Adjuvant radiotherapy of endometrial cancer: role of 18 F-FDG-PET/CT in treatment modulation

Objective : Residual disease after surgery is related to an unfavorable prognosis in patients with endometrial cancer (EC). An early diagnosis and treatment of this condition could improve patients’ outcome. Aim of this study was to define the role of postoperative 18 F-Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography ( 18 F-FDG-PET/CT) in patients with high risk of residual disease after EC surgery. Methods : Patients operated for EC, with one or more risk factors, who underwent 18 F-FDG-PET/CT before adjuvant treatment were included in this observational study. The primary endpoint was the rate of patients in whom 18 F-FDG-PET/CT changed the treatment strategy and/or the radiotherapy (RT) planning. Results : Our analysis included 58 patients (median age: 67.5 years, range: 48.0–86.0) with the following risk factors: lymphadenectomy not performed (26 patients; 44.8%), inadequate lymphadenectomy (23 patients; 39.7%), and high risk of residual disease due to advanced stage (nine patients; 15.5%). Postoperative 18 F-FDG-PET/CT imaging was positive in 18 patients (31%) in the following sites: pelvic extra-nodal disease (one patient), pelvic and/or paraaortic lymph nodes (12 patients), distant metastases (one patient), or combination of previous sites (four patients). Based on these results, the adjuvant therapeutic strategy was changed in five patients, three of whom were referred to chemotherapy alone due to distant metastases and two of whom were referred to nodal-directed treatment due to lymph node metastases (lymphadenectomy and pelvic chemoradiation plus boost, respectively). Furthermore, based on the 18 F-FDG-PET/CT results, the RT plan was modified in 13 patients (addition of a boost on residual pelvic/abdominal disease in 12 and target modification in one, respectively). Therefore, based on postoperative 18 F-FDG-PET/CT findings, the therapeutic strategy and the RT plan were changed in 5 patients (8.6%) and 13 patients (22.4%), respectively. Conclusion : In this analysis, the adjuvant treatment was modified after post-operative 18 F-FDG-PET/CT in about one third of patients. Further studies are needed to better define the risk factors (or their combinations) correlated with higher probability of residual disease after radical hysterectomy-adnexectomy for EC.


Introduction
Endometrial cancer (EC) is the fourth most frequent tumor in the female European population and its incidence is increasing in developed countries [1]. In 2018, as reported in the International Agency for Research on Cancer (IARC) database [2], the crude incidence rate was 31.6/100,000/year. Most cases are diagnosed at an early stage [3] with 90-95% five-year survival rate. However, these rates decrease to 57% in patients with pelvic lymph-nodes (LNs) metastases and to 49% in case of abdominal LNs metastases [4].
When feasible, surgery is the mainstay of EC treatment [5] and the standard option is total hysterectomy with bilateral salpingo-oophorectomy [6]. The main aims of surgery are complete tumor resection and knowledge of the prognostic factors needed to guide adjuvant treatments. In fact, the EC staging is based on a surgical-pathologicalbased system, according to the International Federation of Gynecology and Obstetrics (FIGO) [7]. In particular, pelvic lymphadenectomy provides important prognostic information and can guide the choice of postoperative treatments. While sentinel-lymph-node mapping may be used in uterine-confined disease especially in centers with specific expertise, pelvic and paraaortic lymphadenectomy is recommended in patients with higher risk of treatment failure (G3, non-endometrioid histological type, stage ≥IB) [6,8].
In case of high risk of residual disease due to high pathological stage or incomplete surgical staging, postoperative imaging is recommended by international guidelines [8,9]. The latter, in order to detect both local or metastatic residual disease, suggest both chest/abdominal/pelvic Computed Tomography (CT) (or abdominal/pelvic Magnetic Resonance Imaging (MRI)) and chest CT without contrast. 18 F-FDG-Positron Emission Tomography/Computed Tomography ( 18 F-FDG-PET/CT) is recommended only in selected cases, mainly to clarify ambiguous findings [8].
However, the multidisciplinary oncologicalgynecological team of our institution introduced an operative protocol including postoperative 18 F-FDG-PET/CT in EC patients with high-risk features.
Therefore, the aim of this retrospective analysis was to define the impact of 18-Fluorodeoxyglucose -PET/CT in the adjuvant management of high-risk EC patients.

Study design and inclusion criteria
This is an observational study approved by the local Ethical Committee (ESTHER study, code CE 973/2020/Oss/AOUBo). We included EC patients who underwent postoperative 18 F-FDG-PET/CT in our institution before any adjuvant treatment due to one or more of the following risk factors: grade 3, stage ≥IB, histological types other than adenocarcinoma, and inadequate surgery based on the tumor stage.

Endpoints
Endpoints of the study were: (i) percentage of patients with modified therapeutic strategy (i.e., omitted RT) after 18 F-FDG-PET/CT; (ii) percentage of patients with modification of adjuvant RT dose (in particular delivery of a boost) and/or RT target definition after 18 F-FDG-PET/CT; (iii) outcomes (local control (LC), disease-free survival [DFS] and OS) in patients with positive 18 F-FDG-PET/CT; (iv) comparison between these outcomes and those of patients with negative postoperative 18 F-FDG-PET/CT. In order to assess the rate of patients with modified treatment based on 18 F-FDG-PET/CT, we recorded the adjuvant therapy settled by the multidisciplinary group only on the basis of the pathologic assessment of the surgical specimen.

PET/CT
Whole body 18 F-FDG-PET/CT was performed using the same positioning systems used in the planning and delivery of adjuvant pelvic RT. Three MBq/kg of 18 F-FDG were intravenously injected and after an uptake time of 60 min, images were acquired on a 3D tomograph, for two min per bed position, after defining the isocenter slice. In order to obtain anatomical information and for the attenuation correction, a low-dose CT scan (120 kV, 80 mA) was also performed. 18 F-FDG-PET/CT images were reconstructed using an iterative 3D ordered subsets expectation maximization method with two iterations and 20 subsets, followed by smoothing with CT-based attenuation, scatter, and random coincidence event correction. Then, three skin tattoo marks were then made in the abdominal region to ensure set-up reproducibility and treatment accuracy.

External beam
Pelvic external beam RT (EBRT) was planned and delivered using intensity modulated RT (IMRT) technique. The dose was prescribed according to the International Commission on Radiation Units and Measurements (ICRU) 83 report [10]. In patients with negative 18 F-FDG-PET/CT, the total dose was 45 Gy (1.8 Gy/fraction) and the clinical target volume (CTV) was defined including surgical bed, vaginal cuff, and pelvic lymph-nodes (obturator, external iliac, internal iliac, and presacral). The planning target volume (PTV) was defined as the CTV plus an isotropic margin of 1 cm. Patients with a pelvic nodal or extra-nodal residual disease at 18 F-FDG-PET/CT were treated by adding a simultaneous integrated boost (SIB) (boost dose: 14-20 Gy based on the dose delivered to the surrounding organs at risk [OARs]). In case of residual vaginal disease deemed amenable to endovaginal BRT, a local boost after EBRT was delivered (26 Gy in 4 fractions). In patients with paraaortic nodal metastases on 18 F-FDG-PET/CT, the prophylactic CTV was extended to the upper border of L1 and metastatic LNs were treated with SIB (14 Gy). Concurrent and adjuvant chemotherapy (CHT) was considered in 18 F- FDG-PET/CT positive patients and prescribed taking into account any comorbidities. All patients with para-aortic metastatic nodes received both concomitant and adjuvant CHT.

Brachytherapy
After pelvic RT, most 18 F-FDG-PET/CT negative patients underwent BRT boost to the vaginal cuff, delivered with high dose rate (HDR). Vaginal cylinder or ovoids were placed after local anesthesia. The BRT treatment plan was calculated with 3D-technique on a CT scan, after OARs delineation. The dose was prescribed 5 mm from the external surface of the applicator at the mid-point of the activated dwell positions length [11]. The BRT dose, in 18 F-FDG-PET/CT-negative patients, ranged between 6 and 21 Gy in one-three fractions. Six Gy in one fraction or 10 Gy in two fractions were delivered after postoperative pelvic EBRT, while 21 Gy in three fractions were prescribed in postoperative exclusive BRT.

Chemotherapy
Concurrent cisplatin-based (40 mg/m 2 ) CHT was administered weekly by intravenous infusion. Adjuvant CHT was prescribed according to patients' characteristics and EC stage (loco-regional or metastatic) with Carboplatin and Taxanes being the most used drugs.

Statistical analysis
Demographic and clinical data were analyzed with descriptive statistics and presented as numbers and percentages. Continuous variables were presented in terms of medians and ranges. The Chi-square test and the Mann-Whitney U test were used to compare categorical and continuous variables, respectively. Survival curves were calculated with the Kaplan-Meier method and compared with the log rank test. A value of p < 0.05 was used to define statistical significance. Data were analyzed using SPSS for Windows (version 20.0; SPSS Inc., Chicago, IL, USA).

Results
Fifty-eight patients were included in the analysis (median follow-up: 41 months; range: 5-146) and 18 (31.0%) of them had a positive postoperative 18 F-FDG-PET/CT (median SUV-max: 13.7, range: 1.5-27). Patients referred to our institution for adjuvant treatment between 2009 and 2018, coming from different local centres. Table 1 reports patient characteristics with a comparison between positive versus negative 18 F-FDG-PET/CT cases. The only statistically significant difference between the two groups was the FIGO stage (p = 0.049), moreover we described the higher FIGO stage III percentage in 18 F-FDG-PET/CT positive patients (72.2% versus 37.5%).
The site of macroscopic residual disease and both planned and delivered treatment are shown in Table 2. Briefly, the postoperative 18 F-FDG-PET/CT was positive in the following sites: residual tumor on the vaginal cuff  (one patient), pelvic and/or para-aortic lymph nodes (12 patients), distant metastases (one patient), or combination of previous sites (four patients with nodal disease and pelvic residual tumor and/or distant metastases). Based on these results, the adjuvant therapeutic strategy was changed in five patients: three were referred to chemotherapy alone due to distant metastases and two were referred to nodaldirected treatment due to lymph node metastases (lymphadenectomy and pelvic chemoradiation plus boost, respectively). Furthermore, based on the 18 F-FDG-PET/CT results, the RT plan was modified in 13 patients (addition of a boost on residual pelvic/abdominal disease in 12 and target modification in one, respectively). Therefore, based on postoperative 18 F-FDG-PET/CT findings, the therapeutic strategy and the RT plan were changed in 5 patients (8.6%) and 13 patients (22.4%), respectively. However, the treatment established on the basis of 18 F-FDG-PET/CT was not performed in two patients, in one case due to postoperative complications and in the other for the same reason and due to early tumor progression ( Table 2).
The main possible high-risk features in PET-positive patients were shown in Table 3: eight patients did not performed lymphadenectomy, mainly for clinical comorbidi-ties; 11 patients had grade 3 or undifferentiated tumors; a non-endometrioid or mixed histological type was reported in seven cases.
Three-year OS was 55.6% and 57.1% in the whole group of 18 F-FDG-PET/CT-positive patients including subjects with hematogenous distant metastases and in the sub-group of 18 F-FDG-PET/CT-positive patients with residual disease only in the pelvis or in extra-pelvic lymph nodes, respectively. Furthermore, three-year DFS in the same patient populations was 22.9% and 27.8%, respectively. Compared to 18 F-FDG-PET/CT-positive patients, OS was significantly higher in 18 F-FDG-PET/CT-negative patients (p < 0.001) (Fig. 1). Even when compared with 18 F-FDG-PET/CT-positive patients but without hematogenous metastases, OS was significantly higher in 18 F-FDG-PET/CTnegative patients (p < 0.001) (Fig. 2). Among 18 F-FDG-PET/CT-positive patients, 11 had died at the last observation: nine due to tumor progression and two due to noncancer-related reasons.

Discussion
In our real-life analysis on patients with high-risk EC, evaluated with postoperative 18 F-FDG-PET/CT, the latter was positive in 31.0% of patients leading to a change of the treatment planned based on the pathological examination of the surgical specimen in all subjects.
We showed that, in patients with inadequate surgical staging, postoperative 18 F-FDG-PET/CT can frequently and drastically change the adjuvant therapy plan. Table 3 reports the main risk factors in our PET-positive patients: lack of lymphadenectomy, Grade 3, non-endometrioid histological type, and advanced FIGO stage (≥IIIC). Lymphadenectomy is recommended in these high-risk patients (Grade 3, FIGO stage ≥IB, non-endometrioid histology) and in cases without surgical nodal staging, the European guidelines suggest intensified adjuvant treatment strategies [6]. This recommendation is particularly important in patients with metastatic pelvic and/or abdominal lymph nodes, in whom the risk of residual cancer is higher. However, in our series many patients did not undergo lymphadenectomy, due to their advanced age and/or relevant comorbidities. Furthermore, due to its still controversial role, not even sentinel lymph node analysis was performed. Finally, particularly in patients with nodal residual disease at 18 F-FDG-PET/CT, an intensified local treatment (lymphadenectomy followed by chemotherapy ± chemoradiation) would probably have been justified, at least in selected patients.
In most cases, nodal uptake was detected on postoperative 18 F-FDG-PET/CT in patients with: (i) disease con-fined only to the uterus (based on initial imaging) but surgically inadequately staged, especially at the nodal level, or (ii) with metastatic lymph nodes on pathological examination and in which 18 F-FDG-PET/CT was performed to optimize target volumes and doses of postoperative radiotherapy. In fact, in some of these patients, preoperative imaging raised the suspicion of nodal involvement but clearly, even in patients undergoing lymphadenectomy, there is a risk of incomplete removal of metastatic lymph nodes. It is impossible to predict the outcome of positive 18 F-FDG-PET/CT patients if they had received the previously planned treatment. However, considering that in the case of macroscopic residual tumor the international guidelines recommend doses of 60-70 Gy [8], if these patients had received only a "prophylactic" dose (45-50 Gy) it is easily conceivable that adjuvant radiotherapy would have been ineffective. It should be noted that, despite an early detection of residual disease, the outcome of positive 18 F-FDG-PET/CT patients was very poor, being the 3-year DFS rate below 25.0%. Considering that most tumor progressed with distant metastases, it can be speculated that a more intensive use of systemic treatments could improve the prognosis in this setting. Indeed, excluding three patients with distant metastases, only four out of 15 patients with positive 18 F-FDG-PET/CT underwent CHT. Similarly, in patients with 18 F-FDG-PET/CT showing distant metastases, in all but one cases with ≤5 lesions, the question might be: "would a metastasis-directed therapy (such as stereotactic RT), applied based on current evidence [12], have improved the outcome?". Finally, particularly in patients with nodal residual disease at 18 F-FDG-PET/CT, an intensified local treatment (lymphadenectomy followed by chemoradiotherapy) would probably have been justified, at least in selected patients.
Our analysis has obvious limitations. First, the retrospective study design leads to unavoidable risks of bias. Moreover, the patients included in our analysis were referred for adjuvant treatment to our institution from very heterogeneous centres, where the preoperative staging protocols and the experience and surgical volumes in oncological gynaecology were strongly different. Furthermore, the small sample size limits the possibility of clearly identifying patients to be candidates for postoperative 18 F-FDG-PET/CT. Moreover, the rate of patients not undergoing lymphadenectomy can also be surprisingly high in our case series, limiting their generalizability in terms of frequency of positive PET/CT. However, it should be considered that our study was based on the selection of patients with a high risk of residual cancer after surgery, which at least partially justifies the high percentage of patients without surgical staging. Furthermore, 18 F-FDG-PET/CT positive patients were treated inhomogeneously, particularly in terms of systemic therapies. Moreover, the results of this analysis would likely have been different if international guidelines [8] on clinical management of EC had been followed, particularly regarding the recommendation to perform preoperative lymph node staging in high-risk patients. Therefore, this aspect also limits the generalizability of our results since they were recorded in a partially incorrect therapeutic setting. However, even in this case, we must stress that our analysis was aimed at assessing the impact of PET/CT in a real-world clinical situation not uncommon in our experience. Finally, the only postoperative imaging examination was 18 F-FDG-PET/CT, thus preventing the comparison between the results of the latter with those of other imaging techniques (CT or MRI) currently suggested by international guidelines in this setting [8].
However, our study is original being one of the very few evidence on the role of 18 F-FDG-PET/CT in detecting post-surgical residual EC. In fact, to the best of our knowledge, only Simcock et al. [13] analyzed the results of PET/CT in 48 patients with intermediate or high-risk EC after total hysterectomy and bilateral salpingo-oophorectomy ± lymphadenectomy. In their analysis, 18 F-FDG-PET/CT was positive in 35% of patients and led to a change in the planned treatment in 31% of cases. Therefore, the results of that series are surprisingly similar to those of our analysis and seem to confirm the usefulness of 18 F-FDG-PET/CT in selected patients with resected EC.
Instead, other studies tested PET/CT in patients with suspected EC recurrence during the post-surgical followup. OzcanKara et al. [14] reported higher sensitivity, specificity, and diagnostic accuracy of 18 F-FDG-PET/CT, compared to conventional imaging (CT, MRI, ultrasound) and CA-125 levels, in the evaluation of post-treatment EC with suspected recurrence. Moreover, Sharma et al. [15] reported higher specificity (96% versus 62%), accuracy (92% versus 76%), and comparable sensitivity (89% versus 85%) of 18 F-FDG-PET/CT, compared to CT and MRI, in EC with suspected recurrence. Finally, the systematic reviews by Kadkhodayan et al. [16] and Bollineni et al. [17], showed a high accuracy of 18 F-FDG-PET/CT in detecting EC recurrences and a clear impact on their treatment. More generally, Crivellaro et al. [18] reported that, in high-risk EC, 18 F-FDG PET/CT demonstrated moderate sensitivity but high specificity and accuracy for the nodal status assessment.
Taken together, the available evidence seems to suggest that in adequately early-stage radiologically and surgically staged patients, postoperative 18 F-FDG-PET/CT could be omitted, unless suspicious residual disease at surgery. Moreover, preoperative staging is recommended in all patients, with more attention if known high-risk features are present at diagnosis [8], in order to make an oncological adequate surgical staging, but none of the imaging modality can totally replace the surgical act [19]. Furthermore, the surgical staging and the possible postoperative imaging become even more important considering that nodal disease can not be excluded also if preoperative imaging was negative [20].
In conclusion, 18 F-FDG-PET/CT seems useful both in the post-surgical restaging of selected patients and in the whole body restaging of patients with proven or suspected local EC recurrence. Furthermore, the results of our analysis suggest that, in daily clinical practice, postoperative 18 F-FDG-PET/CT may be justified at least in some categories of patients with high risk of undertreatment. In particular, postoperative 18 F-FDG-PET/CT seems particularly useful in patients with both incomplete surgical nodal staging and advanced FIGO stage (≥FIGO III), or both incomplete surgical nodal staging and grading 3 EC. In fact, in our analysis, the rates of 18 F-FDG-PET/CT positive patients within these two subgroups were 47.4% and 41.0%, respectively.

Conclusions
The results of our analysis warrant further studies in this setting. Such analyses could have the purpose of: (i) a more precise definition of the patient populations to be referred to postoperative 18 F-FDG-PET/CT; this objective could be pursued through the collection of clinical, pathological, bio-molecular and radiomics data from large patient series in order to develop predictive models of postsurgical residual disease; (ii) to test more intensive and possibly more personalized integrated treatments in order to improve the outcome of these patients with poor progno-sis; this intensification could involve the administration of systemic therapies in patients with increased risk of distant metastasis and the use of dose-escalated RT in patients with increased risk of local failure (dose-escalated SIB, boost delivered with stereotactic RT).

Ethics approval and consent to participate
This is an observational study approved by the local Ethical Committee (AVEC -Comitato Etico Area Vasta Emilia Centro): ESTHER study, code CE 973/2020/Oss/AOUBo. All subjects gave their informed consent for inclusion in this study.