**3. Discussion**

In this study, the biphasic [68Ga]Ga-PSMA-11 PET/CT of 233 patients was retrospectively studied. A total of 178 prostatectomized patients and 37 irradiated patients as well as 18 pre-therapy patients were assessed, and their data (e.g. tumor uptake, biphasic LPR) were separately evaluated. As reported in other studies, we also found high LPR from baseline [68Ga]Ga-PSMA-11 PET/CT for the staging (72%) and restaging (68%) of PCa patients [11,13,18].

A recently published prospective, randomized, multi-center study from Australia [9] including 300 men with biopsy-proven PCa found that [68Ga]Ga-PSMA PET/CT yielded 92% accuracy in identifying those with distant metastatic or pelvic nodal disease compared with 65% accuracy from traditional imaging (CT, bone scan). Furthermore, conventional imaging had more equivocal findings, fewer management changes, and higher radiation doses (19.2 mSv vs. 8.4 mSv; *p* < 0.001) [9]. In addition to improving detection, PSMA PET/CT will have a significant impact on a patient's treatment plan and disease management in future guidelines [9,11].

Several acquisition protocols with different acquisition times, including early dynamic to 3 h p.i. imaging, have been proposed for [68Ga]Ga-PSMA PET/CT studies [18–22]. Time activity curves acquired from PCa lesions showed a continuously increasing tracer accumulation during early dynamic PET acquisition, which also supports the essential role of additional late imaging [22]. The addition of delayed scans has been considered to offer substantial advantages for the discrimination of PCa versus non-PCa lesions, as the malignant foci usually show a further increase in tracer accumulation on the late scans. Benign lesions, on the other hand, usually show a decrease in SUV [18–22]. The optimal time point for the various currently available tracers for PSMA PET/CT imaging and the potential of additional late images have been and are currently being investigated. In the present

study, we evaluated the incremental value of [68Ga]Ga-PSMA-11 PET delayed imaging, especially abdominopelvic imaging.

As described by previous studies, there is a significant increase in SUVmax of PCa lesions on delayed images when evaluating dual time point [68Ga]Ga-PSMA-11 PET/CT imaging [18,21,23,24]. Beheshti et al. reported that this increase relates to suspicious lesions (*p* < 0.001) in the prostate bed (11.6 ± 8.2 to 14.8 ± 1.0) as well as to LNs (9.7 ± 5.9 to 12.3 ± 8.8) [23]. Nevertheless, lesions' tracer accumulation on early imaging has been sufficient for diagnosis [23], which is consistent with our results.

Afshar-Oromieh et al. reported a mixed pattern of tracer behavior (increase/decrease of SUVmax in metastases) in the same patient in 11.6% (8/69 patients), whereas six of 69 patients (8.7%) showed a consistent decrease in metastatic uptake [20]. We did not find similar results in our patient group. Another study showed a SUVmax decrease in 26 out of 157 lesions [23]. Beheshti et al. reported [23] an increase of SUVmax over time in most lesions, which was also in agreement with our findings. Delayed images could confirm malignancy of 33 moderately PSMA-avid lesions, which were suspicious of being malignant on early scans, due to the increase of tracer uptake. However, some of the findings were ambiguous (11 lesions only detected in delayed scan, seven lesions only detected in early scan), but all of them were characteristic for PCa in follow-up (such as PSMA PET/CT, PSMA PET/MRI, CT, MRI), which support the results of a previous study that demonstrated [68Ga]Ga-PSMA PET/CT to be a helpful tool to determine malignancy in ambiguous lesions [24]. These different results might be explained by various tumor cell biologies. Afshar et al. suspect that individual lesions may have a decreased rate of internalization of the PSMA ligand [20]. We speculate that miscellaneous and mixed patient populations may also account for at least some of the different findings that have been reported in the literature. In the large patient group with recurrent PCa, the role of primary treatment (e.g., prostatectomy, radiation therapy) could be an important factor as well.

In our study, the clinically most important group (P-P) of recurrent PCa-patients (*n* = 114/178) showed no significant difference in LPR (551 lesions early vs. 554 late). However, a statistically significant increase of tumor uptake in PCa lesions detected by baseline PET/CT compared with delayed PET/CT was shown. As described by previous PSMA PET studies, the standardized uptake values (SUVs) of LNs are significantly higher 3 h p.i. than 1 h p.i., and nearly all LNM of PCa show high PSMA expression [18,25]. The increase of SUVmax values between baseline and delayed scan in our study also did not significantly raise the number of pathological findings in the P-I and in the P-T-groups, especially for LN lesions, which are of important clinical interest for therapy planning. Due to the fact that microscopic LNs uploaded with metastatic tumor cells are frequently non-enlarged, LN staging and restaging by CT and MRI alone is limited [8], and PSMA PET/CT or PET/MRI is preferred [9,11,12,26]. In a series of PCa patients, the authors found 72% [25] of [68Ga]Ga-PSMA-avid LNs to be metastatic in normal-sized LNs (<1 cm) [25,26]. Our data demonstrated nine LNM with high PSMA avidity on PET, which showed no signs of malignancy on CT alone. Aside from the overexpressed PSMA avidity in prostate tumor cells, the LPR of microscopic LNs could improve in the near future as the next generation of scanners (including time-of-flight technique) results in increased spatial resolution, which—compared to the older scanner systems—leads to a higher contrast as well as a higher intrinsic sensitivity [24,27].

A just published comprehensive literature search [28], including nine retrospective and two prospective studies, reported detection rates of [68Ga]Ga-PSMA-PET in recurrent patients for PSA <0.2 ng/mL, for 0.2–0.49 ng/mL, and for PSA 0.5 to <1.0 ng/mL ranged from 11% to 50%, 20% to 73%, and 25% to 88%. Our results match those of Luiting et al. We had LPR values of 27% for 0.2 to <0.5 ng/mL and 32% for 0.5 to <1.0 ng/mL. The subgroup of patients with PSA < 0.2 ng/mL was excluded in our study, because they do not belong to BC patients per definition [15]. The authors [28] observed high specificity rates of [68Ga]Ga-PSMA-PET imaging for pelvic LNM detection in primary staging as well as in restaging, while sensitivity was modest, and they concluded that [68Ga]Ga-PSMA PET has a high impact in patient management concerning the salvage setting [11,28,29]. In our study, we found LNM in 70% of the PSMA-avid metastases, 78% of which were pelvic LNM in restaging patients (P-P). In primary tumor staging, LNM were detected in 31%, of which pelvic LNM were shown in 75%. Previous studies report the dynamic uptake of PSMA in ganglia (e.g., celiac ganglia) [18,30]. However, none of the patients in the present study showed PSMA-positive celiac ganglia, neither 1 h p.i. nor 3 h p.i.

The impact of delayed imaging in our patient groups (P-P, P-I, P-T) was limited due to the lack of significantly increased rates of pathological findings 3 h p.i. Our findings were consistent with the results of a study by Derlin et al. [31] using [68Ga]Ga-THP-PSMA, who also found that delayed imaging did not increase the number of detected metastases significantly (two out of 99 patients). In contrast, Afshar-Oromieh et al. [32], using [68Ga]Ga-PSMA-11, found a 3-h delay as an optimal time point for imaging, as the majority of cancer lesions could be detected then. However, their patient cohort was very small (*n* = 4). It is known from early pharmacokinetic studies that the background activity decreases significantly between 1 and3hp.i., resulting in an improved tumor/background (T/B) ratio [5]. However, in the present study, these higher T/B ratios or contrasts to lesions' tracer accumulation did not result in significantly higher LPR on delayed images, compared to other authors [21]. It must be taken into account here that due to the 68Ga's relatively short half-life of 68 minutes, the count statistics in the 3-hour measurement are significantly lower. Most PC lesions (97%) in our study were already detected early in the imaging process, which sheds some doubt on the need for a second late examination in clinical routine. However, sometimes, it can also be useful to perform late images, e.g., if the effect of urinary activity in assessing pelvic PCa lesions remains unclear in baseline scans [26,33]. In this setting, imaging with 18F-labeled compounds (PSMA-based radiopharmaceuticals such as [ 18F]PSMA-1007) should be considered [34], since it offers advantages in late imaging due to the longer half-life of 110 minutes and the significantly lower positron range with improved resolution and the detection of small LNs. However, with regard to a theranostic approach, therapies with [ 177Lu]Lu-PSMA make a pre-therapeutic PET/CT with [68Ga]Ga-PSMA appear more meaningful [6,35]. From our theranostic point of view, dual-time point imaging definitely has an important teaching value. The results of our study do not really support a routine performance of supplementary late images for every PSMA PET/CT examination. However, delayed imaging is useful to confirm or rule out a suspicious abnormality seen in early images in individual cases. An important point to emphasize is that additional late images are useful for clearing up unclear lesions whose signs of malignancy would lead to a change in the therapeutic approach.
