PSMA-positive nodal recurrence in prostate cancer

Between February 2014 and December 2018, 100 consecutive patients underwent ⁶⁸Ga-PSMA PET/CT prior to SLNRT (67 patients) or SLND (33 patients) with evidence of PET-positive pelvic and/or para-aortic lymph nodes after RPE. Patients who underwent previous adjuvant radiotherapy of the pelvic lymphatic pathways or ADT prior to SLNRT/SLND were excluded. Patients were subgrouped according to the D’Amico criteria [12] incorporating tumour-stage, PSA level and Gleason score (GS) (Table 1). All patients provided written informed consent to undergo ⁶⁸Ga-PSMA PET/CT. This retrospective analysis was performed in compliance with the principles of the Declaration of Helsinki and its subsequent amendments [13].

Radiolabelling of PSMA-HBED-CC was performed with ⁶⁸Ga³⁺ from a ⁶⁸Ge/⁶⁸Ga generator system (GalliaPharm®, Eckert & Ziegler AG, Berlin, Germany) using an automated synthesis module (GRP, Scintomics GmbH, Munich, Germany) and prepacked cassettes (ABX GmbH, Radeberg, Germany) as described previously for a different PSMA ligand by Weineisen et al. [14]. ⁶⁸Ga-PSMA PET/CT imaging was performed according to current guidelines [15] with a Siemens Biograph 64 or GE Discovery 690 PET/CT camera. Phantom studies based on the National Electrical Manufacturers Association NU2-2001 standard were conducted to allow valid pooling of the results, and standardized uptake value (SUV) conversion factors were calculated [16]. ⁶⁸Ga-PSMA PET/CT scans were performed with a diagnostic CT scan (reference mAs, 200–240; 120 kV) and obtained with intravenous injection of iodine-containing contrast agent (Ultravist 300, Bayer Pharma AG, Berlin, Germany; or Imeron 300, Bracco, Konstanz; 2.5 mL/s; in portal venous phase) around 60 min after intravenous administration of ⁶⁸Ga-PSMA. In absence of contraindications, 20 mg furosemide was injected almost simultaneously with ⁶⁸Ga-PSMA injection to lower tracer retention in the bladder. PET images were reconstructed with an axial 168 × 168 matrix based on the TrueX algorithm (3 iterations, 21 subsets; Biograph 64) or on the VUE Point FX algorithm (2 iterations, 36 subsets; Discovery 690).

PET/CT images were interpreted by a consensus read of two nuclear medicine physicians and two radiologists in the sense of a clinical report-based analysis. At least two of the readers had more than 10 years PET/CT experience. The location of each lesion was determined by CT. PET-positive lesions were visually identified by ⁶⁸Ga-PSMA uptake above background level and not associated with physiologic uptake [15].

All patients received ENRT by intensity-modulated radiotherapy (IMRT) or volumetric arc therapy (VMAT) and image-guided radiotherapy techniques (IGRT, 2–5 times/week). Radiotherapy dose regimens were normo- or slightly hypofractionated and a boost to the PET-positive local recurrences within prostatic fossa and lymph nodes was applied simultaneously. The prostatic fossa was treated with a median of 66 Gy (range, 60–67.20 Gy) in single doses of 2 Gy (range, 1.8–2.12 Gy) and the elective lymphatic pathways with 50.4 Gy (range, 45–52.28 Gy) in single doses of 1.8 Gy (range, 1.6–1.8 Gy). PSMA PET-positive local recurrences within in the prostatic fossa were irradiated with 70.0 Gy (range, 68–70 Gy) and PSMA PET-positive lymph nodes with 61.6 Gy (range, 50.4–66 Gy). Planning CT was done in supine position with 3 mm slice thickness. Patients were advised to have a full bladder and empty rectum. Target delineation was performed according to the Radiation Therapy Oncology Group (RTOG) atlas for salvage PCa [17] and for lymphatic pathway delineation [18].

The standard SLND procedures at our department have been described before [7, 19]. Briefly, an open approach through an abdominal midline incision was used and standard extended SLND was performed based on specific regions according to the most recent PSMA PET findings. Dissected lymph nodes were classified based on the respective anatomic region. Routinely, dissected lymph nodes were immediately sent for histopathologic analysis and evaluated according to standard protocols (serial sectioning, 200 μm slices).

Treatment management following PSMA PET was documented for each patient. Follow-up time was defined as the interval in months between SLND/SLNRT and the last recorded PSA. Follow-up examination was first performed 6 weeks to 3 months after SLND/SLNRT and then every 6–12 months. Patients were regarded free of ADT influence after a minimum time-period of 5 months after last application of ADT.

BRFS (PSA ≤0.2 ng/ml) was defined as the primary study endpoint. For statistical analysis, SPSS Statistics 25 (IBM, New York, NY, USA) was used. Demographic data were analysed using descriptive statistics and χ² test. Time to event data was calculated using the Kaplan–Meier method. Differences between subgroups were compared using log rank test with a p value of <0.05 considered statistically significant. The χ² test, Mann–Whitney U test and binary logistic regression were performed to determine the influence of SLND vs. SLNRT, GS ≤6–7 vs. GS 8–9, tumour stage T2 vs. T3–4, nodal stage N0/N x vs. N1, D’Amico stage low/intermediate vs. high, PSA at salvage therapy, PSA persistence vs. PSA recurrence and PET-positive pelvic vs. pelvic and/or retroperitoneal lymph nodes on bRFS.

Most patients (87/100; 87%) had a high-risk PCa at time of RPE with no significant difference between the SLND and SLNRT cohort. SLNRT patients had significantly more tumours with locally advanced pT3 or pT4 disease (82% vs. 67%; p = 0.006) compared to the SLND cohort with significantly more tumours of pT2 category (33% vs. 18%). Furthermore, there were significantly more patients in the SLNRT cohort with evidence of pathologically involved lymph nodes (45% vs. 27%; p = 0.001) and positive surgical margins (54% vs. 12%; p = 0.001) at time of RPE. Overall, there were 58 patients with PSA persistence and 42 patients with PSA recurrence after RPE. SLNRT for PET-positive lymph nodes was significantly more often applied in patients with PSA persistence (73%) than PSA recurrence (27%). On the other hand, SLND was performed significantly more often in patients with PSA recurrence (73%) than PSA persistence (27%; p = 0.001) (Table 1). Time between RPE and SLND/SLNRT was longer in patients with PSA recurrence than PSA persistence (6 vs. 44 months and 5 vs. 69 months). Patients with SLND had only evidence of PET-positive lymph nodes (100%). Patients with SLNRT had PET-positive lymph nodes with (24%) or without (76%) local recurrence within prostatic fossa (p = 0.001). There was no significant difference in the distribution of PET-positive lymph nodes between the two cohorts: Only few patients had evidence of para-aortic lymph nodes (24% in the SLND group and 19% in the SLNRT group; p = 0.063).

Median PSA prior SLND was higher than prior SLNRT (3.07 ng/ml vs. 1.3 ng/ml; p = 0.393). Overall, a median of 8 (1–44) lymph nodes were removed at time of SLND. Evidence of PCa was pathologically confirmed in a median of 1 (0–16) lymph node. In total, there were 4 patients with a Clavien Grade II (2 × paralytic ileus, 2 × lymphorrhea) and 2 patients with a Clavien Grade IIIa (2 × pulmonary artery embolism) complication (Table 2).

ADT was recommended to all patients with SLNRT due to the evidence of PET-positive lymph nodes for 2 years [20, 21]. Consequently, 59/67 (88%) patients were started on ADT before initiation of SLNRT, 42/67 patients (63%) discontinued after a median time of 7 (2–41) months due to patients’ preferences. Median time between end of ADT and last follow-up was 27 months (range 0–48) in those patients with discontinued ADT. In all, 8/67 (12%) patients refused ADT. Patients received radiotherapy treatment as depicted above. Acute grade 2 gastrointestinal and urogenital toxicity were each observed in 19/67 (28%) patients consisting primarily of diarrhoea and increased urinary frequency. Acute grade 3 urogenital toxicity occurred in 1/67 (2%) patients with evidence of urethral stenosis. Late grade 2 toxicity was overall seen in 24/67 (36%) patients with mainly signs of erectile dysfunction and increased urinary frequency. Late grade 3 toxicity was present in 25/67 (37%) patients with mainly erectile dysfunction (36%) (Table 3).

Median follow-up was 17 months (range, 6–53 months) in patients with SLND and 31 months (range, 3–56 months) in patients with SLNRT (p = 0.027). Median posttreatment PSA was significantly higher in patients with SLND (0.47 ng/ml, range <0.03–9.61 ng/ml) than in patients with SLNRT (0.05 ng/ml, range <0.03–19 ng/ml) (p = 0.003). At last follow-up, median PSA was 2.50 ng/ml (range <0.03–68 ng/ml) in patients with SLND and 0.05 ng/ml (range <0.03–268 ng/ml) in patients with SLNRT (p = 0.025). Overall, the 2‑year bRFS was significantly higher in the SLNRT cohort compared to the SLND cohort (92% vs. 30%; p = 0.001) irrespective of ongoing androgen deprivation therapy at last contact (Fig. 1; Table 4). This resulted in a significantly higher rate of distant metastases (52% vs. 21%; p = 0.002) and secondary treatments (39% vs. 15%; p = 0.011) in patients with SLND. There was one tumour-related death in the SLNRT group.

In multivariable analysis stratified for Gleason score at RPE and PSA prior to PSMA PET/CT, SLND vs. SLNRT was significantly associated with bRFS (regression coefficient 1.473, hazard ratio 4.360, 95% CI 1.665–11.418; p = 0.003) (Table 5 and Supplemental Table 1).