- Email: [email protected]
The dependence of prostate postimplant dosimetric quality on CT volume determination
Int. J. Radiation Oncology Biol. Phys., Vol. 44, No. 5, pp. 1111–1117, 1999 Copyright © 1999 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/99/$–see front matter
PII S0360-3016(99)00137-6
CLINICAL INVESTIGATION
Prostate
THE DEPENDENCE OF PROSTATE POSTIMPLANT DOSIMETRIC QUALITY ON CT VOLUME DETERMINATION GREGORY S. MERRICK, M.D.,*† WAYNE M. BUTLER, PH.D.,* ANTHONY T. DORSEY, M.S.,* JONATHAN H. LIEF, PH.D.‡
AND
*Schiffler Oncology Center, Wheeling Hospital, Wheeling, WV; †The George Washington University Medical Center, Division of Radiation Oncology and Biophysics, Washington, DC; and ‡Wheeling Jesuit University, Wheeling, WV Purpose: The postoperative evaluation of permanent prostate brachytherapy requires a subjective determination of the implant volume. This work investigates the magnitude of the effect that various methods of treatment volume delineation have on dosimetric quality parameters for a treatment planning philosophy that defines a target volume as the prostate with a periprostatic margin. Methods and Materials: Eight consecutive prostate brachytherapy patients with a prescribed dose of 145 Gy from 125I as monotherapy comprised the study population. The prostate ultrasound volume was enlarged to a planning volume by an average factor of 1.8 to encompass probable extracapsular extension in the periprostatic region. For this cohort, the mean pretreatment parameters were 30.3 cm3 ultrasound volume, 51.8 cm3 planning volume, 131 seeds per patient, and 42.9 mCi total activity. On CT study sets obtained less than 2 hours postoperatively, target volumes were drawn using three methods: prostate plus a periprostatic margin, prostate only which excluded the puborectalis muscles, the periprostatic fat and the periprostatic venous plexus, and the preplanning ultrasound magnified to conform to the magnification factor of the postimplant CT scan. Three sets of 5 dosimetric quality parameters corresponding to the different volumetric approaches were calculated: V100, V150, and V200 which are the fractions of the target volume covered by 100, 150, and 200% of the prescribed dose, and D90 and D100, which are the minimal doses covering 90 and 100% of the target volume. Results: The postoperative CT volume utilizing the prostate plus margin technique was comparable to the initial planning volume (mean 55.5 cm3 vs. 51.8 cm3, respectively) whereas those determined via superimposing the preplan ultrasound resulted in volumes nearly identical to the initial ultrasound evaluation (mean 32.4 cm3 vs. 30.3 cm3). The prostate only approach resulted in volumes approximately 25% larger than the ultrasound volume approach. Despite the volume determinations being markedly different, no significant differences between the approaches were appreciated for V100, V150, V200, and D90. Large variations seen in D100 were uncorrelated to any of the other parameters and make D100 unsuitable as a quality indicator. Conclusions: In terms of a logarithmic measure, the variation between volumetric approach for V100, V150, V200, and D90 was less than one-fifth the variation of the CT volumes. These results which indicate relative independence of postimplant CT volume determination and dosimetric quality are only valid for a planning philosophy that includes the prostate with a periprostatic margin as the target volume. © 1999 Elsevier Science Inc. Prostate brachytherapy, Dosimetry, CT volume, Quality.
INTRODUCTION Transperineal ultrasound-guided prostate brachytherapy has emerged as a viable option in the management of earlystage carcinoma of the prostate gland. Measured in terms of local control and freedom from biochemical failure, the results of brachytherapy have been found to be as favorable as the most positive radical prostatectomy series (1–7). The widespread acceptability and reproducibility of the favorable results cited will be dependent on appropriate patient selection, assignment to the appropriate treatment regimen, and the ability of brachytherapists to reproduce the technical aspects of the procedure. Postoperative computed tomogra-
phy (CT) based dosimetric analysis provides detailed information regarding the coverage and the uniformity of an implant. It also affords the ability to compare various intraoperative techniques and provides a sound basis for future improvement (8 –19). The dose–response curve for 125I ultrasound-guided prostate brachytherapy reported by Stock et al. (19) provides a rationale for evaluating implants in terms of specific quality parameters. The determination of prostate volume via a postimplant CT remains a subjective endeavor and potentially can influence the reported dosimetric quality. Some investigators determine postimplant CT volume via a prostate only volume (20), others utilize the preplanning ultrasound magni-
Reprint requests to: Gregory S. Merrick, M.D., Wheeling Hospital, Schiffler Oncology Center, 1 Medical Park, Wheeling, WV
26003-6300. E-mail: [email protected] Accepted for publication 30 March 1999. 1111
1112
I. J. Radiation Oncology
●
Biology
●
Physics
fied to conform to the postimplant CT scan (21) and others assign a prostate volume with a periprostatic margin (11– 13). Herein, we report a detailed dosimetric evaluation using each of the three mentioned approaches for 8 consecutive 125I monotherapy implants. MATERIALS AND METHODS The study population consisted of 8 consecutive 125I monotherapy patients who underwent prostate brachytherapy from April 7, 1998 –May 18, 1998 for clinical T1/T2 carcinoma of the prostate gland. Because of significant inaccuracy in Gleason grading, all cases originating from outside institutions were reviewed prior to formulation of a treatment plan (22). Calculation algorithms and seed parameters used in preplans and postoperative dosimetry were those recommended by the American Association of Physicists in Medicine (AAPM) Task Group 43 (TG-43) (23). All preplanning techniques and intraoperative procedures have previously been described (11–13, 24). A preplanning transrectal ultrasound volumetric study of the prostate gland was obtained for all patients at 5-mm intervals extending from the proximal seminal vesicles/base of the prostate gland to the apex. Our preplanning philosophy mandated a target volume be obtained via enlargement of each ultrasound slice with a variable margin to account for likely extracapsular extension. The ultrasound images were enlarged approximately 3 mm in the lateral and anterior dimensions at midgland and up to 7 mm in all dimensions at the base and apex to result in a target volume approximately 1.8 times the ultrasound volume. A preplan was then generated to deliver the prescribed minimal peripheral dose (mPD) to greater than 99.5% of the planning volume with margin. Seeds were routinely planned in the periprostatic region. Great care was taken to limit the volume of the implant receiving 150% of the prescribed dose (V150) to approximately 35% in order to deliver a relatively homogeneic dose to the target volume (11–13). This planning philosophy which prescribes the dose and implant seeds to an enlarged planning volume with margin to deliver a periprostatic dose is justified based on the fact that prostate cancers with favorable pretreatment prostate-specific antigen (PSA) determinations, low Gleason scores, and early clinical T-stages possess a significant probability of extracapsular extension (25–27). Other investigators also utilize enlargement philosophies comparable to ours (28). In the operating room, the radioactive seeds were implanted transperineally via ultrasound and fluoroscopic guidance through predetermined template apertures according to the preplan worksheet. In the periprostatic tissue, the implanted 125I seeds (Nycomed Amersham, Arlington Heights, IL) were embedded in a vicryl suture (Rapid Strand). At the completion of the implant, approximately 6% extra seeds were placed into either potential cold spots (via ultrasound or fluoroscopy) or in the region of the positive biopsies.
Volume 44, Number 5, 1999
A CT scan was obtained for postoperative dosimetric evaluation within 2 hours of implantation. Spiral CT scans were acquired at 5 mm thickness and 5 mm spacing extending 2 cm above and below the most superior and inferior implanted seeds. A set of CT images was printed at a magnification of 1.5 using soft tissue windows to identify the prostate and surrounding structures and a second set of images using bone windows to better distinguish seeds from calcifications and eliminate redundant seed images. A urinary catheter and 10 mm diameter rectal obturator were in place for the identification of the urethra and the anterior rectal mucosa, respectively (11, 24). For this cohort of patients, the actual dose distribution to the prostate gland was then determined by each of three methods. A prostate only approach, referred to subsequently as method B, outlined only the prostate volumes via the CT scan with exclusion of the puborectalis muscles, the periprostatic fat, and the periprostatic venous plexus. Method A outlined the prostate plus a periprostatic margin, and method C used the preplanning ultrasound of the prostate magnified to conform to the magnification factor of the postimplant CT scan. The magnified ultrasound images were superimposed on the CT images by registering the posterior border of the prostate for each ultrasound/CT scan slice. Left–right alignment of the ultrasound image to the CT image placed the catheter marked urethra in the same vertical plane. The actual dose distribution was generated via a dedicated treatment planning computer (Prowess 3000, SSGI, Chico, CA). In addition, the quality parameters D90 (the minimal dose covering 90% of the defined prostate volume), D100 (the minimal dose covering 100% of the volume), V100 (the fraction of the defined prostate volume receiving at least 100% of the prescribed mPD), V150 and V200 (the fraction of the defined prostate volume receiving at least 200% of the prescribed mPD) were calculated. All prostate volumes including the preplanning ultrasound and postimplant CT were determined by one investigator (G.S.M.). RESULTS Table 1 summarizes patient characteristics for 8 consecutive 125I monotherapy patients. Via our planning philosophy of accounting for extracapsular extension, the planning volume was on average 1.8 times greater than the transrectal ultrasound volume determined for the prostate. The specific activity ranged from 0.70 –1.18 mCi/cm3 (average 0.83 mCi/cm3). Smaller glands tend to have a higher specific activity since a given enlargement in millimeters has a greater effect proportionately on a small gland than a large gland and the number of seeds required of given strength is proportionate to volume. Table 2 details for each volume drawing approach the CT target volume plus the quality evaluation parameters D90, V100, V150, and V200 determined via day 0 dosimetry. The postoperative CT volume utilizing the prostate plus margin technique (method A) was comparable to our initial
Implant quality vs. CT volume
● G. S. MERRICK et al.
1113
Table 1. Pretreatment patient characteristics (I-125 implant only, 145 Gy, TG-43)
Patient
Clinical T-stage
Pre-Tx PSA
Gleason score
Ultrasound volume (cm3)
Planning volume (cm3)
Enlargement factor plan vol/US vol
Seed activity (mCi)
Number of seeds
Total activity (mCi)
Specific activity (mCi/cm3)
1 2 3 4 5 6 7 8 Avg.
T2a T1c T1c T2b T1c T2a T1c T1c —
7.3 6.4 5.0 7.5 6.2 6.6 4.4 6.3 6.2
6 (3 ⫹ 3) 6 (3 ⫹ 3) 7 (3 ⫹ 4) 6 (3 ⫹ 3) 6 (3 ⫹ 3) 5 (3 ⫹ 2) 6 (3 ⫹ 3) 6 (3 ⫹ 3) 6
34.3 47.4 43.5 12.0 33.4 29.8 14.2 28.0 30.3
55.8 71.7 68.4 25.3 56.2 55.2 30.5 51.5 51.8
1.63 1.51 1.57 2.11 1.68 1.85 2.15 1.84 1.79
0.316 0.320 0.336 0.313 0.332 0.336 0.317 0.336 0.326
141 156 158 95 135 128 102 137 131
44.56 49.92 53.09 29.73 44.82 43.01 32.33 46.03 42.94
0.80 0.70 0.78 1.18 0.80 0.78 1.06 0.89 0.83
planning volume (55.5 cm3 vs. 51.8 cm3, respectively). The comparability of these 2 volumes suggests that our postimplant evaluation included the prostate gland with margin. The CT volume determined by method A is expected to be greater than the preplan target volume due to edema which usually results in about a 20% volume increase (12). In addition, postoperative dosimetric volume determinations via the ultrasound approach resulted in a volume nearly identical to the initial ultrasound evaluation (32.4 cm3 vs. 30.3 cm3). An exact correlation between these 2 volumes is
not expected because of differences in volume calculation algorithms between the ultrasound and the treatment planning software and errors inherent to the transfer of hard copies. Postoperative prostate volume determinations via a prostate only volume resulted in a volume approximately 25% larger than the ultrasound volume approach. If the defined CT volumes are expressed as ellipsoids, the average ellipsoid dimension was 4.73 cm, 4.21 cm, and 3.96 cm for delineation methods A, B, and C, respectively. Despite the volume determinations being markedly different, no signif-
Table 2. Dosimetric results vs. target volume approach* CT Volume (cm3) Patient
A
B
C
69.3 1
D90 (Gy) A
B
67.6
154 20.3
61.8 5
41.7 51.9
6
179
34.6 7
163
60.9 8
148
Avg.
55.5
42.2
19.5 43.3
45.8 97.1
165
20.9 19.9
93.9 94.3
162 32.4
22.0 48.0
91.5
155 39.2
48.4 41.0
153
26.4 25.5
99.7 89.9
31.4
24.5 61.3
98.4
145 45.0
59.0 51.0
169
18.9 23.8
99.6 97.6
17.0
19.8 47.9
99.2
161 24.6
50.9 54.3
184
26.7 19.6
96.9 97.3
33.6
28.8 57.0
98.3
170 37.9
55.9 46.4
158
20.8 29.4
99.7
162
20.4 52.7
98.5 94.3
38.8
12.0
50.6 52.6
178 153
13.3 22.2
98.4
168
13.9
28.9 50.3
93.7
13.9
14.1
33.4
98.4
17.6 21.3
47.5 97.5
C
14.4
96.1 97.9
B
41.1 33.1
168
A 15.2
95.3
165
C
39.7
97.9
163
B
V200 (%)
37.7
155
43.9
A
94.2
159
29.4
C
92.9
43.6 51.2
V150 (%)
90.5
149
68.7
B
155
41.6
4
A
153 36.8
3
C
146 51.4
2
V100 (%)
20.3 47.5
19.7
* I-125 implant only, 145 Gy mPD, day 0 dosimetry. Target volume approach: A ⫽ CT prostate volume plus margin; B ⫽ CT prostate volume; C ⫽ pretreatment ultrasound volume superimposed on CT.
1114
I. J. Radiation Oncology
●
Biology
●
Physics
Table 3. Descriptive and correlation statistics Parameter Postoperative volume V100 V150 V200 D100 D90
Methods compared
Mean
Range
Vector correlation
A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C
1.42 1.24 1.75 0.971 0.995 0.966 0.967 1.009 0.975 1.051 1.040 1.094 1.007 1.209 0.943 0.958 0.982 0.941
0.28 0.52 0.73 0.046 0.044 0.045 0.142 0.196 0.260 0.188 0.180 0.319 2.956 3.593 2.839 0.074 0.080 0.106
0.966 0.854 0.931 0.821 0.851 0.887 0.960 0.978 0.948 0.967 0.968 0.921 0.089 ⫺0.376 ⫺0.204 0.889 0.919 0.837
icant differences were appreciated for any of the evaluated dosimetric parameters via the three approaches. Based on the cohort averages in the bottom row in Table 2, there is a trend toward higher D90 values via the ultrasound approach. Table 3 summarizes pair-wise comparisons between the three volumetric approaches for the various parameters. Except for postoperative volume, the means of these ratios are all close to 1.00; however, the maximum–minimum range for D100 is quite large. The last column in this table is the linear correlation coefficient obtained by plotting the parameters for one approach against those of another. All the parameters exhibit a high degree of correlation except for the pair-wise plots of D100. These coefficients are independent of the order in which the two quantities are taken. Figure 1 further illustrates the marked differences in volume determination and D100 and emphasizes the modest variation between the other evaluated dosimetric parameters. Data in this figure was obtained by taking the cohort mean of the absolute value of the natural log of the pairwise ratios for a given parameter such as D90 (method A): D90 (method B) for each patient. For a given dosimetry parameter, Q, and CT volume delineation approach A, B, or C, this is expressed as
1n
冉冊 X Y
冘冏 8
⫽
冉
冊冏
parameter Q, approach X 1 䡠 1n 8 i⫽1 parameter Q, approach Y
i
where i is the patient number. Since the log of a ratio has the same magnitude regardless of how the ratio is written, this construct highlights not only deviations from unity but also the spread of the values. For postoperative volume, the average of this log measure is 0.376 (range 0.23– 0.55) while the average variation for the other parameters except for D100 in terms of the log measure is less than one-fifth of this: V100 ⫽ 0.026 (0.014 – 0.034), V150 ⫽ 0.057
Volume 44, Number 5, 1999
(0.047– 0.077), V200 ⫽ 0.077 (0.058 – 0.112), D90 ⫽ 0.045 (0.030 – 0.062). Since some of the voxels contributing to D100 lie in high-dose gradient regions, this parameter is too sensitive to local conditions to serve as a quality measure regardless of the volumetric approach. This excessive variation of D100 that makes it unacceptable in evaluating quality has been observed by others (29, 30), and D100 is included here as a contrast to the usual quality parameters whose log ratio measures indicate that they deviate only slightly from unity and cluster tightly about that value for individual patients. DISCUSSION Over the past decade, prostate brachytherapy has emerged as a viable option for the treatment of early-stage carcinoma of the prostate gland. Recent publications have reported that in terms of local control and freedom from biochemical failure, the results of conformal brachytherapy are as favorable as the most positive radical prostatectomy series (1–7). For 125I monotherapy, Stock et al. (18, 19) have reported a dose threshold for biochemical NED (no evidence of disease) survival. Their dosimetric analysis reported in terms of a D90 revealed that a dose of 140 –160 Gy (TG-43 dosimetry) obtained on postimplant day 30 correlated with optimal biochemical NED survival (19). CT-based postoperative dosimetric analysis provides detailed information regarding the dose distribution to the prostate/periprostatic region, urethra, and rectum (8, 9, 11– 13, 24). Recently, the American Brachytherapy Society (ABS) recommended that prostate implants be defined in terms of a D90, V100, and V150 (31). Merrick et al. (13) recently reported dosimetric evaluation for 60 consecutive patients in terms of ABS guidelines which found that the three quality parameters evaluated are correlated with one another. One area of subjectivity in postoperative dosimetry involves the determination of postimplant CT prostate/target volumes. Approaches to determine postimplant CT target volumes include prostate only volumes (20), prostate plus margin volumes to ensure an adequate extracapsular dose (11–13), and volumes based on the preplanning ultrasound (21). Our data indicate that for a planning philosophy that enlarges the preplanning ultrasound volume approximately 80% and routinely plans seeds in the periprostatic region, the determination of postimplant volume via CT scans and its relationship on dosimetric quality is somewhat academic. The data in Fig. 1 show that while the defined volume ratios and D100 ratios either deviate considerably from unity or have a wide spread of values, the commonly used quality parameters have ratios that cluster tightly about unity. A schematic rationale for this conclusion is illustrated in Fig. 2. When the mPD isodose is just outside the largest defined target volume (except for occasional low-dose incursions), coverage indexes such as V100 and D90 will be relatively insensitive to CT volume delineation. This is because lowdose incursions into the smallest defined prostate volume
Implant quality vs. CT volume
Fig. 1. The mean of the absolute value of the natural log of each of the listed parameters evaluated pair-wise for each patient between the three methods. The variance in the ratios of post-op volume and D100 is large compared to the variance in the other parameters.
(approach C) will occupy a proportionately similar fraction of the larger volumes studied. This result should apply rigorously only to implants using a planning philosophy that prescribes the dose and implants seeds to an enlarged planning volume with margin. Furthermore, homogeneity measures such as V150 may not be invariant to volume drawing technique when the implant approach is less dosimetrically
Fig. 2. A schematic of a volume planned and implanted with periprostatic margins to encompass extracapsular disease. The outermost oval represents the volume encompassed by the prescribed mPD isodose (planning volume with margin) while the heavy line traces the isodose in the post implant plan corresponding to prescribed mPD. A typical low dose incursion due to inadequate execution of the preplan is marked by the shaded region. The lighter ovals labeled A, B and C represent the three methods of drawing the target volume as summarized in Table 2. The fractional volume covered by the mPD (V100) is relatively insensitive to CT volume drawing technique in this case. Other parameters such as V150 will also be similar among the 3 target definitions if the implant is dosimetrically homogeneous.
● G. S. MERRICK et al.
1115
homogeneous such as peripheral or uniform seed loading. We took great care to limit the V150 of the planning volume to approximately 35%. This results in a relatively homogeneic dose to the target volume with margin and results in relative independence between dosimetric quality and volume determination within the planning volume. As such, within the prostate gland, there is no actual point dose escalation, but our approach does result in a greater integral dose to the patient due to significant dose to the extracapsular region. A 5-center multi-institutional review using implant philosophies comparable to ours also reported the relative independence between V80 (the fraction of the treatment volume covered by 80% of the mPD), V100, and postimplant volume determination (32). We do not believe these results are transferable to an implant philosophy that exclusively utilizes high activity intraprostatic seeds. Although these implants also irradiate the prostate with a 3–5 mm periprostatic margin, a steep dose gradient exists in the periprostatic region, and as such, dosimetric quality may be highly dependent on CT volume determination. Our data reveals significant differences in CT-based volumes depending on each of the three investigated approaches. However, the D90, V100, and V150 as well as V200 do not differ significantly while D100 is found to be too erratic to be of any use in quality evaluations. This is due to the fact that some of the voxels contributing to D100 lie in high-dose gradient regions. Roy et al. (29) and Yu et al. (30) reported D100 to be overly sensitive to differences in target volumes to serve as a measure of dosimetric quality. The postimplant CT volumes determined via the prostate plus margin technique are comparable to our preplanning volume while the postimplant volumes determined via the ultrasound approach are comparable to the planning ultrasound volume. As such, these approaches appear to be highly reproducible in terms of volume determination. The prostate only approach resulted in a volume determination intermediate between the two extremes. Regardless of which of the three postimplant CT volume determinations are utilized, the difference in the average dimension of the CT-determined volume calculated as an ellipsoid is less than 10 mm in dimension for any patient. Via our planning approach, the average dimensional enlargement from preplan ultrasound volume to planning volume lies within this range. Recently Waterman et al. (17) reported postoperative dosimetry for 10 patients evaluated via serial CT scans on day 0 and weeks 1, 3, 7, and 15. Edema affected postoperative dosimetry in 7 of the 10 cases. In the three postoperative dosimetry cases where the V100 was greater than 90% on day 0, no significant improvement in V100 determinations was noted with time. All three of these cases were preplanned with a prostate plus margin technique (33). These three cases result in day 0 dosimetry comparable to what we report in this paper and in prior publications (11–13, 24). The other seven cases in Waterman’s publication were planned with a prostate only volume. Waterman et al. (17) indicated that edema is markedly less important in assessment of dosimetric quality for a preplanning philos-
1116
I. J. Radiation Oncology
●
Biology
●
Physics
ophy that entailed prostate plus margin compared to a planning volume that includes prostate only. Dosimetric coverage to the prostate gland improves as the time interval between brachytherapy and the postimplant CT scan increases (15). As edema resolves, all dosimetric quality parameters increase numerically (except for V## parameters such as V80 that may already be at 100%), and this is true regardless of which approach to CT volume determination is used. Therefore, brachytherapists who perform CT dosimetry on day 14 or day 30 will also find that quality is relatively insensitive to their drawing technique.
Volume 44, Number 5, 1999
In summary, a planning philosophy that includes the prostate with margin in the target volume is relatively independent of postimplant CT volume determination in terms of dosimetric quality. This effect of a planning margin stabilizing quality indicators is implied in the report by Waterman et al. (17), who found that the effect of edema is also lessened in appropriately executed implants that were planned with margin. As such, via the implant philosophy that we utilize, the determination of postimplant prostate volume has relatively little effect on ultimate reported dosimetric quality.
REFERENCES 1. Blasko JC, Wallner K, Grimm PD, Ragde H. Prostate specific antigen based disease control following ultrasound guided iodine-125 implantation for stage T1/T2 prostatic carcinoma. J Urol 1995;154:1096 –1099. 2. Wallner K, Roy J, Zelefsky M, Fuks Z, Harrison L. Short term freedom from disease progression after I-125 prostate implantation. Int J Radiat Oncol Biol Phys 1994;30:405– 409. 3. Wallner K, Roy J, Harrison L. Tumor control and morbidity following transperineal I-125 implantation for stage T1/T2 prostatic carcinoma. J Clin Oncol 1996;14:449 – 453. 4. Dattoli M, Wallner K, Sorace R, Koval J, Cash J, Acosta R, Brown C, Etheridge J, Binder M, Brunelle R, Kirwan N, Sanches S, Stein D, Wasserman S. Pd-103 brachytherapy and external beam radiation therapy for clinically localized high risk prostatic carcinoma. Int J Radiat Oncol Biol Phys 1996; 35:875– 879. 5. Ragde H, Blasko JC, Grimm PD, Kenny GM, Sylvester JE, Hoak DC, Landin K, Cavanagh W. Interstitial I-125 radiation without adjuvant therapy in the treatment of clinically localized prostate cancer. Cancer 1997;80:442– 453. 6. Dattoli MJ, Wallner KE, Cash JC, Ross RE, Koval JM, Sorace RA. Pd-103 brachytherapy for clinical T1/T2 prostate carcinomas. Int J Radiat Oncol Biol Phys 1997;39:221. 7. Ragde H, Blasko JC, Grimm PD, Kenny GM, Sylvester J, Hoak DC, Cavanagh W, Landin K. Brachytherapy for clinically localized prostate cancer: Results at 7 and 8 year followup. Sem Surg Onc 1997;13:438 – 443. 8. Willins J, Wallner K. CT based dosimetry for transperineal I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:347–353. 9. Wallner K, Roy J, Harrison L. Dosimetry guidelines to minimize urethral and rectal morbidity following transperineal I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1995;32:465– 471. 10. Willins J, Wallner K. Time-dependent changes in CT-based dosimetry of I-125 prostate brachytherapy. Rad Onc Invest 1998;6:157–160. 11. Merrick GS, Butler WM, Dorsey AT, Walbert HL. Prostatic conformal brachytherapy: I-125/Pd-103 postoperative dosimetric analysis. Rad Onc Invest 1997;5:305–313. 12. Merrick GS, Butler WM, Dorsey AT, Walbert HL. Influence of timing on the dosimetric analysis of transperineal ultrasound-guided prostatic conformal brachytherapy. Rad Onc Invest 1998;6:182–190. 13. Merrick GS, Butler WM, Dorsey AT, Lief JH. The potential role of various dosimetric quality indications in prostate brachytherapy. Int J Radiat Oncol Biol Phys 1999;44:717– 724. 14. Prestidge BR, Bice WS, Prete JJ, Buchholz TA. A dose
15.
16.
17.
18.
19. 20.
21. 22.
23.
24.
25.
26.
27.
volume analysis of permanent transperineal prostatic brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:238. Prestidge BR, Bice WS, Kiefer ET, Prete JJ. Timing of computed tomography based post-implant assessment following permanent transperineal prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;40:1111–1115. Moerland MA, Wijrdeman HK, Beersma R, Bakker CJG, Battermann JJ. Evaluation of permanent I-125 prostate implants using radiography and magnetic imaging. Int J Radiat Oncol Biol Phys 1997;37:927–933. Waterman FM, Yue N, Corn BW, Dicker AP. Edema associated with I-125 or Pd-103 prostate brachytherapy and its impact on post-implant dosimetry: An analysis based on serial CT acquisitions. Int J Radiat Oncol Biol Phys 1998;41:1069 – 1077. Stock RG, Stone NN, DeWyngaert JK, Lavaginini P, Unger PD. Prostate antigen findings and biopsy results following interactive ultrasound guided transperineal brachytherapy for early stage carcinoma. Cancer 1996;77:2386 –2392. Stock RG, Stone NN, Tabert A, Iannuzzi C, DeWyngaert JK. A dose response study for I-125 prostate implants. Int J Radiat Oncol Biol Phys 1998;41:101–108. Wallner K, Blasko J, Dattoli MJ. Implant evaluation. In: Wallner K, Blasko J, Dattoli MJ, editors. Prostate brachytherapy made complicated. Seattle: Smart Medicine Press; 1997. p. 9.1–9.20. Grimm PD. Personal communication. Merrick GS, Butler WM, Farthing WH, Dorsey AT, Adamovich E. The impact of Gleason score accuracy as a criterion for prostate brachytherapy patient selection. J Brachyther Int 1998;14:113–121. Nath R, Anderson LL, Luxton G, Weaver KA, Williamson JF, Meigooni AS. Dosimetry of interstitial brachytherapy sources: Recommendations of the AAPM Radiation Therapy Committee Task Group No. 43. Med Phys 1995;22:209 –234. Merrick GS, Butler WM, Dorsey AT, Lief JH, Walbert HL, Blatt HJ. Rectal dosimetric analysis following prostate brachytherapy. Int J Radiat Oncol Biol Phys 1999;43:1021– 1027. Partin AW, Kattan MW, Subong ENP, Walsh PC, Wojno KJ, Oesterling JE, Scardino PT, Pearson JD. Combination of prostate-specific antigen, clinical stage and Gleason score to predict pathological stage of localized prostate cancer. JAMA 1997;277:1445–1451. Cookson MS, Fleshner NE, Soloway SM, Fair WR. Prognostic significance of prostate-specific antigen in stage T1c prostate cancer treated by radical prostatectomy. Urol 1997;49: 887– 893. Wills ML, Sauvageot J, Partin AW, Gurganus R, Epstein JI. Ability of sextant biopsies to predict radical prostatectomy stage. Urol 1998;51:759 –764.
Implant quality vs. CT volume
28. Wallner K, Blasko JC, Dattoli MJ. Implant design. In: Wallner K, Blasko J, Dattoli MJ, editors. Prostate brachytherapy made complicated. Seattle: Smart Medicine Press; 1997. p. 6.1– 6.42. 29. Roy JN, Wallner KE, Harrington PJ, Ling CC, Anderson LL. A CT-based evaluation method for permanent implants: application to prostate. Int J Radiat Oncol Biol Phys 1993;26: 163–169. 30. Yu Y, Waterman FM, Suntharalingam N, Schulsinger A. Limitations of the minimum peripheral dose as a parameter for
● G. S. MERRICK et al.
1117
dose specification in permanent 125I prostate implants. Int J Radiat Oncol Biol Phys 1996;34:717–725. 31. Nag S. ABS Prostate brachytherapy consensus statement. American Brachytherapy Society Annual Meeting, Albuquerque, New Mexico, May 31, 1998. 32. Bice WS, Prestidge BR, Grimm PD, Friedland JL, Feygelman V, Roach M, Prete JJ, Dubois DF, Blasko JC. Centralized multiinstitutional postimplant analysis for interstitial prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;41:921–927. 33. Waterman F. Personal communication.
PII S0360-3016(99)00137-6
CLINICAL INVESTIGATION
Prostate
THE DEPENDENCE OF PROSTATE POSTIMPLANT DOSIMETRIC QUALITY ON CT VOLUME DETERMINATION GREGORY S. MERRICK, M.D.,*† WAYNE M. BUTLER, PH.D.,* ANTHONY T. DORSEY, M.S.,* JONATHAN H. LIEF, PH.D.‡
AND
*Schiffler Oncology Center, Wheeling Hospital, Wheeling, WV; †The George Washington University Medical Center, Division of Radiation Oncology and Biophysics, Washington, DC; and ‡Wheeling Jesuit University, Wheeling, WV Purpose: The postoperative evaluation of permanent prostate brachytherapy requires a subjective determination of the implant volume. This work investigates the magnitude of the effect that various methods of treatment volume delineation have on dosimetric quality parameters for a treatment planning philosophy that defines a target volume as the prostate with a periprostatic margin. Methods and Materials: Eight consecutive prostate brachytherapy patients with a prescribed dose of 145 Gy from 125I as monotherapy comprised the study population. The prostate ultrasound volume was enlarged to a planning volume by an average factor of 1.8 to encompass probable extracapsular extension in the periprostatic region. For this cohort, the mean pretreatment parameters were 30.3 cm3 ultrasound volume, 51.8 cm3 planning volume, 131 seeds per patient, and 42.9 mCi total activity. On CT study sets obtained less than 2 hours postoperatively, target volumes were drawn using three methods: prostate plus a periprostatic margin, prostate only which excluded the puborectalis muscles, the periprostatic fat and the periprostatic venous plexus, and the preplanning ultrasound magnified to conform to the magnification factor of the postimplant CT scan. Three sets of 5 dosimetric quality parameters corresponding to the different volumetric approaches were calculated: V100, V150, and V200 which are the fractions of the target volume covered by 100, 150, and 200% of the prescribed dose, and D90 and D100, which are the minimal doses covering 90 and 100% of the target volume. Results: The postoperative CT volume utilizing the prostate plus margin technique was comparable to the initial planning volume (mean 55.5 cm3 vs. 51.8 cm3, respectively) whereas those determined via superimposing the preplan ultrasound resulted in volumes nearly identical to the initial ultrasound evaluation (mean 32.4 cm3 vs. 30.3 cm3). The prostate only approach resulted in volumes approximately 25% larger than the ultrasound volume approach. Despite the volume determinations being markedly different, no significant differences between the approaches were appreciated for V100, V150, V200, and D90. Large variations seen in D100 were uncorrelated to any of the other parameters and make D100 unsuitable as a quality indicator. Conclusions: In terms of a logarithmic measure, the variation between volumetric approach for V100, V150, V200, and D90 was less than one-fifth the variation of the CT volumes. These results which indicate relative independence of postimplant CT volume determination and dosimetric quality are only valid for a planning philosophy that includes the prostate with a periprostatic margin as the target volume. © 1999 Elsevier Science Inc. Prostate brachytherapy, Dosimetry, CT volume, Quality.
INTRODUCTION Transperineal ultrasound-guided prostate brachytherapy has emerged as a viable option in the management of earlystage carcinoma of the prostate gland. Measured in terms of local control and freedom from biochemical failure, the results of brachytherapy have been found to be as favorable as the most positive radical prostatectomy series (1–7). The widespread acceptability and reproducibility of the favorable results cited will be dependent on appropriate patient selection, assignment to the appropriate treatment regimen, and the ability of brachytherapists to reproduce the technical aspects of the procedure. Postoperative computed tomogra-
phy (CT) based dosimetric analysis provides detailed information regarding the coverage and the uniformity of an implant. It also affords the ability to compare various intraoperative techniques and provides a sound basis for future improvement (8 –19). The dose–response curve for 125I ultrasound-guided prostate brachytherapy reported by Stock et al. (19) provides a rationale for evaluating implants in terms of specific quality parameters. The determination of prostate volume via a postimplant CT remains a subjective endeavor and potentially can influence the reported dosimetric quality. Some investigators determine postimplant CT volume via a prostate only volume (20), others utilize the preplanning ultrasound magni-
Reprint requests to: Gregory S. Merrick, M.D., Wheeling Hospital, Schiffler Oncology Center, 1 Medical Park, Wheeling, WV
26003-6300. E-mail: [email protected] Accepted for publication 30 March 1999. 1111
1112
I. J. Radiation Oncology
●
Biology
●
Physics
fied to conform to the postimplant CT scan (21) and others assign a prostate volume with a periprostatic margin (11– 13). Herein, we report a detailed dosimetric evaluation using each of the three mentioned approaches for 8 consecutive 125I monotherapy implants. MATERIALS AND METHODS The study population consisted of 8 consecutive 125I monotherapy patients who underwent prostate brachytherapy from April 7, 1998 –May 18, 1998 for clinical T1/T2 carcinoma of the prostate gland. Because of significant inaccuracy in Gleason grading, all cases originating from outside institutions were reviewed prior to formulation of a treatment plan (22). Calculation algorithms and seed parameters used in preplans and postoperative dosimetry were those recommended by the American Association of Physicists in Medicine (AAPM) Task Group 43 (TG-43) (23). All preplanning techniques and intraoperative procedures have previously been described (11–13, 24). A preplanning transrectal ultrasound volumetric study of the prostate gland was obtained for all patients at 5-mm intervals extending from the proximal seminal vesicles/base of the prostate gland to the apex. Our preplanning philosophy mandated a target volume be obtained via enlargement of each ultrasound slice with a variable margin to account for likely extracapsular extension. The ultrasound images were enlarged approximately 3 mm in the lateral and anterior dimensions at midgland and up to 7 mm in all dimensions at the base and apex to result in a target volume approximately 1.8 times the ultrasound volume. A preplan was then generated to deliver the prescribed minimal peripheral dose (mPD) to greater than 99.5% of the planning volume with margin. Seeds were routinely planned in the periprostatic region. Great care was taken to limit the volume of the implant receiving 150% of the prescribed dose (V150) to approximately 35% in order to deliver a relatively homogeneic dose to the target volume (11–13). This planning philosophy which prescribes the dose and implant seeds to an enlarged planning volume with margin to deliver a periprostatic dose is justified based on the fact that prostate cancers with favorable pretreatment prostate-specific antigen (PSA) determinations, low Gleason scores, and early clinical T-stages possess a significant probability of extracapsular extension (25–27). Other investigators also utilize enlargement philosophies comparable to ours (28). In the operating room, the radioactive seeds were implanted transperineally via ultrasound and fluoroscopic guidance through predetermined template apertures according to the preplan worksheet. In the periprostatic tissue, the implanted 125I seeds (Nycomed Amersham, Arlington Heights, IL) were embedded in a vicryl suture (Rapid Strand). At the completion of the implant, approximately 6% extra seeds were placed into either potential cold spots (via ultrasound or fluoroscopy) or in the region of the positive biopsies.
Volume 44, Number 5, 1999
A CT scan was obtained for postoperative dosimetric evaluation within 2 hours of implantation. Spiral CT scans were acquired at 5 mm thickness and 5 mm spacing extending 2 cm above and below the most superior and inferior implanted seeds. A set of CT images was printed at a magnification of 1.5 using soft tissue windows to identify the prostate and surrounding structures and a second set of images using bone windows to better distinguish seeds from calcifications and eliminate redundant seed images. A urinary catheter and 10 mm diameter rectal obturator were in place for the identification of the urethra and the anterior rectal mucosa, respectively (11, 24). For this cohort of patients, the actual dose distribution to the prostate gland was then determined by each of three methods. A prostate only approach, referred to subsequently as method B, outlined only the prostate volumes via the CT scan with exclusion of the puborectalis muscles, the periprostatic fat, and the periprostatic venous plexus. Method A outlined the prostate plus a periprostatic margin, and method C used the preplanning ultrasound of the prostate magnified to conform to the magnification factor of the postimplant CT scan. The magnified ultrasound images were superimposed on the CT images by registering the posterior border of the prostate for each ultrasound/CT scan slice. Left–right alignment of the ultrasound image to the CT image placed the catheter marked urethra in the same vertical plane. The actual dose distribution was generated via a dedicated treatment planning computer (Prowess 3000, SSGI, Chico, CA). In addition, the quality parameters D90 (the minimal dose covering 90% of the defined prostate volume), D100 (the minimal dose covering 100% of the volume), V100 (the fraction of the defined prostate volume receiving at least 100% of the prescribed mPD), V150 and V200 (the fraction of the defined prostate volume receiving at least 200% of the prescribed mPD) were calculated. All prostate volumes including the preplanning ultrasound and postimplant CT were determined by one investigator (G.S.M.). RESULTS Table 1 summarizes patient characteristics for 8 consecutive 125I monotherapy patients. Via our planning philosophy of accounting for extracapsular extension, the planning volume was on average 1.8 times greater than the transrectal ultrasound volume determined for the prostate. The specific activity ranged from 0.70 –1.18 mCi/cm3 (average 0.83 mCi/cm3). Smaller glands tend to have a higher specific activity since a given enlargement in millimeters has a greater effect proportionately on a small gland than a large gland and the number of seeds required of given strength is proportionate to volume. Table 2 details for each volume drawing approach the CT target volume plus the quality evaluation parameters D90, V100, V150, and V200 determined via day 0 dosimetry. The postoperative CT volume utilizing the prostate plus margin technique (method A) was comparable to our initial
Implant quality vs. CT volume
● G. S. MERRICK et al.
1113
Table 1. Pretreatment patient characteristics (I-125 implant only, 145 Gy, TG-43)
Patient
Clinical T-stage
Pre-Tx PSA
Gleason score
Ultrasound volume (cm3)
Planning volume (cm3)
Enlargement factor plan vol/US vol
Seed activity (mCi)
Number of seeds
Total activity (mCi)
Specific activity (mCi/cm3)
1 2 3 4 5 6 7 8 Avg.
T2a T1c T1c T2b T1c T2a T1c T1c —
7.3 6.4 5.0 7.5 6.2 6.6 4.4 6.3 6.2
6 (3 ⫹ 3) 6 (3 ⫹ 3) 7 (3 ⫹ 4) 6 (3 ⫹ 3) 6 (3 ⫹ 3) 5 (3 ⫹ 2) 6 (3 ⫹ 3) 6 (3 ⫹ 3) 6
34.3 47.4 43.5 12.0 33.4 29.8 14.2 28.0 30.3
55.8 71.7 68.4 25.3 56.2 55.2 30.5 51.5 51.8
1.63 1.51 1.57 2.11 1.68 1.85 2.15 1.84 1.79
0.316 0.320 0.336 0.313 0.332 0.336 0.317 0.336 0.326
141 156 158 95 135 128 102 137 131
44.56 49.92 53.09 29.73 44.82 43.01 32.33 46.03 42.94
0.80 0.70 0.78 1.18 0.80 0.78 1.06 0.89 0.83
planning volume (55.5 cm3 vs. 51.8 cm3, respectively). The comparability of these 2 volumes suggests that our postimplant evaluation included the prostate gland with margin. The CT volume determined by method A is expected to be greater than the preplan target volume due to edema which usually results in about a 20% volume increase (12). In addition, postoperative dosimetric volume determinations via the ultrasound approach resulted in a volume nearly identical to the initial ultrasound evaluation (32.4 cm3 vs. 30.3 cm3). An exact correlation between these 2 volumes is
not expected because of differences in volume calculation algorithms between the ultrasound and the treatment planning software and errors inherent to the transfer of hard copies. Postoperative prostate volume determinations via a prostate only volume resulted in a volume approximately 25% larger than the ultrasound volume approach. If the defined CT volumes are expressed as ellipsoids, the average ellipsoid dimension was 4.73 cm, 4.21 cm, and 3.96 cm for delineation methods A, B, and C, respectively. Despite the volume determinations being markedly different, no signif-
Table 2. Dosimetric results vs. target volume approach* CT Volume (cm3) Patient
A
B
C
69.3 1
D90 (Gy) A
B
67.6
154 20.3
61.8 5
41.7 51.9
6
179
34.6 7
163
60.9 8
148
Avg.
55.5
42.2
19.5 43.3
45.8 97.1
165
20.9 19.9
93.9 94.3
162 32.4
22.0 48.0
91.5
155 39.2
48.4 41.0
153
26.4 25.5
99.7 89.9
31.4
24.5 61.3
98.4
145 45.0
59.0 51.0
169
18.9 23.8
99.6 97.6
17.0
19.8 47.9
99.2
161 24.6
50.9 54.3
184
26.7 19.6
96.9 97.3
33.6
28.8 57.0
98.3
170 37.9
55.9 46.4
158
20.8 29.4
99.7
162
20.4 52.7
98.5 94.3
38.8
12.0
50.6 52.6
178 153
13.3 22.2
98.4
168
13.9
28.9 50.3
93.7
13.9
14.1
33.4
98.4
17.6 21.3
47.5 97.5
C
14.4
96.1 97.9
B
41.1 33.1
168
A 15.2
95.3
165
C
39.7
97.9
163
B
V200 (%)
37.7
155
43.9
A
94.2
159
29.4
C
92.9
43.6 51.2
V150 (%)
90.5
149
68.7
B
155
41.6
4
A
153 36.8
3
C
146 51.4
2
V100 (%)
20.3 47.5
19.7
* I-125 implant only, 145 Gy mPD, day 0 dosimetry. Target volume approach: A ⫽ CT prostate volume plus margin; B ⫽ CT prostate volume; C ⫽ pretreatment ultrasound volume superimposed on CT.
1114
I. J. Radiation Oncology
●
Biology
●
Physics
Table 3. Descriptive and correlation statistics Parameter Postoperative volume V100 V150 V200 D100 D90
Methods compared
Mean
Range
Vector correlation
A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C A:B B:C A:C
1.42 1.24 1.75 0.971 0.995 0.966 0.967 1.009 0.975 1.051 1.040 1.094 1.007 1.209 0.943 0.958 0.982 0.941
0.28 0.52 0.73 0.046 0.044 0.045 0.142 0.196 0.260 0.188 0.180 0.319 2.956 3.593 2.839 0.074 0.080 0.106
0.966 0.854 0.931 0.821 0.851 0.887 0.960 0.978 0.948 0.967 0.968 0.921 0.089 ⫺0.376 ⫺0.204 0.889 0.919 0.837
icant differences were appreciated for any of the evaluated dosimetric parameters via the three approaches. Based on the cohort averages in the bottom row in Table 2, there is a trend toward higher D90 values via the ultrasound approach. Table 3 summarizes pair-wise comparisons between the three volumetric approaches for the various parameters. Except for postoperative volume, the means of these ratios are all close to 1.00; however, the maximum–minimum range for D100 is quite large. The last column in this table is the linear correlation coefficient obtained by plotting the parameters for one approach against those of another. All the parameters exhibit a high degree of correlation except for the pair-wise plots of D100. These coefficients are independent of the order in which the two quantities are taken. Figure 1 further illustrates the marked differences in volume determination and D100 and emphasizes the modest variation between the other evaluated dosimetric parameters. Data in this figure was obtained by taking the cohort mean of the absolute value of the natural log of the pairwise ratios for a given parameter such as D90 (method A): D90 (method B) for each patient. For a given dosimetry parameter, Q, and CT volume delineation approach A, B, or C, this is expressed as
1n
冉冊 X Y
冘冏 8
⫽
冉
冊冏
parameter Q, approach X 1 䡠 1n 8 i⫽1 parameter Q, approach Y
i
where i is the patient number. Since the log of a ratio has the same magnitude regardless of how the ratio is written, this construct highlights not only deviations from unity but also the spread of the values. For postoperative volume, the average of this log measure is 0.376 (range 0.23– 0.55) while the average variation for the other parameters except for D100 in terms of the log measure is less than one-fifth of this: V100 ⫽ 0.026 (0.014 – 0.034), V150 ⫽ 0.057
Volume 44, Number 5, 1999
(0.047– 0.077), V200 ⫽ 0.077 (0.058 – 0.112), D90 ⫽ 0.045 (0.030 – 0.062). Since some of the voxels contributing to D100 lie in high-dose gradient regions, this parameter is too sensitive to local conditions to serve as a quality measure regardless of the volumetric approach. This excessive variation of D100 that makes it unacceptable in evaluating quality has been observed by others (29, 30), and D100 is included here as a contrast to the usual quality parameters whose log ratio measures indicate that they deviate only slightly from unity and cluster tightly about that value for individual patients. DISCUSSION Over the past decade, prostate brachytherapy has emerged as a viable option for the treatment of early-stage carcinoma of the prostate gland. Recent publications have reported that in terms of local control and freedom from biochemical failure, the results of conformal brachytherapy are as favorable as the most positive radical prostatectomy series (1–7). For 125I monotherapy, Stock et al. (18, 19) have reported a dose threshold for biochemical NED (no evidence of disease) survival. Their dosimetric analysis reported in terms of a D90 revealed that a dose of 140 –160 Gy (TG-43 dosimetry) obtained on postimplant day 30 correlated with optimal biochemical NED survival (19). CT-based postoperative dosimetric analysis provides detailed information regarding the dose distribution to the prostate/periprostatic region, urethra, and rectum (8, 9, 11– 13, 24). Recently, the American Brachytherapy Society (ABS) recommended that prostate implants be defined in terms of a D90, V100, and V150 (31). Merrick et al. (13) recently reported dosimetric evaluation for 60 consecutive patients in terms of ABS guidelines which found that the three quality parameters evaluated are correlated with one another. One area of subjectivity in postoperative dosimetry involves the determination of postimplant CT prostate/target volumes. Approaches to determine postimplant CT target volumes include prostate only volumes (20), prostate plus margin volumes to ensure an adequate extracapsular dose (11–13), and volumes based on the preplanning ultrasound (21). Our data indicate that for a planning philosophy that enlarges the preplanning ultrasound volume approximately 80% and routinely plans seeds in the periprostatic region, the determination of postimplant volume via CT scans and its relationship on dosimetric quality is somewhat academic. The data in Fig. 1 show that while the defined volume ratios and D100 ratios either deviate considerably from unity or have a wide spread of values, the commonly used quality parameters have ratios that cluster tightly about unity. A schematic rationale for this conclusion is illustrated in Fig. 2. When the mPD isodose is just outside the largest defined target volume (except for occasional low-dose incursions), coverage indexes such as V100 and D90 will be relatively insensitive to CT volume delineation. This is because lowdose incursions into the smallest defined prostate volume
Implant quality vs. CT volume
Fig. 1. The mean of the absolute value of the natural log of each of the listed parameters evaluated pair-wise for each patient between the three methods. The variance in the ratios of post-op volume and D100 is large compared to the variance in the other parameters.
(approach C) will occupy a proportionately similar fraction of the larger volumes studied. This result should apply rigorously only to implants using a planning philosophy that prescribes the dose and implants seeds to an enlarged planning volume with margin. Furthermore, homogeneity measures such as V150 may not be invariant to volume drawing technique when the implant approach is less dosimetrically
Fig. 2. A schematic of a volume planned and implanted with periprostatic margins to encompass extracapsular disease. The outermost oval represents the volume encompassed by the prescribed mPD isodose (planning volume with margin) while the heavy line traces the isodose in the post implant plan corresponding to prescribed mPD. A typical low dose incursion due to inadequate execution of the preplan is marked by the shaded region. The lighter ovals labeled A, B and C represent the three methods of drawing the target volume as summarized in Table 2. The fractional volume covered by the mPD (V100) is relatively insensitive to CT volume drawing technique in this case. Other parameters such as V150 will also be similar among the 3 target definitions if the implant is dosimetrically homogeneous.
● G. S. MERRICK et al.
1115
homogeneous such as peripheral or uniform seed loading. We took great care to limit the V150 of the planning volume to approximately 35%. This results in a relatively homogeneic dose to the target volume with margin and results in relative independence between dosimetric quality and volume determination within the planning volume. As such, within the prostate gland, there is no actual point dose escalation, but our approach does result in a greater integral dose to the patient due to significant dose to the extracapsular region. A 5-center multi-institutional review using implant philosophies comparable to ours also reported the relative independence between V80 (the fraction of the treatment volume covered by 80% of the mPD), V100, and postimplant volume determination (32). We do not believe these results are transferable to an implant philosophy that exclusively utilizes high activity intraprostatic seeds. Although these implants also irradiate the prostate with a 3–5 mm periprostatic margin, a steep dose gradient exists in the periprostatic region, and as such, dosimetric quality may be highly dependent on CT volume determination. Our data reveals significant differences in CT-based volumes depending on each of the three investigated approaches. However, the D90, V100, and V150 as well as V200 do not differ significantly while D100 is found to be too erratic to be of any use in quality evaluations. This is due to the fact that some of the voxels contributing to D100 lie in high-dose gradient regions. Roy et al. (29) and Yu et al. (30) reported D100 to be overly sensitive to differences in target volumes to serve as a measure of dosimetric quality. The postimplant CT volumes determined via the prostate plus margin technique are comparable to our preplanning volume while the postimplant volumes determined via the ultrasound approach are comparable to the planning ultrasound volume. As such, these approaches appear to be highly reproducible in terms of volume determination. The prostate only approach resulted in a volume determination intermediate between the two extremes. Regardless of which of the three postimplant CT volume determinations are utilized, the difference in the average dimension of the CT-determined volume calculated as an ellipsoid is less than 10 mm in dimension for any patient. Via our planning approach, the average dimensional enlargement from preplan ultrasound volume to planning volume lies within this range. Recently Waterman et al. (17) reported postoperative dosimetry for 10 patients evaluated via serial CT scans on day 0 and weeks 1, 3, 7, and 15. Edema affected postoperative dosimetry in 7 of the 10 cases. In the three postoperative dosimetry cases where the V100 was greater than 90% on day 0, no significant improvement in V100 determinations was noted with time. All three of these cases were preplanned with a prostate plus margin technique (33). These three cases result in day 0 dosimetry comparable to what we report in this paper and in prior publications (11–13, 24). The other seven cases in Waterman’s publication were planned with a prostate only volume. Waterman et al. (17) indicated that edema is markedly less important in assessment of dosimetric quality for a preplanning philos-
1116
I. J. Radiation Oncology
●
Biology
●
Physics
ophy that entailed prostate plus margin compared to a planning volume that includes prostate only. Dosimetric coverage to the prostate gland improves as the time interval between brachytherapy and the postimplant CT scan increases (15). As edema resolves, all dosimetric quality parameters increase numerically (except for V## parameters such as V80 that may already be at 100%), and this is true regardless of which approach to CT volume determination is used. Therefore, brachytherapists who perform CT dosimetry on day 14 or day 30 will also find that quality is relatively insensitive to their drawing technique.
Volume 44, Number 5, 1999
In summary, a planning philosophy that includes the prostate with margin in the target volume is relatively independent of postimplant CT volume determination in terms of dosimetric quality. This effect of a planning margin stabilizing quality indicators is implied in the report by Waterman et al. (17), who found that the effect of edema is also lessened in appropriately executed implants that were planned with margin. As such, via the implant philosophy that we utilize, the determination of postimplant prostate volume has relatively little effect on ultimate reported dosimetric quality.
REFERENCES 1. Blasko JC, Wallner K, Grimm PD, Ragde H. Prostate specific antigen based disease control following ultrasound guided iodine-125 implantation for stage T1/T2 prostatic carcinoma. J Urol 1995;154:1096 –1099. 2. Wallner K, Roy J, Zelefsky M, Fuks Z, Harrison L. Short term freedom from disease progression after I-125 prostate implantation. Int J Radiat Oncol Biol Phys 1994;30:405– 409. 3. Wallner K, Roy J, Harrison L. Tumor control and morbidity following transperineal I-125 implantation for stage T1/T2 prostatic carcinoma. J Clin Oncol 1996;14:449 – 453. 4. Dattoli M, Wallner K, Sorace R, Koval J, Cash J, Acosta R, Brown C, Etheridge J, Binder M, Brunelle R, Kirwan N, Sanches S, Stein D, Wasserman S. Pd-103 brachytherapy and external beam radiation therapy for clinically localized high risk prostatic carcinoma. Int J Radiat Oncol Biol Phys 1996; 35:875– 879. 5. Ragde H, Blasko JC, Grimm PD, Kenny GM, Sylvester JE, Hoak DC, Landin K, Cavanagh W. Interstitial I-125 radiation without adjuvant therapy in the treatment of clinically localized prostate cancer. Cancer 1997;80:442– 453. 6. Dattoli MJ, Wallner KE, Cash JC, Ross RE, Koval JM, Sorace RA. Pd-103 brachytherapy for clinical T1/T2 prostate carcinomas. Int J Radiat Oncol Biol Phys 1997;39:221. 7. Ragde H, Blasko JC, Grimm PD, Kenny GM, Sylvester J, Hoak DC, Cavanagh W, Landin K. Brachytherapy for clinically localized prostate cancer: Results at 7 and 8 year followup. Sem Surg Onc 1997;13:438 – 443. 8. Willins J, Wallner K. CT based dosimetry for transperineal I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:347–353. 9. Wallner K, Roy J, Harrison L. Dosimetry guidelines to minimize urethral and rectal morbidity following transperineal I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1995;32:465– 471. 10. Willins J, Wallner K. Time-dependent changes in CT-based dosimetry of I-125 prostate brachytherapy. Rad Onc Invest 1998;6:157–160. 11. Merrick GS, Butler WM, Dorsey AT, Walbert HL. Prostatic conformal brachytherapy: I-125/Pd-103 postoperative dosimetric analysis. Rad Onc Invest 1997;5:305–313. 12. Merrick GS, Butler WM, Dorsey AT, Walbert HL. Influence of timing on the dosimetric analysis of transperineal ultrasound-guided prostatic conformal brachytherapy. Rad Onc Invest 1998;6:182–190. 13. Merrick GS, Butler WM, Dorsey AT, Lief JH. The potential role of various dosimetric quality indications in prostate brachytherapy. Int J Radiat Oncol Biol Phys 1999;44:717– 724. 14. Prestidge BR, Bice WS, Prete JJ, Buchholz TA. A dose
15.
16.
17.
18.
19. 20.
21. 22.
23.
24.
25.
26.
27.
volume analysis of permanent transperineal prostatic brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:238. Prestidge BR, Bice WS, Kiefer ET, Prete JJ. Timing of computed tomography based post-implant assessment following permanent transperineal prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;40:1111–1115. Moerland MA, Wijrdeman HK, Beersma R, Bakker CJG, Battermann JJ. Evaluation of permanent I-125 prostate implants using radiography and magnetic imaging. Int J Radiat Oncol Biol Phys 1997;37:927–933. Waterman FM, Yue N, Corn BW, Dicker AP. Edema associated with I-125 or Pd-103 prostate brachytherapy and its impact on post-implant dosimetry: An analysis based on serial CT acquisitions. Int J Radiat Oncol Biol Phys 1998;41:1069 – 1077. Stock RG, Stone NN, DeWyngaert JK, Lavaginini P, Unger PD. Prostate antigen findings and biopsy results following interactive ultrasound guided transperineal brachytherapy for early stage carcinoma. Cancer 1996;77:2386 –2392. Stock RG, Stone NN, Tabert A, Iannuzzi C, DeWyngaert JK. A dose response study for I-125 prostate implants. Int J Radiat Oncol Biol Phys 1998;41:101–108. Wallner K, Blasko J, Dattoli MJ. Implant evaluation. In: Wallner K, Blasko J, Dattoli MJ, editors. Prostate brachytherapy made complicated. Seattle: Smart Medicine Press; 1997. p. 9.1–9.20. Grimm PD. Personal communication. Merrick GS, Butler WM, Farthing WH, Dorsey AT, Adamovich E. The impact of Gleason score accuracy as a criterion for prostate brachytherapy patient selection. J Brachyther Int 1998;14:113–121. Nath R, Anderson LL, Luxton G, Weaver KA, Williamson JF, Meigooni AS. Dosimetry of interstitial brachytherapy sources: Recommendations of the AAPM Radiation Therapy Committee Task Group No. 43. Med Phys 1995;22:209 –234. Merrick GS, Butler WM, Dorsey AT, Lief JH, Walbert HL, Blatt HJ. Rectal dosimetric analysis following prostate brachytherapy. Int J Radiat Oncol Biol Phys 1999;43:1021– 1027. Partin AW, Kattan MW, Subong ENP, Walsh PC, Wojno KJ, Oesterling JE, Scardino PT, Pearson JD. Combination of prostate-specific antigen, clinical stage and Gleason score to predict pathological stage of localized prostate cancer. JAMA 1997;277:1445–1451. Cookson MS, Fleshner NE, Soloway SM, Fair WR. Prognostic significance of prostate-specific antigen in stage T1c prostate cancer treated by radical prostatectomy. Urol 1997;49: 887– 893. Wills ML, Sauvageot J, Partin AW, Gurganus R, Epstein JI. Ability of sextant biopsies to predict radical prostatectomy stage. Urol 1998;51:759 –764.
Implant quality vs. CT volume
28. Wallner K, Blasko JC, Dattoli MJ. Implant design. In: Wallner K, Blasko J, Dattoli MJ, editors. Prostate brachytherapy made complicated. Seattle: Smart Medicine Press; 1997. p. 6.1– 6.42. 29. Roy JN, Wallner KE, Harrington PJ, Ling CC, Anderson LL. A CT-based evaluation method for permanent implants: application to prostate. Int J Radiat Oncol Biol Phys 1993;26: 163–169. 30. Yu Y, Waterman FM, Suntharalingam N, Schulsinger A. Limitations of the minimum peripheral dose as a parameter for
● G. S. MERRICK et al.
1117
dose specification in permanent 125I prostate implants. Int J Radiat Oncol Biol Phys 1996;34:717–725. 31. Nag S. ABS Prostate brachytherapy consensus statement. American Brachytherapy Society Annual Meeting, Albuquerque, New Mexico, May 31, 1998. 32. Bice WS, Prestidge BR, Grimm PD, Friedland JL, Feygelman V, Roach M, Prete JJ, Dubois DF, Blasko JC. Centralized multiinstitutional postimplant analysis for interstitial prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;41:921–927. 33. Waterman F. Personal communication.