was to combine both modalities as done by the NSGO/EORTC [33] and MaNGO-ILIADE-III trials [34]. 534 patients with operated endometrial cancer in FIGO stage I–III with high-risk factors but no residual tumor were randomly allocated to adjuvant radiotherapy with or without sequential chemotherapy. Pelvic radiation was given according to “departmental guidelines” (total dose ≥44 Gy). Optional vaginal brachytherapy had to be decided upon before randomization. Sequential chemotherapy for those patients in the RT + CT arm consisted of four courses of doxorubicin/epirubicin 50 mg/m2 + cisplatin 50 mg/m2 every 4 weeks. A later amendment allowed alternative regimens, including paclitaxel 175 mg/m + epirubicin 60 mg/m2 or doxorubicin 40 mg/m2 + carboplatin AUC 5 or paclitaxel 175 mg/m2 + carboplatin AUC 5–6 every 3 weeks (Table 3).

Table 3 Studies on concomitant chemoradiation in endometrial cancer
Full size table

None of the studies showed significant differences in terms of overall survival in favor of combination of radiation and chemotherapy. In the combined analysis of both studies, there was a trend toward improved overall survival (HR 0.69, CI 0.46–1.03; P = 0.07). Surprisingly, the combined analysis confirmed the positive trend for patients with endometrioid histology only. Patients with serous/clear cell tumors had no benefit at all [34]. For the subgroup of patients with histologically confirmed lymph node metastases (FIGO IIIC), a recently published comparative analysis [37] showed that either radiation or chemoradiation is superior to chemotherapy in terms of distant recurrence-free and overall survival.

Over the past decade, the sequence of chemotherapy and radiation has been a matter of debate. There is no evidence from randomized trials. Alvarez et al. [38] reported on patients with stage III and IV EC who were “optimally” debulked in 83%. The cohort is comparable to those patients included into the GOG-122 trial. Patients benefited from chemoradiation significantly in terms of survival compared to radiation alone or chemotherapy alone. A smaller subset of patients was analyzed for the sequence of treatment. Those patients who received “sandwich” therapy (chemo-radiation-chemo) had the longest 3‑year OS (91%) and PFS (69%) compared to those treated with either radiotherapy or chemoradiation followed by chemotherapy.

More data were expected impatiently as described for the recently published studies.

PORTEC-3 study

The randomized PORTEC-3 intergroup trial [35] was initiated to investigate the potential benefit of concomitant chemoradiation followed by adjuvant chemotherapy versus pelvic radiotherapy alone for women with high-risk parameters (FIGO stage I grade 3 with deep myometrial invasion and/or LVSI, stage II or III, or serous/clear cell histology). Between 2006 and 2013, 103 centers included 686 women who were randomly allocated to radiation (RT) with 48.6 Gy in 1.8 Gy fractions, or chemoradiation (two cycles of cisplatin 50 mg/m2 in week 1 and 4 of RT, followed by four cycles of carboplatin AUC5 and paclitaxel 175 mg/m2 at 3‑week intervals; CRT-Ch). Primary endpoints were overall survival (OS) and failure-free survival (FFS). Median follow-up time was 5 years. Three- and five-year OS for CRT-Ch vs. RT was 84.4% versus 83.9%, and 81.8% versus 76.7%, respectively (P = 0.183). Patients with stage III EC had the greatest benefit from the combined treatment: 5‑year FFS for stage III was 69.3% for CRT-Ch versus 58.0% for RT (P = 0.032), and 5‑year OS for stage III was 78.7% versus 69.8% (P = 0.114). Also, patients >70 years, with FIGO stage III and/or serous histology had the greater benefit.

As in many other studies on adjuvant treatment in EC, the problem of the PORTEC-3 study is the inappropriate selection of patients. Included were pooled patients in “FIGO stage III” endometrioid carcinoma, which is a very heterogeneous group of patients with, e.g., adnexal involvement (IIIA), infiltration of the parametrium (IIIB), vaginal infiltration, and last but not least, histologically confirmed pelvic (IIIC1) and/or paraaortic lymph node metastases (IIIC2) with considerable differences in prognosis [39, 40]. There is no information available on the subgroups of stage III. Less than 45% in both arms underwent lymphadenectomy (LNE). That means that 55% of all patients with a considerable risk of lymph node metastases could not be exactly staged as IIIC1 or IIIC2, because there were no data on lymph node involvement. This implies a considerable risk for bias. If those patients also underwent LNE, about 30% of them are stage IIIC, which could have led to other results and interpretations. Todo et al. [39] demonstrated the dramatic impact on prognosis of pelvic and/or paraaortic LNE in patients with endometrial cancer. Worst prognosis was associated with pelvic LNE and histologically confirmed lymph node metastases. Favorable prognosis was associated with full (pelvic AND paraaortic) LNE, but only pelvic lymph nodes. Out of consideration for the participating centers and the accrual of patients, LNE was not mandatory in patients with a high risk for lymph node metastases, which is the main weakness of present data.

Negative results in terms of OS for all patients, but positive results for subgroups of the PORTEC-3 study have to be weighed against the treatment-related toxicity: Grade ≥2 events were doubled in the combined arm (94% versus 44% after CRT versus RT). Grade ≥3 events were found in 61% versus 13% (CRT versus RT, (P < 0.0001). At 24 months, grade ≥2 sensory neuropathy persisted in 10% versus 1% of the patients (CRT versus RT, P < 0.0001) [41]. The chemotherapy in the PORTEC 3 trial could not reduce the high rate of distant metastases (28% with radiation, 22% with chemoradiation). As expected, local control was excellent in both arms.

GOG-258/NRG oncology study

The contrast program to the abovementioned PORTEC-3 study was the randomized phase III study on chemotherapy, only versus chemoradiation. 407 and 406 patients were assigned to either chemoradiation (EBRT n = 174; or EBRT ± brachytherapy with concomitant cisplatin 50 mg/m2 d1, d29, n = 204) followed by carboplatin AUC5 plus paclitaxel 175 mg/m2 q21 × 4 (CRT-Ch) versus chemotherapy only (Ch; carboplatin AUC6 and paclitaxel 175 mg/m2 q21 × 6) in “optimally” debulked (<2 cm residual tumor) locally advanced endometrial carcinoma [42]. Stage III/IVA endometrioid and FIGO stage I/II serous cancers were eligible. Pelvic ± paraaortic “lymph node sampling,” whatever that means, was optional.

Radiation + chemotherapy led to a dramatic reduction of pelvic and vaginal recurrences compared to chemotherapy (19 versus 10%; 7 versus 3%; respectively). Nevertheless, 5‑year OS estimation is 70% versus 73% for CRT-Ch versus Ch. No difference in recurrence-free survival could be demonstrated. After 5 years, the rate of distant metastases was slightly lower in the Ch-arm with 21%, versus 27% in the CRT-Ch-arm. Neither the reduction in distant metastases by chemotherapy nor the reduction of local recurrences could be translated into an oncologic benefit of one of the treatment arms. No significant difference in terms of toxicity was reported. Two consequences can be postulated:

  • Radiation led to a dramatic reduction of local recurrences and should be not replaced by chemotherapy only.

  • Data underlie the need for more aggressive systemic treatment in selected patients. Distant metastases rate is comparable to PORTEC-3, but too high.

Chemotherapy schedule and dose density and endometrial cancer

In the earlier studies, anthracycline-containing regimes were used. Later, paclitaxel plus carboplatin was a commonly used drug combination in endometrial cancer, with response rates of 60–70% in advanced or recurrent endometrial cancer. Neither adjuvant nor concomitant schedules were tested. In contrast to cervical cancer, head and neck cancer, and ovarian cancer, there is no evidence of an optimal treatment regimen nor for dose responsiveness in endometrial cancer in vivo.

According to cervical cancer treatment, cisplatin was used in the concomittant treatment. In RTOG-9708 [36], cisplatin was used for the first time in a simultaneous (2 × 50 mg/m2) and concomitant setting (4 × 50 mg/m2). Because of the lack of evidence, PORTEC-3 used the simultaneous chemotherapy part from the abovementioned RTOG-9708 and adopted the adjuvant part from cisplatin to carboplatin in a comparable dose. The simultaneous part of the study contained a lower dose density with 100 mg/m2 in five weeks of cisplatin a lower dose density than usually administered (200 mg/m2 in five weeks). Less cisplatin has been compensated by the adjuvant part with carboplatin AUC5, 4 cycles, every 4 weeks. Nevertheless, focusing on distant metastases, PORTEC-3 presented a disappointing rate of distant metastases of 22% in the combined arm compared with 28% in the radiation only arm.

Unfortunately, two eagerly awaited studies could not clarify the old question on the optimal adjuvant treatment for high-risk endometrial cancer patients. For daily clinical practice, radiation treatment remains the basis of adjuvant treatment in high-risk endometrial cancer. Whether and for whom the combination of simultaneous chemoradiation provides a benefit, remains unclear. It is likely that stage III patients may have a benefit. In this light, the recently published optimistic interpretation of PORTEC-3 seems to be premature [43].

In conclusion, patients with low-risk endometrial cancer are currently treated with hysterectomy and BSO followed by either no further treatment for low-risk (FIGO IA G1,2) or brachytherapy for FIGO IAG3. External beam radiation significantly reduced pelvic and vaginal recurrences in FIGO stage I and II [23, 24, 26, 44].

Although randomized trials were underpowered for overall survival benefit, subgroups of patients with FIGO stage I, II (G3, older age, FIGO ≥IB) might benefit from external beam radiation in terms of survival [6, 45]. Consensus is that patients with locally advanced endometrial cancer, older patients, G3, deep myometrial infiltration are at a higher risk for local recurrence and distant metastases. This implies the use of both radiation and chemotherapy [37, 38, 46]. There are no data on high-risk endometrial cancer demonstrating a superiority of chemotherapy over radiation [34], except GOG-122, which cannot be considered as a contemporary adjuvant study [32]. The combination and schedule of radiation and/or chemotherapy has been a bone of contention for many years. Two studies were initiated to clarify the situation: GOG 258 compared adjuvant chemotherapy versus chemoradiation, PORTEC-3 compared radiation against chemoradiation in patients with high-risk disease. Both studies missed their primary endpoints. PORTEC-3 study showed a better local control in both arms with radiation. Without radiation, 20% of the patients in GOG 258 experienced local failures with chemotherapy only. Patients in both arms did not benefit in terms of distant metastases disease-free survival from potentially toxic chemotherapy.

As a result, radiation remains the standard of care in patients with high-risk disease and FIGO stage III. The conclusion of the recently updated German guideline for endometrial cancer can therefore no longer be considered as an evidence-based recommendation. Nearly 93% of the experts stated that for FIGO stage III, IVA patients, radiation therapy “could” be given additionally to chemotherapy. After the final publication of PORTEC-3, the evidence-based decision should be for adjuvant radiation. Chemotherapy “could” be given additionally, but “should be weighed against the severity and duration of toxicity of combined treatment, especially since overall survival was not significantly improved” with the addition of chemotherapy, as the PORTEC authors stated [35].