Patients

From May 2021 to May 2022, eligible LACC patients were recruited for the RCT study, who received pelvic IMRT and cisplatin-based concurrent chemotherapy, followed by brachytherapy.

The Institutional Review Board of the Second Affiliated Hospital of Xi’an Jiao Tong University approved the clinical trial (approval number: 2021-021), and the trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines [32]. Prior to the intervention, all patients were thoroughly informed about the study design and potential risks, and provided signed informed consent. The trial was registered on Chinese Clinical Trial Registry with the identifier ChiCTR2200066485.

Eligibility criteria included: (1) Patients aged between 18 and 75 years with Karnofsky performance status (KPS) ≥ 70. (2) Cervical biopsy-proven squamous cell carcinoma, adenocarcinoma, or adenosquamous carcinoma of the cervix. (3) International Federation of Gynecology and Obstetrics (FIGO, 2018) stage IB3, IIA2-IVA. (4) Absence of serious underlying medical conditions, including hematopoietic abnormalities and abnormal heart, lung, liver, kidney function, or immunodeficiency.

Exclusion criteria included: (1) Patients with a history of hematologic diseases or malignancies other than cervical cancer. (2) Patients who have previously received surgical treatment for cervical cancer, pelvic irradiation, interventional therapy, chemotherapy, targeted therapy, or immunotherapy. (3) Patients who were pregnant or breastfeeding. (4)Patients presenting with uncontrolled vaginal bleeding, a risk of vaginal fistula formation, or ureteral obstruction. (5) Patients concurrently enrolled in other clinical trials.

External beam radiotherapy (EBRT)CT-based simulation

A Pelvic fixation plate and a thermoplastic mold were utilized to ensure repeatability of the EBRT procedure, and patients were immobilized in the horizontal supine position with their arms raised above their heads. A contrast-enhanced CT scan with a slice thickness of 5 mm was performed from the T12-L1 interspace to 5 cm below ischial tuberosity.

Definition of target volume and OARs

The CT image datasets were imported into the treatment planning system (TPS). The target volume and OARs were delineated by a consistent deputy chief physician following the International commission on radiation units and measurements (ICRU) related reports [33]. The gross tumor volume node (GTVnd) was defined as a lymph node with short-axis diameter ≥ 1 cm on imaging. The clinical target volume (CTV) included the gross tumor, the cervix, the entire uterus, the parauterine, the vagina, and pelvic lymph node drainage areas. The pelvic lymph node drainage areas included a 7 mm region extending beyond the edge of the blood vessel, encompassing the common iliac, extra ilia, intra ilia as well as the obturator and the presacral regions. The upper boundary of CTV was the bifurcation of the common iliac artery and extending to the renal vascular level when the para-aortic lymph nodes were incorporated into GTVnd. The lower boundary of CTV was set 3 cm beneath the most inferior vaginal involvement. The planning target volume (PTV) of the CTV was expanded by 7 mm laterally and 15 mm axially from the primary CTV; the PTV of the GTVnd (PGTVnd) was enlarged by 7 mm laterally and 15 mm axially on primary GTVnd. OARs included rectum, bladder, femoral head, spinal cord, small intestine and colon. When the upper limit of CTV extended to the renal vascular level, kidney and liver were also considered as OARs.

BM contouring

In the BMS group, for each patient, the outer contour of the pelvic bone, lumbar spine, and femoral heads were delineated as a surrogate for BM, based on the CT simulation positioning images in the TPS. The delineation extended from 4 levels above PTV to the end of the ischial nodule.

Previously, dual EBRT plans were designed for patients, using pelvic dose gradients, specifically V10, V20, V25, V30 and V40. It was observed that limiting the dose to V40 < 25% resulted in plan conformity and uniformity comparable to that of non-BMS plans, while still meeting the requirements for PTV and OARs (Fig. 1). Consequently, the BM V40 < 25% was prescribed for BMS group. In addition, it was considered that dose to the pelvic BM from intracavitary brachytherapy (ICBT) was negligible.

Fig. 1

A dual-plan DVH plot of an example patient. The small triangle solid line represented the BMS group, and the small square solid line represented the control group. Abbreviation CTV, clinical target volume; DVH, the dose volume histograms; PTV, plan target volume; GTVnd, plan gross tumor volume node

EBRT planning and evaluation

IMRT treatment plans for all patients were developed using TPS. Treatment was administered using Varian 21EX or Trilogy linear accelerators with 6 MV X-rays. Fixed-field static reverse IMRT was implemented with a field number of 5 or 7 coplanar fields. The prescribed dose ranged from 45.0 to 50.4 Gy delivered in 25–28 daily fractions of 1.8-2 Gy per day, five days per week, and PTV was optimized to ≥ 95% of the V(100%). For patients with positive lymph nodes, simultaneous integrated boost (SIB) was used, increasing the dose for PGTVnd to 55-61.6 Gy delivered in 25–28 daily fractions of 2.2 Gy per day. For OARs, it was prescribed that small intestine and colon V40 ≤ 30%, V50 ≤ 10%, rectal and bladder V50 ≤ 50%, left and right femoral heads V50 ≤ 5% respectively, left and right kidneys Dmean ≤ 13 Gy respectively, liver D30% ≤20 Gy, spinal cord V40 ≤ 0%.

The EBRT plan for each patient was designed by a consistent medical physicist, and was evaluated using data from dose volume histograms (DVHs). We also assessed the plan’s conformity and homogeneity. We calculated conformity index (CI) and homogeneity index (HI) as follows: CI = (VPTV, ref/VPTV)× (VPTV, ref /Vref), where VPTV, ref represents the volume of reference isodose surface encompassing the target, VPTV represents the target volume, and Vref represents the volume of reference isodose surface. An optimal conformity is indicated by a CI value closer to 1. HI = [D2% - D99% ]/D50%, where D2%, D50% and D99% correspond to the doses received by 2%, 50% and 99% of the target volume of the. Optimal homogeneity is reflected by a lower HI value.

ICBT

ICBT was administered to patients using either a two-dimensional therapy technique or image-guided brachytherapy (IGBT) technology. ICBT initiated after the delivery 40–50 Gy of EBRT. A high-dose-rate (HDR) 192iridium after-loading treatment machine was utilized with a dose fractionation schedule of either 6 Gy × 5 fractions or 7 Gy × 4 fractions, 1 to 2 times per week. EBRT and ICBT were typically completed within a duration of 6–8 weeks. In the two-dimensional therapy, point A was defined as a location 2 cm above the ectocervix and 2 cm lateral to the midline. Rectal and bladder were prescribed as Dmax ≤ 60-70%. In the IGBT, High risk CTV (HR-CTV) included the cervix, the vagina and any remaining tumor. It was prescribed that HR-CTV V(100%) ≥ 90%, rectum and bladder D2cm3 ≤ 75%. The equivalent dose in 2 Gy fractions (EQD2) for point A and HR-CTV was 80–85 Gy, increasing to 87 Gy or higher when the tumor diameter of the is 4 cm or greater.

Chemotherapy regime

Concurrently with radiotherapy, patients were administered weekly cisplatin-based chemotherapy (40 mg/m2) for a duration of 6 weeks. Patients with FIGO stage III and IVA received 2 to 4 cycles of adjuvant chemotherapy following the completion of radiotherapy. Adjuvant chemotherapy consisted of TP regimen (cisplatin 60 mg/m2 d1-2 + paclitaxel 135 mg/m2 d1), once every 21 days. Chemotherapy and radiotherapy were typically held if the white blood cell count (WBC) was < 2.0 × 109/L, the absolute neutrophil cell count (ANC) was < 1.0 × 109/L, or the platelet (PLT) count was < 50 × 109/L. Chemotherapy was additionally held if patients developed febrile neutropenia, renal failure, grade ≥ 2 neurotoxicity, or grade ≥ 3 nausea lasting > 24 h. Acute toxicity was graded using the National Cancer Institute Common Terminology Criteria for Adverse Events (CTCTE, version 4.03) [34].

Study endpoint

The primary endpoint of this phase II randomized trial was HT during EBRT. Blood tests including WBC, ANC, hemoglobin (HGB), PLT, and lymphocyte (LYM), were conducted at least weekly throughout the treatment course, with the nadir observed during EBRT used to assess HT. Acute HT was graded according to CTCAE (version 4.03, Table 1s) [34].

Secondary endpoint included dosimetric parameters from EBRT plan’s DVHs, including PTV V(100%), CI, HI, small intestine V40 (%) and V50 (%), rectum V50(%), bladder V50(%), left and right femoral head V30 (%) and V50 (%), left and righ kidney Dmean (Gy), liver D(30%), spinal cord V40(%), and BM V10(%) to V50(%).

Study design

This study was conducted as an open-label, single-center, prospective, randomized clinical trial. Participant enrollment was achieved within a timeframe of 12 months, followed by a 3-month follow-up period. Prior to the initiation of treatment, eligible participants were randomized into two groups (BMS and control), utilizing a computer-generated random number list. Based on previously published studies, we estimated an incidence of grade ≥ 3 acute HT in the BMS group of 40%. The incidence rate of grade ≥ 3 acute HT in the control group was estimated at 62% according to a pretest in our center. To achieve a power of 90% and a two-sided type I error of 0.05, the required sample size for each group was 104 according, based on a two-sided test. Considering an anticipated maximum loss to follow-up rate of approximately 20% for this trial, the final sample size was adjusted to 127 participants for each group.

Statistical analysis

SPSS (IBM, Chicago, IL, version 28.0), R (version 4.1.0, http://www.rproject.org/) and GraphPad Prism (version 9.0) software were used for statistical analysis and to generate graphs for the study. Continuous variables were compared with parametric methods when a normal distribution was confirmed. For non-normally distributed variables and categorical data, nonparametric tests were utilized for comparisons. Univariate and multivariate analyses were conducted using a binary logistic regression model. The receiver operating characteristic (ROC) curve was used for predictive value analysis. The “rms” package in R was used to generate the nomograms. Unless otherwise stated, all analyses were performed with a 2-sided significance level of P = 0.05.