Clinical factor‐based risk stratification for precision therapy in locally advanced squamous cell carcinoma of the uterine cervix

Abstract Background Concurrent chemoradiotherapy (CCRT) is the standard of care for locally advanced cervical cancer. In this study, we analyzed the pretreatment clinical and 18F‐fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT) characteristics of patients with locally advanced cervical squamous cell carcinoma (SCC) to develop a scoring prototype for risk stratification. Methods Two cohorts were constructed in this study. Cohort 1 comprised patients with cervical SCC with 2009 FIGO stage III‐IVA or stage I–II with positive pelvic or para‐aortic lymph node (PALN) on PET/CT from AGOG09‐001 trial. Cohort 2 comprised patients with similar characteristics who had received adequate therapy in our hospital between 2016 and 2021. Pretreatment patient characteristics and PET/CT parameters including maximum standardized uptake value (SUVmax) and metabolic tumor volume (MTV) of primary tumor and nodal SUVmax were assessed for cancer‐specific survival (CSS) using multivariate Cox regression. Results Analysis of combined data from cohorts 1 (n = 55) and 2 (n = 128) indicated age ≥ 66 years, primary tumor MTV ≥87 mL, and positive PALN on PET/CT to be independently significant adverse predictors for CSS (p < 0.001, p = 0.014, and p = 0.026, respectively) with a median follow‐up duration of 51 months. Assigning a score of 1 to each adverse predictor, patients with cumulative risk scores of 0, 1, 2, and 3 were discovered to have a 5‐year CSS of 86.9%, 71.0%, 32.2%, and 0%, respectively (p < 0.001). Conclusion Age, primary tumor MTV, and positive PALN on PET/CT may serve as independent predictors of poor survival in patients with locally advanced cervical SCC. Our findings indicate that patients without any adverse factors can receive standard CCRT, whereas those with at least one adverse factor can consider novel combination therapies or clinical trials.


| INTRODUCTION
Cervical cancer is the fourth most common cancer in women worldwide. 1Since the publication of benchmark clinical trials and the issuance of the National Cancer Institute's clinical alert in 1999, concurrent chemoradiotherapy (CCRT) has become the standard of care for patients with locally advanced cervical cancer (LACC) without distant metastasis. 2 A phase III randomized clinical trial (RCT) conducted by Dueñas-González et al. indicated that survival could be improved in the treatment of International Federation of Gynecology and Obstetrics (FIGO) stage IIB-IVA cervical cancer by supplementing CCRT with weekly gemcitabine and two cycles of post-CCRT adjuvant chemotherapy.However, this regimen has not been accepted as a standard of care. 35][6] In the Asian Gynecologic Oncology Group (AGOG) 09-001 trial, gemcitabine was added to cisplatin during CCRT in comparison to single-agent cisplatin in patients with 2009 FIGO stage III-IVA or any stage with pelvic lymph node (PLN) or para-aortic lymph node (PALN) metastasis.However, this regimen failed to improve the 3-year progressionfree survival (PFS) (71.0% vs. 65.1%,p = 0.71) and overall survival (OS) (85.9% vs. 74.1%,p = 0.89). 4The phase III OUTBACK trial added 4 courses of paclitaxel and carboplatin after CCRT in a similar group of patients, but it also failed to improve PFS and OS. 5 In 2014, the US Food and Drug Administration (FDA) approved the use of bevacizumab, an angiogenesis inhibitor, for the treatment of recurrent/persistent and advanced cervical cancer.The Gynecologic Oncology Group 240 RCT revealed that the addition of bevacizumab improved OS in persistent/recurrent or metastatic cervical cancer with a hazard ratio (HR) of 0.77. 7Although the role of bevacizumab in LACC remains uncertain, its administration at a dose of 10 mg/kg every 2 weeks for 3 cycles in addition to weekly administration of cisplatin at a dose of 40 mg/m 2 during CCRT led to a 3-year PFS of 81.3% and OS of 68.7% in the phase II Radiation Therapy Oncology Group 0417 trial. 8Another angiogenesis inhibitor, Endostar, which works through the inhibition of VEGFR2, has been demonstrated to improve the efficacy of CCRT in LACC with VEGFR2 expression. 9embrolizumab, in combination with chemotherapy, has been approved by the US FDA for the treatment of patients with persistent/recurrent or metastatic cervical cancer with PD-L1 expression due to the results of phase III KEYNOTE-826 trial. 10This RCT demonstrated that the addition of pembrolizumab improved OS with a HR of 0.64.Thus, the addition of immune checkpoint inhibitors to the therapeutic regimen of patients with LACC is likely to be beneficial.However, the phase III CALLA trial, which used durvalumab in combination with and following CCRT, failed to improve survival. 6Another phase III RCT (ENGOT-cx11/KEYNOTE-A18) for evaluating the benefits of using pembrolizumab with CCRT in the treatment of patients with LACC is currently ongoing. 11With the advancement of novel therapies, reliable strategies are required for risk stratification in patients with LACC.Although the cancer prognosis may be favorable and standard CCRT may be adequate for some patients, others may have an unfavorable prognosis; these patients should consider novel combination therapies.
Molecular imaging utilizing 18 F-fluorodeoxyglucose ( 18 F-FDG) and positron emission tomography (PET), which is usually integrated with computed tomography (CT), is widely performed for examining patients with cervical cancer. 12In addition to facilitating tumor staging, pretreatment 18 F-FDG PET/CT offers functional insights into tumors; for example, it provides information and p = 0.026, respectively) with a median follow-up duration of 51 months.

Conclusion:
Age, primary tumor MTV, and positive PALN on PET/CT may serve as independent predictors of poor survival in patients with locally advanced cervical SCC.Our findings indicate that patients without any adverse factors can receive standard CCRT, whereas those with at least one adverse factor can consider novel combination therapies or clinical trials.

K E Y W O R D S
18 F-FDG PET/CT, concurrent chemoradiotherapy, locally advanced cervical cancer, risk stratification, squamous cell carcinoma regarding the maximum standardized uptake value (SUV max ) and volumetric measures such as metabolic tumor volume (MTV).The prognostic value of SUV max in cervical cancer has been previously assessed in a meta-analysis. 13n the present study, we developed a scoring prototype for risk stratification in patients with locally advanced primary cervical squamous cell carcinoma (SCC).This was achieved by analyzing the clinical and PET characteristics of the patients.

| Study cohorts
In the prospective, randomized AGOG09-001 trial, patients with locally advanced primary cervical SCC were enrolled between 2009 and 2012 to participate in an investigation of whether the addition of gemcitabine to CCRT can improve survival. 4A parallel imaging study involving 55 patients demonstrated that 18 F-FDG PET/ CT performed before and during treatment can help predict treatment failure and OS. 14 However, duringtreatment parameters are unavailable for clinical cases in general.Only pretreatment predictors are available for risk stratification for patients with LACC.For the current study, we updated the survival data in the AGOG09-001 study (cohort 1) and added patients with similar clinical status who underwent pretreatment PET/CT and received adequate therapy in our institution between 2016 and 2021 (cohort 2) for evaluation.
Both cohorts comprised of cervical SCC patients with either 2009 FIGO stage III-IVA or stage I-II with positive PLN or PALN on PET/CT. 15In cohort 1, the following patients were excluded: patients aged <35 or > 70 years; those with inadequate bone marrow or impaired pulmonary, liver, or renal function; those with an Eastern Collaborative Oncology Group performance status of >1; and those with a history of chemotherapy or pelvic radiotherapy.In cohort 2, patients who did not receive adequate therapy were excluded since inadequate therapy might exert confounding effects on survival.This study was approved by the Chang Gung Medical Foundation Institutional Review Board (202200908B0).The requirement for written informed consent was waived because of the retrospective nature of this study.

| 18 F-FDG PET/CT imaging
Two PET/CT machines (Discovery ST16, GE Healthcare; Biograph mCT, Siemens Healthcare) were used.The patients fasted for 6 h before undergoing examination.Imaging was performed 50 min after an intravenous injection of 18 F-FDG (370 MBq ± 10%).A nonenhanced CT scan from the head to the thighs was performed; this was followed by PET imaging.All images were acquired with the patient in the supine position.On the basis of the CT data, the PET data were corrected for attenuation by using an ordered subsets expectation maximization algorithm.

| Image interpretation and parametric quantification
PET/CT images were analyzed using parameters quantified on a Syngo MI Workplace software platform (Siemens Healthcare).The presence of positive PLN and/ or PALN was investigated.The PET voxel standardized uptake value (SUV) was calculated by normalizing the image-derived radioactivity concentration to the wholebody concentration (i.e., injected dose divided by patient body weight).SUV max was defined as the maximum SUV in the volume of interest (VOI).MTV was defined as the volume of estimated tumor voxels with increased radioactivity uptake.The VOIs for the primary tumor and positive lymph nodes were manually drawn on the images displayed on the workstation; SUV max was recorded for each VOI.The MTV of the cervical tumor was measured using a fixed boundary SUV of 3.0.Nodal SUV max was defined as the maximum SUV max of all positive PLNs and PALNs.

| Systemic CCRT regimens
CCRT was administered as described previously. 4In brief, external-beam radiotherapy was delivered to the whole pelvis with a radiation dose of 45 Gy.Irradiation fields were extended to the abdominal para-aortic region if necessary.Patients may undergo brachytherapy with six fractions of 4.3 Gy delivered to point A or intensitymodulated radiotherapy.The primary tumor and nodal lesions were treated up to a total dose of 72 Gy and 57.6 Gy, respectively.Latest techniques, such as image-guided brachytherapy, were not used.In cohort 1, systemic therapy consisted of weekly cisplatin (40 mg/m 2 ) either with or without (1:1 randomization) gemcitabine (125 mg/m 2 ) administered during the course of radiotherapy lasting up to 6 cycles.In cohort 2, systemic therapy consisted of single-agent cisplatin or platinum-based doublet/triplet combinations.The "single agent" protocol involved weekly administration of cisplatin (40 mg/m 2 ) during radiotherapy.Diverse protocols were used for the "combination regimen."In general, cisplatin or carboplatin was used as the backbone.The drug was selected mainly on the basis of the attending physician's preference and the patient's clinical condition.The concurrent therapy began during the first week of radiation.The single-agent cisplatin was administered every week for 6 weeks.The combination regimens were generally administered for up to six cycles.In the absence of disease progression, durvalumab, pembrolizumab, and bevacizumab were administered after CCRT.

| Follow-up and survival analyses
Follow-up examinations were performed every month for 3 months after the completion of CCRT.Further follow-up examinations were arranged every 3 months for the first 2 years, every 4 months for the third year, and every 6 months thereafter.In addition to physical examinations, the levels of serum tumor markers, including SCC antigen and carcinoembryonic antigen, were routinely measured.Imaging studies were arranged every 3-6 months for 2 years or whenever relapse was suspected.Biopsy or further imaging correlation was arranged for patients exhibiting positive or suspicious results.Cancer-specific survival (CSS) was calculated from the date of pretreatment PET/CT to that of death from cervical cancer or censored to the date of the most recent follow-up or death unrelated to cervical cancer.PFS was calculated from the date of pretreatment PET/ CT to that of an examination indicating cancer progression/recurrence or censored to the same date as that for CSS.

| Development of a scoring prototype for risk stratification
To determine CSS and PFS, we analyzed available data regarding clinical and imaging predictors, including age, 2009 FIGO stage (I-II vs. III-IVA), tumor grade, PLN status, PALN status, primary tumor SUV max , primary tumor MTV, and nodal SUV max .The effect of systemic therapy protocol (single agent vs. combination regimen) was also evaluated.Independent factors that were found to be significant in multivariate analysis performed to identify the predictors of CSS were used for developing a scoring prototype for risk stratification, with each independent predictor assigned a risk score of 0 or 1.This scoring prototype was used to classify the patients according to their cumulative risk scores.Next, survival curves were generated for patients with different cumulative scores.

| Comparison between patients receiving systemic therapy with single agent or combination regimen
Because the number of patients receiving advanced combination regimens in cohort 2 was small, we performed propensity score matching (PSM) to compare the survival between the patients who received CCRT with standard single-agent chemotherapy and those who received CCRT with combination regimens.A subset of patients receiving CCRT with single-agent chemotherapy was constructed by selecting patients exhibiting characteristics similar to those of patients receiving combination regimens.

| Statistical analysis
Continuous data were analyzed using the Student's t test or Mann-Whitney U test, whereas categorical data were analyzed using the Fisher's exact test or chi-squared test.Due to the presence of both continuous and categorical predictors, univariate Cox regression analysis was first conducted to identify the significant predictors of CSS.Kaplan-Meier survival analysis and the log-rank test were used for assessing 5-year CSS and PFS, with a treebased method used for determining the optimal cutoff values for significant continuous variables for CSS. 16ultivariate and stepwise Cox regression analysis was then performed for identifying the independent predictors of survival.To evaluate the systemic therapy impact, patients receiving single-agent cisplatin were matched (2:1) with those receiving combination therapy through multivariate logistic regression. 17All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc.) and R (version 3.6.1).

| Characteristics of the study cohorts
Cohort 1 comprised 55 patients, whereas 144 patients were retrospectively enlisted in cohort 2. From cohort 2, 5, and 11, patients were excluded because of treatment discontinuation and inadequate therapy, respectively.Figure 1 presents a flow diagram of depicting the selection of patients for the study cohorts and PSM subset.

| Findings related to follow-up and survival
The median follow-up duration for the combined cohort was 51 (range, 2-164) months.A total of 49 patients died, of whom 2 died for reasons unrelated to cervical cancer.Cancer progression or recurrence occurred in 58 patients.

| Determination of significant predictors and cutoff values
The results of the univariate Cox regression analysis are presented in Table 2. Age, FIGO stage, PALN status, and primary tumor MTV were significant predictors of CSS and selected for further analyses.The 5-year CSS and PFS rates assessed by Kaplan-Meier analysis and the optimal cutoff values for continuous variables are presented in Table 3.All the selected predictors remain significant, and the optimal cutoff values for age and primary tumor MTV are 66 years and 87 mL, respectively.The statistical results of systemic therapy protocol for survival were also appended for reference.

| Results of univariate and multivariate Cox regression analyses
The results of the Cox regression analysis performed to identify the predictors of CSS are presented in a p values for the difference between cohorts 1 and 2; between-cohort comparison was performed using the chi-square or Fisher's exact test.
T A B L E 1 Clinical characteristics of the study cohorts.
III-IVA, positive PALN, and primary tumor MTV ≥87 mL were revealed to be significant predictors.In the multivariate analysis, FIGO stage was found to be a nonsignificant predictor.Stepwise Cox regression revealed that the most significant predictor was age (p < 0.001), followed by primary tumor MTV (p = 0.014) and PALN status (p = 0.027).The results of Cox regression analyses for PFS are presented in Table 5.The most significant predictor was age (p = 0.001), followed by primary tumor MTV (p = 0.004) and PALN status (p = 0.004).

| Scoring prototype for prognostic stratification
On the basis of the aforementioned results, the following factors were selected for use in the development of the scoring prototype for risk stratification: age ≥ 66 years, primary tumor MTV ≥87 mL, and positive PALN; each factor was assigned a score of 1. Cumulative risk scores of 0, 1, 2, and 3 corresponded to 5-year PFS and CSS rates of 86.2% and 86.9%, 67.1% and 71.0%, 33.4% and 32.2%, and 0% and 0%, respectively.The survival curves for patients with different cumulative risk scores varied significantly (p < 0.001; Figure 2).Two representative cases are presented in Figure 3.

| Comparison between patients receiving systemic therapy with single agent and combination regimen
In cohort 1, 31 patients received cisplatin with gemcitabine, and 24 patients received cisplatin only.In cohort 2, 117 patients received weekly cisplatin or carboplatin, and 11 patients received combination therapy.The combination regimens were as follows: cisplatin-topotecan, carboplatintopotecan, cisplatin-topotecan-pembrolizumab, cisplatinbevacizumab, cisplatin-nivolumab, cisplatin-durvalumab, cisplatin-pembrolizumab, and cisplatin-bevacizumabpembrolizumab.The baseline characteristics of the patients receiving combination therapy are presented in Table S1.The patients receiving combination therapy in cohort 2 were significantly older and had a higher rate of positive PALN than those in cohort 1.Two PET parameters, primary tumor SUV max and nodal SUV max , were also significantly different.Two matching variables, FIGO stage and PALN status, were employed to reduce the effects of confounding during PSM subset selection.The PSM cohort comprised 126 patients; their baseline characteristics are summarized in Table S2.No significant differences were noted between patients receiving CCRT with single-agent chemotherapy and combination regimens except for the primary tumor SUV max .The corresponding survival curves are presented in Figure 4. Survival rates did not vary significantly different between these two PSM groups (CSS, p = 0.299; PFS, p = 0.274); nevertheless, the combination regimens tended to confer mild survival benefits.The comparison between CCRT with single agent versus combination regimen in the combined cohort revealed no significant difference (Figure S3).In the AGOG09-001 trial, no significant difference was noted in PFS and CSS between the treatment arms, even after extended follow-up (Figure S4).For patients with different cumulative risk scores in the combined cohort, the comparison between CCRT with single agent and combination regimen is presented in Figures S5-S8, respectively.In the patients with cumulative risk scores 2, the benefit of combination regimen approached borderline significance.In the meta-analysis of LACC without distant metastasis, CCRT compared with radiotherapy only was found to confer a 6% benefit in terms of OS. 2 Post-CCRT adjuvant chemotherapy with carboplatin and paclitaxel was investigated in the OUTBACK trial, but no obvious survival benefit was discovered. 5In the CALLA trial, durvalumab or placebo was added to cisplatin during CCRT and post-CCRT maintenance therapies.At a median follow-up of 18.5 months, durvalumab group did not show a statistically significant improvement in PFS (HR 0.84, p = 0.174). 6Use of platinum doublet regimens as neoadjuvant chemotherapy before CCRT has also been attempted with controversial results. 18,19n the randomized INTERLACE trial, weekly carboplatin-paclitaxel was employed in the neoadjuvant setting with pending results (Clini calTr ials.gov identifier: NCT01566240). 20More intricate prognostic predictors are required for the effective identification of patients for whom standard CCRT may be inadequate.
To identify such predictors, we conducted the present study.Our previous trials and the present study included only patients with SCC. 4,14This is because patients with adenocarcinoma or adenosquamous carcinoma generally have significantly worse outcomes than do patients with SCC, regardless of whether they undergo primary radical surgery or primary radiotherapy. 21,22 study was conducted using 8538 patients aged ≥50 years who were selected from the Surveillance, Epidemiology, and End Results database.The results of the study reveal that age at diagnosis significantly affected disease prognosis. 23Compared with the risk of adverse outcomes in patients aged 50-60 years, the HR for patients aged 60-70 years was 1.229, and that for those aged >70 years was 1.625.Older patients often have degenerative or chronic diseases; these patients may exhibit poor tolerance to surgery and experience serious side effects following therapy.Complications in older patients are more difficult to manage and thus result in poor prognosis and higher mortality rates.The Kaplan-Meier analysis performed in our study revealed the age of ≥66 years to be the optimal cutoff age for predicting adverse CSS; the corresponding HR derived from the stepwise Cox regression analysis was 3.262.However, a study reported that patients aged ≤30 years have poor prognosis. 24In the present study, patients aged <35 years had been excluded in cohort 1 and there were only three patients aged <35 years in cohort 2; two patients aged <30 years remained disease-free with a follow-up duration of 49 and 67 months, respectively, and another patient aged 32 remained disease-free with a follow-up duration of 29 month.
Metabolic tumor volume estimated by PET imaging, represents a 3-dimensional value and can be easily measured.Chung et al. 25 identified MTV to be a significant prognostic factor for survival in 63 patients with a FIGO stage ranging from IB to IIA before radical hysterectomy.We previously reported a significant association between pretreatment primary tumor MTV and local failure and poor OS. 14 In a retrospective analysis of 508 patients with radically treated cervical cancer, primary tumor MTV, but not primary tumor SUV max , was discovered to be a significant factor influencing survival. 26This finding corroborates that of our study.
Para-aortic lymph node metastasis is a strong predictor of poor prognosis.The incorporation of nodal involvement in the FIGO 2018 staging definition for cervical cancer highlighted the need for assessing PALN status. 27In a study involving 618 patients with locally advanced cervical SCC, PALN metastasis was found to be an independent prognostic factor for PFS and OS. 28In a study involving patients with PALN metastasis, SUV max of the involved PALN metastasis and nodal volume or total lesion glycolysis were used to stratify patients in terms of survival. 29n our study, the rates of 5-year PFS and CSS in the combined cohort were 66% and 69%, respectively.Age (≥ 66 years), MTV (≥ 87 mL), and PET/CT-defined PALN metastasis were found to be independent predictors of poor survival outcomes.On the basis of these risk predictors, the patients were assigned a cumulative risk score of 0-3.We recommend that patients with a cumulative risk score of 0 receive CCRT with single-agent cisplatin and that for others, novel combination therapies or clinical trials be considered.We performed PSM to compare the survival between patients receiving CCRT with a single agent and those receiving combination regimen.In cohort 1, patients were randomly selected for combination therapy.By contrast, in cohort 2, only 11 patients received combination therapy.Because of the limited sample size of the PSM cohort, the survival difference between the two treatment groups did not achieve statistical significance.However, a mild trend favoring the combination regimen was noted.
The strengths of the present study are as follows: we included a sufficient number of patients who received adequate treatment for LACC and underwent pretreatment 18 F-FDG PET/CT for analysis, and we unveiled the independently significant predictors of poor survival outcomes in this patient population.Our study has some limitations.First, both perspective and retrospective data were analyzed.Second, data were obtained from a single center and from patients belonging to a single ethnicity, which limits the generalizability of our findings.Third, we lacked information regarding biological markers.Finally, the number of patients receiving combination regimens was small in cohort 2, and the regiments were heterogeneous.Nevertheless, our scoring prototype for risk stratification may guide clinicians in selecting an effective therapeutic regimen for patients with locally advanced cervical SCC.

| CONCLUSIONS
A novel scoring protocol has been proposed for patients with advanced cervical SCC with FIGO stage III/IVA or nodal metastasis.Age (≥66 years), primary tumor MTV (≥87 mL), and PET/CT-defined PALN metastasis may serve as independent predictors of poor survival outcomes.Patients with a cumulative risk score of 0 may receive single-agent CCRT with satisfactory outcomes; these patients may be excluded from trials involving the administration of novel combination therapies.However, for patients with a cumulative risk score of ≥1, more intensive treatment or clinical trials may be considered.Validation of the present study in a large multicenter trial is warranted.

F I G U R E 1
Flow diagram depicting the selection of patients for the study cohorts and PSM subset.

F I G U R E 2
Survival curves for patients with different cumulative risk scores.(A) Cancer-specific survival.(B) Progression-free survival.F I G U R E 3 Representative cases of different survival rates in patients with different cumulative risk scores.(A) This is the case of a 79-year-old patient with FIGO 2009 stage IIIB cervical cancer.Pretreatment PET/CT revealed a bulky primary tumor (SUV max = 14.8;MTV = 94 mL) with small positive pelvic and para-aortic nodes (arrows).This patient had a cumulative risk score of 3 and received CCRT with cisplatin and bevacizumab.However, PET/CT imaging performed 3 months after the completion of CCRT revealed supraclavicular nodal metastases; nevertheless, cervical tumor and pelvic/para-aortic lymph nodes indicated prominent regression.The patient died after another 3 months.(B) This is the case of a 57-year-old patient with FIGO stage IIIB cervical cancer.Pretreatment PET/CT revealed a cervical tumor with parametrial extension (SUV max = 13.9;MTV = 85 mL) and a positive left external iliac lymph node (arrow head).This patient had a cumulative risk score of 0 and received standard CCRT with weekly cisplatin.The patient remained disease-free for 3 years as of the latest follow-up.

F I G U R E 4
Survival curves for propensity score-matched patients receiving single-agent concurrent chemoradiotherapy or combination therapy.(A) Cancer-free survival.(B) Progression-free survival.| 11 of 12 LIU et al.

Table 1
presents the patients' baseline clinical characteristics, including their age, 2009 FIGO stage, tumor grade, PLN status, PALN status, systemic therapy for CCRT, and PET parameters.Cohorts 1 and 2 varied significantly in terms of age distributions, PLN status, systemic therapy for CCRT, and primary tumor SUV max .

Table 4 .
In the univariable analysis, age ≥ 66 years, FIGO stage Results of the univariate Cox regression analysis for CSS and PFS in the combined cohort.Results of the Kaplan-Meier analysis for CSS and PFS in the combined cohort.
T A B L E 2 T A B L E 3Abbreviations: C, combination regimen; CSS, cancer-specific survival; FIGO, International Federation of Gynecology and Obstetrics; MTV, metabolic tumor volume; PALN, para-aortic lymph node; PFS, progression-free survival; S, single agent.
Results of the Cox regression analysis for CSS in the combined cohort.Results of the Cox regression analysis for PFS in the combined cohort.
T A B L E 4Abbreviations: CSS, cancer-specific survival; FIGO, International Federation of Gynecology and Obstetrics; MTV, metabolic tumor volume; PALN, para-aortic lymph node.T A B L E 5