Histopathologic Features that Influence the Detection of Occult Lymph Node Metastasis Using PET/CT Imaging with Anti-PSMA 89Zr-Df-IAB2M in Newly Diagnosed High Risk Prostate Cancer Patients

Objectives Appropriate treatment recommendations for men with newly diagnosed high-risk prostate cancer (HRPC) requires an accurate knowledge of pelvic lymph node (LN) disease. Up to 50% of HRPC patients have occult disease at diagnosis as defined by negative conventional imaging (CI) studies. We have been using a novel minibody PET agent, 89Zr-Df-IAB2M (an engineered recombinant antibody fragment that targets the extracellular domain of PSMA) for detection of prostate cancer (PC). An ongoing phase II, single center trial comparing 89Zr-Df-IAB2M PET to 111In Capromab Pendetide (CP) SPECT imaging in HRPC patients with negative CI revealed superiority of 89Zr-Df-IAB2M over CP and CI for LN mets using histopathology as truth standard (WMIS 2015, Honolulu,Hi). However, the biologic and cellular properties of PC deposits influencing the visualization of LN on 89Zr-Df-IAB2M PET/CT is not completely understood. We report here the histologic properties of LN disease and other measures that may account for 89Zr-Df-IAB2M uptake on PET/CT scans in the first 10 enrolled subjects.

Methods IRB approved enrollment of the first 10 HRPC patients with expected extracapsular disease > 15% (Briganti nonogram) or Gleason score 蠅 9 underwent sequential CP SPECT/CT (5mCi 111In, 3 days p.i.) and 89Zr-Df-IAB2M PET/CT (2.5mCi, 1-2 days p.i.) up to 28 days prior to surgery. Surgery consisted of radical prostatectomy with standard lymph node dissection leading to the removal of pelvic lymph nodes from 6-8 nodal stations per subject. PC LN involvement was confirmed by routine H&E staining. All positive nodes were sent to a central laboratory to evaluate PSMA expression. An H-score (range 0-300) was assigned based on staining intensity (0-3) as well as on the % of positive cells at each staining intensity. Central review of all images was performed with the reader blinded to surgical pathology results, screening PSA and Gleason score. The location of any PET or CP positive LNs were recorded along with SUVmax and nodal size based on screening CI studies. Statistical correlations were performed using ROC analysis and Spearman’s correlation with significance p <0.05.

Results Six of 10 (60%) of subjects had pathologically confirmed positive LN disease. The 89Zr-Df-IAB2M PET/CT was positive in 4/6 (67%) subjects while 0 (0%) were positive on CP or CI. There was a good correlation between screening PSA (ρ= 0.890; p=0.006), anti-PSMA H score (ρ= 0.737; p=0.024) and SUVmax at 48h but not microscopic tumor cellularity (ρ= 0.089; p=ns) or nodal size (ρ= 0.667; p=0.102). ROC analysis of 89Zr-Df-IAB2M scan performance based on PSMA staining scores >200 showed good sensitivity for LN detection on PET (75.0% sensitivity, 85.7% specificity, 0.784 AUC, p =0.004) indicating that PSMA density is an important determinant of LN detectability.

Conclusions Early data suggests that the 89Zr-Df-IAB2M PET detection of occult LN metastasis in HRPC is related to PSMA expression but not PC nodal cellularity. As expected, patients with higher screening PSA levels and H-scores have greater uptake on 89Zr-Df-IAB2M PET. Thus, 89Zr-Df-IAB2M remains a promising imaging agent for the detection of occult PSMA positive nodal disease.