MESORECT : Influence of bladder depletion on mesorectal movements during radiation therapy of locally advanced rectal cancers: a prospective comparative study

Highlights

• •Anterior mesorectal movements decrease from upper to lower mesorectum.
• •A “full bladder” portocol is less reproducible than an “empty bladder” protocol.
• •Empty bladder slightly increases OAR doses, with respect to international constraints.
• •Anisotropic margins optimize dosimetric accuracy safety, awaiting for adaptive radiotherapy.

Abstract

Neoadjuvant radiotherapy (RT) is a cornerstone in the management of locally advanced rectal cancer (LARC). With the increasing use of intensity-modulated radiation therapy (IMRT), accurate target delineation and treatment reproducibility are critical. Bladder filling may influence anterior mesorectal displacement. This study prospectively evaluated anterior mesorectal motion under “full” versus “empty” bladder conditions. Forty-eight patients were included. Anterior mesorectal distances relative to bony landmarks were measured at six pelvic levels on planning CT (pCT) and seven cone-beam CTs (CBCTs) acquired during treatment under both bladder protocols. Inter-individual means of intra-individual differences and their standard deviations (SD) were calculated to assess variability. Secondary endpoints included bladder volume reproducibility, an in silico dosimetric comparison, and calculation of anisotropic anterior planning target volume (PTV) margins. Anterior displacements were similar between protocols, with consistent forward motion. Variability decreased from upper to lower rectum. In the upper rectum, bladder filling had little effect (SD: empty 6.48 mm vs. full 5.19 mm). In mid and lower rectum, variability was lower with an empty bladder (mid: 4.60 vs. 5.36 mm; lower: 3.03 vs. 3.45 mm). Bladder volume was poorly reproducible under full-bladder conditions (mean 243 cm³, SD 164 cm³) but consistent with empty bladder instructions (mean 73 cm³, SD 58 cm³). Dosimetric analysis showed slightly higher organ-at-risk doses with an empty bladder, remaining within national constraints. An empty bladder protocol with anisotropic margins improves treatment reproducibility while maintaining dosimetric safety in LARC patients.

1. Introduction

Colorectal cancer (CRC) is the third most frequently diagnosed cancer and the second leading cause of cancer-related mortality worldwide, according to the 2022 GLOBOCAN report [1].

Approximately 35% of CRC arise from the rectum, and nearly 40% are diagnosed at a localized or locally advanced stage [2], [3]. In rectal cancer, the proximity of adjacent pelvic organs and the absence of a protective serosal layer contribute to a substantial risk of local recurrence. Historically, when surgery alone was performed, local relapse accounted for nearly one-third of treatment failures [4], [5], prompting increasing interest in factors influencing local control [6].

Neoadjuvant radiation therapy (RT) has since become a cornerstone of rectal cancer management, allowing tumor downstaging, improved resectability, and a significant reduction in local recurrence rates [7], [8], [9], [10].

These benefits appear consistent across radiation protocols, whether delivered as long-course or short-course [11]. More recently, Total Neoadjuvant Therapy (TNT), combining induction chemotherapy, chemoradiotherapy (CRT), and subsequent Total Mesorectal Excision (TME), has gained widespread acceptance due to improved oncologic outcomes, including higher pathological complete response rates and prolonged disease-free survival (DFS) [12], [13], [14], [15], [16], [17]. Updated results from the PRODIGE-23 trial confirmed these benefits and found significantly improved overall survival (OS: 81.9% vs. 76.1%, P = 0.033) [18] and TNT is now incorporated into most international guidelines [3].

Intensity-modulated radiotherapy (IMRT) allows for more conformal dose delivery and reduced exposure to organs at risk (OARs), but precision in target delineation and reproducibility across sessions are critical [19], [20]. In the case of LARC, mesorectum represents the main component of the Clinical Target Volume (CTV). Several studies have demonstrated that mesorectal motion is greatest in its middle and upper portions [21], [22], [23], [24], [25]; where it lies in direct contact with the posterior bladder wall [26]. To enhance treatment precision and displace small bowel loops from the irradiated volume, patients are routinely instructed to maintain a consistent bladder filling level during planning CT (pCT) and each treatment fraction. However, substantial inter-fraction variability in bladder volume is frequently observed, despite standardized protocols [27], [28], [29]. This variability may be exacerbated in patients with chronic renal insufficiency and further aggravated during treatment by the development of urinary symptoms such as pollakiuria [30]. To our knowledge, there is no prospective data on the influence of bladder filling on anterior mesorectal movements. To assess whether IMRT delivered with an empty bladder could reduce bladder volume variability and anterior mesorectal motion, we conducted a prospective, non-randomized, monocentric study comparing two bladder management protocols: “empty bladder” and “full bladder” (standard practice). The primary objective was to evaluate variability in anterior mesorectal displacement between the two protocols. Secondary objectives included assessing bladder filling reproducibility, evaluating dosimetric feasibility with respect to OAR and planning target volume (PTV) constraints, and proposing anisotropic anterior PTV margins based on anatomical level while accounting for bladder-related variations.

2. Methods

2.1. Trial design and patients

We conducted a monocentric, open-label, prospective, non-randomized study sponsored by and performed at the Institut de Cancérologie de l’Ouest (Nantes and Angers, France). The study protocol was approved by the institutional ethics review board. The trial was registered on ClinicalTrials.gov (NCT04323722) and conducted in accordance with the principles of the Declaration of Helsinki.

Eligible patients were aged 18 years or older, with histologically confirmed rectal adenocarcinoma staged as T2 N + or T3, regardless of nodal status, and considered suitable for long-course neoadjuvant chemoradiotherapy delivered in 25 fractions. An Eastern Cooperative Oncology Group (ECOG) performance status of 0–2 was required. To limit potential confounding related to renal function and bladder filling capacity, a creatinine clearance greater than 50 mL/min was mandatory.

Exclusion criteria included a history of pelvic radiotherapy, active inflammatory bowel disease, use of a urinary catheter or urinary diversion, prior cystectomy, hip replacement, prostatectomy, or hysterectomy, and a urinary incontinence severity score greater than 2. Given the potential impact of tumor invasion of adjacent pelvic organs on mesorectal mobility, patients with T4 tumors were excluded.All participants provided written informed consent prior to enrollment and retained the right to withdraw consent at any time.

2.2. Trial procedures

The pCT was performed with 3 mm slices in both conditions (full and empty bladder), in supine position. For the full-bladder protocol, patients were instructed to void their bladder and then drink 300 cm³ of water one hour before pCT acquisition and each treatment session. For the empty-bladder protocol, pCT was performed within a maximum of 15 min following bladder voiding. Rectal evacuation was recommended one hour before pCT and each fraction. Compliance with bladder preparation instructions was assessed daily using pre-treatment CBCT.

All target volumes, including gross tumor volume (GTV) when identifiable, clinical target volume (CTV), and OARs were delineated by the same radiation oncologist on both pCT datasets to limit inter-observer variability. The CTV included the entire mesorectum defined on CT and MRI, presacral lymph nodes with a 7-mm margin, and internal iliac nodes, with external iliac node inclusion in case of nodal involvement [26]. The planning target volume (PTV) was generated by applying a uniform isotropic margin of 10 mm. For each patient, IMRT treatment plans were generated on both full- and empty-bladder pCTs. Only the “full” bladder plan was used for treatment, with daily IGRT.

Optimization criteria for PTV and organs at risk followed the RECORAD [31] and ICRU83 [32] recommendations. OAR definitions followed the Male/Female RTOG Normal Pelvis Atlas [33]; using the BLADDER and BOWEL BAG structures and dosimetry was analyzed using QUANTEC constraints [34]: BOWEL BAG mean,max, V45 < 150 cm³ and V15 < 830 cm³. Bladder mean and maximum doses were also recorded. PTV coverage was assessed using V98%.

Patients received concomitant CT with oral capecitabine at a dose of 1600 mg/m²/day administered twice daily, at fixed times on radiation therapy treatment days. Screening for DPD deficiency was performed prior to initiation of chemotherapy, in accordance with the french recommendations [35].

During the treatment course, cone-beam computed tomography (CBCT) scans under empty-bladder conditions were acquired on treatment days 1, 2, and 3, and then weekly thereafter (Fig. 1). A minimum of seven CBCT scans were obtained for each patient in both bladder conditions. Empty-bladder CBCTs were performed within 15 min of bladder voiding.

Fig. 1.

Flowchart. LARC: locally advanced rectal cancer; TAP: thoracic abdominal pelvic; BMI: body mass index; ECOG: eastern cooperative oncology group; CT: computed tomography.

Electronic portal imaging devices (EPID) embarked on the Clinac allows for 1024x768 pixels whereas the EPID on Halcyon and Truebeam allow for 1280 x 1290 pixels. Tomographic sections were 3 mm thick for both pCT and CBCTs.

For analysis, all pCT and CBCT images were imported into RayStation® treatment planning software. Weekly CBCTs were rigidly registered to the corresponding pCT (full or empty bladder) using automatic bone fusion, excluding the femoral heads due to their mobility.

Measurements were obtained as distances between six predefined anterior bony landmarks of the sacrum and coccyx and the anterior border of the mesorectum. To best reflect mesorectal motion, measurements were performed in the axial mid-plane. In cases of marked anterior mesorectal wall irregularity on a single axial slice—commonly observed in women due to uterine anatomy—measurements were taken within ± 10 mm laterally to capture the most anterior position. To isolate internal organ motion from external setup variations, measurements were obtained at six anatomical levels defined by sacral and coccygeal intervertebral landmarks: S3–S4, S4–S5, S5–C1, C1–C2, the tip of the coccyx, and 15 mm below the coccygeal tip (five slices)(supplementary figure S1). The upper rectum corresponded to S3–S4 and S4–S5, the mid-rectum to S5–C1 and C1–C2, and the lower rectum to the coccygeal tip and the five slices below it [36]. Bladder delineation was performed on all image sets using atlas-based segmentation, and bladder volume was automatically calculated from these contours.

2.3. Outcomes

The primary endpoint was to evaluate the reliability of anterior mesorectal positional measurements under empty- versus full-bladder protocols by comparing group systematic errors. These were defined as the inter-individual mean of intra-individual mean differences between anterior mesorectal distances measured on pCT and CBCT at six predefined anatomical levels. Positive and negative values indicated anterior and posterior displacement, respectively. Motion variability during treatment was quantified using the standard deviation (SD) of these systematic errors (Fig. 2).

Fig. 2.

Variations of the bladder volume during treatment depending on filling instructions. On the ordinate, the volume in cm³ and on the abscissa, the CBCT scans from 1 to 7, where 0 corresponds to pCT.

Secondary endpoints included bladder volume variability, calculated as the inter-individual mean of intra-individual SDs of bladder volume differences between pCT and CBCT. Dosimetric feasibility of both bladder protocols was assessed through comparison of dose–volume histogram parameters for organs at risk and PTV coverage. Gaussian modeling of anterior mesorectal displacements, stratified by anatomical level, was performed to estimate margins required to encompass 90%, 95%, and 98% of patients.

2.4. Statistical analysis

Qualitative variables were summarized as counts and percentages, and quantitative variables as mean ± SD or median (range), as appropriate. Differences between mesorectal distances on pCT and CBCT were visualized using boxplots and spaghetti plots for each bladder protocol. Group systematic errors and SDs of random and systematic errors were calculated according to the van Herk formalism. Group systematic errors were compared using paired Student’s t tests, while SDs of systematic errors were assessed with the Pitman-Morgan test. Dosimetric comparisons were performed using paired Student’s t tests. The assumption of normality of within-subject differences (paired t-test) and bivariate normality of paired observations (Pitman–Morgan test) were assessed using graphical methods. Bladder volume evolution was analyzed using patient-specific linear regression. Linear mixed-effects models were applied with bladder protocol as a fixed effect. For visualization purposes, distributions of anterior mesorectal displacements were smoothed using Gaussian kernel density estimation at each predefined anatomical level under both bladder filling conditions. Margins correspond to percentiles computed directly from the original (non-smoothed) empirical data.

3. Results

3.1. Patient characteristics

Between September 2020 and March 2024, 48 patients were enrolled. Baseline characteristics are summarized in Table 1. The sex ratio was 3.4. Tumors were predominantly located in the mid and lower rectum, defined as 2–7 cm and 0–2 cm from the level of the levator ani, respectively. Most patients presented with stage III disease, predominantly T3N + . Nearly 80% received intensified neoadjuvant chemotherapy. A total of 758 imaging datasets (pCT and CBCT) were analyzed, yielding 4,479 anterior mesorectal distance measurements.

Table 1.

Demographic and clinical characteristics of the study population (N = 48).

Characteristics	Mean	SD
Age	65.6	11.3
	N	%
Sex
Women	11	22.9
Men	37	77.1
Tumor location
Lower rectum	21	43.7
Middle rectum	25	52.1
Upper rectum	2	4.2
Tumor size
T2	4	8.3
T3	44	91.7
Node
N0	11	22.9
N+	36	75.0
Nx	1	2.1
Neoadjuvant chemotherapy
No	10	20.8
Yes	38	79.2

3.2. Variability of anterior mesorectal distance

No statistically significant differences in group systematic errors were observed between bladder filling protocols (Table 2). Inter-individual mean displacements were predominantly positive, indicating an anterior shift of the mesorectal anterior border relative to the pCT, except in the lower rectum under full-bladder conditions.

Table 2.

Average distribution of systematic measurement deviations.

Mesorectal level	Empty baldder (mm)	Full bladder (mm)	T-Student
High	2.13	1.79	p = 0.69 (NS)
Medium	1.42	1.02	p = 0.58 (NS)
Low	0.43	−0.16	p = 0.21 (NS)

To further characterize motion variability, the standard deviations (SD) of systematic errors were analyzed for each anatomical level (Table 3). In the upper rectum, variability was high and comparable between protocols (SD 5.19 mm for empty bladder vs. 6.48 mm for full bladder, p = 0.28). In contrast, variability was significantly reduced under empty-bladder conditions in the mid-rectum (SD 4.60 mm vs. 5.36 mm, p = 0.005) and in the lower rectum (SD 3.03 mm vs. 3.45 mm, p = 0.04). A spaghetti plot representation of those results is available in the supplementary material (Fig. S2).

Table 3.

Standard deviations of Distribution in systematic measurements.

	Empty bladder (mm)	Full bladder (mm)	Morgan Pitman test
High	5.19	6.48	p = 0.28
Medium	4.60	5.36	p = 0.005
Low	3.03	3.45	p = 0.04

3.3. Bladder volume variations

Bladder volume was measured on all pCT and CBCT scans under both bladder conditions, resulting in 768 measurements. Under empty-bladder instructions, bladder volume remained stable throughout treatment, with no significant temporal variation (p = 0.8989) (Fig. 4). Female patients exhibited significantly lower post-void residual volumes than male patients, with a mean difference of − 27 cm³ over the observation period (p = 0.0025). Conversely, under full-bladder instructions, marked inter-fraction variability was observed throughout treatment (Fig. 5). The mean bladder volume was 243 cm³ (SD 164 cm³), compared with 73 cm³ (SD 58 cm³) under empty-bladder conditions. A progressive decrease in bladder volume was observed over time, with an average reduction of approximately 9 cm³ per week (p = 0.1044), corresponding to a median total decrease of 26 cm³ (12.4%) between the planning CT and the final week of treatment (seventh CBCT). No statistically significant differences in full-bladder volumes were observed between male and female patients.

3.4. Dosimetric comparison

Dosimetric parameters at planning for OARs and PTV coverage are presented in Table 4. The maximum bladder dose was significantly lower under empty-bladder conditions, whereas mean bladder dose did not differ significantly between protocols. For the small bowel, both mean and maximum doses were significantly higher with the empty-bladder protocol. No significant differences were observed for small bowel V45Gy or V15Gy between bladder conditions, although slightly higher values were noted under empty-bladder conditions, with large SDs. No D_(max) values exceeded 105% of the prescribed dose under either full- or empty-bladder conditions. The RECORAD V15Gy constraint (830 cc) was exceeded in 1/48 empty-bladder plans and 3/48 full-bladder plans. The V45Gy constraint (150 cc) was exceeded in 2/48 empty-bladder plans and 1/48 full-bladder plans. Subgroup analyses did not modify these results. PTV coverage did not differ significantly between protocols.

Table 4.

Comparison of dosimetric parameters according to bladder filling instructions in rectal cancer radiotherapy. SD represents standard deviation, PTV planning target volume.

	Empty bladder		Full bladder
	Mean	SD	Mean	SD	p (paired student test)
Dose to organ at risk
Bladder
Mean dose(Gy)	21.9	7.2	21.9	5.0	0.84
Maximal dose(Gy)	44.0	6.3	47.5	3.7	<0.01
Small bowel
Mean dose(Gy)	13.8	6.0	11.9	5.1	<0.01
Maximal dose(Gy)	44.2	9.2	41.9	9.1	<0.01
Dose V45(cm3)	40.4	51.7	32.1	63.5	0.45
Dose V15(cm3)	359.7	247.3	302.6	248.2	0.08
PTV 98%(%)	90.0	3.8	90.8	4.1	0.29

3.5. Margins

Empirical distributions were obtained from the 4,479 anterior mesorectal distance measurements for each predefined anatomical level under both bladder filling conditions.

Under empty-bladder conditions, anterior margins required to achieve population coverage probabilities of 98%, 95%, and 90% were 16, 13.5, and 9.5 mm, respectively, for the upper mesorectal segment (S3–S4 and S4–S5). Corresponding margins for the mid-rectal segment (S5–C1 and C1–C2) were 13.5, 10.5, and 8 mm, while margins for the lower segment (coccygeal tip and five axial slices caudal to the tip) were 7, 5, and 4 mm.

Under full-bladder conditions, the required margins for the upper segment were 15.5, 12.5, and 10 mm for 98%, 95%, and 90% coverage, respectively. For the mid-rectal segment, margins were reduced to 15.5, 13, and 10.5 mm, and for the lower segment to 5.5, 5, and 3.5 mm.

Distribution models are available in the supplementary material (Fig. S3 and S4).

4. Discussion

The rectum and mesorectum are dynamic anatomical structures, subject to positional shifts and volumetric changes related to peristalsis and variations in organ filling [25], [37]. The mesorectum can be subdivided into three anatomical levels with decreasing mobility from cranial to caudal regions: upper, middle and lower mesorectum.

Despite standardized instructions, bladder volume remained highly inconsistent during RT In the present study, full-bladder conditions were associated with marked variability (SD = 164 cm³) and a progressive decrease in volume over the course of treatment. Similar findings have been reported by Sithamparam et al., and Chang et al. described reductions in median bladder volume of up to 59.3% at treatment completion, with substantial inter-fraction variability [28], [38]. This downward trend may reflect the development of urinary symptoms during treatment, such as pollakiuria, which can make sustained bladder filling increasingly challenging.

In contrast, the empty-bladder protocol appeared more reproducible, with no significant temporal variation in bladder volume. No gender-related differences were observed under full-bladder conditions. However, under empty-bladder conditions, women exhibited significantly lower post-void residual volumes than men (–27 cm³, p = 0.0025), in agreement with recent reports [39]; possibly explained by the prostatic sphincter contributing to dysuria.

Strict bladder-filling instructions are often stressful for patients and vulnerable to logistical delays, which can induce volume variations of up to 40% for 15 min delay [40].

Beyond patient comfort, full-bladder protocols were associated with greater intra-patient variability in anterior mesorectal position. Variability was significantly reduced in the middle and lower mesorectum with an empty bladder, while the upper mesorectum remained the most mobile, as previously reported [21], [22], [24], [25]. Improved reproducibility at bladder level may reduce the risk of underdosage in the anterior mesorectum, a common site of local recurrence [41].

Our dosimetric analysis provides reassuring evidence regarding the feasibility and safety of an empty-bladder approach. No significant differences in mean bladder dose were observed between protocols. Data correlating bladder dose with clinical toxicity remain limited; however, a QUANTEC analysis pooling 13 studies reported no risk of grade ≥ 2 bladder toxicity below an equivalent dose of 50 Gy [42] a threshold above the conventional rectal dose range of 45–50 Gy used in France [31].Gastrointestinal tolerance could potentially be affected, as slightly higher doses to the small bowel were observed with the empty-bladder protocol. Sithamparam et al. reported a significant association between reduced bladder volume and grade 2 diarrhea (P = 0.045) [28]. Nevertheless, QUANTEC analyses identified bowel bag V15 as the most predictive parameter for gastrointestinal toxicity. In our study, mean V15 was 360 cm³ under empty-bladder conditions versus 317 cm³ with a full bladder (p = 0.08), a non-significant difference with large SDs likely reflecting inter-observer contouring variability. Moreover, the QUANTEC analysis included patients treated with concomitant 5-FU and 3D-CRT, both associated with higher gastrointestinal toxicity compared with modern IMRT and capecitabine-based protocols. Small bowel V45 also did not differ significantly between protocols. Prospective studies incorporating clinical toxicity assessment are required to further clarify the potential impact of empty-bladder protocols on gastrointestinal tolerance.

The large number of anterior mesorectal distance measurements obtained in this study (n = 4,479) enabled a detailed evaluation of anterior PTV margins by anatomical level. Required margins decreased progressively from superior to inferior segments. These findings are consistent with those of Boldrini et al., who recommended anisotropic margins based on three mesorectal segments in MRI-guided adaptive radiotherapy [43] and Eijkelenkamp et al. (2021) Eijkelenkamp et al., who demonstrated the superiority of direction-specific margins over isotropic expansion in accounting for interfraction motion [44]. The proposed margins are exploratory and are intended solely to provide an analytical indication of margin requirements as a function of the desired treatment adequacy rate within the studied population. Nevertheless, these data emphasize the clinical relevance of anisotropic, segment-specific margins to ensure target coverage while limiting unnecessary irradiation of organs at risk.

Study limitations include the assessment of mesorectal motion in a single spatial dimension, justified by prior evidence identifying the anterior direction as the predominant axis of displacement [21], [22], [24], [25]. While three-dimensional motion analysis would provide a more comprehensive assessment, prior CBCT-based studies have similarly reported dominant anteroposterior displacement, particularly in the upper mesorectum [45]. Image quality represents another limitation, as CBCT resolution may introduce measurement uncertainty; however, the use of high-resolution Halcyon EPID imaging likely reduced this effect. Future studies using MRI-guided radiotherapy may further refine motion characterization.

More broadly, our findings highlight the importance of adaptive volume management during rectal cancer radiotherapy. In this context, daily online adaptive radiotherapy appears particularly promising. A recent trial by Wang et al. demonstrated that treatment margins could be safely reduced to 3 mm using daily online adaptation [46]. Pending broader access to adaptive radiotherapy, treatment under empty-bladder conditions represents a pragmatic and reproducible approach for patients with locally advanced rectal cancer.

CRediT authorship contribution statement

Joly Alexandre: Conceptualization, Methodology, Investigation, Writing – original draft. Blanc-Lapierre Audrey: Formal analysis, Methodology. Llagostera Camille: Data curation. Le Denmat Dominique: Data curation, Investigation. Guimas Valentine: Writing – review & editing. Rio Emmanuel: Writing – review & editing, Supervision, Project administration. Libois Vincent: Writing – review & editing.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: This work was supported by the Institut de cancérologie de l’Ouest. The authors declare no other competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data 1

It contains a description of the measurement method and supplementary infography.