Radiotherapy for postoperative vaginal recurrences of cervical squamous cell carcinoma: analysis of dosing and prognosis

Abstract Squamous cell carcinoma (SCC) is the most common type of vaginal recurrence in cervical cancer patients, and the role of salvage radiotherapy on these patients remains unclear. This study aimed to investigate the efficacy of salvage radiotherapy for vaginal recurrence of SCC in patients who previously underwent surgery and to explore prognostic factors associated with survival. Ninety-seven patients with histologically proven SCC who were treated for vaginal recurrence at Peking Union Medical College Hospital were identified. All patients had previously undergone surgery and received salvage radiotherapy. Factors predictive of overall survival (OS), progression-free survival (PFS), and local control (LC) were investigated. The median follow-up time was 42.5 months. The estimated 5-year OS, PFS, and LC rates were 84%, 79%, and 91%. On multivariate analysis, inguinal lymph node metastasis was significantly associated with poor OS; a tumour size ≤4 cm was associated with longer PFS (p < 0.05); the recurrence pattern was an independent predictor of LC (p < 0.05). In the 45 patients with recurrences that were paravaginal or invasive of surrounding organs, biologically equivalent doses in 2 Gy fractions of ≥72.6 Gy were independently predictive of longer LC (p < 0.05). RT is an effective treatment for postoperative vaginal recurrence in patients with cervical SCC. For patients with extravaginal recurrence, a salvage dose of ≥72.6 Gy appears to be optimal. Impact statement What is already known on this subject? Radiotherapy plays a critical role in treating recurrent cervical cancer, but the effectiveness of RT for vaginal recurrence in patients who previously underwent surgery remains limited. Few studies have focussed on the effect of RT dose on patient survival. What do the results of this study add? This study investigated the efficacy of RT in patients with cervical squamous cell carcinoma who experienced postoperative recurrence. Lymph node metastasis, tumour size and recurrence pattern were significantly associated with survival. Moreover, an EQD2 ≥ 72.6 Gy was independently predictive of longer LC. What are the implications of these findings for clinical practice and/or further research? RT is an effective treatment for postoperative vaginal recurrence in patients with cervical squamous cell carcinoma. For patients with extravaginal recurrence, a salvage dose of ≥72.6 Gy appears to be optimal.


Introduction
Cervical cancer is the fourth most common cancer among women (Bray et al. 2018) and the second most commonly diagnosed malignancy in developing countries (D.Li et al. 2020).Squamous cell carcinoma (SCC) is the most common pathological type of cervical cancer, accounting for approximately 80-90% of cases (Small et al. 2017); its prognosis differs from those of other pathological types (Davy et al. 2003, Galic et al. 2012, Lee et al. 2011, Nakanishi et al. 2000).
Radical hysterectomy with pelvic lymphadenectomy is the standard recommendation for patients with early-stage cervical cancer (Ramirez et al. 2018); however, the pelvic recurrence rate is 10-20% in patients with this disease after primary treatment, and the disease-free survival rate remains poor at 45% in patients with recurrent locoregional cervical cancer (Thomas et al. 1993).The vagina is the most common site of cervical cancer recurrence (Mahdi et al. 2015).Treatment for recurrent cervical cancer that is confined to the upper vagina can be curative (Cohen et al. 2019) yet remains challenging given the lack of a consistent standard treatment after the primary intervention.Pelvic exenteration is one of the main surgical methods for local recurrence; however, the perioperative mortality rate is relatively high, and the procedure may also reduce the patients' quality of life (Benn et al. 2011).Salvage radiotherapy (RT) with or without concurrent chemotherapy is currently recommended owing to its effectiveness and tolerable adverse events (Abu-Rustum et al. 2020, S. W. Kim et al. 2017, Kobayashi et al. 2019).External beam RT (EBRT) with or without chemotherapy and/or brachytherapy (BT) is conventionally prescribed to patients with cervical cancer who experience local recurrence, especially those without a history of RT.
Although emerging data suggest that RT can play a critical role in treating recurrent cervical cancer, studies on the effectiveness of RT for vaginal recurrence in patients who previously underwent surgery, as well as on survival, remain limited (Haasbeek et al. 2008, Jain et al. 2007).Moreover, few studies have focussed on the effect of RT dose on patient survival; whether significantly higher doses would provide more effective treatment would guide future formulations of RT regimens.
In the present study, we retrospectively analysed the data of patients with cervical cancer who experienced vaginal recurrence after hysterectomy and were treated with salvage RT at the Peking Union Medical College Hospital (PUMCH).The study's first aim was to investigate the efficacy of salvage RT in patients with cervical SCC who experienced vaginal recurrence; the second was to identify prognostic factors, including those related to target volume and RT dose.

Patient selection
Medical records of 177 women with cervical cancer who experienced vaginal recurrence and were treated with salvage RT at Peking Union Medical College Hospital (PUMCH) between July 2011 and November 2019 were reviewed.The following inclusion criteria were applied: (1) both initial treatment and recurrence were histologically proven SCC-subtype cervical cancer, (2) the patient had undergone hysterectomy for cervical cancer, and (3) the patient had experienced vaginal recurrence and was consequently treated with salvage RT.We excluded patients who (1) underwent hysterectomy for benign disease or did not complete surgery, (2) had a history of RT, (3) had undergone other treatments for vaginal recurrence before or after salvage RT such as tumorectomy or drugs, and (4) had experienced recurrence in the form of metastasis to distant organs.Ultimately, 97 patients met the inclusion criteria and were included in our study.The cases were restaged according to the 2018 International Federation of Gynaecology and Obstetrics (FIGO) staging system, and there were 49, 38 and 10 patients with stage IA, IB and IIA cervical cancer at primary diagnosis, respectively.Median disease-free interval from primary surgery was 26.1 months (range, 4.6 months to 250.0 months).

Treatment
All patients with recurrence received salvage RT as EBRT, BT, or both at PUMCH.EBRT and BT combined were delivered to 76 of the 97 participants.Fifteen patients received only EBRT, while BT alone was performed in 6 patients.Cisplatin-based concurrent chemoradiotherapy was administered to 54 patients.
All patients who received pelvic EBRT were treated with volumetric modulated arc therapy or helical tomotherapy (n ¼ 91).The Clinical Target Volume (CTV) included all gross tumour, the cervix, paracervical tissue, upper vagina, presacral region and pelvis lymphatic drainage area.The inguinal lymph node drainage area was included when the lower 1/3 of vagina or inguinal lymph node was involved.The retroperitoneal lymphatic drainage area was included when common iliac and/or retroperitoneal lymph node metastasis occurred.The CTV was treated to a median dose 50.4 Gy (range 45.0-50.4Gy) given as 1.8 Gy-fractions, while the median gross tumour volume (GTV) was 50.4 Gy (range, 45.0-80.2Gy) boosted by EBRT.The median primary GTV cumulative dose in EQD2 was 67.7 Gy (range, 45 Gy-85.2Gy) in patients who received only EBRT, while median EQD2 was 70.0 Gy (range, 37.5 Gy-88.9Gy) in EBRT and BT combined.Twenty-three patients received simultaneously lymph node boosts with doses ranging from 54.0 to 70.0 Gy (median 60.2 Gy), with a median EQD2 of 60.95 Gy.
When calculating the total dose of EBRT and BT, the biologically equivalent dose in 2 Gy fractions (EQD2) was utilised; the total EQD2 was the sum of the EQD2 values for EBRT and BT.The EQD2 dose and number of fractions for EBRT were based on the gross tumour volume data, and the minimum dose covering 90% of the CTV (CTV D90) in EQD2 was used as the representative dose of image-guided BT.

Follow-up and statistical analysis
The follow-up was carried out in January 2021.The patients' condition was obtained by telephone follow-up consultation and medical records retrieval.The primary endpoint was overall survival (OS), while the secondary endpoints were progression-free survival (PFS) and local control (LC).OS was calculated from the date of recurrence to that of death from any cause or that of the last clinical follow-up.PFS was defined as the interval between the date of starting salvage RT and that of any recurrence, disease progression, or death.LC was calculated from the start of salvage RT to the date on which local tumour progression was detected, which encompassed local control only.OS, PFS, and LC were estimated using the Kaplan-Meier method, and differences in prognosis between subgroups were compared using log-rank tests via univariate analyses.Multivariate analyses were conducted using the Cox proportional hazards regression model to identify independent prognostic factors.All statistical tests were two-sided, and statistical significance was set at a two-sided p-value <0.05.All data analyses were performed using SPSS for Windows (version 23.0;IBM Corp., Armonk, NY, USA).

Patient and treatment characteristics
The patients' detailed demographic and clinical features are summarised in Table 1.Ninety-seven patients with a median age of 53 years were retrospectively reviewed.Patients were classified into 3 groups according to the recurrence pattern: endovaginal recurrence (n ¼ 52); paravaginal recurrence, in which the tumour invaded paravaginal tissues or extended from the top of the vagina towards the pelvic cavity (n ¼ 34); and invasive recurrence, in which the tumours involved surrounding organs such as the bladder, rectum, and pelvic wall (n ¼ 11).The tumour sizes of most patients were �4 cm (n ¼ 82), whereas bulky tumours larger than 4 cm were observed in 15 patients.The median interval between the date of primary hysterectomy and that of recurrence was 26.4 months (range, 2.5-238.0months).Salvage RT was performed with a median EQD2 of 70.0 Gy (range, 37.5-88.9Gy).Median EQD2 of endovaginal recurrence, paravaginal recurrence and invasive recurrence were 67.7 Gy (range, 37.5 Gy-84.5 Gy), 72.3 Gy (range, 45.0 Gy-87.9Gy), 70.2 Gy (range, 50.4 Gy-88.9Gy), respectively.

Clinical outcomes
The median follow-up time was 42.5 months (range, 2.0-110.9months).The estimated OS, PFS, and LC rates were 88%, 84%, and 91%, respectively, at 3 years and 84%, 79%, and 91%, respectively, at 5 years (Figure 1); moreover, 12 patients died.Ten of them died of tumour relapse, one patient died of other disease, and one patient died of complications.Seventeen patients experienced disease progression during the follow-up period; 4 had local recurrence, 9 had distant metastasis, and 4 experienced both.
Tumour size, lymph node metastasis and recurrence pattern were found to be significantly associated with OS on univariate analysis, and inguinal lymph node metastasis was significantly associated with poor OS on multivariate analysis (Table 2).However, both univariate analysis and multivariate analyses results demonstrated that a tumour size �4 cm (p < 0.05) were significantly associated with longer PFS (Table 2).Lastly, only the recurrence pattern (P < 0.05) was an independent predictor of LC on univariate and multivariate analyses (Table 2).

Effect of dose on prognosis
Since the recurrence pattern was a significant independent predictor of LC in our study, we stratified patients based on this factor to identify a dose cut-off.Endovaginal recurrence was associated with prolonged survival (the estimated 5-year LC rate was 96%), while recurrences that were paravaginal and those that invaded surrounding organs were negative prognostic factors, with an estimated 5-year LC rate of 85%.Forty-five patients with recurrences that were paravaginal or invasive of the surrounding organs were evaluated for the aforementioned prognostic factors as well as dose.On univariate analysis, a dose �72.6 Gy was significantly associated with improved LC (p ¼ 0.034; Figure 2), while none of the other factors were found to be significantly prognostic.

Toxicity and late complications
Twelve patients (12.4%) experienced grade �2 late lower gastrointestinal and urinary tract complications.Although the rates of grade �2 late complications were similar between patients treated with a combination EBRT and BT and those who only underwent EBRT (13.2% and 13.3%, respectively), one patient in the EBRT-only group experienced a grade 5 late complication (a rectovaginal fistula), whereas no such complications were observed in patients who received BT alone (Table 3).Among the 45 patients with recurrent tumours that were paravaginal or invasive of surrounding organs, there was no obvious difference in the rates of grade �2 late complications between those who received EQD2 � 72.6 Gy and those who received <72.6 Gy (Table S1).

Discussion
RT has been shown to be an effective treatment for patients with cervical cancer who experience locoregional recurrence after surgery.In a study by Kim et al., the 5-year local failurefree survival and OS rates of patients with cervical cancer who underwent RT after experiencing postoperative recurrence were 63.9% and 66%, respectively (H.J. Kim et al. 2018).In another study, Kim et al. found that the 5-year PFS and OS rates after RT were 62.7% and 60.1%, respectively (S. W. Kim et al. 2017).These studies showed far better survival rates than did those performed in the previous decade in which the 5-year OS rates were below 50% (Haasbeek et al. 2008, Jain et al. 2007).Our study focussed on RT-naïve patients with cervical SCC; these individuals achieved excellent estimated 5-year OS, PFS, and LC rates of 84%, 79%, and 91%, respectively.These superior clinical outcomes can be partially attributed to the pathological tumour type, higher RT dose, and lack of previous RT.
Although surgeries such as pelvic exenteration are also treatment options for patients with cervical cancer who experience central pelvic relapse without pelvic wall involvement or extrapelvic spread, RT has its own advantages.Pelvic exenteration is usually indicated as a curative therapeutic approach for patients with postoperative central pelvic relapse, but it is associated with high surgery-related morbidity and mortality rates (Chao et al. 2021); the survival rate is less than 50%, while mortality is nearly 5% (Benn et al. 2011, Morley et al. 1989).Other surgical methods such as laterally extended pelvic/endopelvic resection carry certain risks of injury to the nerve, bladder, or rectum, leading to impaired urination and defaecation.Moreover, a shortened vaginal length after surgery greatly reduces the quality of life as well as the female sexual function index (Zhang et al. 2020).As such, RT can address these limitations of surgery.
We found that the factors that significantly influenced the effectiveness of RT for patients with recurrence were tumour size, lymph node metastasis and recurrence pattern.In this study, an initial tumour size �4 cm was a significant prognostic which was associated with prolonged PFS.Previous studies have shown that an initial tumour diameter �4 cm was a predictor of favourable LC in patients with cervical cancer who experienced advanced vaginal recurrence (Weitmann et al. 2006).In Zolciak-Siwinska et al.'s study, a recurrent tumour diameter >3 cm was a significant prognostic indicator as it negatively influenced OS, DFS, and LC in patients with recurrences in the cervix or vagina who underwent re-irradiation (Zolciak-Siwinska et al. 2014).
In our study, cases with the involvement of inguinal lymph nodes had a poorer OS than other statuses of lymph node metastasis.Few studies have discussed the relationship between inguinal lymph nodes and the prognosis of cervical cancer patients.A study on postoperative patients with stage IB to IIB cervical cancer showed that patients with metastasis to the common iliac lymph nodes had a poorer prognosis than involvement in iliac nodes, obturator nodes, or inguinal lymph nodes (Xie et al. 2012).Since metastasis of retroperitoneal lymph nodes is significantly associated with poor survival, cervical cancer with retroperitoneal lymph node metastases is categorised as stage IIIC2 in the 2018 FIGO staging system (Bhatla and Denny 2018, Kilic et al. 2021, Mukai et al. 2021).However, the staging of lymph nodes in FIGO 2018 remains controversial, since there was substantial heterogeneity among stage IIIC cases, which might lead to a better prognosis in stage IIIC disease than stage IIIA/IIIB disease (Shin et al. 2021).Distant lymph node metastases are still categorised as stage IVB, but the prognosis of cervical cancer patients with only distant lymph node metastases was better than patients with organ metastases (Im et al. 2015, Mukai et al. 2021).In patients with vulvar cancer, inguinal lymph node metastasis is the most significant prognostic factor for survival and categorised separately as stage III in FIGO 2009 staging system (Burger et al. 1995, J. Li et al. 2015).More deliberative discussion is needed to interpret how lymph node status should be classified in the FIGO staging system for cervical cancer patients.
In this study, endovaginal recurrence was a significant predictor of more favourable OS, PFS, and LC.Since recurrent tumours that are paravaginal or invasive of surrounding organs are associated with a relatively poor prognosis, it was an important objective of our study to explore treatment that could improve survival in these patients.Through stratified analysis, we discovered that the RT dose significantly affected LC in patients with extravaginal recurrence; this was a most novel finding of this study.
Few investigations have elucidated the relationship between RT dose and survival; hence, ours was the first to show that a higher EQD2 (�72.6 Gy) was significantly associated with prolonged LC.Presently, EBRT and BT combined constitute the first radical concurrent chemoradiotherapy choice for patients with locally advanced cervical cancer; however, there is no consensus on the optimal mode of RT for central recurrent cervical cancer, or on whether EBRT and BT should be administered (Abu-Rustum et al. 2020, Cohen et al. 2019, Hille et al. 2003, H. J. Kim et al. 2018, S. W. Kim et al. 2017).Compared to 3-dimensional conformal radiation therapy, intensity-modulated RT reduces the doses too (and irradiation volumes of) the small intestine, rectum, and bone marrow (Mell et al. 2017, Shih et al. 2016), reducing the risk of acute and chronic gastrointestinal and urinary toxicity (Klopp et al. 2018, Yeung et al. 2020).However, intensitymodulated RT is not a substitute for BT (Assenholt et al. 2008).The latter allows for a higher dose of radiation to the tumour while minimising toxicity to adjacent tissues.Moreover, our study found that a combination EBRT and BT was safer than administering EBRT while increasing the irradiation dose.Moreover, the development of computed tomography-or magnetic resonance imaging-guided 3D intracavitary/interstitial BT and the application of new techniques such as 3D printing are expected to further optimise the dose to the tumour region while reducing toxicities and complications (Fokdal et al. 2016, Jastaniyah et al. 2016, Y. J. Kim et al. 2020, Liu et al. 2021, Sturdza et al. 2016).Notably, higher irradiation doses have not been definitively shown to lead to improved OS or PFS, indicating that RT has limited systemic therapeutic benefits despite excellent LC.Chemotherapy is often recommended for patients with extrapelvic metastases or recurrent diseases who are not candidates for RT or exenterative surgery.Multivariate analysis of data from that study did not find that concurrent chemotherapy was superior to RT alone.However, it has been shown that the tumour volume before BT is an important prognostic factor in locally advanced cervical cancer patients treated with radical RT and that their prognosis is more favourable when the high-risk CTV is �30 cc (Mazeron et al. 2015, Tanderup et al. 2016).With the increase in volume, a higher RT dose is required to achieve an LC rate over 90% (Mazeron et al. 2015, Tanderup et al. 2016).Platinum-based concurrent chemotherapy could increase sensitivity and reduce distant metastasis (Rose 2002) and might be able to shrink the tumour volume faster and achieve potential survival benefits, especially in patients with extravaginal recurrence.Immunotherapy and/or bevacizumab have also been included in the first-line systemic treatment for recurrent metastatic cervical cancer (Abu-Rustum et al. 2020).Cases in this study were relatively early, so immunotherapy or bevacizumab was not involved during the initial RT treatment and was only applied to patients after the failure of the initial RT.Since endovaginal recurrence is associated with less lymph node metastasis rate, as well as excellent OS and PFS rates, RT combined with immunotherapy and/or bevacizumab is of little significance.However, for patients with extravaginal recurrence, these treatments may further improve the curative effect.
The limitations of this study include its retrospective, nonrandomized design and inconsistent management of treatment owing to the long duration of the study.The long time span also added difficulty to patient enrolment due to the incomplete medical and operation records, and we couldn't ideally include only patients with operable early-stage cervical SCC treated with proper surgery with low-risk pathologic features justifying no adjuvant therapy.Second, there have been continuous advances in RT technology over the time span of the study, especially BT which has gone through 2D-BT to 3D-BT, intracavitary to intracavitary combined interstitial technique, and cylinder applicator to 3Dprinted personalised applicator.The type of RT administered as well as the dose distribution and fractionation of EBRT and BT were designed according to both the tumour characteristics and RT instrumentation available at the time.

Conclusions
In conclusion, we found salvage RT to be an effective treatment for vaginal recurrences in patients with pathologically confirmed cervical SCC who had previously undergone surgery.The recurrence pattern and tumour size were significant prognostic factors.For patients with extravaginal recurrence, a dose of �72.6 Gy is recommended; moreover, combined EBRT and BT might be a more suitable option.

Figure 2 .
Figure 2. The local control rates of patients with paravaginal recurrence or recurrence invading surrounding organs according to equivalent dose of 2 Gy per fraction <72.6 Gy vs. �72.6Gy.

Table 1 .
Patient and treatment characteristics.

Table 2 .
Factors predictive of overall survival, progression-free survival and local control.

Table 3 .
Late complications in patients who underwent various radiotherapy regimens.