Int. J. Radiation Oncology Biol. Phys., Vol. 56, No. 5, pp. 1480 –1487, 2003 Copyright © 2003 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/03/$–see front matter
doi:10.1016/S0360-3016(03)00527-3
PHYSICS CONTRIBUTION
ISODOSE PATTERNS IN PATIENTS WITH INADEQUATE PROSTATE BRACHYTHERAPY COVERAGE
WAYNE
DANIEL R. REED, D.O.,*† KENT E. WALLNER, M.D.,*†‡ GREGORY MERRICK, M.D.,§ BUTLER, PH.D.,§ BEN H. HAN, M.D.,* STEVE G. SUTLIEF, PH.D.,*† AND PAUL S. CHO, PH.D.*
*Department of Radiation Oncology, University of Washington, Seattle, WA; †Radiation Oncology, Puget Sound Health Care System, Department of Veterans Affairs, Seattle, WA; ‡Radiation Oncology, Group Health Cooperative, Seattle, WA; §Schiffler Cancer Center, Wheeling, WV Purpose: The development of a practical, real-time dosimetry system should result in improved implant dose distributions and higher prostate cancer control rates. Our purpose here is to demonstrate that intraoperative isodose reconstruction in relation to the seed distribution, even without accurate registration with the prostatic volume, can likely identify an inadequate implant intraoperatively and guide corrective seed placement. Methods and Materials: A total of 102 Pd-103 implants performed by standard techniques, using a modified peripheral loading pattern, were studied. A postimplant computed tomography (CT) scan was obtained 2– 4 h after the implant. The contoured images and sources were entered into a Varian Variseed 7.0 treatment planning system. Dosimetric parameters analyzed included the percent of the postimplant prostate or rectal volume covered by the prescription dose (V100), and the dose that covers 90% of the postimplant prostate volume (D90). Isodose patterns were analyzed at midprostate, and for the entire prostate. Adverse isodose patterns were defined as gaps, holes, islands. Isodose gaps are subprescription intervals between the prostatic margin and the prescription isodose. Isodose holes are regions of subprescription dose within the prostate. Isodose islands are isolated regions >prescription dose inside the prostatic margins. Results: Characteristic isodose patterns were predictive for V100 values. Midprostatic isodose holes were seen in 55% of patients with a V100 < 80%, 5% of patients with a V100 of 80 –90%, and only 1% of patients with a V100 > 90%. When analyzing the entire prostate, isodose holes were seen in 55% of patients with a V100 < 80%, 18% of patients with a V100 of 80 –90%, and 9% of patients with a V100 > 90%. Midprostatic isodose islands were seen in 55% of patients with a V100 < 80%, 5% of patients with a V100 of 80 –90%, and no patient with a V100 > 90%. When analyzing the entire prostate, isodose islands were seen in all patients with V100 < 80%, 36% of patients with a V100 of 80 –90%, and only 1% of patients with a V100 > 90%. The likelihood of a V100 less than 80% was best predicted by the presence of isodose holes or islands at midprostate. Patients with either finding had an 86% chance of having a V100 < 80%. Conclusion: These semiquantitative findings can provide practical guidelines for intraoperative dosimetry, to provide a more rational guide to intraoperative postimplant assessment and modification. If isodose holes or islands are seen within the implanted volume, additional seeds are added to the affected region to obtain a V100 > 80%. © 2003 Elsevier Inc. Prostate, Brachytherapy, Dosimetry.
INTRODUCTION
moderate degree of unpredictability in postimplant quality parameters achieved in the execution of a computergenerated preplan, due to minor source placement errors and an unpredictable degree of postimplant prostatic swelling (4 – 6). V100s typically range from 75% to 95%, even in experienced hands (3, 7–10). Higher D90s and V100s have both been correlated with higher biochemical control rates (3). For simplicity, only a V100 ⬍ 80% will be considered suboptimal here, because the authors believe that V100s are easier to conceptualize (7).
Transperineal ultrasound– guided prostate brachytherapy is widely used for early-stage prostate cancer, with increasing evidence that the likelihood of cancer eradication is partly related to postimplant dosimetric parameters (1–3). The principal quality parameters used are the dose delivered to 90% of the prostate (D90) and the percent of the postimplant prostate volume covered by the prescription dose (V100). Unfortunately, there is a Reprint requests to: Kent Wallner, M.D., Radiation Oncology (174), Puget Sound Health Care System, Department of Veterans Affairs, 1660 S. Columbian Way, Seattle, WA 98108 –1597. Tel: (206) 768-5356; Fax: (206) 768-5331; E-mail: kent.wallner@
med.va.gov Received Nov 11, 2002, and in revised form Mar 24, 2003. Accepted for publication Apr 14, 2003. 1480
Isodose patterns
● D. REED et al.
1481
Fig. 1. Examples of marginal isodose gaps extending inside of the prostatic margins.
Because postimplant dosimetric parameters partly predict the likelihood of biochemical control, it seems desirable to assess them intraoperatively, and to make modifications (adding sources) to optimize parameters and maximize the chance for cure (11). Ideally, imperfections in the implant should be corrected with the patient still under anesthesia, without requiring a repeat trip to the procedure room. Although intraoperative postimplant assessment and modification is intuitively appealing, its implementation is hindered by the need for intraoperative computed tomography (CT) or magnetic resonance imaging to accurately identify the prostate gland and localize implanted seeds. The two readily available intraoperative imaging modalities—transrectal ultrasound (TRUS) and fluoroscopy—are currently not designed to provide both simultaneous image sets. However, combining their images by spatial co-registration (fusion) offers the potential for a practical intraoperative dosimetric assessment (12). We are working to streamline TRUS and fluoroscopic co-registration, to make it a practical intraoperative technique (13). In the process, we have noted that certain isodose patterns are typical of patients with inadequate dosimetric indices, and that such patterns could be readily discerned from a qualitative perusal of isodose lines overlaid on the seed distribution alone, even without the availability of CT or TRUS prostatic images. Using a modified peripheral source loading pattern, with most sources placed near the prostatic capsule, patients with high postimplant V100s typically have minimal underdosed (cold) regions within the prostate— cold areas are nearly always limited to the prostatic periphery (Fig. 1). In contrast, two distinct intraprostatic isodose patterns are strongly
associated and specific for V100s less than 80%. The first are intraprostatic isodose gaps more than 1 cm from the prostatic margin (Fig. 1). The second pattern is isodose holes within the prostate (Fig. 2). The third pattern is the presence of isodose islands inside the prostatic margins (Fig. 3). To quantify what we noted in casual observation of adequate vs. inadequate implants, we reviewed 102 consecutive Pd-103 patients to correlate isodose patterns with suboptimal dosimetric parameters. Our purpose in doing so is to demonstrate that intraoperative isodose reconstruction in relation to the seed distribution, even without accurate registration with the prostatic volume, can likely identify an inadequate implant intraoperatively.
METHODS AND MATERIALS The 102 patients reported here were treated consecutively at the Puget Sound Veterans Affairs on two randomized protocols, with a planned combined accrual of 1200 (14). Patients implanted between June 2001 and June 2002 were studied. Implants were performed by standard techniques, using a modified peripheral loading pattern (15, 16). Low-risk patients, with Gleason score 5– 6 and prostatespecific antigen (PSA) 4 –10 ng/mL, were randomized to implantation with I-125 (144 Gy, TG-43) vs. Pd-103 (125 Gy, NIST-99). Only those patients randomized to Pd-103 (Theraseed, Norcross, GA) are shown here. Pd-103 source strength was 2.5 U/source in all cases (NIST-00). Intermediate-risk patients, with Gleason score 7 or higher and/or PSA 10 –20 ng/mL, were randomized to implantation
1482
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 5, 2003
Fig. 2. Examples of patients with isodose holes of increasing maximum dimension. Open holes communicate with extraprostatic tissue receiving less than prescription dose.
with Pd-103 (90 vs. 115 Gy [NIST-1999]), with 44 vs. 20 Gy external beam radiation therapy, respectively. Pd-103 source strength was 2.5 U/source in all cases (NIST-00). A postimplant CT scan was obtained 2– 4 h after the implant. The CT-derived postimplant target volume was determined as previously described, using 5-mm images at every 5 mm (17). The contoured images and sources were entered into a Varian Variseed treatment planning system (Charlottesville, VA). A redundancy check was performed on seed localization to prevent seed duplication. Prostatic dose–volume histograms were calculated using the prostatic edge identified on CT scan. Dosimetric parameters analyzed included the percent of the postimplant prostate or rectal volume covered by the prescription dose (V100), and the
dose that covers 90% of the postimplant prostate volume (D90). Isodose patterns were analyzed at midprostate, and for the entire prostate. Isodose gaps, holes, and islands were defined as illustrated in Figs. 1–3.
RESULTS Of the 102 patients studied, 11 had a V100 ⬍ 80%, 22 had a V100 of 80 –90%, and 68 had a V100 ⬎ 90% (Fig. 4). An inadequate V100 (⬍80%) was more likely in patients with larger degrees of postimplant volume increase, but there was substantial overlap in the postimplant to preim-
Isodose patterns
● D. REED et al.
1483
Fig. 3. Examples of isodose islands.
plant prostate volume ratio in patients with a V100 greater or less than 80% (Fig. 5). Midprostatic isodose peripheral gaps ⬎1.0 cm were seen in 1 of 11 of patients (9%) with V100 ⬍ 80%, 2 of 22 of patients (9%) with a V100 of 80 –90%, and 1 of 69 patients (1.4%) with a V100 ⬎ 90% (Fig. 6). When analyzing the entire prostate, peripheral isodose gaps ⬎1.0 cm were seen in all 11 patients with a V100 ⬍ 80%, 50% of patients with a V100 of 80 –90%, and 12% of patients with a V100 ⬎ 90% (Fig. 6).
Midprostatic isodose holes were seen in 55% of patients (6 of 11) with a V100 ⬍ 80%, 5% of patients (1 of 22) with a V100 of 80 –90%, and only 1 patient (1.4%) with a V100 ⬎ 90% (Fig. 7). When analyzing the entire prostate, isodose holes were seen in 55% of patients with a V100 ⬍ 80%, 18% of patients with a V100 of 80 –90%, and 9% of patients with a V100 ⬎ 90% (Fig. 7). Midprostatic isodose islands were seen in 55% of patients with a V100 ⬍ 80%, 5% of patients with a V100 of 80 –90%, and no patient with a V100 ⬎ 90% (Fig. 8). When analyzing the entire prostate, isodose islands were seen in all patients with V100 ⬍ 80%, 36% of patients with a V100 of 80 –90%, and only 1 patient (1.4%) with a V100 ⬎ 90% (Fig. 8). All patients with a V100 ⬍ 80% had at least 1 unfavorable isodose event, whether the entire prostate or only the midprostatic image was analyzed (Fig. 9). The likelihood of a V100 less than 80% was best predicted by the presence of isodose holes or islands at midprostate. Patients with either finding had an 86% chance of having a V100 ⬍ 80% (Fig. 10).
DISCUSSION
Fig. 4. 102 Pd-103 patients, arranged in order of increasing V100. Eleven patients (11%) had a V100 ⬍ 80%.
We have shown here that with a modified peripheral seed loading placement, certain isodose patterns are characteristic of low V100s. All patients with a V100 less than 80% show at least 1 adverse isodose event at midprostate—a marginal gap ⬎1.0 cm, isodose holes, or isodose islands.
1484
I. J. Radiation Oncology
● Biology ● Physics
Volume 56, Number 5, 2003
Fig. 5. V100 plotted against the ratio of the postimplant to preimplant volume.
Fig. 6. Likelihood of isodose gaps ⬎1.0 cm in patients with V100 ⬍ 80%, 80 –90%, or ⬎90%.
Fig. 7. Likelihood of isodose holes in patients with V100 ⬍ 80%, 80 –90%, or ⬎90%.
Isodose patterns
● D. REED et al.
Fig. 8. LIkelihood of isodose islands in patients with V100 ⬍ 80%, 80 –90%, or ⬎90%.
Fig. 9. LIkelihood of any unfavorable isodose event in patients with a V100 ⬍ 80%, 80 –90%, or ⬎90%.
Fig. 10. Likelihood of V100 ⬍ 80% when any adverse pattern event is seen at midprostate (left) or the entire prostate (right).
1485
1486
I. J. Radiation Oncology
● Biology ● Physics
Conversely, only 2 patients (3%) with a V100 ⬎ 90% had an adverse event (Fig. 9) These semiquantitative findings can provide substantial practical guidelines for intraoperative dosimetry, to provide a more rational guide to intraoperative postimplant assessment and modification (11, 13). Future work regarding the interpretation of adverse isodose patterns should explore the use of D90 vs. V100, and how much of the prostate needs to be analyzed routinely (i.e., midslice, central one-third, central two-thirds, etc.) Another area in need of clarification is the effect of postimplant image acquisition timing on adverse isodose patterns vs. Quality Assessment parameters (18). Development of a practical real-time dosimetry system should result in improved implant dose distributions and higher cancer control rates. We have shown that TRUS and fluoroscopy fusion can provide such a solution, and we are working to verify its accuracy and to implement it fully in our clinical practice (13). In the meantime, based on the data summarized here, we have begun routinely using intraoperative isodose interpretation alone in relation to the seed distribution to make a quick quality assessment, and to guide the placement of additional sources. To do so, isodose reconstruction from three fluoroscopic views is viewed transversely, with anterior-posterior and right-left orientation. We qualitatively review the intraoperative isodose reconstruction for underdosed areas within the seed distribution. If isodose holes or islands are seen within the source cluster, additional seeds are added to the affected region or regions. The other isodose event—a marginal gap ⬎1.0 cm— cannot be used reliably without a prostatic contour overlay. It should be noted that the data presented here are for
Volume 56, Number 5, 2003
Pd-103 only. It is the authors’ impression that fewer isodose holes occur with I-125, due to less rapid dose fall-off (19). If the concepts described here are put into widespread clinical use, a separate detailed look at I-125 isodose patterns is warranted. At the present time, we have yet not implemented the superposition of the prostatic contour on the intraoperative isodose reconstruction. We are, however, able to orient the isodose reconstruction and implanted volume in relation to the superior-inferior and left-right prostatic edges. Assuming that the seeds were properly placed around the prostatic periphery, isodose holes or islands within the seed cluster likely represent underdosed intraprostatic regions that should be supplemented with additional seeds to bring the V100 above 80%. Although the dosimetric information gleaned from isodose patterns alone is insufficient to replace CT-based analysis, it is a substantial improvement over the current norm of qualitative, visual assessment of intraoperative fluoroscopic and TRUS images. In the future, we expect to show that supplemental seed placement using intraoperative isodose reconstruction and visual identification of adverse isodose events as described here will result in a lower likelihood of inadequate dosimetric parameters. While higher D90s and V100s are associated with higher tumor control rates, it is also imperative to be cognizant that higher urethral and rectal dose incurred in the process of implant augmentation could increase the likelihood of complications (11). Accordingly, the authors are intensifying our own efforts to allow for global evaluation of the effect of intraoperative implant modification (20, 21).
REFERENCES 1. Stock RG, Stone NN, Lo YC, et al. Postimplant dosimetry for 125-I prostate implants: Definitions and factors affecting outcome. Int J Radiat Oncol Biol Phys 2000;48:899–906. 2. Potters L, Cao Y, Calugaru E, et al. A comprehensive review of CT-based dosimetry parameters and biochemical control in patients treated with permanent prostate brachytherapy. Int J Radiat Oncol Biol Phys 2001;50:605–614. 3. Wallner K, Merrick G, True L, et al. I-125 versus Pd-103 for low risk prostate cancer: Preliminary biochemical outcomes from a prospective randomized mulitcenter trial. Submitted. 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 acquisition. Int J Radiat Oncol Biol Phys 1998;41:1069–1077. 5. Chen Z, Yue N, Wang X, et al. Dosimetric effects of edema in permanent prostate seed implants: A rigorous solution. Int J Radiat Oncol Biol Phys 2000;47:1405–1419. 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. Nag S, Bice W, DeWyngaert K, et al. The American Brachytherapy Society recommendations for permanent prostate brachytherapy postimplant dosimetric analysis. Int J Radiat Oncol Biol Phys 2000;46:221–230. 8. Bice WS, Prestidge BR, Grimm PD, et al. Centralized multi-
9. 10. 11. 12.
13. 14. 15. 16.
institutional postimplant analysis for interstitial prostate brachytherapy. Int J Radiat Oncol Biol Phys 1998;41:921–927. Bice WS, Prestidge BR, Sarosody MF. Sector analysis of prostate implants. Med Phys 2001;28:2561–2567. Sidhu S, Morris WJ, Spadinger I, et al. Prostate brachytherapy postimplant dosimetry: A comparison of prostate quadrants. Int J Radiat Oncol Biol Phys 2002;52:544–552. Mueller A, Wallner K, Merrick G, et al. Modification of prostate implants based on post-implant treatment margin assessment. Med Phys 2003;29:2782–2787. Haworth A, Ebert M, Joseph D, et al. Registration of prostate volume with radiographically identified Iodine-125 seeds for permanent implant evaluation. J Brachy Intl 2000;16:157– 167. Gong L, Cho P, Han B, et al. Registration of prostate brachytherapy seeds with prostate anatomy for postimplant dosimetry. Int J Radiat Oncol Biol Phys 2002;54:1322–1330. Wallner K, Merrick G, True L, et al. I-125 versus Pd-103 for low risk prostate cancer: Morbidity outcomes from a prospective randomized multicenter trial. Ca J Sci American 2002;8:67–73. Merrick GS, Butler WM. Modified uniform seed loading for prostate brachytherapy: Rationale, design, and evaluation. Techniques in Urology 2000;6:78–84. Mueller A, Wallner K, Corriveau J, et al. A reappraisal of local anesthesia for prostate brachytherapy. Radiother Oncol. In press.
Isodose patterns
17. Badiozamani KR, Wallner KE, Cavanagh W, et al. Comparability of CT-based and TRUS-based prostate volumes. Int J Radiat Oncol Biol Phys 1999;43:375–378. 18. Speight JL, Shinohara K, Pickett B, et al. Prostate volume change after radioactive seed implantation: Possible benefit of improved dose volume histogram with perioperative steroid. Int J Radiat Oncol Biol Phys 2000;48:1461– 1467. 19. Dicker AP, Lin C-C, Leeper DB, et al. Isotope selection for
● D. REED et al.
1487
permanent prostate implants? An evaluation of 103-Pd versus 125- based on radiobiological effectiveness and dosimetry. Semin Urol Oncol 2000;18:152–159. 20. Smith S, Wallner K, Merrick G, et al, Interpretation of preversus post-implant TRUS images. Med Phys 2003;30:920 – 924. 21. Han B, Wallner K, Merrick G, et al. Prostate brachytherapy seed identification on post-implant TRUS images. Med Phys 2003;30:898 –900.