postoperative complications that were reported in agreement with the standardized Martin criteria. METHODS: We identified 236 PCa patients with a Gleason score of 3 þ 3 or 3 þ 4, clinical stage T2b or less, prostate specific antigen (PSA) 15 ng/dl or less, unilateral disease and life expectancy greater than 10 years treated with partial gland ablation between 2009 and 2016 at a single tertiary care referral center. For partial gland ablation, 188 (79.7%) and 48 (20.3%) patients underwent high intensity focused ultrasound (HIFU) and cryotherapy, respectively. 30-day postoperative complications and length of stay (LoS) were analyzed. Postoperative complications were prospectively recorded according to Martin criteria and categorized according to Clavien-Dindo system. RESULTS: Overall, 188 (79.7%) and 48 (20.3%) had Gleason 3þ3 and 3þ4, respectively. Overall, 186 (78.8%), 43 (18.2%) and 7 (3%) had clinical stage T1c, T2a and T2b, respectively. The median PSA and prostate volume were 6.8 ng/ml and 38 ml, respectively. Nine out of 10 Martin criteria were full satisfied except for risk factors assessment. The overall rate of postoperative complications was 17.4% (n¼41). Of these, 15.7% (n¼37), 0.4% (n¼1) and 1.3% (n¼3) were graded II, IIIb and IVa according to Clavien Dindo, respectively. Patients treated with HIFU had higher rate of complications relative to those treated with cryotherapy (19.7 vs. 6.4%, respectively; p¼0.001). The most common postoperative complications were acute urinary retention (n¼32), followed by fever requiring antibiotics (n¼8), urinary tract infection (n¼6), hematuria (n¼2), sepsis (n¼2), cerebrovascular acci- dent (n¼1), renal colic (n¼1). No postoperative mortality was reported. Median long of stay was 3 days (IQR: 1-4). CONCLUSIONS: Applying the standard criteria proposed by Martin et al. to report complications after surgery, we provided evidence that FT is a relatively safe procedure in high volume center. Source of Funding: None PD34-07 PSA FAILS TO PREDICT TREATMENT FAILURE IN FOCAL HIGH- INTENSITY FOCUSED ULTRASOUND THERAPY IN PROSTATE CANCER Philipp M. Huber*, Bern, Switzerland; Naveed Afzal, Dorset, United Kingdom; Manit Arya, London, United Kingdom; Silvan Boxler, Bern, Switzerland; Susan Charman, London, United Kingdom; Andrew Cornaby, Dorset, United Kingdom; Tim Dudderidge, Southampton, United Kingdom; Mark Emberton, Stephanie Guillaumier, London, United Kingdom; Richard J. Hindley, Basingstoke, United Kingdom; Lucas Leemann, Zurich, Switzerland; Henry Lewi, Chelmsford, United Kingdom; Neil McCartan, Caroline M. Moore, London, United Kingdom; Raj Nigam, Surrey, United Kingdom; Chris Ogden, London, United Kingdom; Raj Persad, Bristol, United Kingdom; Karishma Shah, London, United Kingdom; George N. Thalmann, Bern, Switzerland; Jaspal Virdi, Harlow, United Kingdom; Mathias Winkler, Hashim U. Ahmed, London, United Kingdom INTRODUCTION AND OBJECTIVES: To determine if PSA kinetics following focal high-intensity focused ultrasound (HIFU) for the treatment of non-metastatic prostate cancer can predict treatment failure. METHODS: Retrospective analysis of our prospectively main- tained HIFU (Sonablate 500) database, 598 patients were identified who underwent a focal HIFU (03/2007 to 11/2016). Follow-up with 3-monthly clinical visits and PSA testing in the first year. Thereafter PSA 6-monthly or annually at least. Routine and for-cause mpMRI followed by biopsies were offered. Treatment failure was defined by any sec- ondary treatment (ADT/chemotherapy, cryotherapy, EBRT, RRP, or re- HIFU), metastasis from prostate cancer without further treatment, tumour recurrence with Gleason score 7 on prostate biopsy, or prostate cancer-related mortality. We evaluated a whole series of »nadir plus XX« thresholds (with XX from 0.1 to 2.0) for predicting failure using sensitivity and specificity, and Receiver Operating Characteristic (ROC) curve (statistics using R-language). RESULTS: Median Gleason (range) 7 (6-9), in 80% Gleason 3þ4¼7. Tumours staged as localised T1c-T2c in 522/596 (88%) cases, 74/596 (12.4%) were radiological T3a/b. Baseline median (IQR) PSA was 7.80ng/ml (5.96-10.45) in those with failure and 6.77ng/ml (2.65-9.71) in those without failure. Using ASTRO criteria, sensitivity was 18.7%, specificity 68.3%. Evaluating other PSA nadirþXX thresh- olds, the highest sensitivity of 61.2% was shown for nadirþ0.1ng/ml; the highest specificity of 59.1% for nadirþ2.0ng/ml. All definitions of PSA failure incorporating nadirþXX thresholds led to significant false posi- tives with the ROC curve shifted below the 50% line (Figure 1). CONCLUSIONS: PSA kinetics following focal HIFU therapy occur differently following whole-gland therapy. Using any »nadir plus« definition leads to significant rates of false positives which might lead to unnecessary diagnostic procedures. Source of Funding: NONE PD34-08 DETERMINANTS OF ERECTILE DYSFUNCTION FOLLOWING FOCAL ABLATIVE THERAPY FOR LOCALISED PROSTATE CANCER. Eoin Dinneen*, Neil McCartan, Hashim U. Ahmed, Mark Emberton, Richard Hindley, Clare Allen, Louise Dickinson, Suks Minhas, Tet Yap, London, United Kingdom INTRODUCTION AND OBJECTIVES: Erectile dysfunction (ED) is a recognized adverse effect of focal ablative therapy, although the mechanistic basis to the effects of such treatments on erectile function is largely unknown, with both neuropathic and vascular factors impli- cated. We aimed to evaluate the determinants of ED following focal ablative therapy using HIFU. METHODS: 119 men undergoing focal ablative therapy using HIFU enrolled in three registered studies were included (Clinical- Trials.gov NCT00561314, NCT00561262, NCT00988130). 97 were evaluable with complete MRI and functional data. IIEF-15 questionnaire scores were collected at baseline and every 3 months up to 12 months post-operatively. Two radiologists independently reviewed the pre- and early (1-3 week) post-treatment multi-parametric (mp)MRI scans, whilst blinded to IIEF-15 scores. Pre-treatment prostate mpMRI characteris- tics included prostate volume, index tumour volume where measurable (T2 sequence), position of ablated tumour and extra-capsular extension (ECE) of tumour. Post-focal MRI characteristics included necrosis vol- ume, and treatment effect extending beyond the edge of prostate, including in the region of one or both neurovascular bundles. Logistic regression analysis was used to determine associations between changes in pre- and post-treatment total IIEF-15 scores and MRI characteristics (SPSS version 25, USA). e658 THE JOURNAL OF UROLOGY â Vol. 199, No. 4S, Supplement, Saturday, May 19, 2018