Abstracts

RETURN TO ABSTRACT LISTING


2470. Added Value of Fluciclovine PET/CT to Clinical Nomograms in Staging High-Risk Prostate Cancer
Authors * Denotes Presenting Author
  1. Akinyemi Akintayo *; Emory University
  2. Olayinka Abiodun-Ojo; Emory University
  3. Dattatraya Patil; Emory University
  4. David Schuster; Emory University
Objective:
Clinical nomograms are used for selection of patients with prostate cancer (PCa) that will benefit from radical prostatectomy with extended pelvic lymph node dissection (RP+EPLND). We assessed the added value of fluciclovine PET/CT in predicting lymph node metastasis (LNM) in patients with high-risk PCa.

Materials and Methods:
57 patients (mean PSA: 24.5±28.6 ng/ml) with high-risk PCa underwent fluciclovine PET/CT 10.6±7.8 days before RP+EPLND. Abnormal fluciclovine uptake in the prostate and lymph nodes were noted. Histopathology was the primary reference standard for LNM. Prediction of LNM with Clinical nomograms: Yale formula and Partin table was done. Association between clinical parameters including age, PSA, National Comprehensive Cancer Network (NCCN) risk category, Grade Group and LNM were determined. Multivariable logistic regression model including clinical parameters and nodal status on PET was built. Efficacy of the model in predicting LNM was compared with clinical nomograms using receiver operating characteristics (ROC) analysis.

Results:
31/57 patients had LNM on histology. Fluciclovine PET/CT correctly detected LNM in 17 patients, with sensitivity of 55.3% and specificity of 84.8%. On univariate logistic regression analysis, detection of nodal disease on PET was significantly associated with LNM (p=0.004). There was no association between age, PSA, Grade Groups, risk category and LNM. Prediction derived from Yale Formula (p=0.026) and Partin table (p=0.029) were significantly associated with LNM. Multivariate logistic regression model using clinical parameters and nodal status on PET showed nodal disease on PET significantly predicted LNM (Odds ratio (95%CI): 7.00 (1.62-30.26); p=0.009). The ROC of the model including clinical parameters and PET had AUC (95%CI): 0.79 (0.67-0.91), Yale formula had AUC (95%CI): 0.68 (0.54-0.82) and Partin table had AUC (95%CI): 0.67 (0.53-0.81).

Conclusion:
Our model combining clinical parameters and nodal status on fluciclovine PET/CT performed better than currently used clinical nomograms with significantly higher odds ratio for LNM. Addition of fluciclovine PET/CT to preoperative assessment of patients with high-risk PCa may be valuable in treatment planning.