Optimal bladder and rectal filling for external beam radiation therapy for prostate cancer is a key topic of interest (11–13). Full bladder protocols have the potential to reduce OAR dose exposure via displacement of the small bowel and bladder away from the target volume, with the caveat of increased volume variability (14–16). Similarly, smaller rectal volumes can reduce dose exposure, but are more difficult to reproduce. Although we could infer from the findings of conventional EBRT cohort evaluation, to our knowledge, there has been no prior evaluation or report of interfraction OAR motion and the dosimetric implications in patients undergoing image-guided prostate SBRT.
Despite instructions to maintain a consistently full bladder at simulation and during treatment, we observed volume variability and a systematic decline in bladder volume through the course of therapy with the largest reduction at the 5th fraction. We hypothesize that this could be related to either radiation-related bladder capacity changes due to acute inflammation, increased efficiency of patient setup, and/or diminished adherence to the filling protocol as the treatment course progressed. Within the cohort analyzed, pre-treatment urinary dysfunction reflected by IPSS score was not associated with decline in bladder volume during treatment.
Several past evaluations of patients treated with conventionally fractionated EBRT with full bladder protocols have reported similar results (11, 17, 18). For instance, a prospective study by Nakamura et al. (2010) showed a significant decline of 38% in bladder volume from fraction #1 to #30 (17). To develop a feedback mechanism for consistent interfraction bladder volumes, Stam et al. (2006) used daily bladder scans to provide patients with verbal feedback on fluid consumption to improve bladder filling (18). While there was a moderate improvement in volume reduction in the bladder scan group (19% decline) compared to the control group (31%), the authors concluded that the difference was not sufficient to deem their technique clinically applicable. In a series of 41 patients undergoing conventionally fractionated EBRT, O’Doherty et al. (2006) found that using written instructions resulted in more consistent bladder filling throughout treatment, with an association between patient’s subjective bladder fullness and actual bladder volume (19). Thus, written instructions may help patients maintain a moderately consistent bladder volume for prostate SBRT as well.
Our results, however, call into question the need for such strict bladder filling instructions when using modern radiotherapy techniques that employ more accurate MRI-based target delineation and smaller PTVs facilitated by advanced inter- and intra-fraction image guidance. While significant, it is uncertain if our observation of increased mean dose exposure to the bladder with a decline in bladder filling would ultimately translate to clinically significant GU sequelae in a prospective setting. Our retrospective assessment of the available acute and late GI and GU toxicities data certainly did not reveal a clear correlation to bladder volume and displacement variations.
To date, there have been relatively few defined dose-volume relationships for prostate SBRT that have predicted greater significant GU toxicity. Repka et al. (2014) found bladder wall D15.5% >32.6 Gy to be significantly associated with acute urinary toxicity (20). Therefore, the volume of the bladder receiving high doses might be more relevant than the mean dose. In the current study, bladder D2cc did not change significantly over the course of treatment, suggesting that the area of bladder exposed to high doses was similar to what we expected based on the DVHs generated during treatment planning. Additionally, anatomical sub-sites of the bladder or urethra (i.e., bladder trigone/neck or membranous urethra) could be relatively more important to avoid with high doses of radiation (21, 22). Analysis of the bladder trigone was of specific interest given our previously published analyses of patients treated with conventionally fractionated EBRT and low-dose rate brachytherapy, which revealed a strong association between urinary toxicity and bladder neck dose exposure (21, 22). Overall bladder volume reduction over the course of SBRT observed in the current study was accompanied by a displacement of the trigone anteriorly and away from areas of highest dose exposure. This phenomenon may be explained by collapsing of the overextended bladder walls in the axial plane with a reduction in urine volume between fractions, as illustrated by the inferior shift in the center of mass of the bladder from simulation to the end of treatment. Anterior displacement of the bladder trigone may therefore represent a surrogate for changes in the surface area of the posterior bladder wall. While previous studies regarding bladder filling emphasized the importance of reproducibility of a full bladder to reduce urinary toxicity, the bladder volume reduction observed here and increased mean bladder dose may minimally contribute to toxicity if the D2cc remains unchanged and the trigone is simultaneously spared potentially deleterious dose exposure (23, 24). Clearly further exploration of the optimal bladder filling for simulation and treatment in the setting of a highly precise treatment, such as prostate SBRT, is needed.
Although not as substantial as bladder volume reductions, rectal volume also showed a decreasing trend over the course of the 5 fraction SBRT, with a significant but small increase in mean dose exposure to the rectal wall but decrease in D2cc. These findings are corroborated by studies in conventionally fractionated EBRT involving full bladder protocols, in which rectal volume reductions were observed through the course of therapy (12, 15, 25). While the literature concerning rectal motion in SBRT patients is limited, a recent dosimetric study suggested that rectum volume receiving 75% of the prescribed dose can increase significantly during SBRT due to interfraction organ motion (26). As in bladder dose volume studies, prior studies of rectal motion in relation to dose exposure involved conventionally fractionated EBRT employing more generous PTV margins and substantially greater number of fractions over several weeks and therefore may not be applicable to more precise ultra-hypofractionated treatments, especially considering the increased use of rectal spacers to mitigate rectal dose. Furthermore, while there are specific dose constraint guidelines for the rectum in conventional EBRT and brachytherapy, prospectively validated dose constraint guidelines for ultra-hypofractionated SBRT have yet to be established (27, 28).
Based on an evaluation of rectal tolerance in a Phase I/II trial, Kim et al. (2019) suggested that limiting 35% rectal circumference exposure to less than 39 Gy over 5 fractions may reduce the risk for delayed rectal toxicity (29). Therefore, while studies have reported acceptable levels of rectal toxicity with prostate SBRT and our current assessment yielded nonexistent-to-exceedingly low rates of GI toxicities, ongoing prospective trials with long-term follow up must be evaluated before we can fully understand the dose-volume relationships and the potential impact of interfractional organ motion (1, 5, 8, 30–33).
The key strengths of this study include its use of an objectively verified and consistent process of organ delineation, as well as its analysis of a large number of CT images. A limitation of this study was the inferior image quality produced by the CBCTs, which might have resulted in errors in organ delineation. Similarly, other artifacts due to motion and metal may have also resulted in contouring errors. Provided that a single, trained reviewer performed the organ delineations with a high degree of consistency, any major errors were likely mitigated. Another limitation is the retrospective nature of the study and the absence of detailed toxicity outcomes to analyze in the context of our organ motion and dosimetric data. In terms of measuring organ motion, an indirect measurement of center of mass changes to contoured volumes had to be employed due to limitations in geometric data generated on the treatment planning system.
Lastly, even with consistent organ delineation by a single reviewer and a large number of CT images analyzed, the post hoc nature of using clinical CBCTs invariably represents an important limitation. Namely, the field of view (FOV) of a small number of the CBCTs were optimized for visualization of fiducial markers and pelvic bony landmarks, while minimizing superior and inferior borders to limit image acquisition time and unwarranted radiation exposure. Inevitably, the superior borders of bladders in 69 of 425 CBCTs (16.2%) throughout all fractions and 10 of 85 fraction #5 CBCTs (11.8%) were out of FOV to be precisely contoured. This was due to the FOV not extending superiorly to account for overfilled bladders at the time of image acquisition (average volumes of bladder out of CBCT FOV: 414.5 ± 116.8 mL; average volumes of bladder within CBCT FOV: 195.8 ± 105.1 mL). The degree of bladder overfilling in these CBCTs were remarkable enough to limit any major changes to the overall pattern of bladder volume decline, COM, or dosimetry to warrant exclusion of patients from analyses.