Outcomes of External-Beam Radiation Therapy Boost with Conventional Fractionation in Cervical Cancer: A Retrospective Analysis about 133 Cases

Introduction: External-beam radiation therapy boost is a treatment option in cervical cancer when brachytherapy is not feasible. Though less effective than brachytherapy, some encouraging results have been reported from some institutions experiences. We conducted this study to assess outcomes of EBRT boost for our patients at National Institute of Oncology in Rabat. Patients and Methods: We collected data from patients treated for cervical cancer between January 2012 and December 2015. Patients, tumor and treatment characteristics were collected. Overall survival (OS), disease-free survival (DFS) and prognostic factors influencing DFS were assessed. Results: One hundred and thirty-three patients were enrolled. Median age was 52 years. Patient haemoglobin level ranged from 3.9 to 15.5 g/dl [mean: 11.2 g/dl]. Most tumors were classified stage III/IVA (63.2%) according to the FIGO classification. Regional lymph node metastases (pelvic and or para-aortic) were observed in 45.1%. Median total dose to tumor was 69.6 Gy (ranging from 66 to and >11 g/dl) were the two significant and independent prognostic factors influencing DFS (HR: 1.86; p = 0.01 for the former) (HR: 0.59; p = 0.03 for the latter). Conclusion: Our study showed that EBRT boost in conventional fractionation was an acceptable treatment option for cervical cancer unamenable to brachytherapy, especially in the two subgroups of patients that are those without pelvic and/or para-aortic lymph node metastasis and those with haemoglobin level above 11 g/dl.


Introduction
Concurrent chemoradiotherapy followed by intracaviatry brachytherapy is the mainstay of the treatment of locally advanced cancer of the cervix [1]. The treatment, generally, follows two steps. Firstly, External-Beam Radiation Therapy (EBRT) is used to give a radiation dose (45 to 50 Gy) to a large volume that includes whole the genital tract and regional lymph node areas. This, obviously, will result in tumor shrinkage. Secondly, brachytherapy is used to deliver a high dose in a reduced and limited tumor volume (boost dose) [1] [2]. The goal is to attain a high total dose in the tumor (up to 85 or 90 Gy). However, in some situations, brachytherapy can be either not feasible or not available. Therefore, using EBRT only for the whole treatment course constitutes an acceptable option. In the last two decades, some institutions have published the results of their practices. Although it is known that the EBRT boost resulted in fewer outcomes than brachytherapy, some encouraging results had been reported [3] [4] [5].
According to GLOBOCAN 2018 statistics, cervical cancer is the second most common malignancy in women in Morocco. In 2018, 3388 new cases and 2465 death were registered in the country. It is the second cancer leading cause of death in women [6]. The National Institute of Oncology is the referral hospital of adult and solid malignancy in the Rabat district. Many cervical cancers, especially the locally advanced, are treated each year in the department of radiotherapy of our institution. No study has previously addressed the concern of EBRT boost in cervical cancer. Therefore, we conducted this study to assess outcomes of our practice and determine prognostic factors associated with survival in cervical cancer treated with EBRT boost.

Patients' Selection
This study is a retrospective one. Data were collected from patients treated for cervical cancer in the department of radiotherapy of NIO (National Institute Oncologists (FIGO) staging system, after a workup that included gynaecologic examination and thorax-abdomen-pelvic CT-scan or pelvic MRI plus thorax-abdomen CT-scan. Complete blood cells count and kidney function assessment (urea, creatinine, and clearance) were obtained for all patients. Patients treated by EBRT boost were selected based on two manners: • After the whole pelvic radiation therapy (WPRT) course, if brachytherapy or surgery was not feasible, the treatment was completed by the EBRT boost. • At the initial evaluation, some patients with very large and extensive tumors were selected, at the physician discretion, to receive an EBRT boost immediately after the whole pelvic radiation therapy (WPRT) course.

Treatment Description
The treatment plan used a 3-Dimensional Conformal Radiation Therapy (3DCRT) technique for all cases. All patients had a 3D CT-scan (3Dimensional Computed Tomography scan) simulation in supine position, arms above the chest, empty rectum, and a full bladder. CT slices thickness was 5 mm. Intravenous contrast enhancement was mandatory and realized for all patients who had no kidney dysfunction.
Whole Pelvic Radiation Therapy (WPRT) field covered the gross primary tumor, the total genital system (cervix, uterine corpus, vaginal portion, parametria, and adnexa) and the regional pelvic node areas, including the common, internal and external iliac, obturator and presacral lymph nodes. The field was extended to cover para-aortic lymph nodes if patients had para-aortic lymph node metastasis. EBRT boost volume included primary tumor, cervix, uterine corpus, bilateral parametria, and a portion of vagina. Patients were not re-simulated after WPRT; the same treatment plan was used for both the WPRT and the boost. A 7 -10 mm geographical margin was added to the volume described above to obtain the planning target volume. Radiation therapy was delivered in convention- Organization) performance status, blood cells count and renal clearance were evaluated or performed. The cycle was precluded if patients had PS grade 3 or 4, haemoglobin level below 8 g/dl, granulocytes number less than 1500/mm 3 , platelet number < 100,000/mm 3 or creatinine clearance < 60 ml/minute.

Follow-Up
Patients were examined every 3 or 4 months during the first 2 years after completing treatment, thereafter every 6 months for the 3 subsequent years, then annually. Treatment response, disease recurrence, and complications were recorded. Tumor response was assessed using clinical examination. Imaging was not performed systematically. An MRI or CT-scan was required for patients who presented suspicious symptoms. All recurrences were histologically confirmed if the site was accessible for biopsy.

Survival and Statistical Analysis
The endpoints of our study included disease-free survival (DFS) and overall survival (OS). DFS was calculated from the date of the start of treatment to the date of any sign of tumor relapse may it be local, regional or remote. The OS was calculated from the date of the beginning of treatment to the day of death (of any cause); any patient loss of follow-up after failure or disease progression was assumed dead from the disease.
Survivals were estimated by the Kaplan-Meier method. Predictor variables, including age, general condition, tumor size (the greatest axis measurement), lymph node metastasis, number of concurrent chemotherapy cycles, haemoglobin level (measured at diagnosis, before any treatment or blood transfusion) and overall treatment time (OTT), were analyzed to assess their impact on disease control.
The differences in DFS in univariate analysis were assessed with the log-rank test.
For multivariate analysis, a Cox proportional hazards model was developed using forward stepwise regression for all variable showing a trend (p < 0.15) on univariate analysis. Results were considered significant for p values less than 0.05.

Results
According to our eligibility criteria, 133 patients were enrolled in the study.  Table 3.
The 5-years' overall survival (OS) and disease-free survival (DFS) were 47% (median = 49.15 months) and 44% (median = 28.49 months) respectively ( Figure   1 and Figure 2). The prognostic effects of variables on DFS were presented    in Table 4. On univariate analysis, haemoglobin level and regional lymph nodes (RLN) metastasis influenced significantly the 5-years DFS (Figure 3 and Figure   4). Patients who had haemoglobin levels above 11 g/dl had significant higher

Discussion
Our study was a retrospective one that aimed overall survival, disease-free sur-  [3]. In Haeyong et al. study the DFS and OS were 54.7% and 75% respectively [5]. Mastsuura et al. reported that 3 years' local control and overall survival were 75% and 43.8% respectively [4].
Apart from the absence of brachytherapy, others tumor characteristics can influence treatment results. These characteristics include patient age and general condition (WHO performance status), tumor size and stage, haemoglobin level, regional lymph nodes status, and overall treatment time OTT. Saibishkumar et al. found that the response to EBRT was the significant factor associated with better DFS and OS [7]. According to Haeyoung et al., achieving complete treatment response at a first follow-up visit and shorter OTT were both significantly associated with favourable local failure-free survival [5]. In Kei et al. study, FIGO stage and tumor size were the significant factors associated with survival [8]. In our study, patients having tumor size equal or less than 6 cm, FIGO stage I/II and overall treatment time more than 60 days, had better DFS than their counterparts (size > 6 cm, FIGO stage III/IV and OTT ≤ 60 days) but the differences were not significant.
Haemoglobin level and regional lymph node metastasis have emerged as the two significant and independent prognostic factors. Haemoglobin level and tumor size are two characteristics reflecting tumor oxygenation that is a main radiobiology parameter influencing tumor sensitivity to radiation. Anaemia results in less oxygenation of the tumor, and the bulky tumor is characterized by increased ischemic and necrotic areas. Small tumors are the most oxygenated and are more sensitive to radiation. Conversely, large tumors, in which many ischemic and necrosis areas occur, are less sensitive. Consequently, these large tumors benefit from fractionation, as progressive tumor-shrinking results in progressive re-oxygenation of ischemic areas. Therefore, fractionation can overcome the tumor size effect. Anaemia, conversely, have a permanent effect. Blood and oxygen supplies are permanently reduced, creating a "chronic hypoxia" state that cannot be overcome by fractionation. Low haemoglobin level is also worsened, in the time, with chemotherapy due to hematologic toxicity. Many authors had previously demonstrated that a low haemoglobin level was associated with bad outcomes [11] [12] [13]. Theoretically, blood transfusion can restore blood and oxygen supplies, but studies that focused on the effect of transfusion are conflicting [14] [15] [16].
The second predictive factor is lymph node metastasis. The presence of pelvic Ultimately, we can conclude that prognosis in locally advanced cervical cancer unamenable to undergo brachytherapy is related to tumor extent (reflected by the local extent and regional lymph node involvement) and tumor oxygenation (reflected by haemoglobin level and tumor size).
In the new radiotherapy era, some studies addressed the role of high conformal radiotherapy technics such as intensity-modulated radiation therapy (IMRT) and stereotaxic body radiation therapy (SBRT) in management of cervix able fact is that the sites of the pelvic lymph node, the number and the size of these pelvic lymph nodes were not precisely described. Therefore, it was impossible to identify patients at high risk that would have relapse in absence of para-aortic prophylactic irradiation. Further studies should evaluate the role of prophylactic para-aortic irradiation in high-risk patients.

Conclusion
Although most patients in our study experienced recurrence within 5 years following treatment, our result showed that the EBRT boost in conventional fractionation is valuable in two subgroups of patients including those without regional lymph node metastasis and those with haemoglobin levels above 11 g/dl.