Salvage Radiotherapy Strategy and Its Prognostic Signi�cance for Patients With Locally Recurrent Cervical Cancer After Radical Resection: a Retrospective 10-year Analysis

Background: Salvage radiotherapy modes for treating patients with local cervical cancer recurrence after radical surgery are controversial. Therefore, we aimed to evaluate the clinical e�cacy and prognostic signi�cance of two radiotherapy modes—involved-eld radiotherapy combined with regional lymph nodes (regional radiotherapy) and involved-eld radiotherapy alone (local radiotherapy)—in these patients. Methods: We retrospectively enrolled patients with local recurrence who underwent radical surgery without radiotherapy for early-stage cervical cancer from January 2010 to January 2020. Clinical outcomes were analyzed using the Kaplan–Meier method and a Cox proportional hazards model. Results: Forty-four patients underwent intensity-modulated radiotherapy (IMRT)-based salvage treatment. The 5-year overall survival and progression-free survival rates were 64% and 60.2%, respectively. Sixteen of 18 patients with stump recurrence and 15 of 26 patients with pelvic and abdominal cavity recurrence received regional radiotherapy, while others received local radiotherapy. Univariate analysis showed that patients with stump recurrence, who underwent regional radiotherapy, and with a lower systemic in�ammation response index (SIRI) had better prognosis than their counterparts. Patients aged < 51 years, with stump recurrence, recurrence time ≤ 24 months, recurrence site=1, and a lower SIRI who received regional radiotherapy had a better prognosis than patients who received local radiotherapy. SIRI correlated with the recurrence site and radiotherapy mode. Conclusion: Locally recurrent cervical cancer treated with IMRT-based salvage therapy has a good prognosis. Recurrence site, SIRI, and the radiotherapy mode signi�cantly in�uenced prognosis. Regional radiotherapy may be suitable for patients with stump recurrence, recurrence time ≤ 24 months, and one recurrence site.


Background
Cervical cancer is the most common gynecologic cancer in China and has a higher incidence than ovarian and endometrial cancers [1].Radical surgery is the standard treatment for early-stage cervical cancer.Most patients have good prognosis; however, approximately 25% of all patients experience recurrence [2].The most common recurrence site is the pelvic cavity, followed by the retroperitoneal and distant lymph nodes and other organs [3].For patients with local recurrence after radical surgery who have not received previous radiotherapy (RT), salvage concurrent chemoradiotherapy (CCRT) is recommended [4].However, several aspects remain to be further studied.
First, the development of RT technology has led to an era of intensity-modulated RT (IMRT), which is superior to conventional RT in many aspects [5].Thus, the prognosis of patients with local recurrence treated with IMRT-based salvage therapy should be explored.Second, the RT mode to be selected for CCRT in patients with local recurrence of cervical cancer remains undetermined.At present, there are two common RT modes: involved-eld RT combined with regional lymph nodes (regional RT) and involvedeld RT alone (local RT) [2,6,7].Many studies have shown that RT covering both the regional lymph node area and gross tumor is an appropriate RT mode [8,9].However, this mode of RT confers considerable side effects in patients.Another RT mode involving radiation to the gross tumor only has been shown to lead to a good prognosis (4-year overall survival [OS]: 50.1%; 4-year local control [LC]: 67.4%) and result in few side effects, especially for patients with recurrent lesions measuring less than 17 mm; it is also considered a good alternative for patients with local recurrence [10].However, the two RT modes are still controversial.
Thus, the optimal RT mode for cervical cancer patients with local recurrence needs to be investigated.
Third, ongoing research is focused on the identi cation of new prognostic factors.There is increasing evidence that in ammation plays an important role in the development of various tumors [11].The platelet-to-lymphocyte ratio (PLR) and systemic in ammation response index (SIRI), based on peripheral blood cell counts, are in ammatory response biomarkers.Many studies have associated these biomarkers with the prognosis of several malignant tumors, including operable cervical cancer [12][13][14].
However, the effect of SIRI and PLR on prognosis of patients with locally recurrent cervical cancer remains unknown.
This study was conducted with an aim to evaluate the survival prognosis of patients treated with IMRTbased salvage therapy.Additionally, we analyzed potential prognostic factors of the two RT modes for patients with local recurrence.

Patients
A cohort of patients diagnosed with cervical cancer were retrospectively examined at the Sencond Xiangya Hospital of Central South University from January 2011 to January 2021.The following patients were included in the study: (1) those with a pathological diagnosis of cervical cancer, (2) those with local recurrence (≤ 3 lesions) who were initially treated with radical surgery without RT for early-stage cervical cancer, (3) those whose follow-up data were complete, and (4) those with recurrence time of more than 3 years and diagnosis made via puncture biopsy.The following patients were excluded from the study: (1) those with other tumors, infectious diseases, hematological diseases, or severe liver or renal dysfunction; (2) those who received adjuvant RT; (3) those who had distant metastasis; or (4) those for whom local treatment was not available.The last follow-up was conducted on June 1, 2021.Recurrence was con rmed based on clinical, imaging, and pathological evidence.During the routine follow-up, patients with new lymph node involvement measuring more than 1 cm detected on computed tomography (CT) or based on abnormal uorodeoxyglucose uptake on positron emission tomography (PET) were considered to have recurrence.The newly punctured pathological specimen was veri ed with the previous pathology and determined as recurrent cervical cancer.The disease stage was de ned according to the International Federation of Gynecology and Obstetrics (FIGO) 2009 staging system [15].

Radiotherapy
There are two RT modes.Regional RT refers to irradiation of the gross tumor and prophylactic irradiation of the lymphatic drainage area.Local RT refers to irradiation of the gross tumor alone.In general, the RT mode is selected according to the location and timing of recurrence.Vaginal stump recurrence generally necessitated regional RT including brachytherapy.Patients with pelvic and abdominal recurrence received corresponding RT strategies according to their physical condition, recurrence time, and location.Elderly patients or patients with postoperative complications generally received local RT.Patients with a prolonged recurrence time (more than 2 years) generally received local treatment.Otherwise, regional RT was administered.IMRT was used for external irradiation, which was planned with the Varian Eclipse Treatment Planning System 11.0 (Varian Medical Systems, Palo Alto, CA, USA) and delivered with 6-MV X-rays using a Varian 23EX.For regional RT, the gross tumor volume (GTV) included all detectable involved areas of recurrent disease con rmed using CT or PET, and the clinical target volume (CTV) included the lymphatic drainage area.For local RT, only the GTV was considered.The planning target volume was delineated by margins of 7-10 mm around the GTV and CTV.Cone-beam CT was performed once a week.
For CTV of the pelvic lymph node eld, the superior margin was set at the L4-L5 intervertebral space, and the inferior margin was set at the lower edge of the obturator.For CTV of the para-aortic lymph node eld, the superior margin was set at the renal vascular level, and the inferior margin was located between L5 and S1.The CTV dose was 45-50 Gy, with 1.8 or 2 Gy administered daily, and the GTV dose was 60-66 Gy.When the vaginal stump was the site of recurrence, brachytherapy (192Ir) was used to boost externalbeam RT doses.According to the depth and size of tumor invasion, a supplement of 18-30 Gy was administered [16].

Data collection
Clinical information, including that of age, initial stage and treatment, pathological pattern, time to recurrence, salvage RT with or without chemotherapy, recurrence sites, number of recurrence sites, RT mode, and results of routine blood tests at the time of recurrence, was collected.PLR and SIRI were calculated using the following formulae: PLR = platelet count/lymphocyte count and SIRI = neutrophil count × monocyte count/lymphocyte count [12,17].

Statistical analysis
The SPSS software (version 26.0; SPSS, Inc., Chicago, IL, USA) was used for all statistical analyses.The major endpoint was OS calculated from the date of diagnosis until the date of death or nal follow-up.Progression-free survival (PFS) was de ned as survival from the date of diagnosis until the date of 1) disease progression, 2) relapse, 3) mortality from any cause, or 4) nal follow-up.
The cutoff values for PLR and SIRI were calculated using receiver operating characteristic (ROC) curves, and the optimal cutoff value was de ned according to the maximum Youden index [18].Fisher's exact test was used to determine whether there was a correlation between two variables.Univariate analysis was performed using the Kaplan-Meier method to assess the 5-year OS and PFS rates.Statistical differences between survival curves were evaluated using the log-rank test.Multivariate Cox proportional hazards models were created using the input selection technique of factors signi cant in the univariate analysis.In all tests, P-values < 0.05 indicated a signi cant difference.Hazard ratios (HRs) and their 95% con dence intervals (CIs) were estimated to assess the magnitude of risk.

Patient and treatment characteristics
A total of 44 patients were included in the analysis.The median follow-up time was 33 months, and patient characteristics are summarized in Table 1.The median age of the included patients was 51 (range, 34-69) years.The median time to recurrence was 20 (4-96) months.There were 23 and 21 patients with a time to recurrence of ≤ 20 and > 20 months, respectively.In many studies, 24 months was used as the cutoff time to recurrence, and according to this cutoff, a signi cant difference in outcomes was found [19,20].The time to recurrence was ≤ 24 months for 26 patients and > 24 months for 18 patients.Forty-one patients had squamous cell carcinoma and three had adenocarcinoma.There were 18 cases of stump recurrence and 26 cases of pelvic and abdominal lymph node recurrence.There was only one recurrence site in 31 patients, two sites in 10 patients, and three sites in 3 patients.In the salvage treatment cohort, 16 patients with stump recurrence received regional RT and 2 patients received local RT, and 15 patients with pelvic and abdominal cavity recurrence received regional RT and 11 patients received local RT.There were 37 patients treated with CCRT and 7 with RT alone for physical reasons.Chemotherapy in CCRT comprised a platinum-based regimen.

Cutoff values for PLR and SIRI
For PFS, the area under the ROC curve (AUC) of PLR was 0.587, whereas that of SIRI was 0.744.The optimal cutoff values were 150.6 and 1.158 for PLR and SIRI, respectively (Online Resource 1).For OS, the AUC of PLR was 0.543, whereas that of SIRI was 0.735.The optimal cutoff values were 150.6 and 1.158 for PLR and SIRI, respectively (Online Resource 1).

Prognostic analysis
Forty-four patients with local recurrence received the IMRT-based salvage treatment, and the 5-year OS and PFS rates were 64.0% and 60.2%, respectively (Figure 1a, b).In the univariate analysis, recurrence site, RT mode, and SIRI signi cantly in uenced the OS and PFS rates (Table 2).The 5-year OS and PFS rates of patients with stump recurrence were 100% and 94.1%, respectively, which are better than those of patients with pelvic and abdominal recurrence, whose 5-year OS and PFS rates were 40.5% and 36.7% (Figure 2a-b), respectively.Patients who received regional RT had better prognosis than those who received local RT (OS: 76.6% vs. 36.9%,p=0.012;PFS: 73.4% vs. 30.8%,p=0.003) (Figure 2c-d).On comparing the prognosis of the two RT modes with different clinical features, it was found that patients aged < 51 years, with stump recurrence, recurrence time ≤ 24 months, recurrence site=1, and a lower SIRI who had received regional RT had better OS and/or PFS rates than those who had received local RT (Table 3).The analysis of SIRI as a prognostic factor (Figure 2e-f) showed that a higher SIRI was associated with a poorer prognosis (OS: p<0.001;PFS: p<0.001).The correlation between SIRI and clinical characteristics of patients was analyzed (Table 4).SIRI was related to the recurrence site (p=0.000)and RT mode (p=0.002), which further con rmed that SIRI was related to prognosis.

Toxicity
The main toxicities were enteritis, cystitis, fatigue, nausea, and vomiting, and most of the toxicity were of grade 1 or 2. Grade 3 toxicities were observed in ve patients, including: enteritis (4.5%), nausea and vomiting (4.5%), fatigue (2.3%).Patients with high dose (≥60 Gy) or RT combined with chemotherapy were more likely to have grade 3 toxicity.No grade 4 or 5 toxicity was observed in any patients, and no treatment-related death occurred.

Discussion
At present, there is no standard treatment for patients with locally recurrent cervical cancer.Thus, CCRT should be considered with the selective use of brachytherapy [21,22].A retrospective study of 50 patients with isolated para-aortic lymph node recurrence of cervical cancer after radical surgery compared the clinical outcomes of different salvage treatment modes, including RT, CCRT, surgery, chemotherapy, and best supportive care.The results showed that CCRT was an effective salvage treatment for isolated paraaortic lymph node recurrence; the 3-year OS and PFS rates associated with CCRT were 85.7% and 71.4%, respectively [2].Another retrospective study of 22 cervical cancer patients with lymph node recurrence who previously underwent radical operation showed that salvage RT with concurrent chemotherapy was a good choice for patients; the 5-year PFS and OS rates for all 22 patients were 32.7% and 30.7%, respectively [23].However, these studies did not explore the prognosis following IMRT for patients with local recurrence.A retrospective study comparing IMRT with conventional RT for patients with recurrent cervical cancer found that IMRT included higher irradiation doses for tumors (61.8 Gy vs. 50.3Gy), had fewer side effects, and resulted in better prognosis (5-year OS: 35.4% vs. 21.4%;5-year PFS: 26.1% vs. 15.1%)than conventional RT [24].In our study, all patients received IMRT for external irradiation.The results showed that IMRT-based salvage treatment resulted in better prognosis and the 5-year OS and PFS rates were 64% and 60.2%, respectively.
The location of recurrence and metastasis is an important factor in uencing prognosis.One study showed that the prognosis of para-aortic lymph node metastasis is better than that of supraclavicular lymph node metastasis and that the prognosis of lymph node metastasis is better than that of hematogenous metastasis [6].In our study, the pelvic cavity was the most common recurrence site, followed by the vaginal stump and extra-pelvic area.Furthermore, the prognosis of patients with vaginal stump recurrence was better.
Whether the time to recurrence is related to prognosis is questionable.Jeon et al. reported that a time to recurrence of > 18 months was associated with better OS and PFS [23].Singh et al. reported that the time to recurrence > 24 months after the initial therapy was a good prognostic factor [20].Another study showed that time to recurrence of > 10 months was not signi cantly associated with OS [7].Kubota et al. reported that a median time to recurrence of > 10 months, as well as other cutoff values including 10, 18, and 24 months, was not signi cantly associated with OS, LC, or PFS [2].This study showed that the cutoff time to recurrence, whether 20 or 24 months, did not in uence prognosis but was related to the prognosis with respect to the RT mode.
There are many types of salvage RT for locally recurrent cervical cancer, and these can be summarized as follows: regional RT and local RT [2,6,7].It is unclear as to which mode is better.Sato et al. compared the effects of the two RT modes on the prognosis of patients with oligo-recurrence; 4 patients received gross tumor RT, whereas 17 patients received RT including the regional lymph node area.No signi cant difference in outcomes was found [9].However, the number of cases in this study was very small.This study demonstrated that regional RT was associated with better prognosis than local RT.Most patients with stump recurrence who received regional RT had a good prognosis.However, there was no signi cant difference between the two modes in the prognosis of patients with pelvic and abdominal lymph node recurrence.It was found that patients aged < 51 years, with recurrence time ≤ 24 months, recurrence site=1, and SIRI less than the cutoff value seem to be more suitable for regional RT.
Both SIRI and PLR are biomarkers of systemic in ammatory response, which can re ect the tumor microenvironment.Many studies have shown that SIRI, a new systemic in ammatory response biomarker, is a better prognostic factor than other biomarkers, and a high SIRI is associated with poor prognosis of patients with various malignancies [13,14,[25][26][27].In addition, SIRI can change dynamically with changes in tumor burden and immune response status in cervical cancer patients, and patients with a decrease in SIRI of > 75% had better prognosis (p < 0.001).SIRI was also a potential marker for therapeutic response monitoring in patients with curable cervical cancer [12].PLR is associated with prognosis but is not an independent prognostic factor in many types of cancers [12,14,27].Furthermore, this study showed that SIRI was a prognostic factor for patients with locally recurrent cervical cancer but not PLR, which is consistent with previous results.
Other factors, such as initial stage and treatment, pathological pattern, and salvage RT with or without chemotherapy, did not in uence prognosis.Salvage CCRT was not shown to improve prognosis when compared with RT alone in this study, which is inconsistent with the results of many other studies [21,23].This may be related to the imbalance in the number of cases between the chemoradiotherapy and RT groups.
Our study had some limitations.First, this study was a retrospective study performed at a single institution.Second, this study involved a relatively small number of patients.Future studies should consider including a larger sample size or employing a randomized clinical study design to validate these preliminary results.
In conclusion, patients with locally recurrent cervical cancer who received IMRT-based salvage RT had good prognosis.Vaginal stump recurrence, regional RT, and a low SIRI predict better survival.Further analysis showed that when the time to recurrence is ≤ 2 years or the recurrence site is the vaginal stump, regional RT is recommended.

List Of Abbreviations
Figures

Figure 2 Overall
Figure 2

Table 2
Univariate analysis of overall survival and progression-free survival rates for all patients

Table 3
Comparison of patient prognosis with the two radiotherapy modes with different clinical features

Table 4
Correlation analysis between systemic in ammation response index and clinical features