Detection level and pattern of positive lesions using PSMA PET/CT for staging prior to radiation therapy

Background To determine the potential role of 68Ga-PSMA positron emission tomography/computed tomography (PET/CT) in radiotherapy (RT) planning for prostate cancer (PCa). Methods One hundred twenty-nine patients (pts) with 68Ga-PSMA PET/CT were retrospectively analysed. Potentially influencing factors (androgen deprivation therapy, amount of 68Ga-PSMA-HBED-CC, PSA doubling time ≤/> 10 months, PSA before PET/CT, T−/N-category and Gleason score) were evaluated by logistic regression analysis. The detection rate of PSMA PET/CT was compared to contrast enhanced CT and its impact on RT management analysed. Results One hundred twenty-nine patients (pts) (20 at initial diagnosis, 49 with PSA relapse and 60 with PSA persistence after radical prostatectomy) received PSMA PET/CT prior to RT. The majority of pts. (71.3%) had PET-positive findings (55.1% of pts. with PSA recurrence, 75% of pts. with PSA persistence and 100% of newly diagnosed pts). Median PSA before PET/CT in pts. with pathological findings (n = 92) was 1.90 ng/ml and without (n = 37) 0.30 ng/ml. PSA level at time of PET/CT was the only factor associated with PET-positivity. In pts. with a PSA ≤ 0.2 ng/ml, the detection rate of any lesion was 33.3%, with a PSA of 0.21–0.5 ng/ml 41.2% and with a PSA of 0.51–1.0 ng/ml 69.2%, respectively. Regarding the anatomic distribution of lesions, 42.2% and 14.7% of pts. with relapse or persistence had pelvic lymph node and distant metastases. In pts. at initial diagnosis the detection rate of pelvic lymph nodes and distant metastases was 20% and 10%. 68Ga-PSMA PET/CT had a high detection rate of PCa recurrence outside the prostatic fossa in pts. being considered for salvage RT (22.4% PET-positive pelvic lymph nodes and 4.1% distant metastases). Compared to CT, PSMA PET/CT had a significantly higher sensitivity in diagnosing rates of local recurrence/primary tumour (10.1% vs. 38%), lymph nodes (15.5% vs. 38.8%) and distant metastases (5.4% vs. 14.0%). This resulted in a modification of RT treatment in 56.6% of pts. Conclusions The detection of PCa is strongly associated with PSA level at time of 68Ga-PSMA PET/CT. PSMA PET/CT differentiates between local, regional and distant metastatic disease with implications for disease management. PSMA PET/CT allows for tumour detection in post-prostatectomy pts. with PSA ≤ 0.5 ng/ml considered for salvage RT.


Background
A precise detection and visualisation of disease extent is of essential importance for indication and target volume definition in radiotherapy of prostate cancer. Due to limitations in sensitivity and specificity, MRI and choline PET/CT are not considered precise enough especially at PSA levels ≤1 ng/ml [1,2], although both imaging modalities are as well capable of modifying treatment approaches [3][4][5][6][7]. Recently, a 68 Ga-labelled PSMA-targeted ligand with high affinity to prostate-specific membrane antigen (PSMA), a cell surface protein overexpressed in prostate cancer cells has been introduced for positron emission tomography (PET) imaging [8][9][10][11]. PET with 68 Ga-PSMA demonstrated a superior tumour-tobackground signal intensity and substantially higher detection rates than have been previously reported for other imaging modalities in patients with newly diagnosed [12][13][14][15] or recurrent [3,[16][17][18][19][20] prostate cancer. From surgical series, it is known that 68 Ga-PSMA positron emission tomography/computed tomography (PET/CT) allows for a highly correct identification of lymph node metastases with a sensitivity and specificity ranging from 66% to 84% and 82% to 99% [14,15], respectively. However, there is a substantial percentage of prostate cancer patients with a negative 68 Ga-PSMA PET/ CT. Bearing this in mind, we tried to determine retrospectively which factors correlate with a positive 68 Ga-PSMA PET/CT in patients with primary, persistent or recurrent prostate cancer. Furthermore, we compared the detection rate of 68 Ga-PSMA PET/CT to contrast enhanced CT and analysed its impact on the decisionmaking process within our department of radiation oncology. Several clinical scenarios may be distinguished: A PET-positive local recurrence within the prostatic fossa might justify a simultaneous integrated boost and an additional antiandrogen therapy. Likewise, PET-positive pelvic lymph node metastases might be treated with an enlargement of the clinical target volume, a focally higher dose and antiandrogen therapy. In addition, the evidence of distant metastases might even cause a non-realisation of a planned radiotherapy treatment.

Study population
68 Ga-PSMA PET/CT has been routinely offered to patients for prostate cancer staging before radiotherapy in our clinic (February 2014-August 2016). A total of 129 patients consecutively received 68 Ga-PSMA PET/CT prior to radiotherapy in different clinical scenarios: 20 patients at initial diagnosis (mostly high-risk patients with suspected extra-prostatic manifestations), 60 with biochemical persistence after radical prostatectomy and 49 with biochemical relapse (Table 1). All patients gave written informed consent to undergo 68 Ga-PSMA PET/ CT. This retrospective analysis is in compliance with the principles of the Declaration of Helsinki and its subsequent amendments [21] and was approved by the Ethics Committee of the LMU Medical Faculty .
PSMA ligand and PET/CT imaging PSMA-HBED-CC was radiolabelled with 68 Ga 3+ from a 68 Ge/ 68 Ga generator system (GalliaPharm ® , Eckert & Ziegler AG, Berlin, Germany) using an automated synthesis module (GRP, Scintomics GmbH, Munich, Germany) and pre-packed cassettes (ABX GmbH, Radeberg, Germany) as described previously for a different PSMA ligand by Weineisen et al. [22]. 68 Ga-PSMA PET/CT images extending from the base of the skull to the mid-thigh were acquired. PET/CT scan was obtained with intravenous injection of iodinecontaining contrast agent (Ultravist 300, Schering, Berlin, Germany; or Imeron 300, Bracco, Konstanz; 2.5 mL/s; in portal venous phase) 60 min after almost simultaneous intravenous administration of 20 mg furosemide and 68 Ga-PSMA (median 189 megabecquerel (MBq), range 87-293). Directly prior to the PET/CT scan, patients were asked to empty their bladder to minimise tracer accumulation.

Image interpretation
PET/CT was interpreted by a consensus read of one nuclear medicine physician and one radiologist. Location of lesions was each determined by CT. PET-positive lesions were identified by 68 Ga-PSMA uptake visually above background and not associated with the physiologic uptake. CT-positive nodes were defined by increased short axis diameter, loss of fatty hilum, or increased contrast enhancement. Bone metastases were detected by suspicious sclerotic lesions, visceral metastases by suspicious hypodense or hyperdense lesions in the respective organ. Based on PET/CT images and reports, stage according to PET or CT was documented separately by one nuclear medicine physician and one radiation oncologist [23].

Statistical analysis
Demographic and tumour characteristics were analysed. The association between positive 68 Ga-PSMA PET/CT findings and possibly interacting variables, like androgen deprivation therapy (ADT) at the time of PET/CT, amount of injected tracer, PSA-level, PSA doubling time, T−/N-category and Gleason score was assessed by univariate and multivariate logistic regression analysis. P values <0.05 were considered statistically significant. PSA before PET/CT was classified in eight steps (≤ 0.2 ng/ml -> 20 ng/ml). Gleason score was grouped in six different classes (GS 5/6, GS 7a, GS 7b, GS 8, GS 9 and GS 10), T-category in four (T1 -T4) and Ncategory in three classes (N0, N1, Nx). Injected tracer amount was evaluated as multiples of 100 MBq. Patients after radical prostatectomy were classified by PSA doubling time ≤ vs. > 10 months. PET/CT positive findings (primary tumour/local recurrence, lymph node and distant metastases) are presented separately according to the respective treatment indication prior to radiotherapy. The detection rate of residual tumour, recurrence or newly diagnosed disease comparing PET/CT to CT exclusively as well as change in radiotherapy management was analysed by Fisher's test.

Results
A total of 129 patients (20 at initial diagnosis, 60 with biochemical persistence and 49 with biochemical relapse after radical prostatectomy) consecutively received 68 Ga-PSMA PET/CT prior to radiotherapy (   (Table 3)

Tumour location and PET versus CT-positive findings
Prior to definitive radiotherapy, 68 Ga-PSMA PET/CT showed suspicious lesions within the prostate gland in 90% of patients and suspicious pelvic lymph nodes and/ or distant metastases in 20%/10% of the patients, respectively. 68 Ga-PSMA PET/CT had a high detection rate of prostate cancer recurrence outside the prostatic fossa in patients being considered for salvage radiotherapy: 22.4% of patients had PET-positive pelvic lymph nodes and 4.1% distant metastases. Both patients with distant metastases had also local recurrence exclusively without evidence of pelvic lymph node metastases. In patients considered for salvage radiotherapy with a PSA < 0.5 ng/ml (25 patients), 68 Ga-PSMA PET/CT detected in 16.0% local recurrences within the prostatic fossa and in 20.0% PET-positive pelvic lymph nodes. In patients with postoperatively biochemical persistence, PET/CT revealed a high number of lymph node metastases (58.3%), distant metastases (23.3%) and macroscopic residual tumour (21.7%). Seventeen of 60 patients with persistent PSA after radical prostatectomy had a pre-PSMA PET/CT PSA <0.5 ng/ml. Of these patients, there was one patient (5.9%) with macroscopic residual tumour, 7 patients with pelvic lymph node metastases (41.2%) and 1 patient with distant metastases (5.9%). The single patient with bone metastases had pelvic lymph node involvement as well. Figure 3 shows a lesion based analysis of PETpositive local recurrence/primary tumour, PET-  Figure 4 demonstrates the integration of 68 Ga-PSMA PET/CT images into radiation therapy planning in three patients.

Discussion
This heterogeneous group of patients with primary, recurrent or persistent prostate cancer represents a typical cohort referred to radiation oncology departments. We are thus confident that our findings are of relevance in clinical practice.
In general, out of 129 patients with high-risk primary, persistent or recurrent prostate cancer, PSMA-PET revealed a least one relevant lesion in 71.3% of the cohort. This rate compares nicely to the number of PET-positive patients (82.8%) in the study of Afshar-Oromieh et al. [16]: a similarly heterogeneous group of patients with progressive disease after initial treatment (e.g. radiotherapy and/or surgery; n = 292), before initiation of local therapy (n = 27) or before therapy with radiolabelled PSMA ligands (n = 38) was analysed.
The clear correlation of PSA level and probability of detection was seen in our analysis (Fig. 1) as well as in the data provided by Ceci et al. [25] (n = 70), Afshar-Oromieh et al. [16] (n = 292) and Eiber et al. [17]. Accordingly, patients with biochemical recurrence having the lowest PSA levels prior to PSMA PET/CT had the lowest detection rate (55.1%) compared to patients with persistent (75%) or primary prostate cancer (100%).
At present, PSA levels ≤0.2 ng/ml are frequently judged in postoperative patients to be non-critical, nevertheless one-third of our patients (33.3%) had PET-positive findings even below a PSA of 0.2 ng/ml. This rate drastically increased up to 69.2% in patients with a PSA level between 0.51-1.0 ng/ml. Thus, PSMA PET/CT detects early local recurrence or metastatic disease and possibly allows for a more effective and early treatment.
Apart from PSA-level before PET/CT, multivariate analysis detected no significant associations between a positive PSMA PET/CT and any of the other factors including ADT, amount of injected tracer, PSA doubling time, and Gleason score. This finding is well in accordance with the observations provided by Afshar-Oromieh et al. [16] regarding PSA doubling time and Gleason score. In contrast, Afshar-Oromieh et al. reported a positive association between detection rate and ongoing ADT. In animal models and in vitro cell culture experiments, ADT increases PSMA expression which might contribute to improved detection rate by 68 Ga-PSMA PET/CT [26][27][28][29][30]. However, at present the complex association between an ADT driven increase of PSMA uptake in the individual tumour lesion, the ADT driven reduction of global tumour load and the final PET signal is not completely understood. In our opinion, the use of ADT in a cohort with mostly high-risk prostate cancer patients as evaluated by Afshar-Oromieh et al. is also a surrogate for high tumour burden and high PSA levels. We therefore counsel our patients to start with showing that PSMA PET/CT is significantly more sensitive than CT-based 3D-volumetric lymph node evaluation [18].
Furthermore, we analysed the question how far an increased detection rate leads to a change in therapeutic management. In our series, PSMA PET/CT-positive findings had a substantial influence on the therapeutic concept in 56.6% of the patients with implication on either adaptation of treatment volumes, dose concepts or the commencement of hormone ablation therapy. This compares nicely to similar analyses [19,[31][32][33] all Fig. 3 Lesion based analysis (PET-positive local recurrence/primary tumour and PET-positive lymph node and distant metastases) according to sub-grouping of patients (patients before definitive RT at initial diagnosis, patients before salvage RT with PSA recurrence, patients before Salvage RT with PSA < 0.5 ng/ml, patients before additive RT with PSA persistence, patients before additive RT with PSA persistence with PSA < 0.5 ng/ml) reporting major modifications of therapeutic management in 33.3% -53.7% of the patients.
At present, PSMA PET/CT is widely regarded as the best modality for lymph node staging in a primary as well as in a postoperative setting [14,34]. Data provided by van Leeuwen et al. [19] suggest that the level of lymph node positivity has been strongly underestimated in the pre-PSMA PET/CT era. In their analysis, lymphatic recurrences outside the prostatic fossa occurred in 20% of patients scheduled for salvage radiotherapy with additional 10% of distant metastases detected in PSMA PET/CT. Bearing in mind that currently only target volumes covering the prostatic fossa are routinely employed [35], a substantial rate of lymph node metastases is not treated. This is supported by our own data showing that even below a PSA level of 0.5 ng/ml, PSMA PET/CT still detected 20.0% of pelvic lymph node metastases in patients with salvage radiotherapy indication.
The documented need to treat PSA relapses as early as possible (even below PSA levels of 0.2 ng/ml) and the fact that the diagnostic sensitivity is optimal above a PSA level of 0.5 ng/ml currently creates an inherent dilemma. Since all data available consistently document that PSA control is significantly better when radiotherapy is commenced as early as possible [36,37], it is not justified to wait until PSA is in an optimal diagnostic range. Our strategy to cope with this dilemma is therefore to employ PSMA PET/CT in patients even around 0.2 ng/ml especially whenever high-risk features are present. Like the analysis by Henkenberens et al. on 29 patients with biochemical recurrence receiving an individualized radiotherapy treatment plan based on PSMA PET/CT findings [38], target volumes are modified accordingly in our institution after critical assessment. Nevertheless, keeping the sensitivity and specificity rates from surgical series in mind, PSMA PET/CT might still underestimate the true extent of disease and should therefore from our point of view not result in an omission of nomogram [39] triggered radiotherapy treatment volumes.
At present, legal constraints prohibit the use of PSMA in truly prospective studies in Germany. Despite the clear limitations of retrospective approaches, the available data from several analyses strongly underline the high value of PSMA-PET/CT for staging as well as for treatment stratification in patients with primary high-risk, persistent or recurrent carcinoma of the prostate.

Conclusions
In this retrospective study, we showed that the detection of prostate cancer is strongly associated with PSA level at time of Ga-PSMA PET/CT. Better than any other imaging modality, PSMA PET/CT differentiates local, regional and distant metastatic disease with considerable implications for disease management. In post-prostatectomy patients with rising PSA ≤ 0.5 ng/ml with indication for salvage radiotherapy, PSMA PET/CT is narrowing the diagnostic gap as until now the gradual PSA increase often occurred long before recurrent disease could be localized clinically or by imaging.