Evaluation of 3T Pelvic MRI Imaging in Prostate Cancer Patients Receiving Postprostatectomy IMRT

S378

International Journal of Radiation Oncology  Biology  Physics

2454

optimal timing and toxicity for post-prostatectomy radiation therapy (PPRT) with IMRT or 3DCRT. Materials/Methods: From 1991-2010, 498 men received PPRT: 101 were treated with ADJ RT for high risk features (pT3, positive margins, or Gleason score [GS] 8-10) and 397 were treated with SALV RT. 3DCRT was used in 134 men and IMRT was used in 364. Men with a rising PSA were stratified by pre-RT PSA levels to identify those who might benefit from early SALV. Median RT doses for ADJ and SALV were 64 Gy and 68 Gy. 119 men received androgen deprivation therapy (ADT) (24 ADJ/ 95 SALV) with median duration of 17.7 mos. PSA recurrence was defined as PSA nadir + 2ng/mL. Kaplan-Meier methods estimated survival functions and were compared using the log-rank test. Acute toxicity was compared using Fisher’s exact test. Cox proportional hazard methods were used for multivariable analyses (MVA). Recursive partitioning analysis (RPA) identified a significant PSA cut-point for initiation of RT. Results: Median follow-up was 84.1 mos. Median age at RT was 59 yrs. RPA identified PSA of 0.7 ng/mL as a significant cut-point for initiation of RT. Outcomes for biochemical control (BC), distant metastasis-free survival (DMFS), cause-specific survival (CSS) and overall survival (OS) are reported in the table. Outcomes did not differ between ADJ vs SALV initiated when PSA <0.4 ng/mL. On MVA, predictors of improved BC were ADJ RT, GS<8 and positive surgical margins. Predictors for increased risk of DM were SALV RT and GS8. GS8 was the only independent predictor of worse CSS. ADJ RT was the only significant predictor of improved OS on MVA (p Z 0.042). IMRT vs 3DCRT did not predict for grade 2+ acute GI toxicity (7.4% vs 11.2%, p Z 0.203) but trended toward improved 4-yr toxicity (4.5% vs 8%, p Z 0.066). Conclusions: ADJ PPRT improves rates of BC, DM and OS when compared to SALV. Differences in outcomes are greatest when PSA approaches 0.7 ng/mL; however, outcomes in men salvaged with PSA<0.4 ng/mL appear similar to ADJ. We therefore recommend ADJ or early IMRT for men with high risk features after prostatectomy. IMRT shows a favorable trend in late GI toxicity compared to men treated with 3DCRT. Author Disclosure: C. Murphy: None. R.G. Uzzo: None. K.J. Ruth: None. R. Viterbo: None. E. Plimack: None. M.K. Buyyounouski: None. A. Kutikov: None. D.Y.T. Chen: None. R.E. Greenberg: None. E.M. Horwitz: None.

Concordance of Gleason Score on Biopsy and Prostatectomy in a Contemporary Cohort of U.S. Veterans Undergoing Radical Prostatectomy M. Agarwal,1 M. Rotman,1 D. Schwartz,2,1 and D. Schreiber2,1; 1SUNY Downstate Medical Center, Brooklyn, NY, 2VA NY Harbor Healthcare System, Brooklyn, NY Purpose/Objective(s): The Gleason Score of Prostate Cancer closely correlates with clinical behavior, with a higher score predictive of a greater likelihood of non-organ confined disease as well as poor outcome. It is a key component of predictive nomograms that often aids in guiding clinical decision making. Biopsy undergrading or overgrading has been reported to be present in up to 35% of specimens from many single institution studies. Recent literature has noted reduced variability and this publication reports correspondence trends between Gleason scores in biopsy and pathological specimens in a contemporary group of men at a Veterans Administration Hospital. Materials/Methods: Retrospective review of 250 consecutive patients who underwent radical prostatectomy between 2003 and 2010 was conducted. Comparison of Gleason score of the needle biopsy with the pathologic Gleason score from surgery was conducted for the whole cohort as well as grouped into Gleason scores of 7, 8-9. The concordance of the respective Gleason scores was analyzed using the coefficient of agreement (Kappa). Univariate and Multivariate logistic regression analysis was performed using SPSS version 20 to determine potential factors (age, year of diagnosis, race, PSA, clinical T-classification, % biopsy cores positive) that may impact concordance of Gleason score. All probability tests were twosided with a significant level set at 0.05. Results: A total of 250 patients were included in the analysis. The biopsy Gleason score and the pathologic Gleason score was identical for 155 patients (62%). Of the remaining patients, 69 (27.6%) were undergraded by the biopsy and 24 (9.6%) were overgraded by the biopsy. The highest proportion of discordance was for patients with a clinical Gleason score of 6. The lambda for Gleason score concordance was 0.37, indicating a poor agreement overall. The concordance rate for biopsy Gleason score 6 was 55.4%, biopsy Gleason 7 was 74.3%, biopsy Gleason 8 was 27.2%, and biopsy Gleason 9 was 55.6%. Univariate and Multivariate logistic regression analysis did not reveal any clinical factors that would further predict for concordance between the clinical and pathologic Gleason scores. Conclusions: Data collected at this single institution appears to correlate well with previously published reports. However, patients with a biopsy Gleason score of 6 are most at risk for pathologic upgrading. Author Disclosure: M. Agarwal: None. M. Rotman: None. D. Schwartz: None. D. Schreiber: None.

2455 Improved Outcomes in Men Treated With Adjuvant or Early Salvage Postprostatectomy IMRT or 3DCRT at a Single Institution C. Murphy, R.G. Uzzo, K.J. Ruth, R. Viterbo, E. Plimack, M.K. Buyyounouski, A. Kutikov, D.Y.T. Chen, R.E. Greenberg, and E.M. Horwitz; Fox Chase Cancer Center, Philadelphia, PA Purpose/Objective(s): Data suggest that men treated adjuvantly (ADJ) have better biochemical and clinical outcomes compared to men treated for a PSA or clinical recurrence (salvage-SALV). We examined outcomes,

Poster Viewing Abstract 2455; Table with PSA cut point of 0.7 ng/mL 7-yr outcome BC DMFS CSS OS

7-year outcomes for ADJ and SALV

ADJ (%)

SALV PSA <0.7(%)

SALV PSA  0.7(%)

Log-rank P value

87.7 95.6 95.8 94.9

66.1 93.5 95.5 94.6

36.8 82.7 96.1 88.2

<.0001 .0416 .1405 .029

2456 Evaluation of 3T Pelvic MRI Imaging in Prostate Cancer Patients Receiving Postprostatectomy IMRT H.A. Gay, V. Verma, L. Chen, J.M. Michalski, Y. Hu, W. Zhang, L. Eschen, S. Fergus, D. Mullen, and S. Strope; Washington University School of Medicine, St. Louis, MO Purpose/Objective(s): The goal of the present study is to evaluate the utility of a 3T pelvic magnetic resonance imaging (MRI) in detecting a local recurrence or nodal involvement in post-prostatectomy prostate cancer patients prior to receiving adjuvant or salvage intensity-modulated radiation therapy (IMRT), and the clinical factors associated with positive MRI findings. Materials/Methods: From January 2006 to May 2012, 90 prostate cancer patients status post prostatectomy with rising PSA had a 3T MRI prior to adjuvant or salvage IMRT. Patients were grouped into those with positive (MRI+) and negative (MRI-) findings on MRI. The following variables were analyzed for their usefulness in predicting positive findings on MRI: initial presenting and initial post-operative PSA, PSA at the time of imaging, PSA velocity, surgical margins, Gleason score, pathological stage, pre-RT digital rectal examination (DRE), and type of surgical prostatectomy. Results: The only significant variable predictive of a positive MRI on multivariate analysis was positive margins. Specifically, 15/46 (33%) of patients with positive margins had a positive MRI, while 5/44 (11%) of patients with negative margins had a positive MRI. Furthermore, in the positive margin group, the location of the positive findings on the 3T MRI

Volume 87  Number 2S  Supplement 2013 corresponded with the description of the positive findings on the pathology report in 9 of 12 (75%) cases. Conclusions: Post-prostatectomy patients with pathologic positive margins are 3 times more likely to have positive findings on a 3T MRI. The optimal treatment of MRI positive patients and the cost-effectiveness of limiting 3T MRIs to patients with positive margins should be further studied. Author Disclosure: H.A. Gay: None. V. Verma: None. L. Chen: None. J.M. Michalski: None. Y. Hu: None. W. Zhang: None. L. Eschen: None. S. Fergus: None. D. Mullen: None. S. Strope: None.

2457 Dynamic Dosimetry Using Ultrasound-Fluoroscopy Registration Identifies Cold Spots Intraoperatively During LDR Prostate Brachytherapy D. Song,1 Y. Le,1 J. Lee,1 N. Kuo,1 A. Robinson,1 A. Deguet,1 E. Burdette,2 G. Fichtinger,3 and J. Prince1; 1Johns Hopkins University, Baltimore, MD, 2Acoustic Medsystems, Inc, Champaign, MD, 3Queens University, Kingston, ON, Canada Purpose/Objective(s): Outcomes after permanent prostate brachytherapy vary depending on whether adequate target coverage is achieved. We have developed a system of intraoperative dosimetry which computes source positions from fluoroscopy images and spatially registers them to ultrasound (RUF). We evaluated whether this system accurately identifies areas of inadequate dose by comparing RUF-computed dosimetry with standard post-implant CTMR dosimetry. Materials/Methods: Fifty patients were enrolled in an IRB-approved study; 38 patients with CT-MR dosimetry available were analyzed. Standard US-based real time needle/source tracking and dosimetry update were utilized intraoperatively, and all patients had US-based dosimetry predicting prostate V100 > 99% at implant completion. Prior to anesthesia withdrawal, 3-6 non-isocentric fluoroscope images and 1 mm thickness TRUS images were captured for prostate contours and end-of-procedure RUF dose calculation. RUF dose was for comparison purposes only and was not utilized to guide source placement. Details of the RUF system and algorithms have been described previously. Day 1 post-implant dosimetry based on fused CT-MRI was utilized as ground truth. Results: Mean CT-MR V100 was 97% (2% SD). Ten of 38 patients with CT-MR V100 of <96% were selected for further analysis. Prostate volume was separated into 3 regions (base, mid-gland and apex) with each region divided into four sectors (L, R, A, P; 12 sectors total). Sector analysis was performed on both RUF and CT-MRI dosimetry to detect sectors with V100 < 80%. Mean RUF-predicted prostate V100 for these patients was 91% (3%) and CT-MR V100 was 94% (2%). Mean TRUS prostate volume was 49 cc (17 cc) and CT-MR prostate volume of 46 cc (10 cc) with mean difference of 3 cc (9 cc). Of 10 cold spots identified on CTMR (average sector V100 Z 60%  24%), RUF dosimetry detected 8 with average sector V100 67%  21%. Of the remaining 110 sectors, RUF dosimetry indicated additional 15 cold spots with V100 Z 72%  9%, while average CT-MR V100 from same 15 sectors was 91%  5%. The majority of false positive cold spots were in either base or anterior part of prostate. The overall sensitivity of RUF detection is 80% and specificity is 86%. Conclusions: Detection of cold spots intraoperatively is achievable using the RUF system. Overestimation of base and anterior of prostate volume in TRUS vs CT-MR contours may lead to overestimation of cold spots by RUF; improvement in post-implant TRUS prostate volume rendering is expected to further improve results. Author Disclosure: D. Song: None. Y. Le: None. J. Lee: O. Patent/License Fee/Copyright; Patent. N. Kuo: None. A. Robinson: None. A. Deguet: None. E. Burdette: P. Ownership Other; ownership. G. Fichtinger: O. Patent/License Fee/Copyright; patent. J. Prince: O. Patent/License Fee/ Copyright; patent.

Poster Viewing Abstracts S379

2458 Gastrointestinal and Genitourinary Toxicity and Quality of Life After Postprostatectomy Radiation Therapy: Are Normal Tissue DVH Parameters for Intact Prostate Cancer Applicable? A. Parekh,1 M.C. Ranck,2 S.K. Chennupati,3 and S.L. Liauw2; 1University of Chicago Pritzker School of Medicine, Chicago, IL, 2University of Chicago Hospitals, Chicago, IL, 3Oregon Health and Science University, Portland, OR Purpose/Objective(s): Dose volume histogram (DVH)-toxicity relationships have been defined for intact prostate cancer but are relatively unexplored for post-prostatectomy radiation therapy (PPRT). Our goal was to identify DVH parameters related to late gastrointestinal (GI) or genitourinary (GU) toxicity and quality of life (QOL) for men treated with post-operative RT. We also evaluated whether intact DVH guidelines would have importance in the post-operative setting. Materials/Methods: Ninety-six men treated with PPRT between 2001 and 2010 were identified in a prospectively maintained database, with DVH and toxicity follow-up available. Median age was 61. All men were treated with Intensity-Modulated RT with a median dose of 66 Gy to the prostate bed. Thirty-eight percent received concurrent hormonal therapy (median 4 mos). Thirty-five percent of men were treated to an initial pelvic field. Late toxicity (>3 months after RT) was defined by CTC criteria v3.0. Patient reported QOL was assessed in 41 men using the EPIC-26 survey at time 0, 2, 6, 12, 18 and 24 months. Global domain scores were generated on a scale of 0-100 (full health). DVH parameters including % of rectum and bladder receiving 70, 65, and 40 Gy were tested for associations with maximal GI and GU toxicity, and QOL by global domain scores and distress at all time points. Toxicity was evaluated according to whether men met “intact DVH guidelines” of V70, 65, and 40 Gy <20%, 40%, and 80% for the rectum, and <30%, 60%, and 80% for the bladder. Median follow-up was 38 months. Results: The rates of late Grade 2+ GI and GU toxicity at 2 years were 7 and 13%, respectively. The median 2-y global urinary incontinence, urinary frequency, and bowel scores were 75, 93, and 94, respectively. 5%, 5% and 3% of patients reported moderate or severe distress at 2 years in urinary continence, urinary frequency, and GI function, respectively. The median V70, 65, and 40 Gy to the bladder was 3%, 27%, and 59% for the bladder, and 3%, 25%, and 60% for the rectum. There were no DVH parameters identified which were associated with late grade 2+ toxicity or QOL for any subset domain. Patients meeting intact prostate normal tissue planning goals (69%) did not have any statistically different outcomes compared to the overall group (median 2-y global scores for incontinence, irritability/obstructive symptoms, and GI symptoms were 73, 94, and 100, respectively). Conclusions: No DVH relationships between toxicity or QOL were identified, possibly due to low rates of toxicity in this cohort, variation in bladder and rectal filling, the use of bladder and rectum DVH sparing goals over this time period, or relatively lower doses for post-prostatectomy RT. It is reasonable to apply intact prostate cancer DVH guidelines in the absence of any established post-operative relationships. Author Disclosure: A. Parekh: None. M.C. Ranck: None. S.K. Chennupati: None. S.L. Liauw: None.

2459 Combining Prostate-Specific Antigen Nadir and Time to Nadir Allows for Early Identification of Patients at Highest Risk for Development of Metastasis and Death Following Salvage Radiation Therapy W.C. Jackson,1 S.B. Johnson,1 Y. Song,1 B. Foster,1 H.M. Sandler,2 G.S. Palapattu,1 F.Y. Feng,1 and D.A. Hamstra1; 1University of Michigan, Ann Arbor, MI, 2Cedars-Sinai, Los Angeles, CA Purpose/Objective(s): Following definitive radiation therapy (RT) for prostate cancer, both PSA nadir (nPSA) and the time to nPSA (TnPSA) have been shown to be prognostic for clinically meaningful patient outcomes. Little is known regarding the prognostic capability of PSA nadir