Vickie Kong

A randomized comparison of interfraction and intrafraction prostate motion with and without abdominal compression

BACKGROUND AND PURPOSE To quantify inter- and intrafraction prostate motion in a standard VacLok (VL) immobilization device or in the BodyFix (BF) system incorporating a compression element which may reduce abdominal movement. MATERIALS AND METHODS Thirty-two patients were randomly assigned to VL or BF. Interfraction prostate motion >3 mm was corrected pre-treatment. EPIs were taken daily at the start and end of the first and last treatment beams. Interfraction and intrafraction prostate motion were measured for centre of mass (COM) and individual markers. RESULTS There were no significant differences in interfraction (p0.002) or intrafraction (p0.16) prostate motion with or without abdominal compression. Median intrafraction motion was slightly smaller than interfraction motion in the AP (7.0 mm vs. 7.6 mm) and SI direction (3.2 mm vs. 4.7 mm). The final image captured the maximal intrafraction displacement in only 40% of fractions. Our PTV incorporated >95% of total prostate motion. CONCLUSIONS Intrafraction motion became the major source of error during radiotherapy after online correction of interfraction prostate motion. The addition of 120 mbar abdominal compression to custom pelvic immobilization influenced neither interfraction nor intrafraction prostate motion.

Prostate delineation using CT and MRI for radiotherapy patients with bilateral hip prostheses.

BACKGROUND AND PURPOSE Prostate cancer patients with bilateral hip prostheses (b-HP) are rare, but pose unique challenges for high-dose radiotherapy. The purpose of this study was to determine the effect of b-HP on prostate delineation using CT and MR. METHODS AND MATERIALS The planning CT and MR of 7 patients with b-HP were co-registered. Four observers delineated the prostate on both data sets, blinded to the volumes delineated by their peers and those on the alternate image set. Inter-modality differences and inter-observer variability were quantified by calculating 3D volume overlaps. The effect of prostheses physical extent on these variables was evaluated. MR geometric distortion was explored using the differences in vector distances between three intra-prostatic markers (IPMs) on CT and MR. RESULTS Prostate volumes on CT were consistently larger than MR volumes in all patients when averaged over the 4 observers (+10% to +46%, p<0.001). The mean inter-modality volume overlap was 1.59, which varied significantly between patients (1.35 to 1.82, p=0.04). There was a trend toward smaller inter-observer variability in the MR volumes (1.95 vs. 1.71, p=0.08). No differences could be identified between the IPM vectors on CT and MR. CONCLUSIONS Although limited by a small sample size, this study found that MR prostate volumes were significantly smaller than those from CT for b-HP patients. The largest differences between CT and MR prostate volumes and the greatest inter-observer variability on CT tended to be seen in patients with the largest prostheses. b-HP do not appear to cause measurable MR geometric distortions in the sequences used.

Patient-specific PTV margins in radiotherapy for bladder cancer – A feasibility study using cone beam CT ☆

PURPOSE To assess the feasibility of using cone beam computed tomography (CBCT) to generate patient-specific PTV margins for bladder cancer patients treated with radiation therapy (RT). METHODS Eleven patients underwent CT simulation and daily RT (full bladder and empty rectum). CBCT was done prior to each fraction, and the whole bladder was contoured off-line. For the first 15 CBCTs of each patient, the bladder was aligned with CT-simulation bladder (pBladder) to create an occupancy volume (OV). A 5mm isotropic margin was added to OV (OV+5). A measurement-based PTV (mPTV) was generated by measuring maximal displacement between pBladder and OV in six directions. OV, OV+5, mPTV, and a standard PTV (2cm isotropic margin) were compared for absolute and relative volume differences. Using the final 10 CBCT of each patient, the ability of each study volume to encompass the entire CBCT bladder was determined. RESULTS 161/165 CBCT images were of adequate quality for contouring. No daily trend in bladder volume variation was noted. The median absolute volumes (cm(3)) were: 221, 271, 426, 440, and 914 for pBladder, OV, OV+5, mPTV, and standard PTV, respectively. The median ratios of the study volumes/pBladder were: 1.4 OV, 2.1 OV+5, 2.4 mPTV, 4.1 standard PTV. OV+5 was smaller than mPTV in 9 patients. There was considerable inter-patient variability in study volumes and no apparent association of the magnitude of margin expansion and pBladder. The bladder was encompassed in 69%, 99%, 99%, and 100% of the final 10 fractions by OV, OV+5, mPTV, and standard PTV, respectively. CONCLUSIONS The use of daily CBCT to generate patient-specific PTV margins is feasible and results in a marked reduction in the irradiated volume compared to population-based margins. As daily bladder volume varied unpredictably with considerable differences between patients, these findings support the use of patient-specific PTV margin expansions for bladder radiotherapy.

Phase 2 trial of guideline-based postoperative image guided intensity modulated radiation therapy for prostate cancer: Toxicity, biochemical, and patient-reported health-related quality-of-life outcomes

PURPOSE The purpose of this study was to characterize treatment-related toxicities, health-related quality of life (HRQOL), and biochemical outcomes in patients treated with postoperative image guided intensity modulated radiation therapy (IMRT) for prostate cancer using a consensus guideline for defining the clinical target volume. METHODS AND MATERIALS Between August 2007 and October 2008, patients considered for radiation therapy (RT) after prostatectomy were enrolled. The clinical target volume (prostate bed) was delineated according to published consensus guidelines, and patients were prescribed a dose of 66 Gy in 33 fractions. Radiation treatment planning prioritized rectal dose constraints over target volume coverage. Treatment was delivered by use of IMRT and daily cone beam computed tomographic guidance. Toxicity (graded according to the National Cancer Institutes Common Terminology Criteria for Adverse Events) and HRQOL assessments according to the Expanded Prostate Cancer Index Composite (EPIC) questionnaire were collected prospectively at baseline, at week 5 (during RT), at 3 months, and at yearly follow-up visits. Cumulative toxicity and biochemical relapse-free rates were calculated by the Kaplan-Meier method. Paired Student t tests with multiple testing correction were used to assess changes in HRQOL. RESULTS A total of 68 men were evaluated, with median follow-up of 5.9 years. Fifty-three patients (77.9%) and 15 patients (22.1%) were treated with salvage and adjuvant RT, respectively. Primary planning objectives were met in most cases (97.1%), but planning target volume coverage was compromised in 40% of cases because of large planning target volumes (mean 347.6 cm(3)). There were no grade 3 or 4 acute toxicities. Cumulative 5-year incidence of late gastrointestinal and genitourinary grade 2 toxicities was 12.3% (95% confidence interval [CI], 11.1%-13.5%) and 10.6% (95% CI, 9.5%-11.6%), respectively. No grade 3 or 4 late toxicities were observed. Transient declines in EPIC gastrointestinal domain summary score (mean 87.66 at 3 months vs 92.76 at baseline; P = .006) and genitourinary irritative subscale (week 5 mean score 83.37 vs 89.45 at baseline; P = .007) were observed. Complete recovery occurred between 3 and 12 months after therapy, remaining stable compared with baseline at 5-year follow-up. Sexual HRQOL remained stable at 5 years, with an improving trend in bother subscale. Biochemical relapse-free rate at 5 years was 72.7% (95% CI, 61.9%-83.5%). CONCLUSIONS Guideline-based postprostatectomy image guided IMRT with rigid rectal dose constraints resulted in favorable toxicity profiles; long-term stability in gastrointestinal, genitourinary, and sexual HRQOL; and expected biochemical control rates. Concerns regarding toxicity and HRQOL should not preclude recommendation for RT after prostatectomy.

Deriving patient-specific planning target volume for partial bladder image guided radiation therapy

PURPOSE Image guided radiation therapy (IGRT) using bony anatomy for bladder cancer requires the use of large population-based planning target volume (PTV) margins to compensate for geometric uncertainties. This may result in a large volume of normal tissue being irradiated unnecessarily. Identification of the clinical target volume (CTV) is also a challenge during target delineation and treatment position verification. This study describes the use of lipiodol (Guerbet, US) and cone beam computed tomography (CBCT) in deriving patient-specific PTV (PS-PTV) for partial bladder IGRT. METHODS AND MATERIALS Twelve patients underwent lipiodol injection into the bladder wall prior to radiation treatment. A PS-PTV was generated by the following: (1) Delineating partial bladder CTV (CTVpb) on 15 CBCTs; (2) registering the CBCTs with the planning CT image using lipiodol; (3) combining the 15 CTVpb to create an occupancy volume (OV); and (4) expanding the OV by 3 mm. Its efficacy in reducing irradiated volume and in providing coverage was assessed by comparing it with a 20-mm population-based PTV (popPTV) and using phase 2 CBCTs. RESULTS The median PS-PTV and popPTV (cm(3)) were 102 (range, 37-336) and 325 (range, 211-631), respectively. Median distance between the CTVpb and the PS-PTV edge (mm) were 6 superior, 6 right, 7 left, 7 anterior, 8 posterior, and 11 inferior. The absolute median reduction in the overlapping volume of rectum, small bowel, and large bowel were 0.3 cm(3), 5.3 cm(3), and 13.0 cm(3), respectively. Despite large reductions in volume and margin compared with popPTV, PS-PTV achieved 100% target coverage. CONCLUSIONS Using lipiodol and CBCT to derive PS-PTV facilitated large reductions in the irradiated normal tissue volume without compromising target coverage.

Comparison of 3 image-guided adaptive strategies for bladder locoregional radiotherapy

The objective of this study was to compare the dosimetric differences of a population-based planning target volume (PTV) approach and 3 proposed adaptive strategies: plan of the day (POD), patient-specific PTV (PS-PTV), and daily reoptimization (ReOpt). Bladder patients (n = 10) were planned and treated to 46 Gy in 23 fractions with a full bladder in supine position by the standard strategy using a population-based PTV. For each patient, the adaptive strategy was executed retrospectively as follows: (1) POD-multiple distributions of various PTV sizes were generated, and the appropriate distribution based on the bladder of the day was selected for each fraction; (2) PS-PTV-population-based PTV was used for the first 5 fractions and a new PTV derived using information from these fractions was used to deliver the remaining 18 fractions; and (3) ReOpt-distribution was reoptimized for each fraction based on the bladder of the day. Daily dose was computed on all cone beam computed tomographies (CBCTs) and deformed back to the planning computed tomography (CT) for dose summation afterward. V95_Accu, the volume receiving an accumulated delivered dose of 43.7 Gy (95% prescription dose), was measured for comparison. Mean V95_Accu (cm3) values were 1410 (standard deviation [SD]: 227), 1212 (SD: 186), 1236 (SD: 199), and 1101 (SD: 180) for standard, POD, PS-PTV, and ReOpt, respectively. All adaptive strategies significantly reduced the irradiated volume, with ReOpt demonstrating the greatest reduction compared with the standard (- 25%), followed by PS-PTV (- 16%) and POD (- 12%). The difference in the magnitude of reduction between ReOpt and the other 2 strategies reached statistical significance (p = 0.0006). ReOpt is the best adaptive strategy at reducing the irradiated volume because of its frequent adaptation based on the daily geometry of the bladder. The need to adapt only once renders PS-PTV to be the best alternative adaptive strategy.

Evaluation of high dose volumetric CT to reduce inter-observer delineation variability and PTV margins for prostate cancer radiotherapy

PURPOSE The aim was to determine whether the enhanced soft tissue contrast provided by high-dose volumetric CT (HDVCT) can reduce inter-observer variability in delineating prostate compared to helical conventional CT (CCT) scans and 3T MRI scans for patients undergoing radical prostate cancer radiotherapy. Secondly, to quantify the potential PTV reduction with decreased inter-observer variability. MATERIALS AND METHODS A 320 slice volumetric CT scanner was used. The wide-detector coverage of 16cm enabled volumetric image acquisition of prostate gland in one rotation. Three imaging studies were performed on ten patients. CCT and HDVCT were performed consecutively at the same coordinate system followed by MRI. Five radiation oncologists delineated the prostate. RESULTS The inter-observer variability is 2.0±0.6, 1.9±0.4 and 1.8±0.4mm for CCT, HDVCT and MR respectively with the maximum at the apex region. Comparing inter-observer difference variability between CCT and HDVCT with MR indicates that observers have larger variations in contouring using CCT than HDVCT especially at apex. Jaccard index of HDVCT is significantly higher than CCT with a mean difference of 0.03 (p=0.011). Both MRI and HDVCT provide the opportunity for a 2mm PTV margin reduction at the apex compared to CCT. CONCLUSION Inter-observer variability in delineation remains an important source of systematic error. HDCTV for treatment planning reduces this error without recourse to MRI and permits a PTV reduction of 2mm at the apex. The margins required to account for residual error with any imaging modality are still greater than are used in typical current practice.

Planned versus 'delivered' bladder dose reconstructed using solid and hollow organ models during prostate cancer IMRT.

BACKGROUND AND PURPOSE All studies to date have evaluated the dosimetric effect of bladder deformation using an organ model that includes the dose to the urine. This research reconstructed bladder dose using both hollow and solid organ models, to determine if dose/volume differences exist. MATERIALS AND METHODS 35 prostate IMRT patients were selected, who had received 78Gy in 39 fractions and full bladder instructions. Biomechanical modelling and finite element analysis were used to reconstruct bladder dose (solid and hollow organ model) using every third CBCT throughout the treatment course. RESULTS Reconstructed dose (ReconDose) was 11.3Gy greater than planned dose (planDose) with a hollow bladder model (p<0.001) and 12.3Gy greater with a solid bladder model (p<0.0001). Median reconstructed volumes within the 30Gy, 65Gy and 78Gy isodoses were 3-4 times larger with the solid organ model (p<0.0001). The difference between planning bladder volume and median treatment volume was associated with the difference between the planDose and reconDose below 78Gy (R(2)>0.61). CONCLUSIONS Substantial differences exist between planned and reconstructed bladder dose, associated with the differences in bladder filling between planning and treatment. Dose reconstructed using a solid bladder model over-reports the volume of bladder within key isodose levels and overestimates the differences between planned and reconstructed dose. Dose reconstruction with a hollow organ model is recommended if the goal is to associate that dose with toxicity.

Evaluation of resource burden for bladder adaptive strategies: A timing study

Interfraction bladder motion is substantial and therefore many different adaptive radiotherapy approaches have been developed to accommodate that motion. Previous studies comparing the efficacy of those adaptive strategies have demonstrated that reoptimization (ReOpt) was dosimetrically superior when compared to Plan of the Day (POD) and Patient‐specific PTV (PS‐PTV). However, the feasibility of clinical implementation is dependent upon assessment of the resource burden. This study assessed and compared the resource burden of three adaptive strategies.

Impact of image registration surrogates on the planning target volume geometry for bladder radiation therapy.

PURPOSE The use of Lipiodol and cone beam computed tomography (CT) has facilitated the generation of patient-specific planning target volumes (PTV) to reduce irradiation of normal tissue. However, injecting Lipiodol to demarcate the target volume is an invasive procedure. Center of bladder (COB) and bladder wall surface (BWS) encompassed by the clinical target volume (CTV) are proposed to be the alternative noninvasive surrogates. This study examines the impact of using these 3 surrogates for image registration on the resultant geometry of patient-specific PTVs. METHODS AND MATERIALS Twenty bladder cancer patients who had Lipiodol injection before planning CT were included. Lipiodol, CTV, and bladder were delineated on the planning CT. In addition, CTVs were delineated on 5 cone beam CT scans from each patient. Cone beam CT scans were registered to planning CT using Lipiodol, COB, and BWS to generate Lipiodol-PTV, COB-PTV, and BWS-PTV. Using Lipiodol as the reference, the difference in the 2-dimensional/3-dimensional displacement values and the geometry of the resultant PTVs were quantified. RESULTS A total of 1200 displacement values and 60 volumes were included for analysis. The overall median and standard deviation (SD) of the 3-dimensional displacement (mm) measured by Lipiodol, BWS, and COB are 25 (SD, 15), 24 (SD, 14), and 21 (SD, 15), respectively. Lipiodol-PTV has the smallest mean volume, followed by BWS-PTV and COB-PTV. BWS-PTV was more geometrically similar to Lipiodol-PTV when compared with COB-PTV. Six of 20 COB-PTVs who had CTV located at either the dome or the base of the bladder were larger than the corresponding Lipiodol-PTV by more than 20%. CONCLUSION Lipiodol is the optimal image registration surrogate for partial bladder radiation therapy. However, for patients who are contraindicated for Lipiodol injection, BWS is the preferred noninvasive surrogate because the derived PTV has a smaller geometric variation from the Lipiodol-PTV when compared with COB, especially when the CTV was located at the dome or the base of the bladder.