Anatomic features predict post-radical prostatectomy continence

August 1, 2011

Certain anatomic features identified on preoperative magnetic resonance imaging independently predict continence recovery after radical prostatectomy but seem to add minimal prognostic value above use of standard preoperative variables.

Washington-Certain anatomic features iden-tified on preoperative magnetic resonance imaging independently predict continence recovery after radical prostatectomy but seem to add minimal prognostic value above use of standard preoperative variables, findings of a pilot study indicate.

Investigators from Memorial Sloan-Kettering Cancer, New York reviewed data from 967 men who underwent prostatectomy with preoperative MRI staging from 2001 to 2004. After excluding men who received neoadjuvant therapy, had preoperative incontinence, or had missing continence follow-up data, 600 men were included in the analysis.

Bony and soft tissue dimensions on the MRI images were determined by two raters blinded to the patients' clinical and pathologic data. The measurements performed were chosen to reflect pelvic size at the height of prostate dissection, prostate size and apical depth in relation to the pelvis, urethral length and volume, and distance between the urethra and levator muscle at the prostate apex. With continence defined as use of no pads or protective material, statistical associations were analyzed between the various MRI variables and recovery of continence at 6 and 12 months. Longer membranous urethral length, larger urethral volume, and closer anatomic relation between the urethra and levator muscle at the apex of the prostate were identified on multivariate analysis as independent predictors of continence recovery.

Next, these features were added into a base model for predicting continence recovery that included age, comorbidities, clinical stage, clinical grade, PSA, and year of surgery. The predictive accuracy of the base model was 0.637 at 6 months, and it was significantly improved with the addition of each MRI variable. However, a full model that included all of the standard and MRI variables had a predictive accuracy of only 0.675 at 12 months.

"Postoperative incontinence continues to be a major concern for patients undergoing RP," said first author Christian von Bodman, MD.

"Surgeons subjectively report that RP is more challenging in men with a narrow pelvis or if the prostate is deep in the pelvis, and data from surgical studies indicate an association between these anatomic features and more adverse functional outcomes. Since MRI is the best imaging technique available for assessing soft tissue or bony structures of the pelvis, we sought to explore whether any features on preoperative MRI might be associated with continence recovery and should be identified to optimize anatomic understanding and surgical technique," added Dr. von Bodman, who was a research fellow at Memorial Sloan-Kettering at the time of the study, working with Jaspreet Sandhu, MD, and colleagues.

"We were surprised to find that prostate size, deep location, and narrow pelvis did not affect continence recovery, and the low additive value of the predictive variables suggests the MRI information should not be implemented into clinical counseling. However, this is exploratory data from a pilot study for which the main limitation is that not all patients who underwent RP during the study period received preoperative staging MRI," said Dr. von Bodman, who is currently at the Ruhr-University of Bochum, Germany.

All men were operated on by one of five experienced surgeons, and the baseline characteristics of the study group were typical for a population of men undergoing RP without neoadjuvant therapy. Median age was 66 years; median body mass index, 27 kg/m2 ; and median PSA, 5.4 ng/mL. The majority had clinical stage T1 (60%) or T2a (19%) disease and biopsy Gleason scores of 6-7.