Does ITV vaginal procedure ensure dosimetric coverage during IMRT of post-operative gynaecological tumours without instructions concerning rectal filling?

Abstract

Aim

To find out whether the internal target volume (ITV) vaginal procedure ensures dosimetric coverage during intensity-modulated radiation therapy (IMRT) of post-operative gynaecological tumours without instructions on rectal filling.

Background

The ITV vaginal procedure does not necessarily include all movements of the bladder, and does not include changes in the rectal volume. We should know if the vaginal ITV is a useful tool in maintaining CTV coverage during treatment.

Materials and methods

A retrospective analysis of 24 patients treated between July 2012 and July 2014 with adjuvant IMRT for gynaecological cancer. All patients underwent empty and full bladder CT on simulation (CT-planning) and three weeks later (CT-control). ITV displacement was measured and the 3D vector was calculated. ITV coverage was then evaluated by comparing the volume covered by the prescription isodose on both CT's. Patients were asked to have full bladder but they did not follow recommendations for the rectum.

Results

The mean 3D vector was 0.64 ± 0.32 cm (0.09–1.30). The mean ITV coverage loss was 5.8 ± 5.7% (0–20.2). We found a significant positive correlation between the 3D vector and the loss of coverage (Pearson correlation, r = 0.493, 95% CI: 0.111–0.748, p = 0.0144). We did not find any significant correlation between the bladder and rectal parameters with the 3D vector and loss of dosimetric coverage. We found a trend between the maximum rectal diameter in CT-planning and 3D vector (r = 0.400, 95% CI: −0.004 to 0.692, p = 0.0529).

Conclusion

ITV vaginal procedure contributed to ensuring a good dose coverage without instructions on rectal filling.

1. Background

The need for adjuvant treatment in high-risk patients treated with hysterectomy for cervical or endometrial cancer is well-established.1, 2, 3, 4, 5, 6, 7

Adjuvant pelvic radiotherapy, however, increases the side effects such as gastro-intestinal (GI), genito-urinary (GU) and hematologic toxicity.1, 2, 3 In 40–60% of cases, whole “conventional” pelvic irradiation of up to 45–50 Gy results in acute GI grade 2 side effects, and in 26% of cases, this procedure is associated with late GI toxicity. Severe complications occur in 3% of cases, mainly in patients who have experienced an acute adverse event (grade 2 or higher). In a randomised study⁵ comparing the addition of cisplatin-based chemotherapy (CT) to pelvic radiation therapy (RT), Grades 3 and 4 hematologic and gastrointestinal toxicity were more frequent in the Radiotherapy plus chemotherapy group with 27 episodes of grade 4 toxicity in 21 of 122 patients, most of which were hematologic.

The use of intensity-modulated radiation therapy (IMRT) in sparing the surrounding normal tissues has been reported by several investigators.8, 9, 10, 11, 12, 13 Two prospective Phase II trials14, 15 together with an international multicentre study¹⁶ confirm that pelvic IMRT reduces GI acute toxicity grade 2 or higher in up to 28% of patients.

The standardisation of the pelvic IMRT has been performed, but with its rapid dose fall-off around the PTV and internal movement of organs during treatment, it may cause a dosimetric coverage loss of the target and unnecessary OAR inclusion into high dose regions.

At the time of defining the vaginal clinical target volume (CTV), in the adjuvant setting for IMRT in gynaecological tumours, we must take into account the movement of pelvic organs between fractions during treatment. The position of the vagina can vary greatly depending on the different volumes of the bladder and rectum. This is why RTOG introduced internal target volume (ITV) in order to individualise the movement of the vagina associated with bladder filling.¹⁷ Even so, this method does not necessarily include all movements of the bladder, and does not include changes in the rectal volume.¹⁸

2. Aim

We investigated the extent of vaginal shifts of our patients treated with postoperative IMRT for gynaecological tumours. We used instructions for the bladder filling during treatment and gave no recommendation regarding the rectum. We assessed whether the movement of the vaginal ITV entails the loss of dosimetric coverage, and we have sought the correlation between that loss of dosimetric coverage, with vaginal motion, and the volumes of the bladder and rectum during treatment.

3. Materials and methods

3.1. Patient characteristics

24 consecutive patients treated with adjuvant pelvic IMRT, after undergoing hysterectomy for endometrial or cervical cancer, were evaluated between July 2012 and July 2014 with institutional review board approval. In accordance with our treatment protocol, the patients were treated with IMRT, 45 Gy in 25 fractions of 1.8 Gy. Adjuvant chemotherapy was given to 10 of 24 patients.

3.2. CT simulation and treatment planning

Patients were instructed to attend the planning-CT (p-CT) session and subsequent daily treatment sessions in full bladder conditions. To accomplish this, patients were asked to empty the bladder and drink 33 cc of water in 15 min, 1 h before the planning-CT. Patients were not given instructions concerning rectal filling. Treatment position was supine decubitus, using a knee cushion to immobilise the hip and the legs. Patients were then scanned from the upper abdomen to below the perineum, with a slice width of 3.75 mm (LightSpeed RT; General Electric, Fairfield, CT, USA) (full bladder CT). After this first scan, patients emptied their bladder and, immediately, a second scan was acquired in the same position (empty bladder CT). Both CT scans were then rigidly registered and vaginal CTV was contoured following RTOG guidelines for each scan.¹⁹ ITV was then constructed on the full bladder CT as the Boolean union of both CTVs. Fig. 1 summarises the whole image procedure followed in the study. Additionally, nodal CTV was also contoured following the consensus guidelines described by Small et al.¹⁷ Planning target volume (PTV) was created by adding a 7 mm margin to ITV and nodal CTV, according to RTOG protocol. The following organs at risk (OAR) were delineated on both CTs: rectum, bladder, femoral heads and peritoneal cavity.

Fig. 1.

Creation of ITV from full and empty bladder CT. p: planning CT; c: control CT; FULL: full bladder CT; EMPTY: empty bladder CT; CTV: clinical target volume; ITV: internal target volume.

Treatment was planned in Eclipse (Varian Medical Systems, Palo Alto, CA) on the full bladder p-CT. All treatment plans consisted of nine equispaced 6 MV photon beams with dynamic intensity modulation, and were calculated to deliver the prescription dose to 97% of the PTV. Version 10 of the Dose Volume Optimizer and the Analytical Anisotropic Algorithm were used for the fluence optimisation and dose calculation, respectively, with a dose resolution of 2.5 mm.

On week 3 of treatment, a second set of full and empty CT scans were acquired, and the volumes were contoured again (control-CT, c-CT) by the same radiation oncologist.

To investigate whether the full/empty bladder CT simulation process prevents tumour loss coverage along the treatment, p-CT and c-CT full bladder scans were rigidly registered in ARIA registration workspace. Afterwards, original plans were recalculated on the full bladder c-CT using the same beam parameters and fluence (Fig. 1).

ITV dose coverage loss was then assessed by subtracting the ITV volume covered by the prescription isodose in p-CT and c-CT.

To quantify the ITV movement between p-CT and c-CT, two parameters were calculated: the coordinates’ difference between the centre-of-mass of both ITVs, and the 3D vector calculated as the square root of the sum of the squares of the coordinates difference.

3.3. Statistical analysis

To know the relationship between the movement of the ITV and its loss of coverage with the main organs at risk, bladder and rectal volumes were calculated from each planning and control scan. The parameters evaluated were: the ratio between the organ volumes in both CTs, and the difference of the volumes between them. Other parameters evaluated were the maximum rectal diameter and the maximum anteroposterior (AP) rectal diameter at the coccyx level in both CTs and, finally, the ratio between the AP rectal diameter at the coccyx level in both CTs.

A descriptive analysis was carried out calculating frequencies and percentages for qualitative variables and mean (standard deviation), median and range for quantitative variables. A Pearson correlation analysis was performed to evaluate the relationship between ITV coverage loss and 3D vector.

Results are presented in a scatter plot, and a linear regression is fitted. All analysis were carried out in Stata 13.1.

4. Results

The mean age was 60 years (range 37–77). The patient and tumour characteristics of the 24 patients are summarised in Table 1.

Table 1.

Patient characteristics.

Characteristic	Number of patients (%)
All patients	24(100)
Age (years)
Mean	60
Range	37–77
Primary tumour
Cervix	3 (12.5)
Endometrium	20 (83.3)
Nodal relapse	1 (4.2)
Histology
Endometriod	15 (62.5)
No endometrioid	7 (29.2)
SCC	2 (8.3)
Chemotherapy
Yes	10 (41.7)
No	14 (58.3)
EBRTa (fractions × Gy)
25 × 1.8	23 (95.8)
28 × 2.12	1 (4.2)
Brachytherapy
Yes	14 (58.3)
No	10 (41.7)

^aEBRT: external beam radiation therapy.

The ITV displacements obtained are shown in Table 2 for the three main axes. Mean 3D vector was 0.64 ± 0.32 cm (range 0.09–1.30).

Table 2.

ITV displacements.

Axis	Mean (cm)	Standard deviation (cm)	Range (cm)
Left/right	0.17	0.12	0.01–0.50
Superior/inferior	0.35	0.28	0.00–0.93
Anterior/posterior	0.41	0.31	0.03–1.26

4.1. ITV coverage loss

The mean ITV coverage loss was 5.8 ± 5.7% (range 0–20.2). 11 of 24 patients (45.8%) showed a 3D vector of <0.7 cm. For these patients, the loss of coverage was less than 10%.

The remaining 13 patients (54%) had a 3D vector of > 0.7 cm, but only 16.7% (4/24) had a loss of coverage higher than 10% (range 16.4–20.2%).

We found a significant positive correlation between the 3D vector and the loss of coverage (Pearson correlation = 0.493, 95% CI: 0.111–0.748, p = 0.0144) (see Fig. 2).

Fig. 2.

ITV dosimetric loss of coverage (%) vs 3D vector (cm).

4.1.1. Bladder volumes

The results of the analysis of bladder volumes can be seen in Table 3.

Table 3.

Bladder and rectal volumes.

	p-CTa				c-CTb
	Bladder vol (ml)		Rectal vol (ml)	Max rectal diameter (cm)	Bladder vol (ml)		Rectal vol (ml)	Max rectal diameter (cm)
	Full	Empty			Full	Empty
Mean	204.8	42.2	72.2	5.1	220.2	49.4	69.4	5
Median	217.7	42.6	65	5	225.6	50.8	61.6	5.4
Max value	536.8	86.1	132	7.3	523.2	97.2	202.6	8.8
Min value	51.26	23.2	37.6	3.8	64.7	22.8	33.4	3

^ap-CT: planning CT.

^bc-CT: control CT.

Despite written and oral instruction, only 7 of 24 (29.2%) patients had consistent bladder volume, with a variability of <50 ml between both full bladder CTs. 12 of 24 (50%) patients achieved bladder volumes in control CT that were greater than the volumes at the initial full-bladder simulation. However, despite this great variability, the median bladder volume in both CTs was similar.

We analysed the correlations between bladder volumes, the ratio and differences between bladder volumes in planning CT and control CT. We did not find any significant correlation between these parameters and the 3D vector or loss of dosimetric coverage.

4.1.2. Rectal volumes

The results of the analysis of rectal volumes can be seen in Table 3.

Parameters such as mean and median of the rectal volume and maximum rectal diameter in both CTs were similar. We did not note a decrease in rectal filling over time, even in those patients who presented GI toxicity.

We did not find any significant correlation between rectal parameters and 3D vector or loss of dosimetric coverage. We only found a trend between the maximum rectal diameter in planning CT and 3D vector that can be seen in Fig. 3 (Pearson correlation 0.400, 95% CI: −0.004 to 0.692, p = 0.0529). However, we did not find correlation between the loss of coverage and the maximum rectal diameter (Fig. 4A and B) in any of the CTs.

Fig. 3.

3D vector (cm) vs maximum rectal diameter in planning CT full bladder (p-CT).

Fig. 4.

ITV dosimetric loss of coverage (%) vs maximum rectal diameter in planning CT full bladder (p-CTFULL) (A) and control CT full bladder (c-CTFULL) (B).

Finally, we did not find any correlation between changes in bladder volume in the planning CT and control CT, and rectal volume change.

5. Discussion

Adjuvant treatment with IMRT for the gynaecological neoplasms has proven to be better than conventional treatment with conformal radiotherapy on the basis of lower morbidity without compromising control of the disease. Even so, with IMRT, the motion of internal organs can cause dose gradients around the CTV, so it may be displaced from the high dose area in some treatment fractions. The vaginal CTV is located between the bladder and rectum and is therefore subject to the movement of these organs during treatment. This movement varies greatly from patient to patient, meaning that it is necessary to customise strategies for managing movement during treatment.

Accordingly, from the first implementation of adjuvant IMRT in pelvic gynaecological cancer on a multi-institutional level, the RTOG recommended creating a vaginal ITV delineated on CT’ with empty and full bladder. With the aim of maintaining constant bladder volume during treatment, patients were instructed to ingest a consistent volume of water 60 min before each treatment session.

With the widespread use of adjuvant IMRT in gynaecological tumours we analysed our data regarding the control of vaginal motion and vaginal ITV strategy.

Jhingran et al.,¹⁸ using the ITV procedure, obtained twice weekly CT scans during treatment of 16 patients and reported median shifts of 0.64 cm in the AP dimension, 0.62 cm in the SI dimension, and 0.23 cm in the right-left plane. In 11 of 16 cases the shifts of the vaginal apex correlated with changes in the volume of the bladder and rectum. We did not find any correlation between bladder volumes and 3D vector as surrogate of ITV vaginal movement.

Rash et al.²⁰ reported the impact of rectal filling on 145 images from 5 patients treated with adjuvant IMRT for endometrial or cervical cancer, and ITV procedure used at simulation. The mean rectal distention was 3.7 cm (SD 1.8 cm), ranging from 2.05 cm to 6.01 cm. They established a correlation between rectal distention and AP movement of the vaginal cuff marker. To prevent rectal motion, some authors have recently used rectal balloons during radiotherapy. Taku et al.²¹ reported the vaginal motion on 18 patients treated with ITV procedure at simulation and placement of an endorectal balloon (ERB) during radiotherapy. They found an inter-treatment fiducial A/P movement with a mean of 2.8 mm (SD = 2.5 mm, range 0–11.8 mm) and suggested that the stability in rectal volumes was provided by ERB placement. They recommend treatment with ERB for patients to benefit from smaller A/P margins and potentially reduced toxicity to the rectum. Our patients, without instructions for rectal filling and without ERB, achieved similar or slightly higher shifts in A/P movement (mean of 4.1 mm, SD = 3.1 mm, range 0.3–12.6 mm). We found a weak correlation between the maximum diameter of the rectum in the planning CT and 3D vector. We did not find correlations with any of the rectal measurements and loss of dosimetric coverage. According to our results, we do not consider it necessary to use the ERB for patients that have a high probability of having small bowel toxicity at any level during radiotherapy.

Some authors solve the problem of organ movements with generous margins, but the PTV margins are another field of controversy. The RTOG recommended 0.7 cm in its Phase II trial, but does not give specifications regarding the margin in the closed Phase III trial and the consensus guidelines recommend either 1–1.5 cm, or for each institution to decide for itself.

Chopra et al.²² reported that 4.1 mm, 10.6 mm and 10.3 mm, for left-right, craniocaudal and AP margins respectively, could encompass 95% of the observed displacements. Shih et al.²³ reported clinical outcomes in endometrial cancer and Folkert MR et al.²⁴ in cervical cancer treated with adjuvant pelvic IMRT. They used vaginal contrast to improve the visualisation of the vaginal cuff allowing a better contour of the organ. It was expanded by 2 cm to generate the vaginal cuff CTV. An additional 1 cm expansion was applied to create the vaginal cuff PTV. On the other hand, in the French study,¹⁵ the PTV was generated from the CTV by adding a 7 mm margin without ITV procedure, which can cause a loss of vaginal coverage. They tried, however, to counterbalance this risk of vaginal infradosification with a brachytherapy boost.

Our study of 24 patients, even understanding the small sample size and the observational study design, allowed us to vary our care practice, allowing patients to go to daily treatment sessions with comfortable bladder and without firm instructions, thus avoiding patient discomfort. The fact that we found no clear relationship between changes in rectal volume and the loss of coverage, has allowed us to continue not giving instructions related to the rectum.

As we found a weak correlation between the maximum diameter of the rectum in the p-CT and 3D vector, we should consider the possibility of applying a non-isotropic margin, in A/P direction, to those patients with major rectal maximum diameter on simulation. Notwithstanding, the conclusions of our study must be considered within the context of its limitations, especially regarding the number of CT scans per patient.

6. Conclusions

Although vaginal movement was observed for the majority of patients, only a few of them showed a significant loss of coverage as a result of this movement. The absence of correlation between bladder volumes and vaginal ITV supports the case for using ITV procedure to ensure adequate dose coverage, even so, we advise patients to maintain comfortable bladder during treatment without fixed instructions, as following a strict bladder-filling protocol does not necessarily achieve stable bladder volumes. Without instructions regarding rectal-filling, we found a weak correlation between rectal volume and 3D vector, allowing us to infer that the diameter of the rectum could be a predictor of ITV vaginal motion.

Conflict of interest

None declared.

Financial disclosure

None declared.