A comparison of preplan transrectal ultrasound with preplan-CT in assessing volume and number of seeds needed for real-time ultrasound-based intra-operative planning in prostate 125I seed implantation

Results The median difference in volume between preimplant imaging and the intra-operative planning ultrasound was 3.59 and 5.2 cc for patients who underwent p-TRUS and p-CT, respectively. p-TRUS volumes more closely correlated with real-time intra-operative volumes ( R = 0.84 in all patients and R = 0.91 in hormone-naïve patients) vs. p-CT ( R = 0.82). The median number of seeds wasted using p-CT was 18 vs. 7 using volumes derived from p-TRUS. Conclusions The number of seeds ordered could be more accurately obtained from p-TRUS volumes, and this translated into less seed wastage. Our findings indicate that p-TRUS is a more accurate and an economically superior alternative to p-CT imaging in the era of real-time ultrasound planning. Keywords Intra-operative planning Preplan imaging Prostate brachytherapy Purpose Low–dose rate brachytherapy is an accepted treatment option for patients with low-risk prostate cancer (1, 2) . Historically, a preplanned technique has been used, requiring the patient to undergo a planning transrectal ultrasound (TRUS), in the simulated implant position before permanent seed placement. From this imaging study, a dosimetric plan was generated (3) . This preplanning technique is limited by the uncertainty in prostate size change, edema, and difficulty in reproducing the patient's preplanning position during the intra-operative procedure (4) . Additionally, preplan techniques do not easily accommodate the adjustment of a plan intra-operatively based on immediate dosimetric feedback, as is the case with real-time planning  (5, 6) . More recently, an intra-operative (real-time) treatment planning technique has been adapted by many institutions where dosimetry and volumetric information are obtained immediately before and during seed placement (7) . Several groups have demonstrated equivocal dosimetric results using real-time planning and it is being used at increasingly more centers (8–10) . Although this technique allows dosimetric planning to be done at the time of the procedure, some form of preplan imaging is still essential. Acquisition of accurate prostate volume is mandatory to identify total activity, activity per seed, and number of seeds needed to ensure adequate coverage of the target volume as well as anatomic limitations, such as pubic arch interference. The number of 125 I seeds needed to deliver the appropriate dose for the target volume can be predicted based on the volume of the prostate gland (11) . The use of MRI, CT, or TRUS to obtain prostate volumes varies among institutions and there is currently no consensus regarding the most appropriate imaging modality to obtain this information. A preplan CT (p-CT) scan is an appealing alternative to preplan transrectal ultrasound (p-TRUS) as postplan dosimetry is done with a CT scan and, theoretically, volumetric information should correlate well with this scan. Unfortunately, as seed placement is done with ultrasound, CT scans tend to overestimate or underestimate the volume of the prostate and therefore may compromise dosimetric parameters. Furthermore, accurately predicting the number of seeds needed for the procedure is dependent on accurate assessment of volume. We therefore conducted a retrospective study to compare how volumes obtained from p-CT scans or p-TRUS correlated with real-time ultrasound and postimplant CT volumes and the difference in accuracy of seed estimation between these techniques. Methods and materials Ninety-two patients with a diagnosis of organ-confined prostate cancer underwent 125 I permanent seed implants at Thomas Jefferson University Hospital for low-risk prostate cancer between February 2002 and August 2008. All patients presented with a favorable risk profile, and no patient exceeded stage T2b, Gleason score greater than 7, or prostate-specific antigen level above 10. Fifty-one patients underwent p-TRUS before intra-operative planning and 41 patients were evaluated by p-CT. Patients underwent p-TRUS or p-CT based on whether they received treatment at one of the two Thomas Jefferson University Hospital locations. Each hospital is staffed by a different radiation oncologist with the same urologist treating patients at both centers. p-CT scans were conducted in the Department of Radiation Oncology, and volumes were generated by the radiation oncologist using CMS FocalSim software (CMS, Saint Louis, MO). p-TRUS volumes were acquired by a urologist. Both procedures were performed approximately 1 month before the implant. The median number of seeds per cubic centimeter was 1.8. Median D 90 was 180.23 Gy, and a median of 97.32% of the prostate received 100% of the dose. No significant differences within these values were noted between the two groups. Volume measurements Patients who underwent prostate volume assessment via p-TRUS were scheduled for TRUS approximately 1 month before the seed implant procedure. After patients were positioned in left lateral decubitus position, the urologist placed the biplane rectal probe (Type 8551A) and acquired transverse images on a BK 1886 ultrasound unit (BK Medical, Herlev, Denmark). Height (anterior–posterior), width (left–right), and length (superior–inferior) dimensions of prostate were determined using the ellipsoid prostate formula to calculate the volume of the prostate (see Fig. 1 ) (12) . CT scans were conducted with the GE Lightspeed Scanner (GE Healthcare, Chalfont St. Giles, UK) using 2.5 mm slice thickness. Images were obtained with patients in the supine position. Images were transferred to FocalSim software, and the prostate volume was determined by a radiation oncologist. The prostate volume was delineated on all slices and volumetric data were acquired from FocalSim structure property statistics. An ellipsoid formula was not used for CT volumes. For both p-CT and p-TRUS methods, seeds were ordered based on prostate volume and total activity as determined from the Wu nomogram (13) . For real-time imaging, the patient was placed in stirrups in the lithotomy position with sufficient hip flexion to avoid the pubic arch. Using the biplane transrectal probe and ultrasound unit, transverse images were captured at 5-mm increments covering the prostate from base to apex. These images were subsequently sent to the Variseed treatment planning system (Varian Medical Systems, Palo Alto, CA) and the prostate was contoured by the radiation oncologist. Approximately 1 month after seed implantation, patients underwent a postimplant CT scan in the supine position to determine whether adequate dosimetric coverage had been achieved. Comparisons were made between p-TRUS volumes and p-CT volumes with the real-time ultrasound volumes and post-CT volumes. The Pearson correlation coefficient was used to evaluate the correlation between p-TRUS and p-CT with real-time ultrasound volumes. Results The median difference in volume between the initial screening imaging and the intra-operative planning ultrasound was 3.59 cc (standard deviation [SD], 3.3 cc) for patients who underwent p-TRUS and 5.2 cc (SD, 6.8 cc) for patients who underwent p-CT. The difference in volume between the real-time planning ultrasound and subsequent postimplant CT scan performed 1 month after implant was 3.8 cc (SD, 3.1 cc) for the p-TRUS and 2.4 cc (SD, 2.9 cc) for p-CT. p-TRUS volumes more closely correlated with real-time intra-operative volumes ( R = 0.84; see Fig 2 ) vs. p-CT ( R = 0.82; see Fig 3 ). Two-tailed paired t  tests conducted between p-TRUS and real-time ultrasound volumes did not indicate a statistically significant difference in volumes, whereas comparison of p-CT with real-time ultrasound volumes showed a statistically significant difference between the two volumes ( p = 0.0005). Nineteen patients received androgen suppression (leuprolide) before implantation. Indications for androgen suppression included Gleason score 7, clinical stage exceeding T2a, and prostate gland size greater than 50 cc on preimplant imaging (p-CT or p-TRUS). Androgen suppression was administered for a 3- to 6-month period, and patients were re-imaged 3–4 weeks before implantation. When patients who received androgen therapy were removed from the analysis, the Pearson coefficient improved to R = 0.91 when comparing p-TRUS with real-time ultrasound and remained the same for p-CT comparisons. Although it was thought that after months of hormone therapy the prostate gland would not change after the second preimplant image, the improved correlation when evaluating only patients who had not received hormone intervention indicates this is something that should be considered in all patients receiving hormones. p-CT was more closely correlated with postimplant CT-based volumes ( R = 0.89) compared with p-TRUS ( R = 0.78). The median number of seeds wasted using p-CT was 18 vs. 7 using volumes derived from p-TRUS. Although many factors are taken into consideration in determining the cost of 125 I seeds such as varying activity levels and cost between different vendors, the average cost for an 125 I seed is approximately US$35 per seed (Healthcare Common Procedure Coding System Outpatient Prospective Payment System C2639). The number of seeds ordered using p-CT resulted in wastage of approximately 18 seeds or approximately US$630 compared with US$245 when using p-TRUS volumes. Discussion Intra-operative planning is an attractive option for prostate seed placement as it allows incorporation of real-time volume and position of the prostate, allowing for improved dosimetric coverage. It also negates the time-consuming process of trying to match a previously generated plan to a patient whose positional differences may make it impossible to reproduce the plan exactly. Several studies have demonstrated that intra-operative plans have comparable dosimetry to preplanning techniques (14–17) and conform to the American Brachytherapy Society–defined doses. In addition, benefits such as reduction in number of seeds used for treatment, decreased procedure times, and improvement in dosimetric variables have been noted. A study by Shanahan et al . (14) attempted to compare preparation and procedure time between a preplanned technique and an interactive real-time technique and determined that a real-time planning technique reduced preplanning time, procedure time, and number of needles used while maintaining dosimetric coverage. Raben et al . (8) demonstrated not only a decrease in seeds implanted but also a dosimetric advantage with real-time techniques with respect to dosing to critical structures, such as the urethra and rectum when tissue constraints were able to be incorporated intra-operatively, as well. Within this paradigm of real-time planning, a method for ascertaining the number of seeds needed to complete the procedure and ensure adequate coverage is needed. Preplanning ultrasound has the ability to detect pubic arch interference as well as the presence or absence of median lobe hypertrophy in the treatment position. It also has the ability to measure three-dimensional size and prostate volume in a cost-efficient and an accurate manner (14) . Several studies have compared TRUS volumes with CT volumes and have demonstrated that CT is less accurate than ultrasound or MRI in delineating prostate volumes. Although some studies have demonstrated a tendency for CT scans to overestimate the size of the gland, others have demonstrated that CT scans are more prone to greater interobserver variability when compared with TRUS or MRI, leading to either overestimation or underestimation of the gland size dependent on the skill or preference of the radiation oncologist (18–20) . Although interobserver variability is inevitable in delineating prostate volumes, TRUS-based volumes are generally regarded to be more reliable than CT-based volumes because borders are more distinct. Our institution has demonstrated in prior reports significant interobserver variability when using CT volumes vs. ultrasound volumes. We found a variation in prostate volumes generated from ultrasound imagery between three different observers that ranged from 1% to 26%, with a median of 11% (21) . Similar findings were demonstrated by Smith et al . (22) , who reported a variation in prostate volumes of 7–32% (median 13%) when using ultrasound. In contrast, the interobserver variability when using CT scans in studies by Lee et al . (23) and Han et al . (24) demonstrated larger variations in volumes from 10% to 21% (median, 16%) and 9% to 29% (median, 17%), respectively, indicating greater interobserver variability with CT volumes compared with ultrasound volumes. The poor correlation between p-CT volumes and real-time ultrasound volumes does not allow for systematic compensation of error, as would be the case if we consistently overestimated the volume by a set amount. Within this patient cohort, interobserver differences could not be assessed. It is, however, apparent from the increased median difference in volumes between p-CT and real-time volumes as well as an inferior correlation coefficient that more variability existed for CT predictions of real-time TRUS values. Although in some cases the gland size was overestimated with CT, in others, the gland was underestimated, leaving no choice but to order additional seeds to avoid the devastating scenario of seed shortage in the operating room. For example, the largest degree of underestimation within the patients who underwent p-TRUS demonstrated an estimated volume of 37.54 cc compared with a real-time volume of 48.52 cc. Using the Wu nomogram, this would correspond to approximately 20% more activity needed than predicted using p-TRUS volumes (43.8 mCi compared with 37.5 mCi). The greatest underestimation in volume for p-CT was 41 cc compared with a real-time volume of 70 cc. This would correspond to 40% more activity needed than predicted by p-CT volumes. In addition, two-tailed paired t  tests conducted between p-TRUS and real-time ultrasound volumes did not indicate a statistically significant difference in volumes, whereas comparison of p-CT with real-time ultrasound volumes showed a statistically significant difference between the two volumes ( p = 0.0005). For this reason, a buffer of 30–40% is used to order seeds when using CT scan preimaging vs. 15–20% with ultrasound preimaging to ensure that there is no shortage of seeds on the day of procedure. To our knowledge, no other study has quantified this uncertainty in preimplant imaging. This translated into more than double the number of seeds wasted per procedure (18 seeds wasted using p-CT vs. 7 wasted when using p-TRUS). There are certainly advantages to using p-CT scans for volume acquisition. A CT scan is more comfortable for the patient with respect to preplan imaging in that the patient only needs to lay flat for a short period of time compared with the discomfort of a TRUS done in the dorsal lithotomy position. However, CT scans still require an additional visit to the hospital, create scheduling delays, and, based on our results, may not accurately provide the correct volume to ascertain the number of seeds needed. In addition to wasted seeds, dose from imaging is another consideration when choosing between p-TRUS and p-CT. Although dose from a pelvic CT scan is minimal, it may be a contributing argument against the use of CT when an equal or a superior nonionizing imaging technique is available. In addition, p-CT increases the cost of the procedure with respect to both the cost of the wasted seeds and the cost of the CT scan, which is currently three times that of an ultrasound. This difference does translate into additional cost to the patient and health care system. Private insurance companies vary widely in how patients are reimbursed; however, Medicare currently reimburses per seed (20) . Although many factors are taken into consideration in determining the cost of 125 I seeds, such as varying activity levels and cost between different vendors, the average cost for an 125 I seed is approximately US$35. We found that the number of seeds ordered using p-CT resulted in wastage of approximately 18 seeds or approximately US$630 compared with US$245 when using p-TRUS volumes. In addition, CT scans used for prostate brachytherapy treatment planning are reimbursed by Medicare at US$300 (HCPCS code 77290) vs. US$100 for a TRUS (HCPCS code 76873). Although the cost of a brachytherapy procedure is currently one of the least expensive options for the treatment of prostate cancer with an estimated cost of US$8300, the additional cost of a p-CT scan in addition to the number of seeds wasted when using this imaging modality compared with the cost when using p-TRUS is approximately US$600. This represents almost 7% of the total procedure cost (20) . Of note, patients do have an ultrasound done at time of biopsy, and it may be worthwhile to obtain more accurate volumes at the time of diagnosis. In our institution, p-TRUS was always a separate study from the biopsy. Volumes were not formally documented within the biopsy reports from Thomas Jefferson University Hospital as biopsy was done before cancer diagnosis. In addition, many patients underwent biopsy at an outside hospital and outside reports infrequently contained volume measurements. When documented, only base × height measurements were collected, making it difficult to compare with a true three-dimensional volume measurement. To our knowledge, no other article has compared volumes obtained during biopsy with those obtained during a planning study and this should be investigated further. Although this places more demand on the urologist at the time of biopsy, as they are now expected to obtain accurate gland volume and assess candidacy for brachytherapy before a diagnosis of prostate cancer, the saved time, expense, and patient discomfort that can be avoided may be invaluable. Alternatively, if images could be digitally captured at this time, they could later be contoured if a patient's biopsy is positive and the patient pursues brachytherapy treatment. Conclusions Although p-CT volumes have stronger correlation with volumes obtained during postimplant analysis, p-TRUS volumes more strongly correlate with volumes obtained during intra-operative planning. We found that the number of seeds ordered could be more accurately obtained from a p-TRUS volume and this translated into less seed wastage. Our findings indicate that p-TRUS is a more accurate and an economically superior alternative to p-CT imaging in the era of real-time ultrasound planning. References [1] J.C. Blasko P.D. Grimm J.E. Sylvester Palladium-103 brachytherapy for prostate carcinoma Int J Radiat Oncol Biol Phys 46 2000 839 850 [2] P.D. Grimm J.C. Blasko J.E. Sylvester 10-year biochemical (prostate-specific antigen) control of prostate cancer with (125)I brachytherapy Int J Radiat Oncol Biol Phys 51 2001 31 40 [3] B.V. Williams J. Horton A. Lawyer Loading technique comparison in permanent 125I prostate implants Med Dosim 24 1999 273 277 [4] Y. Yu L.L. Anderson Z. Li Permanent prostate seed implant brachytherapy: Report of the American Association of Physicists in Medicine Task Group No. 64 Med Phys 26 1999 2054 2076 [5] R.G. Stock N.N. Stone Y.C. Lo Intraoperative dosimetric representation of the real-time ultrasound-guided prostate implant Tech Urol 6 2000 95 98 [6] R. Nath L.L. Anderson G. Luxton Dosimetry of interstitial brachytherapy sources: Recommendations of the AAPM Radiation Therapy Committee Task Group No. 43. American Association of Physicists in Medicine Med Phys 22 1995 209 234 [7] E.M. Messing J.B. Zhang D.J. Rubens Intraoperative optimized inverse planning for prostate brachytherapy: Early experience Int J Radiat Oncol Biol Phys 44 1999 801 808 [8] A. Raben H. Chen A. Grebler Prostate seed implantation using 3D-computer assisted intraoperative planning vs. a standard look-up nomogram: Improved target conformality with reduction in urethral and rectal wall dose Int J Radiat Oncol Biol Phys 60 2004 1631 1638 [9] R.G. Stock N.N. Stone Y.C. Lo Postimplant dosimetry for (125)I prostate implants: Definitions and factors affecting outcome Int J Radiat Oncol Biol Phys 48 2000 899 906 [10] L. Chauveinc T. Flam S. Solignac Prostate cancer brachytherapy: Is real-time ultrasound-based dosimetry predictive of subsequent CT-based dose distribution calculation? A study of 450 patients by the Institut Curie/Hospital Cochin (Paris) Group Int J Radiat Oncol Biol Phys 59 2004 691 695 [11] B. Al-Qaisieh E. Brearley S. St Clair A study of a pretreatment method to predict the number of I-125 seeds required for prostate brachytherapy Int J Radiat Oncol Biol Phys 65 2006 304 307 [12] S. Hilaris D. Nori L.L. Anderson Atlas of brachytherapy 1988 McGraw-Hill Professional New York, NY [13] A. Wu C.C. Lee M. Johnson A new power law for determination of total (125)I seed activity for ultrasound-guided prostate implants: Clinical evaluations Int J Radiat Oncol Biol Phys 47 2000 1397 1403 [14] T.G. Shanahan P.J. Nanavati P.W. Mueller A comparison of permanent prostate brachytherapy techniques: Preplan vs. hybrid interactive planning with postimplant analysis Int J Radiat Oncol Biol Phys 53 2002 490 496 [15] G.S. Merrick W.M. Butler A.T. Dorsey Potential role of various dosimetric quality indicators in prostate brachytherapy Int J Radiat Oncol Biol Phys 44 1999 717 724 [16] S. Nag W. Bice K. DeWyngaert The American Brachytherapy Society recommendations for permanent prostate brachytherapy postimplant dosimetric analysis Int J Radiat Oncol Biol Phys 46 2000 221 230 [17] W.D. D'Souza H.K. Lee M.B. Palmer Is intraoperative nomogram-based overplanning of prostate implants necessary? Int J Radiat Oncol Biol Phys 56 2003 462 467 [18] V. Narayana P.L. Roberson R.J. Winfield Impact of ultrasound and computed tomography prostate volume registration on evaluation of permanent prostate implants Int J Radiat Oncol Biol Phys 39 1997 341 346 [19] S.C. Hoffelt L.M. Marshall M. Garzotto A comparison of CT scan to transrectal ultrasound-measured prostate volume in untreated prostate cancer Int J Radiat Oncol Biol Phys 57 2003 29 32 [20] K.R. Badiozamani K. Wallner W. Cavanagh Comparability of CT-based and TRUS-based prostate volumes Int J Radiat Oncol Biol Phys 43 1999 375 378 [21] J. Xue F. Waterman J. Handler The effect of interobserver variability on transrectal ultrasonography-based postimplant dosimetry Brachytherapy 5 2006 174 182 [22] S. Smith K. Wallner G. Merrick Interpretation of pre- versus postimplant TRUS images Med Phys 30 2003 920 924 [23] W.R. Lee M. Roach III J. Michalski Interobserver variability leads to significant differences in quantifiers of prostate implant adequacy Int J Radiat Oncol Biol Phys 54 2002 457 461 [24] B.H. Han K. Wallner G. Merrick The effect of interobserver differences in post-implant prostate CT image interpretation on dosimetric parameters Med Phys 30 2003 1096 1102