Introduction

Cervical cancer remains a significant global health issue, ranking as the fourth most common malignancy among women globally, with an estimated 604,127 new cases and 341,831 deaths each year1. Particularly in China, the disease burden exhibits striking severity. According to 2022 epidemiological projections, there were 111,820 new diagnoses and 61,579 mortalities—figures that represent nearly 20% of global cervical cancer fatalities2,3. The clinical management approach for non-operable primary cervical cancer has evolved to prioritize definitive radiotherapy, with brachytherapy emerging as the cornerstone modality, delivering 40–60% of the total radiation dose to the tumor targets4.

The unique dosimetric advantages of brachytherapy stem from its capacity to deliver high-intensity radiation through intracavitary or interstitial applicators, achieving rapid dose fall-off gradients that spare adjacent critical structures. This radiobiological superiority, however, is exquisitely dependent on geometric precision. As dose distribution follows the inverse square law, millimeter-level deviations in applicator positioning can precipitate disproportionate dose perturbations—a 2 mm displacement may escalate rectal dose exposure by 15–20%, while simultaneously creating cold spots in tumor regions5. Clinical correlations substantiate these dosimetric observations: comparative analyses reveal a 50% improvement in 5-year survival rates between cohorts with optimal versus suboptimal implant geometry (60% vs 40%)6, underscoring the life-altering consequences of implantation accuracy.

Despite these critical implications, the current academic discourse reveals notable gaps in knowledge. Existing research predominantly focuses on macro-level complications, such as uterine perforation, which has a reported incidence ranging from 4.5 to 17.9%7,8, while overlooking the systematic assessment of more subtle implantation inaccuracies. The literature lacks comprehensive analyses of: (1) dosimetric variance patterns associated with specific error types (e.g., cervical os displacement, uterine axis deviation, tandem depth discrepancy); (2) consequent alterations in TCP and normal NTCP; (3) differential impacts between 2D and 3D image-guided treatment planning approaches.

Our investigation addresses these deficiencies through a dual analytical approach. First, we establish a novel classification system for uterine tandem malpositioning based on spatial deviation vectors. Second, by employing Monte Carlo simulation and biological modeling, we quantify how post-correction replanning in both 2D and 3D paradigms modulates dosimetric parameters (D90, D2cc) and their corresponding clinical endpoints. This methodology not only provides unprecedented insights into the radioclinical consequences of implantation inaccuracies but also establishes an evidence base for precision brachytherapy protocols.

Material and methods

Patient selection and applicator insertion

Cervical cancer patients treated with intracavitary brachytherapy (ICBT) in our hospital from November 2020 to July 2021 were included in this study. All patients received pelvic external beam radiotherapy (EBRT) 45 Gy, according to the situation and a supplementary dose of 5–10 Gy to parametrial or 10–15 Gy to residual enlarged lymph nodes and HDR-BT 6 Gy × 5–6 F. During EBRT, single-agent cisplatin (40 mg/m2) was given weekly for 5–6 times. BT was performed after completion of EBRT, and the entire radiotherapy course was completed within 8 weeks. All patients underwent pelvic gynecological examination and pelvic Magnetic Resonance Imaging (MRI)/Computed Tomography (CT) imaging, both performed at the time of diagnosis and before brachytherapy. Segedin et al. identified several risk factors including " necrotic cervical tumor, cervical polyp, submucosal fibroid, stenosis, or distortions of the cervical canal (particularly due to prior conization), and a retroflexed or extremely anteflexed uterus," all of which are associated with a higher incidence of uterine perforation. However, in the article by M Parvath ES9 additional uterine pathology was not significantly associated with uterine perforation during brachytherapy. In this study, the anatomical location of the uterus before brachytherapy was classified into the following types, as delineated in the literature by Chakrabarti B7and illustrated in Fig. 1.

Fig. 1
figure 1

Different anatomical situations during uterine and cervical brachytherapy (sagittal view). (A) Normal sized and anteverted, (B) Retroverted uterus, (C) Acutely anteflexed uterus, (D) Short uterus, (E) Presence of obstructing cervical mass.

Full size image

All patients were placed with CT compatible applicators under intravenous anesthesia (Fletcher-Suit style tandem-ovoid intracavitary applicators). A CT scan with a slice thickness of 3 mm was conducted using a Philips CT simulator (Philips Healthcare, Andover, MA), covering the range from the vulva to 5 cm above the uterus. The acquired CT scan images were subsequently transferred to the Oncentra Brachytherapy planning system (Version 4.6.0) for 3D treatment planning. At our institution, in the event of imperfect applicator implantation during brachytherapy, re-implantation would be undertaken by a different gynecologic oncologist. If immediate correction was not feasible, the re-implantation procedure was scheduled 72 h later, with all applicator implantation images retained. For this study, all saved image sets from the designated study phase were reviewed to identify and categorize types of intrauterine tube implantation imperfections.

Imperfection implantation was defined based on four conditions observed on CT scan (Fig. 2): (1) Anterior Wall Perforation: perforation of the uterine tube from the anterior ventral side of the uterine axis; (2) Posterior Wall Perforation: perforation of the uterine canal from the dorsal side of the uterine axis; (3) Inadequate Implantation: the uterine tube did not exceed the isthmus of the cervix or was in the uterine cavity and did not exceed 1 cm above the upper edge of the lesion or was more than 2 cm from the bottom of the uterine cavity; (4) Fundus Perforation: visual confirmation of the applicator tip extension beyond the uterine serosa on CT scan. The differences brought about by the various types of applicator implantation dissatisfaction in 2D and 3D plans were analyzed, and clinical records were maintained for each patient.

Fig. 2
figure 2

Intrauterine tube implantation imperfect classification was defined as four conditions on CT scan: (A) Anterior Wall Perforation: perforation of the uterine tube from the anterior ventral side of the uterine axis; (B) Posterior Wall Perforation: perforation of the uterine canal from the dorsal side of the uterine axis; (C) Inadequate Implantation: the uterine tube did not exceed the isthmus of the cervix or was in the uterine cavity and did not exceed 1 cm above the upper edge of the lesion or was more than 2 cm from the bottom of the uterine cavity; (D) Fundus Perforation: visual confirmation of the applicator tip extension beyond the uterine Serosa on CT scan.

Full size image

Contouring and treatment planning

The perfect/imperfect positioning CT images were re-transmitted to the Oncentra Brachytherapy planning system (Version 4.6.0). Gynecologic oncologists delineated the high-risk clinical target volume (HR-CTV) and OARs based on gynecological physical examination, MRI/CT images at diagnosis and before BT. Target volume definition and reporting were adhered to Gynaecological (GYN) GEC-ESTRO recommendations, with mandatory contouring of organs at risk and dose-volume parameter reporting (e.g., the minimal dose to 90% of the clinical target volume [D90], the minimal dose to the most exposed 2 cm3 of the respective organ [D₂cc])10,11. The target volume and OARs were delineated according to IBS-GEC ESTRO-ABS recommendations12. HR-CTV encompassed all cervical and residual tumors, verified through gross examination, physical examination, imaging, and pathology, while OARs included the bladder, rectum, sigmoid colon and small intestine. The bladder was defined as the entire outer wall, with its lower boundary at the beginning of the urethra. The rectum encompassed the entire outer wall, with its lower boundary 1 cm above the anus and upper boundary at the rectosigmoid junction. The sigmoid colon was delineated from the level of flexion of the rectum and sigmoid colon to the level of disappearance of the corpus uteri. Small intestine included the intestinal canal and mesentery, with its lower boundary at the flexion level of the rectum and sigmoid colon, and upper boundary at the disappearance level of the uterine body.

Treatment employed a 192Ir source in Flexitron HDR brachytherapy set-up (Nucletron, an Elekta company, Stockholm, Sweden). Utilizing the Oncentra treatment planning system (Elekta AB, Stockholm, Sweden, version 4.6.0), both 2D and 3D brachytherapy plans were conducted on perfect/imperfect images of the intrauterine implantation, with a prescription dose of 6 Gy in a single fraction. The 2D plans aimed to compare normalization at point A, defined as the point located 2 cm lateral to the central canal of the uterus and 2 cm from the mucous membrane of the lateral fornix in the axis of the uterus. 3D plans were designed to deliver at least 100% of the prescribed dose to 90% volume CTV (HR-CTV D90), with optimization carried out to ensure CTV coverage and reduce dose to OARs as necessary. For patients, the EBRT and HDR-BT radiation doses need to meet specific dose constraints: HR-CTV D90 ≥ 85–90 Gy EQD2, rectum D2cc ≤ 65–75 Gy EQD2, sigmoid D2cc ≤ 70–75 Gy EQD2, and bladder D2cc ≤ 80–90 Gy EQD2, ensuring that two gray equivalent doses (EQD2) (α/β = 10) were delivered to the 100% isodose curve passing through this point. All 2D plans and imperfect 3D plans for intrauterine implantation were solely utilized for dose analysis in this study and were not implemented clinically. Only corrected 3D plans were implemented for the patients.

Parameters calculation

We use a COIN13 to evaluate implant quality and dose specification in brachytherapy.

$${\text{COIN}} = \frac{{{\text{CTV}}ref}}{CTV} \times \frac{CTVref}{{Vref}}$$
(1)

In the formula, CTVref represents HR-CTV volume covered by the 100% prescription dose line, CTV represents the HR-CTV volume, and represents the total volume covered by the 100% prescription dose line. COIN is commonly utilized to quantify the degree of agreement between the prescribed dose line (plane) and the target volume in brachytherapy. An ideal COIN value = 1, with clinical treatment typically requiring COIN > 0.6413.

Regarding OARs, doses at volumes of 2 cc and 1 cc, and the maximum dose for all OARs (rectum, bladder, sigmoid, and intestines), were evaluated. Specifically, doses were assessed on a 2 cc volume to allow for direct comparison with dose tolerance thresholds outlined in the work of GYN-GECESTRO11. Both the physical dose and EQD2 of the OARs were calculated for D2cc using Eq. (2).

$${\text{EQD2}} = {\text{D}}\frac{{d + {\alpha \mathord{\left/ {\vphantom {\alpha \beta }} \right. \kern-0pt} \beta }}}{{2 + {\alpha \mathord{\left/ {\vphantom {\alpha \beta }} \right. \kern-0pt} \beta }}}$$
(2)

Based on the concept of equivalent uniform dose (EUD), Gay, H. A14 proposed the TCP / NTCP model Mathematical models of NTCP and TCP have been proposed for better optimization and evaluation of HDR BT plans. We used the published code of TCP / NTCP, which takes into account the total dose of radiation therapy (EBRT 45 Gy + BT 6 Gy × 5 fractions)15.

$${\text{EUD}} = \sum\limits_{i = 1} {\left( {{\text{V}}_{i} D_{i}^{\alpha } } \right)}^{{\frac{1}{\alpha }}}$$
(3)
$${\text{TCP}} = \frac{1}{{1 + \left( {\frac{{{\text{TCD}}_{50} }}{{{\text{EUD}}}}} \right)^{4\gamma 50} }}$$
(4)
$${\text{NTCP}} = \frac{1}{{1 + \left( {\frac{{{\text{TD}}_{50} }}{{{\text{EUD}}}}} \right)^{4\gamma 50} }}$$
(5)

Vi is a unitless partial volume, Di is the dose received by the voxel volume at i, and the sum of all partial volumes Vi values equals to 1. atumor =  − 10, abladder = 2, arectum = 8.33, aintestines = 6, asigmoid = 615. TCD50 is the irradiation dose received at a tumor control rate of 50%, γ50 = 3; TD50 is the dose tolerated by normal tissue at a complication rate of 50% for OARs (the bladder, rectum, sigmoid and intestines), γ50 = 416.

Statistical analysis

IBM® SPSS® Statistics software version 23 was used for data analysis. Paired sample t test was used to compare parameters generated with or without normalization for intrauterine tandem implantation, both imperfect and corrected perfect, across the two plans. A significance level of 0.05 was used, and p ≤ 0.05 was considered statistically significant. EQD2 calculations were performe using linear quadratic (LQ) spreadsheet available at American Brachytherapy Society’s http://www.americanbrachytherapy.org/guidelines/LQ_spreadsheet.xls website.

Ethics approval and consent to participate

This study was approved by the ethics committee of the Chongqing University Cancer Hospital Review Board (Approval ID:CZLS2023085-A). Informed consent was obtained from all participants and/or their legal guardians. All methods were carried out in accordance with relevant guidelines and regulations.

Results

In our study, among 319 patients undergoing brachytherapy (total 1742 insertions), imaging imperfections were observed in 24 cases (7.52% of patients), corresponding to 38 individual insertion events (2.18% of all procedures) in Table 1. The characteristics of the 24 patients included in the study analysis are shown in Table 2.

Table 1 List of patients who were dissatisfied with intrauterine tube implantation.
Full size table
Table 2 Patient characteristics (N = 24).
Full size table

Anatomical location of uterus before brachytherapy

Among the 24 patients with imperfect uterine tube implantation, the retroverted uterus accounted for 45.83% (11/24), the acutely anteflexed uterus accounted for 37.5% (9/24), and the small uterus accounted for 16.67% (4/24). Additionally, there were five patients with obstructing cervical masses, none of whom had this condition in isolation: one patient had a retroverted uterus, three had an acutely anteflexed uterus, and one had a small uterus.

Radiotherapy parameters

In all patients with an imperfect uterine tube placement, there was good agreement in VOI (p > 0.05), and CTVref was less than VOI. CTVref was smaller in PER-3D group than in PER-2D group. The Vref in the PER-3D group exhibited the smallest magnitude, significantly differing from that in the PER-2D group (p < 0.05); COIN values exceeding 0.64 were exclusively observed in the PER-3D group, contrasting with values below this threshold in other groups. Particularly, the smallest COIN value was recorded in the imperfect 2D (IM-2D) group.

These findings indicate that only the PER-3D group meets the clinical requirements for COIN, achieving target coverage by encapsulating the smallest feasible volume, while the remaining groups attained target coverage by expanding the dose coverage range, as shown in Table 3.

Table 3 Comparison of average dose-volume parameters between perfect and imperfect implantatoin (Mean ± SD). Dates were obtained from 2 and 3D plan in these patients with uniform prescription of 6 Gy.
Full size table

Inadequately implanted

In 38 cases of imperfect intrauterine tube implantation, the proportion of inadequate intrauterine tube implantation was 31.58% (12/38), in Table 1. Inadequate intrauterine tandem implantation increased EQD2 of rectum, sigmoid and small intestine in 3D plans (p < 0.05); however, there was no statistically significant difference in EQD2 of OARs in 2D plans (p > 0.05). as depicted in Table 3A and Fig. 3A. In Table 4 we observed that inadequately implanted reduced TCP, which was more significant in the 2D plans; the NTCP of OARs in the IM-3D group was significantly higher than that in the PER-3D group, with the exception of the bladder (p < 0.05); additionally, the NTCP of the bladder in the IM-2D group was lower compared to the PER-2D group (p < 0.05).

Fig. 3
figure 3

Dose distribution after perfect / imperfect implantation of the uterine tandem in both plans.

Full size image
Table 4 Normal tissue complication probability under the same prescription for uterine tandem imperfect implantation (`x ± s).
Full size table

Anterior wall perforation

Anterior uterine wall perforation accounted for 28.95% of the 38 cases of imperfect intrauterine implantation, in Table 1. When Anterior uterine wall perforation occurred, EQD2 in the rectum, small intestine, and bladder was significantly increased in 3D plans (p < 0.05). EQD2 was also significantly higher in the small intestine in the 2D plans (p < 0.05), while EQD2 of rectum and sigmoid colon in the 2D plans decreased, especially for sigmoid colon (p < 0.05), as shown in Table 3B and Fig. 3B. As depicted in Table 4, anterior wall perforation resulted in a reduction of TCP, with the impact being particularly pronounced in the 3D plan.

Posterior wall perforation

Posterior Wall Perforation occurred in 9 patients, accounting for 23.68% (9/38) of the 38 cases of imperfect intrauterine implantation, in Table 1. When posterior wall perforation occurred in the concatemat uteri, the EQD2 of OARs increased in both 2D and 3D plans, except for a reduction of EQD2 in bladder in the 2D plans,. Notably, significant increases in EQD2 were observed for both rectum and bladder in the 3D plans, and for rectum in the 2D plans (p < 0.05). For details, refer to Table 3C and Fig. 3C.

Fundus perforation

Fundusperforation was observed in 6 patients, constituting 15.79% (6/38) of the 38 cases of imperfect intrauterine implantation, as detailed in Table 1. Table 3D and Fig. 3D show that EQD2 of sigmoid colon was increased when axis perforation occurred (p < 0.05), and Table 4 shows that when axis perforation occurred, TCP of patients was decreased (p < 0.05) and NTCP of OARs was increased (p > 0.05).

Clinical outcomes

During the process of uterine tandem implantation, neither the patients nor the doctors reported any obvious perforation. The possibility of re-perforation was evaluated after perforation was detected via positioning CT scans, and experienced doctors were tasked with re-implanting the uterine tandem. There were no instances of bleeding or damage to surrounding organs during or within 2 h after the operation, and no symptoms of pelvic infection were reported within 1 week after operation.

At the three-month follow-up after treatment, one patient was lost to follow-up, while 18 patients exhibited local complete response and five showed partial response. At the two-year follow-up post-treatment, four patients were lost to follow-up, 18 patients remained in local complete remission, and one patient experienced recurrence and continuous progression.

Discussion

Brachytherapy stands as an indispensable modality for the radical radiotherapy of cervical cancer. The technical precision of brachytherapy implantation has been demonstrated to significantly influence treatment outcomes7,17,18,19. Sandun et al.20 emphasized the clinical significance of even minor shifts in intracavitary (IC) and combined intracavitary/interstitial (IC/IS) applicator positions in cervical cancer treatment. Uterine perforation represents the most studied applicator location in brachytherapy8,18,21,22. However, no prior study has classified uterine position before brachytherapy, categorized dissatisfaction with intrauterine implantation, and conducted dosimetric and clinical outcome predictions.

Literature review indicates that the incidence of uterine perforation following applicator placement assisted by CT scan ranges from 2.8 to 13.7%23,24,25. The rate of uterine perforation per insertion varies between 4.56 and 6.4%, while the rate per patient spans from 7.39 to 17.9%18,26. In our study, the rate of uterine perforation was 2.18% (38/1742) per insertion and 7.52% (24/319) per patient, both of which lower than previous reports. Currently, in addition to CT and MRI, ultrasound is used to guide applicator implantation. With intraoperative US guidance, the reported incidence of perforation was reduced to 0.9–1.4%27,28. It is recommended that ultrasound-guided intrauterine tube implantation should be selected according to the patient’s uterine position in order to reduce the incidence of uterine perforation. Real-time intraoperative ultrasound guidance may be useful to reduce uterine perforation rates and thus increase brachytherapy use. Ultrasound can assist in performing not only intracavitary brachytherapy but also interstitial brachytherapy29,30.

It’s likely that reported incidence rates underestimate the actual occurrence due to underreporting by surgeons. Otter S et al.31, highlighted older age, larger tumors, retroverted uterus, and stenotic cervical as predisposing factors for uterine perforation. Other studies identified factors such as age over 60 years24,32,33, history of prior conization21,32, physician inexperience7, scarring or distortion of the cervical34, and increasing tumor size as risk factors for uterine perforation. A novel scoring system for predicting the risk of uterine perforation: uterine position was included in the risk factors, with sensitivity of 46% and 66.7%, Specificity of 96% and 73% in the test and validation columns, respectively (P < 0.05). Additional uterine pathology was not a risk factor for uterine perforation9. Pre-planning by imaging is required before insertion to assess uterine cavity length, orientation, and position, as well as to determine the location of any residual cervical mass. This pre-planning reduces the likelihood of imperfect brachytherapy implantation7. The study by Onal et al. revealed that preoperative MRI planning reduced the incidence of perforation to 4% from 11% without pre-planning31. The pre-BT MRI image details were utilized for scoring the patients prior to BT9, of the patients in this study, 91.7% underwent MRI before BT treatment.

The anatomical positions of the cervix and uterus before brachytherapy were classified based on pre-planning images, and the types of uterine perforation were manually categorized. Onal C31 enrolled 200 patients and concluded that the most frequent perforation site was the posterior uterine wall (4%), followed by the fundus (2.5%) and anterior wall (2%). In our study of 319 patients, the most common cause of imperfect uterine tube insertion was inadequate uterine tube insertion, followed by anterior uterine perforation. The larger sample size in our cohort may provide more robust data compared to earlier studies.

In our study, different types of intrauterine tube implantation imperfections during brachytherapy were investigated. COIN was utilized to evaluate the degree of conformality in target area13. A COIN value above 0.64 meets clinical requirements and signifies optimal treatment quality. Analysis of radiobiological parameters across all groups revealed that only the PER-3D group achieved a COIN value > 0.64, meeting clinical requirements. These results indicate that 3D plan is the optimal choice for brachytherapy.

Suboptimal placement of the applicator can result in uterine perforation, inadequate dose to the target and excess dose to adjacent normal tissues. The dose variation caused by uterine perforation varies significantly depending on the perforation location. Optimal placement of the tandem and ovoids is essential for an acceptable brachytherapy plan35. Failure to perform a brachytherapy boost correlates with increased local failure risk in patients with cervix cancer, whereas tandem malposition does not correlate with increased risk. In Manon Kissel’s study18, even with suboptimal brachytherapy implantation, optimization prevented inadequate dose coverage or poorer local control. In cases with obstructive cervical mass, good subsequent insertion is possible with acceptable dose volume parameters, if planned dose can be delivered to its 60% volume. Indian institutional experience documented by Chakrabarti et al. revealed that analyzes dosimetric variations after imperfect brachytherapy insertions7. Insertion of adequate length of tandem aligned to uterine axis is warranted for adequate tumor coverage. In the Bikramjit study, anterior perforation resulted in a high bladder dose in 40% of cases and a high rectal dose in 20% cases7. When inserted in the retroverted direction in a retroverted uterus, the tandem generates acceptable dose-volume parameters. Acute anteflexion and mobile retroversion should be corrected whenever detected.

TCP and NTCP radiobiological parameters were used to predict tumor control probability and normal tissue complication probability. The advantage of NTCP and TCP radiobiology lies in their ability to predict and evaluate the ultimate therapeutic effect of a specific treatment regimen, with prediction conclusions aligning closely with clinical observations22,36. When intrauterine tube implantation was imperfect, the varying positions of the uterine tube had distinct dose effects on OARs across different plans, resulting in reduced TCP and increased NTCP for different OARs (see Tables 3, 4). In conclusion, regardless of the site of intrauterine canal implantation imperfection, EQD2 and NTCP of OARs were increased to varying degrees, while TCP was decreased. This underscores the impact of brachytherapy implant quality on patients’ disease-free survival, aligning with conclusions from prior studies17.

One investigation demonstrated that patients diagnosed with cervical cancer at Stages III/IV experienced a reduction of approximately 1% in local tumor control per day of treatment delay beyond 30 days. Moreover, Petereit et al. observed a notable decline in 5-year survival rate from 65 to 54%, when treatment duration exceeded 55 days37,38. A recent study echoed these findings, reporting a 3-year local tumor control rate of over 94% when total radiation therapy was administered within a 7-week timeframe39. Due to the rapid doubling time of cancer cells, current evidence strongly advocates against treatment interruptions during standard therapy protocols. This underscores the critical need to address delays caused by uterine perforation during brachytherapy, as such events may compromise treatment efficacy and survival outcomes. In our study, all cases of uterine perforation were promptly corrected without clinically significant complications. Similar findings have been reported in other studies, where patients completed treatment without significant delay or further complications after correct reinsertion of the tandem21,32,34. Thus, immediate evaluation for possible reinsertion should be pursued upon identifying uterine perforation. Our study demonstrates that prompt correction of applicator placement is feasible and safe in cases of uterine perforation, allowing patients to proceed with brachytherapy without undue delay, thus potentially optimizing treatment efficacy and patient outcomes8.

Conclusions

Compared to the IMP group and the PER-2D group, the PER-3D group demonstrated clinical superiority in cervical cancer brachytherapy with a COIN value exceeding 0.64. When suboptimal placement of the uterine canal occurs, the resulting dose to organs at risk (OARs) fails to meet clinical standards, and the rise in NTCP (Normal Tissue Complication Probability) comes at the expense of a decreased TCP (Tumor Control Probability). Immediate correction of the uterine canal position is strongly recommended to avoid damage to other organs. For patients with abnormal cervix and uterine body positions, ultrasound-guided uterine tube placement is advised to enhance accuracy. Most importantly, the incidence of uterine perforation is significantly lower in patients undergoing image-guided intracavitary insertion for cervical cancer.