Parameters predicting [¹⁸F]PSMA-1007 scan positivity and type and number of detected lesions in patients with biochemical recurrence of prostate cancer

Prostate cancer (PCa) is a frequently diagnosed cancer worldwide, leading to a major burden of morbidity and mortality in men [1]. Biochemical recurrence of PCa after primary treatment with curative intent occurs in 30–50% of patients [2]. If this biochemical recurrence is caused by metastatic disease, systemic therapy using androgen deprivation therapy (ADT) is still considered as a corner stone in the clinical management of these patients; however it can lead to toxicity. Recent use of hybrid imaging modalities has made it possible to differentiate between local recurrence or metastatic disease, hereby offering the possibility to determine the exact number of metastatic lesions. Indeed, treating patients presenting with low-volume PCa (oligometastatic PCa) using metastasis-directed therapy, such as surgery or stereotactic body radiotherapy, is gaining interest worldwide [3‐5]. Over the last few years, [⁶⁸Ga]Ga-PSMA-11-PET/CT has replaced choline-PET/CT imaging in patients with biochemically recurrent PCa [6] and has been included in the European Association of Urology (EAU) guidelines in this setting [7]. [⁶⁸Ga]Ga-PSMA-11-PET/CT is known for its high overall detection efficacy of 63–95%, which increases with rising serum PSA levels [8‐12]. Several studies have shown that PSA, ADT and Gleason score can predict [⁶⁸Ga]Ga-PSMA-11-PET/CT scan positivity [10, 12‐16]. More recently, fluorine-18 labelled PSMA ligands have been introduced, representing several beneficial characteristics over gallium-68: fluorine-18 has a longer half-life, can be produced in large quantities in a cyclotron and has a potentially better imaging resolution due to lower positron energy [17, 18]. High detection rates between 60 and 95% have also been described for these fluorine-18 labelled PSMA ligands, such as [¹⁸F]PSMA-1007 and [¹⁸F]DCFPyL [17, 19‐24]. Contrary to [⁶⁸Ga]Ga-PSMA-11, which is excreted via urinary system, [¹⁸F]PSMA-1007 has the additional advantage that it is primarily excreted via the hepatobiliary system, ultimately leading to better visualization of local recurrences of PCa in proximity to the bladder [20]. In order to help the clinician in selecting suitable candidates for [¹⁸F]PSMA-1007-PET/CT imaging in the setting of biochemical recurrence and avoid negative and therefore unhelpful scans, it is important to identify predictors of a positive [¹⁸F]PSMA-1007-PET/CT scan [14]. To our knowledge, these predictors have not been previously identified in a large group of patients with biochemically recurrent PCa. The aim of this study was therefore to evaluate diagnostic performance of [¹⁸F]PSMA-1007-PET/CT in patients with biochemically recurrent PCa and identify parameters that predict [¹⁸F]PSMA-1007-PET/CT scan positivity as well as the type and number of detected lesions.

Promising studies have been published about the detection of lesions in biochemically recurrent PCa with [¹⁸F]PSMA-1007-PET/CT, showing high detection rates between 60 and 95% [20‐24]. In our study, we aimed to evaluate the diagnostic performance of [¹⁸F]PSMA-1007-PET/CT as well as to identify parameters that independently predict [¹⁸F]PSMA-1007-PET/CT scan positivity as well as the type and number of detected lesions [30].

We found an overall positivity rate of 80% for [¹⁸F]PSMA-1007 PET/CT. This result is in line with the study of Giesel et al., in which a detection efficacy of 81.3% was found in 251 patients with biochemical recurrence of PCa after RP and comparable PSA levels (median PSA 1.2 ng/ml) [21]. Rahbar and coworkers showed a detection efficacy of 95% in 100 patients with biochemical relapse and a median PSA level of 1.34 ng/ml [23]. It must be noted that in their study, patients were scanned 120 min after tracer injection, compared to a mean interval of 81 ± 16 min (mean ± SD) in our study. It is known for [⁶⁸Ga]Ga-PSMA-11 as well as [¹⁸F]PSMA-1007 that a longer uptake time is linked to an increase in tumour-to-background ratio, thereby possibly increasing detection rates [24, 31‐33]. Other studies with smaller patient cohorts have reported detection rates between 60 and 75% at median PSA levels ranging from 0.6 to 1.2 ng/ml [20, 22, 24]. Furthermore, of the 137 included patients included in our study, 35 patients had a follow-up scan and 3 patients had a second follow-up scan. The largest part of these patients had progressive disease on their follow-up scan. However, in two patients we observed a decrease in disease burden between the initial and follow-up scan due to receiving radiation therapy on the bone lesions.

In our study local recurrence was seen in 23% of the scans. A systematic review and meta-analysis by von Eyben et al., reported a local recurrence of 14% on [⁶⁸Ga]Ga-PSMA-11-PET/CT in patients with biochemical recurrence of prostate cancer after primary treatment [34]. Our higher local detection rate is possibly due to the fact that [¹⁸F]PSMA-1007 is primarily excreted via the hepatobiliary system, ultimately leading to better visualization of local recurrences of PCa in proximity to the bladder [20].

Our data showed that PSA velocity and PSA value at the moment of the scan are significant predictors of [¹⁸F]PSMA-1007-PET/CT scan positivity (OR 67.6 and 1.6, respectively). Similarly, Witkowska-Patena et al. described PSA value as a significant predictor for [¹⁸F]PSMA-1007-PET/CT positivity in a cohort of 40 patients who underwent radical treatment and had low rising PSA (0.008 to ≤ 2.0 ng/ml) [22]. PSA value has also been described as a significant predictor of a positive [⁶⁸Ga]Ga-PSMA-11-PET/CT in multiple studies [10, 13‐15]. Moreover, our results show that PSA value is also a significant independent predictor of the presence and the number of bone lesions (OR 1.007 and IRR 1.003, respectively) in multivariable analysis. There are only a few studies that have investigated the influence of various parameters on the detection and number of bone metastases on [⁶⁸Ga]Ga-PSMA-11-PET/CT. The group of Heidelberg has shown that PSA value was significantly correlated with several [⁶⁸Ga]Ga-PSMA-11-associated parameters of bone metastases, such as the degree of tracer uptake [35]. Verburg et al. performed a retrospective study on [⁶⁸Ga]Ga-PSMA-11-PET/CT in 155 patients with recurrent disease and reported a similar detection rate of bone metastases (32%, compared to 33% in our study) [36]. They identified that the proportion of patients with bone metastases increased with increasing PSA values, which is in accordance with the finding in our study. Additionally, PSA doubling time was the only independent predictor of bone metastases in their study. Also, Ceci and coworkers found that PSA doubling time was an independent predictor of the presence of bone metastases [37]. We could not identify PSA doubling time as a predictor of bone lesions, possibly because in patients with low PSA levels, PSA doubling time offers very little advantage, because slight changes in low PSA levels can have a large influence on the PSA doubling time value [12, 38]. Our study did, however, identify PSA velocity, which is also a parameter of PSA kinetics, as a significant predictor of the detection and number of bone lesions in univariable analysis. In addition, our study did not reveal PSA doubling time as a significant predictor of scan positivity or type or number of lymph node or soft tissue lesions. The difference in significancy between PSA doubling time and PSA velocity in our study most likely lies in the distribution of both parameters. PSA velocity has more compact values with a few strong outliers compared to PSA doubling time, leading to a more skewed distribution. Therefore, we hypothesize that PSA velocity has a more linear correlation with the outcome parameters compared to PSA doubling time.

We must take into account that previous studies have proposed that [¹⁸F]PSMA-1007-PET/CT exhibits a higher rate of unspecific focal bone marrow uptake [39].The group of Dietlein also saw this phenomenon when performing an intra-individual comparative study of [¹⁸F]PSMA-1007-PET/CT against 3 renal excreted PSMA‐tracers ([⁶⁸Ga]Ga-PSMA-11, [¹⁸F]DCFPyL or [¹⁸F]JK‐PSMA‐7) [40]. Although the readers in our study were already aware of this drawback and other potential pitfalls of bone uptake [41], we still advice to carefully assess [¹⁸F]PSMA-1007 uptake in bone. A systematic approach when interpreting PSMA PET studies, for example PSMA-RADS or recently published E-PSMA standardized reporting guidelines, can be helpful when reporting these unspecific findings [42, 43].

Furthermore, we did not find a significant association between prior or ongoing use of ADT and [¹⁸F]PSMA-1007-PET/CT scan positivity. Conversely, our study did identify prior ADT as a significant independent predictor of the presence of bone and soft tissue lesions and the number of soft tissue lesions. In addition, ongoing ADT was identified as a significant independent predictor of the number of lymph node lesions. We did not exclude patients based on first, second or higher line ADT exposure. A previous retrospective study with [¹⁸F]PSMA-1007 in 251 patients with biochemical recurrence has shown a higher rate of detection in patients with prior ADT exposure [21]. It must be noted however that in this study, patients who had received second-line ADT or chemotherapy were excluded. Ongoing ADT has also been described as a significant predictor of a positive [⁶⁸Ga]Ga-PSMA-11-PET/CT scan in multiple previous studies [8, 10, 13, 15]. It has been shown preclinically and clinically that ADT can increase PSMA expression in prostate cancer cells and that it can increase the number of lesions visualized by PSMA-PET [44, 45]. But one should be cautious about how to interpret these results, as part of the effect may also be caused by the fact that patients who underwent ADT are more likely to have more advanced PCa than those who did not. Furthermore, multiple in vitro and vivo studies investigated the temporal relationship between the initiation of ADT and PSMA uptake. In our study, the initiation of ADT varied from 3 to 220 months prior to [18F]PSMA-1007 PET/CT scan, which can be considered long-term. Predominantly, it is proposed that short-term ADT use increases PSMA uptake and long-term ADT decreases it [46, 47].

Moreover, the Gleason score has been identified as a predictor of a [⁶⁸Ga]Ga-PSMA-11 scan positivity in several studies [12, 16]. However in our study, Gleason score was only a significant independent predictor of the number of bone lesions on [¹⁸F]PSMA-1007-PET/CT and not of the overall scan positivity. To the best of our knowledge, this finding has not been described in other studies with [¹⁸F]PSMA-1007. Further studies with larger patient cohorts are needed to validate our findings.

We investigated the effect of different types of primary treatment on the presence of local recurrence on [¹⁸F]PSMA-1007 PET/CT. Our study showed that ADT, EBRT and brachytherapy compared to RP were significant predictors of local recurrence on [¹⁸F]PSMA-1007-PET/CT scan. To our knowledge, primary treatment as a significant predictor of local recurrence on [¹⁸F]PSMA-1007 or [⁶⁸Ga]Ga-PSMA-11-PET/CT has not been described previously. Our results on primary treatment as predictive parameter must be interpreted cautiously, as a part of the patients who received RP as primary treatment received adjuvant and/or salvage EBRT afterwards, making adjuvant and/or salvage EBRT a potential confounder in these findings. In accordance to our results, a study with recurrent disease on [⁶⁸Ga]Ga-PSMA-11-PET/CT did not find a significant correlation of the parameters PSA value, PSA doubling time and Gleason score to local recurrence [36].

To the best of our knowledge, this is the first study that investigated the influence of different parameters on type and number of lesions detected on [¹⁸F]PSMA-1007-PET/CT scan. It is useful to know whether a patient is likely to have oligo- or polymetastatic disease, because this can lead to different treatment strategies.

Our study has some limitations. Firstly, this is a retrospective, monocentric study using a patient cohort that is heterogeneous considering the primary treatment that was given to patients, in that way representing a typical patient cohort that would be evaluated during clinical routine practice. Secondly, the PSMA-expression of the primary PCa was not known. As approximately 10% of prostate cancers are PSMA negative [48], false negative scans could have confounded our sample. Thirdly, our study merely describes the presence of positive lesions, without histological validation of all lesions or clinical follow-up. In the 11 patients in whom histological validation was available, all lesions detected on [¹⁸F]PSMA-1007-PET/CT were proven to be recurrent PCa. The effect of [¹⁸F]PSMA-1007-PET/CT on patient management is very difficult to obtain given the retrospective nature of our study and the fact that potential other imaging modalities were used to define the patient management. Nevertheless, various studies have proven that [⁶⁸Ga]Ga-PSMA-11 PET/CT can lead to a change of treatment in a large group of patients with biochemically recurring prostate cancer [49, 50]. Further prospective studies are needed to further investigate the role of [¹⁸F]PSMA-1007-PET/CT on patient management. Fourthly, a standardized reporting classification system for PSMA PET was not used and results of possible additional imaging ordered by clinicians after [¹⁸F]PSMA-1007-PET/CT in doubtful cases were not included in our analysis.

[¹⁸F]PSMA-1007-PET/CT was positive in 80% of patients with biochemical recurrence of prostate cancer after primary treatment. PSA value and PSA velocity were significant predictors of a positive [¹⁸F]PSMA-1007-PET/CT scan as well as of the presence and number of bone and soft tissue lesions and of the number of lymph node lesions in uni- and/or multivariable analysis. Further prospective studies with larger patient cohorts are needed to validate our findings and eventually lead to the development of a prediction nomogram that can help in optimal patient selection for [¹⁸F]PSMA-1007-PET/CT.