- Email: [email protected]
Standardization of prostate brachytherapy treatment plans
Int. J. Radiation Oncology Biol. Phys., Vol. 50, No. 1, pp. 257–263, 2001 Copyright © 2001 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/01/$–see front matter
PII S0360-3016(01)01471-7
PHYSICS CONTRIBUTION
STANDARDIZATION OF PROSTATE BRACHYTHERAPY TREATMENT PLANS ROGER OVE, M.D., PH.D.,* KENT WALLNER, M.D.,†‡ KAS BADIOZAMANI, M.D.,‡ TAMMY KORJSSEON, C.M.D., R.T.T.,† AND STEVEN SUTLIEF, PH.D.†‡ *Department of Radiation Oncology, University of Alabama Medical Center, Birmingham, AL; †Radiation Oncology, Puget Sound Health Care System, Department of Veterans Affairs, Seattle, WA; ‡Department of Radiation Oncology, University of Washington, Seattle, WA Purpose: Whereas custom-designed plans are the norm for prostate brachytherapy, the relationship between linear prostate dimensions and volume calls into question the routine need for customized treatment planning. With the goal of streamlining the treatment-planning process, we have compared the treatment margins (TMs) achieved with one standard plan applied to patients with a wide range of prostate volumes. Methods and Materials: Preimplant transrectal ultrasound (TRUS) images of 50 unselected University of Washington patients with T1–T2 cancer and a prostate volume between 20 cc and 50 cc were studied. Patients were arbitrarily grouped into categories of 20 –30 cc, 30 – 40 cc, and 40 –50 cc. A standard 19-needle plan was devised for patients in the 30- to 40-cc range, using an arbitrary minimum margin of 5 mm around the gross tumor volume (GTV), making use of inverse planning technology to achieve 100% coverage of the target volume with accentuation of dose at the periphery and sparing of the central region. The idealized plan was applied to each patient’s TRUS study. The distances (TMs) between the prostatic edge (GTV) and treated volume (TV) were determined perpendicular to the prostatic margin. Results: Averaged over the entire patient group, the ratio of thickness to width was 1.4, whereas the ratio of length to width was 1.3. These values were fairly constant over the range of volumes, emphasizing that the prostate retains its general shape as volume increases. The idealized standard plan was overlaid on the ultrasound images of the 17 patients in the 30- to 40-cc group and the V100, the percentage of target volume receiving 100% or more of the prescription dose, was 98% or greater for 15 of the 17 patients. The lateral and posterior TMs fell within a narrow range, most being within 2 mm of the idealized 5-mm TM. To estimate whether a 10-cc volume-interval stratification was reasonable, the standard plan generated from the 30- to 40-cc prostate model was applied to 5 patients each from the 20- to 30-cc group and the 40- to 50-cc group. Using the standard plan designed for the 30- to 40-cc group, the TMs were closer to 10 mm than to 5 mm for the smaller volume glands and too small for the larger volume ones, assuming an ideal margin of 5 mm. Conclusion: The application of standardized plans to prostate brachytherapy is feasible. Stratifying the volume in 10-cc intervals appears to be adequate, suggesting that the majority of cases appropriate for treatment with brachytherapy might be treated with three standard plans. While the authors believe that the use of a limited number of standard treatment plans is feasible, practical, and medically acceptable, it should be emphasized that the use of a standard plan should always be previewed by computer-aided application to the particular patient’s planning images. © 2001 Elsevier Science Inc. Brachytherapy, Prostate brachytherapy, Standardized plan, Planning.
scription isodose line (Fig. 1) (2, 3). In most patients, ECE is limited to within 3 mm of the prostatic edge, and a TM of 3 mm should suffice, assuming that the implant is executed perfectly and there is no implant-related change to the prostate volume. In practice, however, implants are not executed perfectly, and a variable degree of implant-related prostatic swelling occurs, such that the ideal TM is not known (4 – 8). Although custom-designed plans are the norm, the relationship between linear prostate dimensions and volume calls into question the routine need for customized treatment
INTRODUCTION With the publication of encouraging clinical series, the number of patients treated with prostate brachytherapy is likely to increase substantially (1). One inefficiency in treating large numbers of patients is the process of custom planning. In practice, custom plans are devised with the goal of treating the prostate plus a margin, to allow for source displacement and potential extracapsular disease extension (ECE). The treatment margin (TM) is considered as the perpendicular distance from the prostatic edge to the preReprint request to: Dr. Kent Wallner, Radiation Oncology, No. 174, Department of Veterans Affairs, 1660 South Columbia Way, Seattle, Washington 98108-1597.
Accepted for publication 10 April 2000.
257
258
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 1. Treated margin (TM) measurement points.
planning. As the prostate volume increases, the linear dimensions change slowly due to the power law relationship between the linear dimension and volume. If the prostate shape is similar between patients, the variability in linear dimensions should be relatively small between patients. With the goal of streamlining the treatment-planning process, we have compared the TMs achieved with one standard plan applied to patients with a wide range of prostate volumes. The analysis provides insight as to the volume range over which a single plan might be applicable, and establishes simple guidelines for excluding patients from the standard plan approach. METHODS AND MATERIALS Preimplant TRUS images of 50 unselected University of Washington patients with T1-T2 cancer and a prostate volume between 20 cc and 50 cc were studied. All images were obtained using a Siemens Sonoline Prima ultrasound scanner (Siemens Medical Group, Inc., Issaquah, WA) and a Winston–Barzell stepper unit (Barzell–Whitmore Maroon Bells, Inc., Sarasota, FL). The prostate image was outlined
Volume 50, Number 1, 2001
by a technologist and checked by a physician. The greatest prostatic width (lateral), length (craniocaudal), and thickness (anterior–posterior) dimensions were determined from the TRUS images. Patients were arbitrarily grouped into categories of 20 –30 cc, 30 – 40 cc, and 40 –50 cc. A standard plan was devised for patients in the 30- to 40-cc range. Averages of the length, width, and thickness for the group were used to synthesize an idealized prostate model, taken to be a modified ellipsoid with diameters equal to the averages from the group. The posterior–lateral border of the ellipsoid was extended slightly to account for the typical prostate shape. A standard plan was generated for the gross target volume (prostate, or GTV), using an arbitrary minimum margin of 5 mm around the GTV, making use of inverse-planning technology to achieve 100% coverage of the target volume with accentuation of dose at the periphery and sparing of the central region. Criteria used to define the objective function were the following: (1) 100% of the GTV covered by 100% dose (144 Gy), (2) minimum 5 mm TMs, (3) urethral dose less than 250 Gy, and (4) the surface of the prostate was prescribed to 120%. The latter two requirements tended to give peripherally loaded implants. The plan was generated using the Varian Medical Systems/MMS TheracPLUS B3DTUI planning system (Palo Alto, CA), using NycomedAmersham Model 6711 125I sources (Nycomed-Amersham, Chicago, IL). Activity was chosen to be 0.4 mCi per source. The plan included 19 needles with 96 sources, using endto-end spacing of 0.0 or 0.5 cm. The TG-43 brachytherapy dose-calculation formalism was used, using an air-kerma strength traceable to the 1985 standard. Minor modifications of the inversely generated plan were made to limit the number of needles and minimize left-to-right asymmetry (Fig. 2). The idealized plan was applied to each patient’s TRUS study, using the treatment-planning software. In each case, the posterior prostate margin was aligned on the same grid row as the idealized plan, without regard for where the anterior prostatic margin fell. The distances (TMs) between the prostatic edge (GTV) and treated
Fig. 2. Seed locations and isodoses for the standard plan, generated for an idealized prostate in the 30- to 40-cc range. Cuts are shown at 0.0, 1.5, and 3.0 cm from the prostatic base.
Standard plans
● R. OVE et al.
259
Fig. 3. Width, thickness (AP), and length measurements vs. volume. Over this volume range, the data can reasonably be fit to a straight line.
volume (TV) were determined perpendicular to the prostatic margin (Fig. 1). RESULTS The orthogonal prostate dimensions increased gradually with increasing prostate volume (Fig. 3). The similar slopes of the dimension vs. volume lines for the width, thickness, and length suggest that the prostate tends to enlarge equally in all directions. The average dimensions for the 17 patients in the 30- to 40-cc group were 4.7 ⫾ 0.4 cm by 3.5 ⫾ 0.3 cm by 3.9 ⫾ 0.6 cm. The average volume of the group was 34.8 cc. The greatest variation in linear dimension was in the length, possibly reflecting uncertainty in defining the prostate base and apex on TRUS images. Average measurements of width, thickness, and length for the 20- to 30-cc group were 4.5 ⫾ 0.5 cm, 3.0 ⫾ 0.2 cm, and 3.4 ⫾ 0.4 cm,
respectively, whereas in the 40- to 50-cc group they were 5.0 ⫾ 0.5 cm, 3.8 ⫾ 0.2 cm, and 4.1 ⫾ 0.5 cm, respectively. The relationship between linear dimensions and volume is illustrated in Figs. 4 and 5. Averaged over the entire patient group, the ratio of width to thickness was 1.4 (⫾ 0.17) while the ratio of width to length was 1.3 (⫾ 0.22). These values were fairly constant over the range of volumes, emphasizing that the prostate retains its general shape as volume increases. Regardless of the exact description of the geometric model representing the idealized prostate, if the shape is invariant as size increases, then volume will be proportional to the cube of any of the linear dimensions. Consequently, when considering the linear dimensions as a function of volume we expect the following relationships to hold: X ⫽ aV1/3 ,
Fig. 4. Volume vs. width, thickness, and length.
Y ⫽ bV1/3 ,
Z ⫽ cV1/3
260
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 50, Number 1, 2001
Fig. 5. Predicted cube root behavior of the linear measurements vs. volume, plotted with the measure data. Length is omitted for clarity. Closed diamonds represent the width and open squares represent the thickness.
The factors a, b, and c are constants, determined to be 1.46, 1.07, and 1.16, respectively by fitting to the measured data. Figure 5 shows the measured and predicted data for X and Y as a function of volume—the agreement of these curves is consistent with simple overall scaling of the gland as volume increases. To test how well individual cases would be covered by a single plan, the idealized standard plan was overlaid on the ultrasound images of the 17 patients in the 30- to 40-cc group. The posterior edge of the gland was arbitrarily aligned with the 2.0 row of the TRUS needle-guide template (2.0 cm from the TRUS probe central axis) as is customarily done for planning and intraoperative patient setup. The most cephalad plane of the standardized plan was aligned with the prostatic base. The V100, the percentage of target vol-
ume receiving 100% or more of the prescription dose as determined from the dose volume histogram, was 98% or greater for 15 of the 17 patients (Fig. 6). Regions of underdosage were located primarily at the apex and base. The lateral and posterior TMs fell within a narrow range, most being within 2 mm of the idealized 5 mm TM (Fig. 7). The posterior margins show the least variability, because all plans were aligned along the 2.0-cm needle-guide template line. The anterior margin showed the greatest variation, due variability in the prostatic thickness and the arbitrary alignment of the posterior margin on the 2.0 row of the template. To estimate whether a 10-cc volume interval stratification was reasonable, the standard plan generated from the 30- to 40-cc prostate model was applied to 5 patients each from the 20- to 30-cc group and the 40- to 50-cc group. In the 20- to
Fig. 6. V100 vs. prostatic dimensions for patients in the 30- to 40-cc group. Both patients with a V100 below 100% had at least one dimension near the upper limits of the group (some points represent more than 1 patient).
Standard plans
● R. OVE et al.
261
Fig. 7. Treatment margins vs. volume, as measured on the central prostatic image with the standardized plan applied to each patient in the 30- to 40-cc group (the left and right lateral margins are grouped together).
30-cc group, all target volumes were covered (V100 ⫽ 100%) (Fig. 8). In the 40- to 50-cc group, 3 cases were grossly undertreated as judged by the V100 and margins, with V100s of 98.9%, 99.1%, 98.1%, 93.7%, and 88.6%. Using the standard plan designed for the 30- to 40-cc group, the TMs were closer to 10 mm than to 5 mm for the smaller volume glands and too small for the larger volume ones, assuming an ideal margin of 5 mm. The cases in the 30- to 40-cc volume range that failed to be covered by the standard plan were predictable based on their linear measurements. In each case that failed, one of the linear dimensions was more than 5 mm greater than the average measurement that defined the standard prostate. In contrast, all of the cases that were well covered had linear dimensions within 5 mm of the averages. In the 40- to 50-cc group, all 5 cases failed to be covered with a 5-mm margin. DISCUSSION Although individualized treatment planning for prostate brachytherapy is currently the norm, custom-made plans
may be unnecessary for most patients. Instead, a limited number of standard plans, chosen on the basis of prostate volume, might suffice and would streamline the planning process. The first feasibility requirement for using a limited number of standardized plans is prostate shape similarity between patients. According to the data presented here, there appears to be little variation in prostate geometry between patients. Our second requirement for limiting prostate brachytherapy to a limited number of standard plans is that the clinical application of a standard plan be in no measurable way dosimetrically inferior to a custom plan. Unfortunately, any effort to evaluate dosimetric coverage by a standard plan(s) is hampered by the lack of rationally established, data-based treatment margins, dosimetric measures of implant quality, and the unpredictability of implant-related prostate swelling (3, 9, 10). The V100 (fraction of prostate volume receiving prescription dose) and the D90 (the minimum dose covering 90% of the target volume) have been commonly used criteria for dosimetric coverage (14). Recommendations vary,
Fig. 8. V100 and treatment margins (TMs) achieved with the standard plan vs. transrectal ultrasound (TRUS) volume. As the volume increased, the target coverage by the standard plan decreased (left) along with the TMs (right).
262
I. J. Radiation Oncology
●
Biology
●
Physics
from comparing the D90 with the prescription dose (D90 ⬎ prescription) to determining if at least 80% of the target volume receives the prescription dose or higher (V100 ⬎ 80%) (10 –14). Tailor-made preimplant plans presumably cover the entire volume to 100%, and, in the majority of cases in the 30- to 40-cc range, the standardized plan achieved the same. Considering the uncertainty introduced by implant procedure-related prostate swelling, minor differences in V100 between custom and standard plans are likely to be insignificant. Probably more meaningful than the V100 or D90 are the TMs achieved, because the real goal of the implant is to cover the prostate with a cancercidal dose, allowing for ECE and imperfection in implant implementation and implantrelated prostate swelling. Unfortunately, what constitutes an adequate TM is clouded by the unpredictable, variable degree of implant-related prostate volume changes. Current evidence supports a treatment margin of 3 mm to cover extraprostatic extension. In prostatectomy patients that would have been candidates for brachytherapy, the maximum (per patient) radial extent beyond the prostate was studied in two recent series and found to be 0.6 –1.4 mm on average (15, 16). Because extraprostatic extension tends to occur primarily posterolaterally, our standard plan was always applied with the posterior porstatic border arbitrarily aligned on the 2.0 template row, to minimize the possibility of an inadequate posterior–lateral margin. A 5-mm rather than 3-mm margin was used to increase the likelihood of achieving adequate margins when applied in practice, where implant-related prostate shape changes may occur. There is a limit to the range of prostate volumes that can be covered by a standard plan. Although one standard plan appears to suffice for the majority of cases within a 10-cc size range, more than one plan would probably be needed if a wider volume range were treated. Applying the standard plan to the 20- to 30-cc group resulted in complete coverage but possibly excessive margins. In the 40- to 50-cc group, 3 of 5 patients were grossly undertreated, and would have required substantial plan modifications to achieve a V100 of 100% and larger TMs. But with most prostates falling into the 20- to 50-cc volume range, a limited number of standard plans would likely suffice for most patients. In addition to streamlining the planning process, there are
Volume 50, Number 1, 2001
several potential dosimetric advantages to using standardized treatment plans. With custom-designed, unique plans, there is the potential for excessive dose inhomogeneity, even given the same prostate volume and shape (17). Though its significance for prostate brachytherapy has not been well documented, dose inhomogeneity may lead to increased complication rates. Inhomogeneity could be minimized with an inverse-planned standard set of plans, and kept constant between patients. A second benefit of standardized plans is the potential to make the most efficient use of sources, to minimize the number used. Finally, the use of standard plans should decrease the potential for operator error in preparing preloaded needles and implementing the plan intraoperatively, because the few plans would be used over and over again. While the authors believe that the use of a limited number of standard treatment plans is feasible, practical and medically acceptable, it should be emphasized that the use of a standard plan should always be previewed by computeraided application to the particular patients’s planning images. To do so, we suggest that the idealized plan be applied to each patient’s TRUS study, with the posterior prostate margin aligned on the same grid row as the idealized plan. The distances (TMs) between the prostatic edge (GTV) and treated volume (TV) should be determined perpendicular to the prostatic margin and compared to some standard, keeping in mind that the proper “standard” is still a subject of some controversy (3). Interactive experimentation with inverse planning software tends to reinforce the notion that the minimum of the objective function is fairly broad. There is an enormous number of solutions that satisfy our objective and subjective criteria for an optimized implant. Much of the current effort in treatment planning involves comparison of similar plans, with the differences between choices unlikely to be clinically significant. A more-efficient use of resources would be the use of top-quality tools to generate a limited number of class solutions that can be applied to most patients, eliminating the need for extensive individualized treatment planning and perhaps allowing for the manufacture of prepackaged implant kits. This strategy could be equally advantageous for preplanned or intraoperatively planned systems.
REFERENCES 1. Ramos CG, Carvalhal GF, Smith DS, et al. Retrospective comparison of radical retropubic prostatectomy and 125-iodine brachytherapy for localized prostate cancer. J Urol 1999; 161:1212–1215. 2. ICRU. Prescribing, recording, and reporting photon beam therapy. ICRU Report 50 1993. p. 1– 8. 3. Han B, Wallner K, Aggarwal S, et al. Treatment margins for prostate brachytherapy. Semin Urol Oncol 2000;18:137–141. 4. Waterman FM, Yue N, Corn BW, et al. Edema associated with I-125 or Pd-103 prostate brachytherapy and its impact on post-implant dosimetry: An analysis based on serial CT acquistion. Int J Radiat Oncol Biol Phys 1998;41:1069 –1077. 5. Willins J, Wallner KE. Time-dependent changes in CT-based
dosimetry of I-125 prostate brachytherapy. Radiat Oncol Invest 1998;6:157–160. 6. Prestidge BR, Bice WS, Kiefer EJ, et al. Timing of computed tomography-based postimplant assessment following permanent transperineal prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;40:1111–1115. 7. Yue N, Chen Z, Peschel R, et al. Optimum timing for imagebased evaluation of 125-I and 103-Pd prostate seed implants. Int J Radiat Oncol Biol Phys 1999;45:1063–1072. 8. Merrick GS, Butler WM, Dorsey AT, et al. Influence of timing on the dosimetric analysis of transperineal ultrasound-guided, prostatic conformal brachytherapy. Radiat Oncol Invest 1998; 6:182–190.
Standard plans
9. Badiozamani K, Wallner K, Sutlief S, et al. Anticipating prostatic volume changes due to prostate brachytherapy. Radiat Oncol Invest 1999;7:360 –364. 10. Merrick GS, Butler WM, Dorsey AT, et al. The dependence of prostate postimplant dosimetric quality on CT volume determination. Int J Radiat Oncol Biol Phys 1999;44:1111– 1117. 11. Nag SN, Beyer D, Friedland J, et al. American Brachytherapy Society (ABS) recommendations for transperineal permanent brachytherapy of prostate cancer. Int J Radiat Oncol Biol Phys 1999;44:789 –799. 12. Stock RG, Stone NN, Tabert A, et al. A dose–response study for I-125 prostate implants. Int J Radiat Oncol Biol Phys 1998;41:101–108. 13. Willins J, Wallner K. CT-based dosimetry for transperineal
● R. OVE et al.
14.
15. 16.
17.
263
I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:347–353. Yu Y, Waterman FM, Suntharalingham N, et al. Limitations of the minimum peripheral dose as a parameter for dose specification in permanent I-125 prostate implants. Int J Radiat Oncol Biol Phys 1996;34:717–725. Davis BJ, Pisansky TM, Wilson TM, et al. The radial distance of extraprostatic extension of prostate carcinoma: Implications for prostate brachytherapy. Cancer 1999;85:2630 –2637. Sohayda C, Kupelian PA, Ciezki J, et al. Extent of extracapsular extension: Implications for planning conformal radiotherapy and brachytherapy. Int J Radiat Oncol Biol Phys 1998;42(Suppl.):132. Anderson LL. A. “natural” volume-dose histogram for brachytherapy. Med Phys 1986;13:898 –903.
PII S0360-3016(01)01471-7
PHYSICS CONTRIBUTION
STANDARDIZATION OF PROSTATE BRACHYTHERAPY TREATMENT PLANS ROGER OVE, M.D., PH.D.,* KENT WALLNER, M.D.,†‡ KAS BADIOZAMANI, M.D.,‡ TAMMY KORJSSEON, C.M.D., R.T.T.,† AND STEVEN SUTLIEF, PH.D.†‡ *Department of Radiation Oncology, University of Alabama Medical Center, Birmingham, AL; †Radiation Oncology, Puget Sound Health Care System, Department of Veterans Affairs, Seattle, WA; ‡Department of Radiation Oncology, University of Washington, Seattle, WA Purpose: Whereas custom-designed plans are the norm for prostate brachytherapy, the relationship between linear prostate dimensions and volume calls into question the routine need for customized treatment planning. With the goal of streamlining the treatment-planning process, we have compared the treatment margins (TMs) achieved with one standard plan applied to patients with a wide range of prostate volumes. Methods and Materials: Preimplant transrectal ultrasound (TRUS) images of 50 unselected University of Washington patients with T1–T2 cancer and a prostate volume between 20 cc and 50 cc were studied. Patients were arbitrarily grouped into categories of 20 –30 cc, 30 – 40 cc, and 40 –50 cc. A standard 19-needle plan was devised for patients in the 30- to 40-cc range, using an arbitrary minimum margin of 5 mm around the gross tumor volume (GTV), making use of inverse planning technology to achieve 100% coverage of the target volume with accentuation of dose at the periphery and sparing of the central region. The idealized plan was applied to each patient’s TRUS study. The distances (TMs) between the prostatic edge (GTV) and treated volume (TV) were determined perpendicular to the prostatic margin. Results: Averaged over the entire patient group, the ratio of thickness to width was 1.4, whereas the ratio of length to width was 1.3. These values were fairly constant over the range of volumes, emphasizing that the prostate retains its general shape as volume increases. The idealized standard plan was overlaid on the ultrasound images of the 17 patients in the 30- to 40-cc group and the V100, the percentage of target volume receiving 100% or more of the prescription dose, was 98% or greater for 15 of the 17 patients. The lateral and posterior TMs fell within a narrow range, most being within 2 mm of the idealized 5-mm TM. To estimate whether a 10-cc volume-interval stratification was reasonable, the standard plan generated from the 30- to 40-cc prostate model was applied to 5 patients each from the 20- to 30-cc group and the 40- to 50-cc group. Using the standard plan designed for the 30- to 40-cc group, the TMs were closer to 10 mm than to 5 mm for the smaller volume glands and too small for the larger volume ones, assuming an ideal margin of 5 mm. Conclusion: The application of standardized plans to prostate brachytherapy is feasible. Stratifying the volume in 10-cc intervals appears to be adequate, suggesting that the majority of cases appropriate for treatment with brachytherapy might be treated with three standard plans. While the authors believe that the use of a limited number of standard treatment plans is feasible, practical, and medically acceptable, it should be emphasized that the use of a standard plan should always be previewed by computer-aided application to the particular patient’s planning images. © 2001 Elsevier Science Inc. Brachytherapy, Prostate brachytherapy, Standardized plan, Planning.
scription isodose line (Fig. 1) (2, 3). In most patients, ECE is limited to within 3 mm of the prostatic edge, and a TM of 3 mm should suffice, assuming that the implant is executed perfectly and there is no implant-related change to the prostate volume. In practice, however, implants are not executed perfectly, and a variable degree of implant-related prostatic swelling occurs, such that the ideal TM is not known (4 – 8). Although custom-designed plans are the norm, the relationship between linear prostate dimensions and volume calls into question the routine need for customized treatment
INTRODUCTION With the publication of encouraging clinical series, the number of patients treated with prostate brachytherapy is likely to increase substantially (1). One inefficiency in treating large numbers of patients is the process of custom planning. In practice, custom plans are devised with the goal of treating the prostate plus a margin, to allow for source displacement and potential extracapsular disease extension (ECE). The treatment margin (TM) is considered as the perpendicular distance from the prostatic edge to the preReprint request to: Dr. Kent Wallner, Radiation Oncology, No. 174, Department of Veterans Affairs, 1660 South Columbia Way, Seattle, Washington 98108-1597.
Accepted for publication 10 April 2000.
257
258
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 1. Treated margin (TM) measurement points.
planning. As the prostate volume increases, the linear dimensions change slowly due to the power law relationship between the linear dimension and volume. If the prostate shape is similar between patients, the variability in linear dimensions should be relatively small between patients. With the goal of streamlining the treatment-planning process, we have compared the TMs achieved with one standard plan applied to patients with a wide range of prostate volumes. The analysis provides insight as to the volume range over which a single plan might be applicable, and establishes simple guidelines for excluding patients from the standard plan approach. METHODS AND MATERIALS Preimplant TRUS images of 50 unselected University of Washington patients with T1-T2 cancer and a prostate volume between 20 cc and 50 cc were studied. All images were obtained using a Siemens Sonoline Prima ultrasound scanner (Siemens Medical Group, Inc., Issaquah, WA) and a Winston–Barzell stepper unit (Barzell–Whitmore Maroon Bells, Inc., Sarasota, FL). The prostate image was outlined
Volume 50, Number 1, 2001
by a technologist and checked by a physician. The greatest prostatic width (lateral), length (craniocaudal), and thickness (anterior–posterior) dimensions were determined from the TRUS images. Patients were arbitrarily grouped into categories of 20 –30 cc, 30 – 40 cc, and 40 –50 cc. A standard plan was devised for patients in the 30- to 40-cc range. Averages of the length, width, and thickness for the group were used to synthesize an idealized prostate model, taken to be a modified ellipsoid with diameters equal to the averages from the group. The posterior–lateral border of the ellipsoid was extended slightly to account for the typical prostate shape. A standard plan was generated for the gross target volume (prostate, or GTV), using an arbitrary minimum margin of 5 mm around the GTV, making use of inverse-planning technology to achieve 100% coverage of the target volume with accentuation of dose at the periphery and sparing of the central region. Criteria used to define the objective function were the following: (1) 100% of the GTV covered by 100% dose (144 Gy), (2) minimum 5 mm TMs, (3) urethral dose less than 250 Gy, and (4) the surface of the prostate was prescribed to 120%. The latter two requirements tended to give peripherally loaded implants. The plan was generated using the Varian Medical Systems/MMS TheracPLUS B3DTUI planning system (Palo Alto, CA), using NycomedAmersham Model 6711 125I sources (Nycomed-Amersham, Chicago, IL). Activity was chosen to be 0.4 mCi per source. The plan included 19 needles with 96 sources, using endto-end spacing of 0.0 or 0.5 cm. The TG-43 brachytherapy dose-calculation formalism was used, using an air-kerma strength traceable to the 1985 standard. Minor modifications of the inversely generated plan were made to limit the number of needles and minimize left-to-right asymmetry (Fig. 2). The idealized plan was applied to each patient’s TRUS study, using the treatment-planning software. In each case, the posterior prostate margin was aligned on the same grid row as the idealized plan, without regard for where the anterior prostatic margin fell. The distances (TMs) between the prostatic edge (GTV) and treated
Fig. 2. Seed locations and isodoses for the standard plan, generated for an idealized prostate in the 30- to 40-cc range. Cuts are shown at 0.0, 1.5, and 3.0 cm from the prostatic base.
Standard plans
● R. OVE et al.
259
Fig. 3. Width, thickness (AP), and length measurements vs. volume. Over this volume range, the data can reasonably be fit to a straight line.
volume (TV) were determined perpendicular to the prostatic margin (Fig. 1). RESULTS The orthogonal prostate dimensions increased gradually with increasing prostate volume (Fig. 3). The similar slopes of the dimension vs. volume lines for the width, thickness, and length suggest that the prostate tends to enlarge equally in all directions. The average dimensions for the 17 patients in the 30- to 40-cc group were 4.7 ⫾ 0.4 cm by 3.5 ⫾ 0.3 cm by 3.9 ⫾ 0.6 cm. The average volume of the group was 34.8 cc. The greatest variation in linear dimension was in the length, possibly reflecting uncertainty in defining the prostate base and apex on TRUS images. Average measurements of width, thickness, and length for the 20- to 30-cc group were 4.5 ⫾ 0.5 cm, 3.0 ⫾ 0.2 cm, and 3.4 ⫾ 0.4 cm,
respectively, whereas in the 40- to 50-cc group they were 5.0 ⫾ 0.5 cm, 3.8 ⫾ 0.2 cm, and 4.1 ⫾ 0.5 cm, respectively. The relationship between linear dimensions and volume is illustrated in Figs. 4 and 5. Averaged over the entire patient group, the ratio of width to thickness was 1.4 (⫾ 0.17) while the ratio of width to length was 1.3 (⫾ 0.22). These values were fairly constant over the range of volumes, emphasizing that the prostate retains its general shape as volume increases. Regardless of the exact description of the geometric model representing the idealized prostate, if the shape is invariant as size increases, then volume will be proportional to the cube of any of the linear dimensions. Consequently, when considering the linear dimensions as a function of volume we expect the following relationships to hold: X ⫽ aV1/3 ,
Fig. 4. Volume vs. width, thickness, and length.
Y ⫽ bV1/3 ,
Z ⫽ cV1/3
260
I. J. Radiation Oncology
●
Biology
●
Physics
Volume 50, Number 1, 2001
Fig. 5. Predicted cube root behavior of the linear measurements vs. volume, plotted with the measure data. Length is omitted for clarity. Closed diamonds represent the width and open squares represent the thickness.
The factors a, b, and c are constants, determined to be 1.46, 1.07, and 1.16, respectively by fitting to the measured data. Figure 5 shows the measured and predicted data for X and Y as a function of volume—the agreement of these curves is consistent with simple overall scaling of the gland as volume increases. To test how well individual cases would be covered by a single plan, the idealized standard plan was overlaid on the ultrasound images of the 17 patients in the 30- to 40-cc group. The posterior edge of the gland was arbitrarily aligned with the 2.0 row of the TRUS needle-guide template (2.0 cm from the TRUS probe central axis) as is customarily done for planning and intraoperative patient setup. The most cephalad plane of the standardized plan was aligned with the prostatic base. The V100, the percentage of target vol-
ume receiving 100% or more of the prescription dose as determined from the dose volume histogram, was 98% or greater for 15 of the 17 patients (Fig. 6). Regions of underdosage were located primarily at the apex and base. The lateral and posterior TMs fell within a narrow range, most being within 2 mm of the idealized 5 mm TM (Fig. 7). The posterior margins show the least variability, because all plans were aligned along the 2.0-cm needle-guide template line. The anterior margin showed the greatest variation, due variability in the prostatic thickness and the arbitrary alignment of the posterior margin on the 2.0 row of the template. To estimate whether a 10-cc volume interval stratification was reasonable, the standard plan generated from the 30- to 40-cc prostate model was applied to 5 patients each from the 20- to 30-cc group and the 40- to 50-cc group. In the 20- to
Fig. 6. V100 vs. prostatic dimensions for patients in the 30- to 40-cc group. Both patients with a V100 below 100% had at least one dimension near the upper limits of the group (some points represent more than 1 patient).
Standard plans
● R. OVE et al.
261
Fig. 7. Treatment margins vs. volume, as measured on the central prostatic image with the standardized plan applied to each patient in the 30- to 40-cc group (the left and right lateral margins are grouped together).
30-cc group, all target volumes were covered (V100 ⫽ 100%) (Fig. 8). In the 40- to 50-cc group, 3 cases were grossly undertreated as judged by the V100 and margins, with V100s of 98.9%, 99.1%, 98.1%, 93.7%, and 88.6%. Using the standard plan designed for the 30- to 40-cc group, the TMs were closer to 10 mm than to 5 mm for the smaller volume glands and too small for the larger volume ones, assuming an ideal margin of 5 mm. The cases in the 30- to 40-cc volume range that failed to be covered by the standard plan were predictable based on their linear measurements. In each case that failed, one of the linear dimensions was more than 5 mm greater than the average measurement that defined the standard prostate. In contrast, all of the cases that were well covered had linear dimensions within 5 mm of the averages. In the 40- to 50-cc group, all 5 cases failed to be covered with a 5-mm margin. DISCUSSION Although individualized treatment planning for prostate brachytherapy is currently the norm, custom-made plans
may be unnecessary for most patients. Instead, a limited number of standard plans, chosen on the basis of prostate volume, might suffice and would streamline the planning process. The first feasibility requirement for using a limited number of standardized plans is prostate shape similarity between patients. According to the data presented here, there appears to be little variation in prostate geometry between patients. Our second requirement for limiting prostate brachytherapy to a limited number of standard plans is that the clinical application of a standard plan be in no measurable way dosimetrically inferior to a custom plan. Unfortunately, any effort to evaluate dosimetric coverage by a standard plan(s) is hampered by the lack of rationally established, data-based treatment margins, dosimetric measures of implant quality, and the unpredictability of implant-related prostate swelling (3, 9, 10). The V100 (fraction of prostate volume receiving prescription dose) and the D90 (the minimum dose covering 90% of the target volume) have been commonly used criteria for dosimetric coverage (14). Recommendations vary,
Fig. 8. V100 and treatment margins (TMs) achieved with the standard plan vs. transrectal ultrasound (TRUS) volume. As the volume increased, the target coverage by the standard plan decreased (left) along with the TMs (right).
262
I. J. Radiation Oncology
●
Biology
●
Physics
from comparing the D90 with the prescription dose (D90 ⬎ prescription) to determining if at least 80% of the target volume receives the prescription dose or higher (V100 ⬎ 80%) (10 –14). Tailor-made preimplant plans presumably cover the entire volume to 100%, and, in the majority of cases in the 30- to 40-cc range, the standardized plan achieved the same. Considering the uncertainty introduced by implant procedure-related prostate swelling, minor differences in V100 between custom and standard plans are likely to be insignificant. Probably more meaningful than the V100 or D90 are the TMs achieved, because the real goal of the implant is to cover the prostate with a cancercidal dose, allowing for ECE and imperfection in implant implementation and implantrelated prostate swelling. Unfortunately, what constitutes an adequate TM is clouded by the unpredictable, variable degree of implant-related prostate volume changes. Current evidence supports a treatment margin of 3 mm to cover extraprostatic extension. In prostatectomy patients that would have been candidates for brachytherapy, the maximum (per patient) radial extent beyond the prostate was studied in two recent series and found to be 0.6 –1.4 mm on average (15, 16). Because extraprostatic extension tends to occur primarily posterolaterally, our standard plan was always applied with the posterior porstatic border arbitrarily aligned on the 2.0 template row, to minimize the possibility of an inadequate posterior–lateral margin. A 5-mm rather than 3-mm margin was used to increase the likelihood of achieving adequate margins when applied in practice, where implant-related prostate shape changes may occur. There is a limit to the range of prostate volumes that can be covered by a standard plan. Although one standard plan appears to suffice for the majority of cases within a 10-cc size range, more than one plan would probably be needed if a wider volume range were treated. Applying the standard plan to the 20- to 30-cc group resulted in complete coverage but possibly excessive margins. In the 40- to 50-cc group, 3 of 5 patients were grossly undertreated, and would have required substantial plan modifications to achieve a V100 of 100% and larger TMs. But with most prostates falling into the 20- to 50-cc volume range, a limited number of standard plans would likely suffice for most patients. In addition to streamlining the planning process, there are
Volume 50, Number 1, 2001
several potential dosimetric advantages to using standardized treatment plans. With custom-designed, unique plans, there is the potential for excessive dose inhomogeneity, even given the same prostate volume and shape (17). Though its significance for prostate brachytherapy has not been well documented, dose inhomogeneity may lead to increased complication rates. Inhomogeneity could be minimized with an inverse-planned standard set of plans, and kept constant between patients. A second benefit of standardized plans is the potential to make the most efficient use of sources, to minimize the number used. Finally, the use of standard plans should decrease the potential for operator error in preparing preloaded needles and implementing the plan intraoperatively, because the few plans would be used over and over again. While the authors believe that the use of a limited number of standard treatment plans is feasible, practical and medically acceptable, it should be emphasized that the use of a standard plan should always be previewed by computeraided application to the particular patients’s planning images. To do so, we suggest that the idealized plan be applied to each patient’s TRUS study, with the posterior prostate margin aligned on the same grid row as the idealized plan. The distances (TMs) between the prostatic edge (GTV) and treated volume (TV) should be determined perpendicular to the prostatic margin and compared to some standard, keeping in mind that the proper “standard” is still a subject of some controversy (3). Interactive experimentation with inverse planning software tends to reinforce the notion that the minimum of the objective function is fairly broad. There is an enormous number of solutions that satisfy our objective and subjective criteria for an optimized implant. Much of the current effort in treatment planning involves comparison of similar plans, with the differences between choices unlikely to be clinically significant. A more-efficient use of resources would be the use of top-quality tools to generate a limited number of class solutions that can be applied to most patients, eliminating the need for extensive individualized treatment planning and perhaps allowing for the manufacture of prepackaged implant kits. This strategy could be equally advantageous for preplanned or intraoperatively planned systems.
REFERENCES 1. Ramos CG, Carvalhal GF, Smith DS, et al. Retrospective comparison of radical retropubic prostatectomy and 125-iodine brachytherapy for localized prostate cancer. J Urol 1999; 161:1212–1215. 2. ICRU. Prescribing, recording, and reporting photon beam therapy. ICRU Report 50 1993. p. 1– 8. 3. Han B, Wallner K, Aggarwal S, et al. Treatment margins for prostate brachytherapy. Semin Urol Oncol 2000;18:137–141. 4. Waterman FM, Yue N, Corn BW, et al. Edema associated with I-125 or Pd-103 prostate brachytherapy and its impact on post-implant dosimetry: An analysis based on serial CT acquistion. Int J Radiat Oncol Biol Phys 1998;41:1069 –1077. 5. Willins J, Wallner KE. Time-dependent changes in CT-based
dosimetry of I-125 prostate brachytherapy. Radiat Oncol Invest 1998;6:157–160. 6. Prestidge BR, Bice WS, Kiefer EJ, et al. Timing of computed tomography-based postimplant assessment following permanent transperineal prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;40:1111–1115. 7. Yue N, Chen Z, Peschel R, et al. Optimum timing for imagebased evaluation of 125-I and 103-Pd prostate seed implants. Int J Radiat Oncol Biol Phys 1999;45:1063–1072. 8. Merrick GS, Butler WM, Dorsey AT, et al. Influence of timing on the dosimetric analysis of transperineal ultrasound-guided, prostatic conformal brachytherapy. Radiat Oncol Invest 1998; 6:182–190.
Standard plans
9. Badiozamani K, Wallner K, Sutlief S, et al. Anticipating prostatic volume changes due to prostate brachytherapy. Radiat Oncol Invest 1999;7:360 –364. 10. Merrick GS, Butler WM, Dorsey AT, et al. The dependence of prostate postimplant dosimetric quality on CT volume determination. Int J Radiat Oncol Biol Phys 1999;44:1111– 1117. 11. Nag SN, Beyer D, Friedland J, et al. American Brachytherapy Society (ABS) recommendations for transperineal permanent brachytherapy of prostate cancer. Int J Radiat Oncol Biol Phys 1999;44:789 –799. 12. Stock RG, Stone NN, Tabert A, et al. A dose–response study for I-125 prostate implants. Int J Radiat Oncol Biol Phys 1998;41:101–108. 13. Willins J, Wallner K. CT-based dosimetry for transperineal
● R. OVE et al.
14.
15. 16.
17.
263
I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys 1997;39:347–353. Yu Y, Waterman FM, Suntharalingham N, et al. Limitations of the minimum peripheral dose as a parameter for dose specification in permanent I-125 prostate implants. Int J Radiat Oncol Biol Phys 1996;34:717–725. Davis BJ, Pisansky TM, Wilson TM, et al. The radial distance of extraprostatic extension of prostate carcinoma: Implications for prostate brachytherapy. Cancer 1999;85:2630 –2637. Sohayda C, Kupelian PA, Ciezki J, et al. Extent of extracapsular extension: Implications for planning conformal radiotherapy and brachytherapy. Int J Radiat Oncol Biol Phys 1998;42(Suppl.):132. Anderson LL. A. “natural” volume-dose histogram for brachytherapy. Med Phys 1986;13:898 –903.