Advertisement

Quality Metric to Assess Adequacy of Hydrogel Rectal Spacer Placement for Prostate Radiotherapy and Association of Metric Score with Rectal Toxicity Outcomes

Open AccessPublished:January 23, 2023DOI:https://doi.org/10.1016/j.adro.2022.101070

      ABSTRACT

      Purpose

      : Although hydrogel spacer placement (HSP) minimizes rectal dose during prostate cancer radiotherapy, its potential benefit for modulating rectal toxicity could depend upon the achieved prostate-rectal separation. We therefore developed a quality metric associated with rectal dose reduction and late rectal toxicity among patients treated with prostate stereotactic body radiotherapy (SBRT).

      Materials and Methods

      : A quality metric consisting of prostate-rectal interspace (PRI) measurements from axial T2-weighted MRI simulation images was applied to 42 men enrolled on a multi-institutional phase II study using HSP with prostate SBRT (45 Gy in 5 fractions). A score of 0, 1, or 2 was assigned to a PRI measurement of <0.3 cm, 0.3-0.9 cm, or ≥1 cm, respectively. An overall spacer quality score (SQS) was computed from individual scores at rectal midline and ±1 cm laterally, located at the prostate base, mid-gland, and apex. Associations of SQS with rectal dosimetry and late toxicity were evaluated.

      Results

      : The majority of the analyzed cohort had an SQS of 1 (n=17; 41%) or 2 (n=18; 43%). SQS was associated with maximum rectal point dose (rectal Dmax; p=0.002), maximum dose to 1 cc of rectum (D1cc; p=0.004), and volume of rectum receiving ≥100% of prescription dose (V45; p=0.046) and ≥40 Gy (V40; p=0.005). SQS was also associated with a higher incidence of (p=0.01) and highest-graded late rectal toxicity (p=0.01). Among the 20 men who developed late grade ≥1 rectal toxicity, 57%, 71%, and 22% had an SQS of 0, 1, and 2, respectively. Men with an SQS of 0 or 1 compared to 2 had 4.67-fold (95% CI: 0.72-30.11) or 8.40-fold (95% CI: 1.83-38.57), respectively, greater odds of developing late rectal toxicity.

      Conclusion

      : We developed a reliable and informative metric for assessing HSP, which appears to be associated with rectal dosimetry and late rectal toxicity after prostate SBRT.

      INTRODUCTION

      Given the anatomic proximity of the prostate to normal tissues such as the rectum, acute and late toxicities are possible with prostate cancer external beam or brachytherapy treatments. One such strategy to minimize toxicity is the insertion of a commercially supplied polyethylene glycol (PEG) hydrogel rectal spacer, a polymerization solution injected transperineally under ultrasound-guidance to temporarily distance the rectum away from the prostate gland. This rectal spacer remains intact throughout a typical course of radiotherapy and is hydrolyzed 3-6 months after placement. The first prospective randomized clinical trial to evaluate the impact of the hydrogel rectal spacer in prostate cancer patients treated with conventionally-fractionated external beam radiotherapy (EBRT) established its ability to increase the mean prostate-rectal separation by 1 cm; this translated into a lower rectal dose and reduced incidence of late grade 1-2 rectal toxicity [1, 2]. A systematic review and meta-analysis of 7 studies comparing over 1000 prostate cancer patients treated with dose-escalated EBRT and/or brachytherapy with or without hydrogel rectal spacer placement (HSP) reported a 66% V70 rectal dose reduction with a corresponding decline in acute (RR 0.72) and late (RR 0.38) grade 1 and higher rectal toxicity, as well as superior patient-reported >3 month bowel quality of life in the presence of the spacer hydrogel [3].
      Attempts have been made to characterize the quality of HSP in order to accurately assess its true dosimetric- and toxicity-reduction impact given the variability of reported benefit across individual studies. However, these published HSP quality metrics [4-6] have not been universally adopted, as they are somewhat cumbersome to use and were not consistently evaluated for their utility in predicting late rectal toxicity. A secondary analysis of HSP quality in the pivotal phase III spacer trial demonstrated that 49% of patients had symmetric hydrogel rectal spacers in all three axial slices analyzed, with a correlation between inferior composite symmetry scores and smaller rectal dose reduction relative to the pre-spacer plan [4]. Although informative, this metric does not assess the actual thickness of the prostate-rectal interspace (PRI) created by the spacer gel, which is more relevant to rectal dosimetry.
      Since clinical observations have suggested the spacer's ability to reduce rectal dose is dependent on the quality of the implant, a universally adopted practical metric is needed. We therefore created a HSP quality metric that quantitatively characterizes the adequacy of the PRI at the prostate base, mid-gland, and apex, which also has the benefit of showing association with reductions in rectal dosimetry and late toxicity in men treated with prostate stereotactic body radiotherapy (SBRT).

      MATERIALS AND METHODS

      Derivation of hydrogel rectal spacer quality metric

      As previously described [7], all patients had a 10 cc polyethylene hydrogel spacer (SpaceOAR System, Boston Scientific, Inc., Bedford, MA) introduced transperineally into the perirectal space between Denonvilliers’ fascia and the anterior rectal wall under ultrasound guidance approximately 1–2 weeks prior to their MRI simulation session. Using the T2-weighted axial slices from the MRI simulation, the interspace between the posterior aspect of the prostate gland and the anterior rectal wall was measured in the anterior-posterior orientation parallel to anatomic midline. This total prostate-rectal interspace (PRI) thickness consisted of natural tissue space (i.e., fat, soft tissue, and/or neurovasculature if present) and the implanted hydrogel rectal spacer (which is hyperintense on T2-weighted images). Measurements were made at rectal midline (along the left-right direction), as well as 1 cm to the right and left, as demonstrated in Figure 1a; note, rectal midline did not necessarily correspond to the middle of the prostate gland (along the left-right direction). To obtain a representation of the PRI throughout the length of the prostate gland, measurements were made at the “base” (axial slice located 0.5–0.8 cm caudal to the most superior slice comprising the prostate base, excluding the median lobe within the bladder), “mid-gland” (axial slice located midway between the prostate “base” and ”apex”), and “apex” (axial slice located 0.5–0.8 cm cranial to the most inferior slice comprising the prostate apex), as illustrated in Figure 1b. Because of the occasional tortuous shape of the rectum, rectal midline was sometimes in a different location at the base, mid-gland, or apex slice.
      Figure 1
      Figure 1(a) Depiction of prostate-rectal interspace (PRI) thickness of hydrogel rectal spacer implant and natural space (i.e., fat, soft tissue, and/or neurovasculature) between the anterior rectal wall and posterior prostate (or seminal vesicles, whichever is more posterior), measured in the anterior-posterior direction on MRI T2-weighted axial slices. PRI measurements were obtained at rectal midline (which may not necessarily be anatomic midline), as well as 1 cm to the right and left of rectal midline. (b) Depiction of base, mid-gland, and apex PRI measurements along rectal midline.
      PRI measurements (rounded to one decimal place) were assigned a value between 0-2 and entered into a 3 × 3 table. These numerical categories were defined as follows: a value of “0” for PRI measurements <0.3 cm, “1” for PRI measurements between 0.3–0.9 cm, and “2” for PRI measurements ≥1.0 cm (Figure 2). (Figure 2. A cut-off of 0.3 cm was chosen because this is a typical minimum planning target volume (PTV) posterior expansion for prostate SBRT treatments; a PRI of less than 0.3 cm would therefore result in overlap of the PTV with the rectum. A minimum threshold of 1.0 cm was used to define PRI category “2” because publications [3] have reported this value as the approximate mean perirectal space created by the hydrogel rectal spacer, and thus should be achievable with an optimal implant. The scores within the 3 × 3 table easily provide a synopsis of PRI symmetry along the left-right axis; adequacy of separation between the prostate and rectum at the base, mid-gland, and apex; and deficiencies of HSP whether there are areas that the hydrogel rectal spacer could have been implanted to increase prostate-rectal separation.
      Figure 2
      Figure 2Example of a summary table of PRI measurements.
      We then created a hydrogel rectal spacer quality metric score for each patient, using the following rules (Figure 3):
      • -
        Step 1, define a PRI summary score for each table row (i.e. base, mid-gland, and apex) of a patient:
        • If at least two cells within a table row have a value of “2”, then assign that table row a summary score of “2”.
        • If at least two cells within a table row have a value of “0”, then assign that table row a summary score of “0”.
        • Otherwise, a summary score of “1” is assigned to the table row.
      • -
        Step 2, define the PRI overall spacer quality score (SQS) for a patient as the lowest value of the three overall row PRI scores.
      Figure 3
      Figure 3Rules for generating an overall spacer quality score (SQS) from the 3 × 3 PRI thickness table.

      Patient cohort

      The association between this SQS and late rectal toxicity was retrospectively evaluated in 42 men with low (11.6%) or intermediate (88.4%) risk prostate cancer enrolled on a multi-institutional phase II study using HSP with prostate SBRT to assess toxicity and tumor control[8]. One of the patients included in Folkert, et al's, original paper[8] was ineligible for this secondary analysis study due to the absence of follow-up data; our study cohort of 42 men had a median follow-up of 48 months. Patients had prostate glands ≤80 cc (2 patients had prior cytoreduction with bicalutamide), baseline American Urological Association Symptom Index (AUA-SI) score ≤18, no prior transurethral resection (TURP) or cryotherapy of the prostate, and no history of inflammatory bowel disease (IBD). Greater than 95% of the PTV (a 0.3 cm isocentric expansion of the contoured prostate gland) received a total dose of 45 Gy over 5 fractions, delivered in 2–3 fractions per week on non-consecutive days; concurrent or adjuvant ADT was not allowed. Physician-reported late rectal toxicity (National Cancer Institute Common Toxicity Criteria for Adverse Events, NCI-CTCAE v4.0) was assessed up to 60 months, and the presence of rectal wall ulcers on anoscopies or sigmoidoscopies were assessed serially through 9 months post-SBRT. We then applied the SQS to all patients’ HSP by measuring the PRI on the MRI simulation scans.

      Statistical analysis

      Using standard procedures, we calculated descriptive statistics, including median and ranges for continuous parameters, as well as percentages and frequencies for categorical variables. Differences in various baseline factors (including dosimetry and late rectal toxicity) between SQS groups were evaluated using Kruskal-Wallis tests for continuous factors and Fisher's exact tests for categorical factors. Simple linear regression modeling was performed to assess associations between rectal dosimetric measures and the SQS. Univariable logistic regression was employed to evaluate the association between late rectal toxicity and the independent variable, SQS. The SQS was analyzed as a three-level variable (0 is worst, 2 is best). Coefficients of determination (R2) were computed for each model. To evaluate the predictive power of the SQS, we generated a univariable (UVA) logistic model for each score under consideration: SQS with three levels (0, 1, and 2), SQS dichotomized as 0-1 versus 2, and SQS dichotomized as 0 versus 1-2. We computed Nagelkerke's adjusted pseudo R-square values to assess model fit, and qualitatively compared them between models. All statistical computations were performed and generated using SAS Software Version 9.4 (The SAS Institute, Cary, NC).

      RESULTS

      Rectal dosimetry and late rectal toxicity

      Among the 42 men whose hydrogel spacer implants were evaluated in this study, 35 (83%) and 7 (17%) were enrolled by the University of Texas Southwestern Medical Center (UTSW) and Memorial Sloan Kettering Cancer Center (MSKCC), respectively. As reported in Table 1, among these patients who were treated to 45 Gy with SBRT in 5 fractions, the median and mean maximum dose to 1 cc of rectum (rectal D1cc) was 3337.5 (range, 1473.4-4620.0) and 3369.8 (SD, 782.8) cGy, respectively. The mean volume of rectum receiving 100% of the prescription dose (V45 Gy) was 0.2 (SD, 0.4) cc. Over a median follow-up of 48 months, 20 men (48%) developed a late rectal toxicity, consisting of 14 and 6 patients with grade 1 and 2 as the highest reported late rectal toxicities, respectively; there were no grade 3 or higher late rectal toxicities.
      Table 1Rectal dosimetry and late rectal toxicity
      Total Cohort n = 42
      Rectal Dmax (cGy)Median (range)4418.0 (2927.6-5089.7)
      Mean (SD)4288.8 (541.0)
      Rectal D0.035cc (cGy)Median (range)4182.8 (2710.0-4990.0)
      Mean (SD)4068.1 (619.4)
      Rectal D1cc (cGy)Median (range)3337.5 (1473.4-4620.0)
      Mean (SD)3369.8 (782.8)
      Rectal V45Gy (cc)Median (range)0.0 (0.0-1.8)
      Mean (SD)0.2 (0.4)
      Rectal V40Gy (cc)Median (range)0.1 (0.0-6.5)
      Mean (SD)0.8 (1.3)
      Rectal V30Gy (cc)Median (range)2.5 (0.0-22.3)
      Mean (SD)3.3 (4.0)
      Presence of late rectal toxicityYes20 (47.6)
      No22 (52.4)
      Late rectal toxicity (highest grade reported)Median (range)0 (0-2)
      Mean (SD)0.6 (0.7)
      022 (52.4)
      114 (33.3)
      26 (14.3)
      Abbreviations: SD = Standard Deviation, Rectal Dmax = maximum rectal point dose (cGy), Rectal D”X”cc = maximum dose (cGy) to X cc of rectum, and Rectal V”X”Gy = volume (cc) of rectum receiving at least “X” Gy.

      Prostate-rectal interspace summary score and overall spacer quality score

      As Table 2 shows, all patients had at least 0.3 cm of space between the prostate and rectum at the level of the prostate mid-gland (i.e., PRI mid-gland summary score of 1-2). The majority of men had a PRI summary score of 2 at the base (74%), mid-gland (83%), and apex (64%). When a composite score consisting of individual PRI summary scores from the base, mid-gland, and apex was computed (defined as the SQS), 7 (17%), 17 (40%), and 18 (43%) men in this cohort had scores of 0, 1, and 2, respectively.
      Table 2Prostate-Rectal Interspace Scores and Overall Spacer Quality Score
      Total Cohort n = 42
      PRI Base summary scoreMedian (range)2 (0-2)
      Mean (SD)1.7 (0.6)
      0, N (col %)2 (4.8)
      1, N (col %)9 (21.4)
      2, N (col %)31 (73.8)
      PRI Mid-gland summary scoreMedian (range)2 (1-2)
      Mean (SD)1.8 (0.4)
      1, N (col %)7 (16.7)
      2, N (col %)35 (83.3)
      PRI Apex summary scoreMedian (range)2 (0-2)
      Mean (SD)1.5 (0.7)
      0, N (col %)5 (11.9)
      1, N (col %)10 (23.8)
      2, N (col %)27 (64.3)
      SQSMedian (range)1 (0-2)
      Mean (SD)1.3 (0.7)
      0, N (col %)7 (16.7)
      1, N (col %)17 (40.5)
      2, N (col %)18 (42.9)
      Abbreviations: Prostate-Rectal Interspace Score = PRI, Overall Spacer Quality Score = SQS, and SD = Standard Deviation.

      Associations between SQS and rectal dosimetric and late toxicity outcomes

      All prostate SBRT rectal dosimetric measures analyzed (rectal Dmax, D1cc, V45Gy, and V40Gy) were associated with SQS, as shown in Table 3. Treatment planning scans with an SQS of 2 exhibited a lower (p=0.002) rectal Dmax (median of 3920.8 cGy, range 2927.6-4787.9) compared to a score of 0 (median of 4683.3 cGy, range 3639.0-4896.2) or 1 (median of 4575.6 cGy, range 3839.6-5089.7). We observed evidence that a higher SQS was associated with superior rectal dosimetry using simple linear regression modeling for rectal Dmax (p <0.001) and D1cc (p =0.002), as shown in Table 4. For example, a patient with an SQS of 1 had a rectal Dmax 661.8 cGy higher (95% CI 362.6-961.0) than a patient with an SQS of 2, and a patient with an SQS of 0 had a rectal Dmax 517.5 cGy higher (95% CI 123.4-911.6) than a patient with an SQS of 2. Similarly, a patient with an SQS of 1 had a rectal D1cc 798.3 cGy higher (95% CI 347.3-1249.3) than a patient with an SQS of 2, and a patient with an SQS of 0 had a rectal D1cc 894.3 cGy higher (95% CI 300.3-1488.3) than a patient with an SQS of 2.
      Table 3Associations between SQS and rectal dosimetry or late rectal toxicity
      Percentages for specific SQS groups are row percentages. Abbreviations: Rectal Dmax = maximum rectal point dose (cGy), Rectal D1cc = maximum dose (cGy) to 1 cc of rectum, and Rectal V”X”Gy = volume (cc) of rectum receiving at least “X” Gy.
      SQS Score
      0 n = 7 (16.7%)1 n = 17 (40.5%)2 n = 18 (42.9%)p-value
      Kruskal-Wallis or Fisher's exact test.
      Rectal Dmax (cGy)Median (range)4683.3 (3639.0-4896.2)4575.6 (3839.6-5089.7)3920.8 (2927.6-4787.9)0.002
      Rectal D1cc (cGy)Median (range)4340.0 (2755.0-4440.0)3720.0 (2493.9-4620.0)2884.0 (1473.4-4290.0)0.004
      Rectal V45Gy (cc)Median (range)0.4 (0.0-0.8)0.0 (0.0-1.8)0.0 (0.0-0.3)0.046
      Rectal V40Gy (cc)Median (range)1.8 (0.0-2.4)0.4 (0.0-6.5)0.0 (0.0-2.5)0.005
      Presence of late rectal toxicityYes4 (57.1)12 (70.6)4 (22.2)0.01
      No3 (42.9)5 (29.4)14 (77.8)
      Late rectal toxicity (highest grade reported)Median (range)1 (0-1)1 (0-2)0 (0-2)0.01
      1 Kruskal-Wallis or Fisher's exact test.
      2 Percentages for specific SQS groups are row percentages.Abbreviations: Rectal Dmax = maximum rectal point dose (cGy), Rectal D1cc = maximum dose (cGy) to 1 cc of rectum, and Rectal V”X”Gy = volume (cc) of rectum receiving at least “X” Gy.
      Table 4Simple linear regression modeling of SQS as the independent variable
      Dependent variableSQS ValueN (%)β (95% CI)p-value
      Overall ANOVA. Abbreviations: Rectal Dmax = maximum rectal point dose (cGy), Rectal D1cc = maximum dose (cGy) to 1 cc of rectum, and Rectal V”X”Gy = volume (cc) of rectum receiving at least “X” Gy
      R2
      Rectal Dmax (cGy)218 (42.9)Reference<0.0010.34
      117 (40.5)661.806 (362.562, 961.049)
      07 (16.7)517.520 (123.393, 911.647)
      Rectal D1cc (cGy)218 (42.9)Reference0.0020.28
      117 (40.5)798.337 (347.342, 1249.332)
      07 (16.7)894.344 (300.348, 1488.340)
      Rectal V45Gy (cc)218 (42.9)Reference0.060.13
      117 (40.5)0.267 (0.017, 0.518)
      07 (16.7)0.323 (-0.007, 0.653)
      Rectal V40Gy (cc)218 (42.9)Reference0.090.12
      117 (40.5)0.820 (-0.004, 1.644)
      07 (16.7)0.990 (-0.095, 2.075)
      1 Overall ANOVA.Abbreviations: Rectal Dmax = maximum rectal point dose (cGy), Rectal D1cc = maximum dose (cGy) to 1 cc of rectum, and Rectal V”X”Gy = volume (cc) of rectum receiving at least “X” Gy
      We observed an association between SQS and incidence of late rectal toxicity (p=0.01), as illustrated in Table 3: 57%, 71%, and 22% of men with a score of 0, 1, and 2, respectively, developed late grade 1-2 rectal toxicity. Further, men with an SQS of 0 or 1 had a 4.67-fold (95% CI: 0.72-30.11) and 8.40-fold (95% CI: 1.83-38.57) greater odds of a late rectal toxicity compared to a score of 2 using univariable logistic regression modelling (Table 5). There was a weaker statistical association between late rectal toxicity and the individual PRI row summary scores at the base (p=0.06), mid-gland (p=0.04), and apex (p=0.15), supporting the utility of this composite SQS which accounts for the entire length of prostate-rectum overlap within the treatment field.
      Table 5Univariable logistic regression of SQS (as binary or 3-tier scores) with late rectal toxicity
      SQSN (# Events)Presence of late rectal toxicity OR (95% CI)p-value
      : Based on UVA logistic regression
      R-square
      : Nagelkerke's adjusted pseudo R-square
      (0 or 1) vs. 2218 (4)Reference0.020.25
      07 (4)4.67 (0.72, 30.11)
      117 (12)8.40 (1.83, 38.57)
      0-1 vs. 2218 (4)Reference0.0060.24
      0-124 (16)7.00 (1.73, 28.34)
      0 vs. 1-21-235 (16)Reference0.580.01
      07 (4)1.58 (0.31, 8.15)
      1 : Based on UVA logistic regression
      2 : Nagelkerke's adjusted pseudo R-square
      Based on these observations that an SQS of 2 (compared to 0 or 1) was associated with superior rectal dosimetry and less late rectal toxicity, we also assessed whether the SQS should be reported as a 3-tier (0 versus 1 versus 2) or binary score (0-1 versus 2, or 0 versus 1-2). Based on qualitative comparison of the Nagelkerke's adjusted pseudo R-square values shown in Table 5, we found that the SQS reported as a 3-tier (R2=0.25) or binary score of 0-1 versus 2 (R2=0.24) were similar in terms of predictive power of grade 1 and higher late rectal toxicity. In comparison, a binary SQS of 0 vs. 1-2 demonstrated weak predictive power of late rectal toxicity (R2=0.01).

      DISCUSSION

      By accounting for the prostate-rectal interspace (PRI) thickness created by the hydrogel rectal spacer at the level of the prostate base, mid-gland, and apex at midline and 1 cm laterally peripherally, we have created an overall spacer quality score (SQS) that is associated with rectal dosimetry and late rectal toxicity after treatment with SBRT for prostate cancer. Forty three percent of men had an SQS of 2, highlighting that the highest SQS score is achievable if the insertion of the hydrogel rectal spacer is performed by experienced users in a methodical fashion. An SQS of 2 not only was associated with lower radiation doses to the rectum (as measured by Dmax, D1cc, V45 Gy, and V40 Gy), but also significantly fewer incidences of late rectal toxicity compared to a score of 0-1 (22% versus 67%).
      Although previously published HSP metrics have been widely adopted, we present a robust and practical instrument that both predicts for reduced rectal dose and late rectal toxicity. Fischer-Valuck, et al, developed a symmetry score in which more asymmetric implants correlated with less reductions in rectal dose compared to the treatment plan before insertion of the rectal spacer [4]. However, hydrogel rectal spacer symmetry (relative to the prostate midline on axial images) does not always equate to the actual thickness of the PRI in the anterior-posterior direction which, as we demonstrate with our metric, is associated with reduced rectal dose. For example, a patient could have a symmetric implant, but only minimal (<0.3 cm) space created between the prostate and rectum and therefore derive negligible dosimetric and toxicity benefit. In the Fischer-Valuck, et al, analysis of HSP quality in the published phase III spacer trial[1, 2], all but the most severe asymmetric implants resulted in decreased rectal dose. This implies that inadequate separation between the prostate and rectum accounted for less dose-sparing to the rectum; less symmetry did result in a smaller rectal dose reduction, however. There was also no attempt by the authors to evaluate association of their symmetry score with rectal toxicity, highlighting the utility of our metric.
      Hwang, et al.[5], utilized a more complicated model that accounted for implant symmetry in a 3-dimensional orientation and hydrogel rectal spacer volume. However, they did not identify an association between their metric and rectal toxicity in men treated with prostate SBRT [5]. The authors attributed this potentially to a small sample size of 22 patients, the moderate SBRT dose of 36.25 Gy delivered in 5 fractions, and inadequate length of follow-up. However, we observed a similar rate of late grade 1-2 rectal toxicity (48% in our versus 30% in the Hwang, et al. studies) in a small sample size of 42 men that was still associated with late rectal toxicity. A mathematical equation consisting of four component equations (hydrogel volume, symmetry in the left-right and cranial-caudal orientations, and prostate-rectal space at mid-gland) was derived by Liu, et al., to quantify HSP quality [6]; no meaningful associations, however, were correlated with this score as well.
      Increasing SBRT doses delivered to the prostate have been correlated with lower 2-year post-treatment positive biopsy rates [9] and superior biochemical control rates [10]. Thus, such measures used to attenuate dose to the rectum by anatomically separating the prostate and rectum via an implantable hydrogel rectal spacer are even more important. In a retrospective study from MSK of 551 men treated with SBRT for prostate cancer, those who had hydrogel rectal spacers implanted prior to treatment developed fewer late grade 2 and higher rectal toxicity (1%) compared to those without an implant (6%) [11].
      Since studies using the hydrogel rectal spacer have not reported on the quality of the implant, it will be difficult to completely understand its benefits if a uniform QA metric is not widely employed. Poor implants, which would be defined by an SQS of 0, will unnecessarily have placed a patient at risk of complications from the procedure without providing a reduced risk of developing toxicity from the actual treatment. Our SQS metric has the benefit of quantifying the total space between the prostate and rectum, which includes both the natural anatomic space and the additional space created by the hydrogel implant. It has been previously reported that fewer post-brachytherapy rectal complications occur with larger natural space present between the prostate and rectum [12]. Therefore, in men who have favorable anatomy in certain areas along the PRI, it could potentially be decided before implantation where the hydrogel rectal spacer should be focally inserted to enhance this separation.
      Further, we anticipate SQS's utility in not only reporting late rectal toxicity with external beam and brachytherapy prostate treatments in the presence of HSP, but also as a self-improvement tool for radiation oncologists to identify areas for improvement in implant technique; visualization of the scores within the 3 × 3 table facilitates assessment of spacer symmetry both in the left-right and superior-inferior orientations. We anticipate that physicians who are more experienced with placement of the hydrogel rectal spacer will on average, have implants with a higher SQS. For example, Pinkawa, et al., assessed a group of 64 consecutive patients with rectal hydrogel spacers inserted at their institution over time and demonstrated that a learning curve existed: the latter half of patients had improved spacer symmetry, larger prostate-rectal separations, moderated rectal dose, and fewer acute rectal toxicity after radiotherapy [13].
      A limitation of this study includes the small sample size, which may have weakened our ability to determine whether the SQS should be reported as a 3-tier or binary score (in which 0-1 are grouped together). Further, analysis of late toxicity did not account for follow-up time because this data was unavailable; therefore, we acknowledge that the results may be biased. Further, we do not have data to assess whether SQS predicts which patients will have resolution of their late rectal toxicity. Another weakness of this study is that most patients who developed a late rectal toxicity, had a grade of 1 (14 of 20 patients with toxicity), which we recognize as generally considered clinically insignificant. However, prostate SBRT, especially when a hydrogel rectal spacer is present, is a well-tolerated treatment with <5% late grade 2 and higher rectal toxicity rates [11]. Similarly, there was no grade 2 or higher late rectal toxicity in men with HSP treated with conventionally fractionated radiation to the prostate in the pivotal phase III spacer study[1, 2]. Therefore, it is reasonable that our SQS metric can predict any graded late rectal toxicity. All of these shortfalls highlight the necessity to externally validate the SQS using a larger sample size, with more complete toxicity data, and with different radiotherapy techniques (i.e., SBRT, moderate hypofractionation, and brachytherapy) to ensure its versatility and utility.
      Although this comprehensive instrument that accounts for hydrogel spacer symmetry and thickness throughout the entire irradiated length of the prostate-rectal interface was associated with both rectal dosimetry and late rectal toxicity in this small sample size, a more simplistic metric might achieve the same goals. Using the longer follow-up data from a published prospective randomized trial that compared rectal toxicity and dosimetry in patients treated with definitive radiation in the presence or absence of a hydrogel rectal spacer[1], we plan to rigorously test the validity of the SQS and compare it to reduced metrics that describe the quality of separation at only one level of the prostate gland.

      CONCLUSION

      We have developed a metric to report on the overall quality of PRI created by HSP, that is significantly associated with rectal dosimetry and late rectal toxicity. This SQS will need to be validated using an external dataset with a larger and more diverse patient cohort to corroborate its utility.

      Author Responsible for Statistical Analysis

      Zhigang Zhang, PhD, Memorial Sloan Kettering Cancer Center, Department of Epidemiology and Biostatistics, 1275 York Ave, New York, NY 10065; zhangz@mskcc.org

      IRB Approval

      This is a secondary analysis of data originally collected under an Institutional Review Board-approved prospective study. Data remained de-identified and compliant with the Health Insurance Portability and Accountability Act.

      Funding

      None.

      Data Availability Statement for this Work

      Research data from this work are stored in an institutional repository and will be shared upon request to the corresponding author.

      REFERENCES

      • 1
        Hamstra, D.A., et al., Continued Benefit to Rectal Separation for Prostate Radiation Therapy: Final Results of a Phase III Trial. Int J Radiat Oncol Biol Phys, 2017. 97(5): p. 976-985.
      • 2
        Mariados, N., et al., Hydrogel Spacer Prospective Multicenter Randomized Controlled Pivotal Trial: Dosimetric and Clinical Effects of Perirectal Spacer Application in Men Undergoing Prostate Image Guided Intensity Modulated Radiation Therapy. Int J Radiat Oncol Biol Phys, 2015. 92(5): p. 971-977.
      • 3
        Miller, L.E., et al., Association of the Placement of a Perirectal Hydrogel Spacer With the Clinical Outcomes of Men Receiving Radiotherapy for Prostate Cancer: A Systematic Review and Meta-analysis. JAMA Netw Open, 2020. 3(6): p. e208221.
      • 4
        Fischer-Valuck, B.W., et al., Hydrogel spacer distribution within the perirectal space in patients undergoing radiotherapy for prostate cancer: Impact of spacer symmetry on rectal dose reduction and the clinical consequences of hydrogel infiltration into the rectal wall. Pract Radiat Oncol, 2017. 7(3): p. 195-202.
      • 5
        Hwang, M.E., et al., A novel model to correlate hydrogel spacer placement, perirectal space creation, and rectum dosimetry in prostate stereotactic body radiotherapy. Radiat Oncol, 2018. 13(1): p. 192.
      • 6
        Liu, H., et al., Proposed Hydrogel-Implant Quality Score and a Matched-Pair Study for Prostate Radiation Therapy. Pract Radiat Oncol, 2020. 10(3): p. 202-208.
      • 7
        Taggar, A.S., et al., Placement of an absorbable rectal hydrogel spacer in patients undergoing low-dose-rate brachytherapy with palladium-103. Brachytherapy, 2018. 17(2): p. 251-258.
      • 8
        Folkert, M.R., et al., A Multi-Institutional Phase II Trial of High-Dose SAbR for Prostate Cancer Using Rectal Spacer. Int J Radiat Oncol Biol Phys, 2021.
      • 9
        Zelefsky, M.J., et al., Five-Year Outcomes of a Phase 1 Dose-Escalation Study Using Stereotactic Body Radiosurgery for Patients With Low-Risk and Intermediate-Risk Prostate Cancer. Int J Radiat Oncol Biol Phys, 2019. 104(1): p. 42-49.
      • 10
        Jackson, W.C., et al., Stereotactic Body Radiation Therapy for Localized Prostate Cancer: A Systematic Review and Meta-Analysis of Over 6,000 Patients Treated On Prospective Studies. Int J Radiat Oncol Biol Phys, 2019. 104(4): p. 778-789.
      • 11
        Zelefsky, M.J., et al., Early Tolerance and Tumor Control Outcomes with High-dose Ultrahypofractionated Radiation Therapy for Prostate Cancer. Eur Urol Oncol, 2020. 3(6): p. 748-755.
      • 12
        Kang, M.H., et al., Difference in the rate of rectal complications following prostate brachytherapy based on the prostate-rectum distance and the prostate longitudinal length among early prostate cancer patients. Korean J Urol, 2015. 56(9): p. 637-43.
      • 13
        Pinkawa, M., et al., Learning curve in the application of a hydrogel spacer to protect the rectal wall during radiotherapy of localized prostate cancer. Urology, 2013. 82(4): p. 963-8.

      Conflicts of Interest

      The authors have no relevant financial disclosures to report. Drs. Folkert and Zelefsky are consultants for Boston Scientific Inc.