99mTc-labelled PSMA ligand for radio-guided surgery in nodal metastatic prostate cancer: proof of principle

Purpose Intraoperative identification of prostate cancer (PCa) lymph node (LN) metastases (LNM) detected by preoperative PSMA PET/CT may be facilitated by PSMA radio-guided surgery (RGS) with use of a γ-probe. Earlier we demonstrated excellent performance of the 111In-labelled PSMA ligand DKFZ-617 ([111In]In-PSMA-617) in RGS for ex situ distinction of LN vs LNM at lymphadenectomy (LA) at a single LN level. In comparison with indium-111, technetium-99m has better physical properties for γ-probe measurements, better availability and lower radiation exposure for patients and medical personnel. Against this background, we evaluated the uptake of 99mTc-PSMA-I&S ligand at the level of single LN and its power to discriminate between unaffected LN and LNM. Methods Six patients with PCa with the suspicion of LNM on preoperative PSMA-PET/CT underwent [99mTc]Tc-PSMA-I&S RGS (4 salvage LA, 2 primary LA) with intravenous injection of [99mTc]Tc-PSMA-I&S 24 h prior to surgery. Resected samples were isolated manually aiming at the level of single LN. Uptake measurements were done ex situ with a high-purity germanium detector. Receiver operating characteristic (ROC) analysis was performed based on [99mTc]Tc-PSMA-I&S uptake expressed as lean body mass standard uptake value (SUL). Results Separation of the tissue samples from 73 subregions resulted in 498 single samples. After final histopathology 356 LN, 160 LNM und 11 non-nodal PCa samples were identified. Median SUL of tumor-free samples (0.26) and samples with cancer (3.5) was significantly different (p < 0.0001). ROC analysis revealed an area under the curve (AUC) of 0.917 (95% CI 0.89–0.95). Using a SUL cutoff of 1.1, sensitivity, specificity, positive predictive value, and negative predictive values were 76.6%, 94.4%, 89.4% and 86.9%. Conclusion Ex situ analysis of [99mTc]Tc-PSMA-I&S uptake at single LN level showed good diagnostic performance for the ex situ distinction of tumor-bearing vs tumor-free LN during RGS.


Introduction
Prostate cancer (PCa) is the most commonly diagnosed cancer in men [1,2]. Pelvic and retroperitoneal lymph nodes (LN) are the first site for metastases [2,3]. Although the impact of LA at radical prostatectomy has not been yet fully clarified, it remains the gold standard of staging PCa and an improvement of the oncological outcome could be suggested [4]. Regretfully, about 15-30% of the patients will develop a biochemical recurrence with elevated PSA level and clinical recurrence (metastases) possible at different anatomical sites [5,6]. Regardless if LA is done at primary therapy or in case of recurrence, an accurate Open Access *Correspondence: michael.mix@uniklinik-freiburg.de 1 Department of Nuclear Medicine, Medical Center -University of Freiburg, Faculty of Medicine, University of Freiburg, Freiburg, Germany Full list of author information is available at the end of the article preoperative identification of LNM by, for instance, positron emission tomography/computed tomography (PET/CT) is a prerequisite for successful surgery [7,8]. PET/CT targeting prostate-specific membrane antigen (PSMA) using 68 Ga-labelled PSMA ligands has demonstrated an excellent ability to detect LNM prior to surgery and is widely used as a tool for staging before primary therapy and for restaging in the setting of biochemical relapse [7,[9][10][11]. Due to several advantages of the positron emitter fluorine-18, more and more frequently 18 [12][13][14][15].
If PET/CT indicates "regional pelvic LNM" as the only finding at clinical recurrence, surgical removal (i.e., "salvage lymphadenectomy", salvage LA) [16,17] of the lymphatic tissue or targeted radiotherapy may be suggested in patients in good general condition [7,18,19], but should be considered to be experimental and individual therapeutic approaches [1,2]. Locating suspected LNM during surgery (e.g., salvage LA) is often very challenging in the case of small LNM and reduced accessibility to the LNM (e.g., because of atypical location of LNM and tissue adhesions). In order to address this issue, radio-guided surgery (RGS) using γ-emitting tracers had been introduced [20][21][22][23][24]. Suspected LNM could be identified during surgery using a γ-probe with acoustic feedback. Accordingly, the surgeon is able to conduct in situ and ex situ measurement of suspected regions and from resected tissue samples [25]. For [ 111 In]In-PSMA-I&T RGS, the use of the γ-probe during salvage LA provided a specimenbased sensitivity of 83.6%, a specificity of 100%, and an accuracy of 93% [26]. Because of several well-known advantages (e.g., costs, availability, physical properties for the γ-probe detection, radiation exposure) PSMA tracers labelled with technetium-99m are currently preferred over those labelled with indium-111 [22,27]. However, RGS is applied differently at the institutions [21,28] and there are no data available about the exact uptake of [ 99m Tc]Tc-PSMA-I&S at a "single LN level" (manual separation of the resected tissue samples into single LN and LNM). By performing a precise ex-situ analysis with a high-purity germanium detector, we investigated the absolute tracer uptake and the performance of [ 99m Tc]Tc-PSMA-I&S in patients who underwent RGS and in whom we did a meticulous manual separation of the LN.

Patients
From 12/2017 to 03/2019, 38 patients with prostate cancer and the suspicion of exclusive LNM (without detectable bone or visceral metastases) on PSMA PET/ CT underwent a LA after application of [ 99m Tc] Tc-PSMA-I&S prior to surgery. In a subset of 6/38 patients a single LN preparation and tracer uptake measurement of the samples was performed, being not a part of the clinical routine but addressing the issue of the current analysis. Two of these six patients underwent extended LA at radical prostatectomy for primary PCa. Four of six patients underwent a salvage LND on a compassionate-use basis because of biochemical recurrence (PSA > 0.2 ng/ml after radical prostatectomy) ( Table 1). The LA in the six patients, regardless if at primary therapy or at the stage of biochemical recurrence was conducted to achieve a maximal tumor reduction. The large number of LN samples was also suitable for a dedicated analysis of tracer uptake, presented here. The sample processing (manual separation after surgery followed by direct measurement of the samples in a high-purity germanium detector) was planned and conducted in a prospective intention. The local ethics committee approved this data analysis (No. 562/15). Informed consent was obtained from each subject. For both tracers we used clinically established imaging protocols and uptake times as described by Jilg et al. [7] and Giesel et al. [15]. [ 68 Ga]Ga-PSMA-PET/CT (n = 4) imaging was done 62 ± 2 min after injection of mean 220 ± 35 MBq and [ 18 F]F-PSMA-PET/CT (n = 2) 124 ± 5 min after injection of mean 321 ± 26 MBq. A PSMA-positive lesion was defined as focal tracer accumulation greater than physiological local background activity. PET/CT study evaluation was done with co-registered PET and CT datasets using predefined PET window settings (inverted grey scale, SUV range: 0-5 g/ml for 68 Ga-PSMA and 0-10 g/ ml for 18 F-PSMA). All patients showed increased focal 68 Ga-or 18 F-PSMA uptake in at least one pelvic and/or retroperitoneal region.

Lymphadenectomy
Open lymphadenectomy was performed as described by Jilg et al. [7]. The extent of LA was determined first by the aim to adhere to a template LA and according to the presence of PET-positive lesions: in the case of a pelvic PET-positive lesion(s), a bilateral pelvic LA was intended whenever possible. In the case of an additional PET-positive lesion in the retroperitoneum, a retroperitoneal LA was conducted. Subregions for a template pelvic bilateral LA were: common iliac vessels, external iliac vessels, obturator vessels, internal iliac vessels (presacral region). Subregions for a template retroperitoneal LA were: aortic bifurcation, aortal and caval region. Whenever permitted by the intraoperative circumstances (deviation from the template, e.g., caused by surgical difficulties) we adhered to this template. Nodal fibrofatty tissue (FFT) from each subregion was collected separately at surgery.
During LAs, γ-probe measurements were performed in situ and ex situ. γ-counts from ex situ measurements provided an assessment if the removed tissue specimen was presumably tumor-bearing or tumor-free. The number of counts from the γ-probe used intraoperatively (ex-situ measurement) gave the surgeon feedback if the suspected tumor tissue was resected or if the LA in this subregion had to be continued which ultimately increased the intensity of LA in this subregion. In cases in which ex situ γ-counts were very low (which indicates that tissue resected was free of tumor) and preoperative PET/CT predicts an LNM in the respective anatomical region, the surgeon extended the search for LNM. Results from γ-probe measurements during surgery were not evaluated in the current analysis because these measurements were done on unprocessed specimens, consisting of nodal FFT, LN and LNM.

Sample processing
Following surgery, the specimens from all subregions were manually separated into single samples under guidance of an experienced pathologist (Fig. 1). Each single sample was weighed and numbered for further analysis. In 60/498 processed single samples, more than one LN and/or LNM was present after final histopathology. These additional LNs and LNMs were not identifiable during manually macroscopic separation. Samples with more than one LN or LNM were excluded in the further analysis.

Analysis of trace uptake
Analysis of tracer uptake was performed analog as described by Mix   Tc-PSMA-I&S radioactivity measurement of each single sample was done with a high-purity germanium detector (Canberra Inc., model GX2018-CP5+), calibrated with a multi-isotope reference source (type VZ-2139/ NG3 from Eckert&Ziegler Nuclitec's, DKD-accredited measurement laboratory in Germany) and cross calibrated for tissue sample geometry. Tracer uptake was calculated as percent injected dose per gram (%ID/g) and as SUL following PET standardized uptake value using lean body mass instead of body weight: with t as delay between patient injection and time of surgery, T 1/2 as half-life of technetium-99m. Lean body mass was calculated according to Janmahasatain et al. [30].

Histopathological analysis
All resected LNs (i.e., the entire LN in the case of LNs ≤ 4 mm, one central slice in the case of LNs > 4 mm) were formalin-fixed and paraffin-embedded. The pathologist was not aware of the PET findings, the clinical estimate of the tissue from the surgeon or any of the tracer uptake measurements. Histopathologic evaluation was performed by one pathologist on hematoxylin and eosin (H&E) stained tissue slides.

Statistics
Descriptive statistics were obtained by calculating mean, standard deviation (SD), median and range. Continuous variables were compared with a two-sided unpaired Mann-Whitney U test. A receiver operating characteristic (ROC) analysis was performed in order to analyse the diagnostic performance of [ 99m Tc]Tc-PSMA-I&S tracer uptake for identification of LNM. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were analyzed at the most appropriate cutoff value established as the one with the highest result of the sum of sensitivity and specificity (Youden index). The 11 tumor samples with non-nodal PCa were handled as LNM in statistical analysis. Prism8 GraphPad was used for statistical calculations.

Lymphadenectomy
Clinical patient characteristics such as age and Gleasonscore and the outcomes from LAs of the 6 patients are summarized in Table 1. The 6 patients were part of a larger cohort that underwent RGS because of suspected LNM in a PSMA-PET/CT and had a high-risk PCa stage . Time from PSMA PET/CT to LA was 2 ± 1.5 months. Median PSA at surgery was 33.3 ng/ml (range 0.99-127 ng/ml) ( Table 1). Figure 1 shows the workflow of the sample processing. Even though only 6 patients underwent surgery a relatively large number of 516 LN from 73 subregions had been removed (mean 86 ± 39 per patient). Median 11.5 (mean 26.7 ± 31.9) LNM were identified per patient. From each subregion, median 7 (mean 6.6 ± 1.7) single samples were identified. The majority of the samples and the achieved single samples originated from pelvic regions (Table 1).

Tracer uptake
Data on weight, tracer uptake (SUL) and percent of injected dose per gram (%ID/g) from single samples removed at [ 99m Tc]Tc-PSMA-I&S RGS are shown in Table 2. Tumor-free samples and tumor-bearing samples showed clearly significantly different values (p < 0.0001 each) for all 3 parameters (weight, tracer uptake, SUL). There was no correlation between tissue weight and tracer uptake or SUL in the tumor-free and tumor-bearing samples. Tracer uptake of the lightest LNM and LN (both 0.01 g) was 1.8 × 10 -4 and 1.0 × 10 -6 %ID/g while the heaviest LNM (6.0 g) and LN (7.8 g) showed a tracer uptake of 6.9 × 10 -5 and 1.5 × 10 -6 %ID/g. The weight of the LNM and LN with the lowest uptake (1.9 × 10 -8 and 1.1 × 10 -9 %ID/g) was 0.05 and 1.3 g while the weight of the LNM and LN with the highest uptake (2.9 × 10 -3 and 8.8  Fig. 3.

Discussion
Conditions for optimal RGS (tracers, protocols) are still in development and under investigation [27]. Available reports on the application of 99m Tc-labelled PSMA ligands for detection of LNM during RGS were based on the analysis of "sample-mixtures" from one anatomical region consisting out of LN, LNM and FFT [22,27]. Consequently, tracer activity uptake measured in such samples represents the sum of tracer distribution in different kind of tissues with different uptake characteristics (LN, LNM, FFT). Just as well, results from intraoperative gamma probe measurements (regardless if in situ or ex situ) might be blurred due to, e.g., multiple surrounding healthy tissue or unspecific tracer uptake in, e.g., healthy lymph nodes or other tissue. Accordingly, the exact performance of the 99m Tc-labelled PSMA ligand and its diagnostic accuracy (performance) on a single LN basis could not be evaluated.
To our best knowledge, the present analysis is the first clinical investigation of tracer uptake for [ 99m Tc] Tc-PSMA-I&S RGS in tumor tissue at the level of single lymph nodes. The large number of LN and LNM samples removed at surgery (406 in 6 patients) indicates that larger tumor formations were not missed. It suggests a comprehensive data record to evaluate the performance of the tracer for the distinction between  tumor-bearing and tumor-free LN. SUL was significantly different between samples with PCa and nonaffected samples ( Table 2 and Fig. 2). To reach the mean tracer uptake in tumor containing LN it would take more than 30-fold of the amount of mean tracer uptake in tumor-free LN (  [22] showed for 99m Tc labelled PSMA RGS in 31 patients undergoing salvage LND a specimen-based sensitivity of 83.6%, a specificity of 100%, a positive predictive value of 100% and a negative predictive value of 89.2%. These results are apparently better than the data in this study but they are hardly to compare because of the different surgical approaches. Based on PET/CT findings, Maurer et al. performed primarily a unilateral targeted RGS-lymphadenectomy with subsequent intraoperative γ-probe measurements to exclude additional lesions. Our approach was to adhere whenever possible to a bilateral template lymphadenectomy, both with subsequent intraoperative γ-probe measurements for verification. The latter approach might be less vulnerable to miss false negative LNM or very small LNM missed by PET/CT or, e.g., by a weak γ-probe signal due to a low [ 99m Tc]Tc-PSMA-I&S accumulation. Furthermore, the object of this study was to analyse a high number of single lymph nodes (prepared out of 498 single samples from 6 patients in this study) but not mixed tissue samples (136 specimens from 31 patients by Maurer et al. [22]). Additional optimization of the well-known widespread pencil γ-probes for an intraoperative situs or for robotic surgery might improve the approach. To address this issue Oosterom et al. developed a first DROP-IN γ-probe with a scanning direction range between 0° and 180° [31]. It can be assumed that the trend towards assisted respective guided surgery such as RGS will potentially improve the results of salvage LA.

Limitations
The main limitation of this study is the small number of patients and the known inhomogeneity of PCa (e.g., Gleason score, PSA at surgery) which makes a comparison to the results with other traces like 111 In-labeled PSMA ligands [23] difficult. Even if there is a high number of single LN and LNM samples (n = 406), in terms of tumor progression (number of LNM), the six patients are also very inhomogeneous. For example, patient no. 6 revealed a dramatically high number of LNM overall (n = 85) and showed a high PSA value, even after surgery (Additional file 3: table 1).
Generally, there may be an additional bias because only patients with suspected LNM on a PSMA-PET/CT and therefore known PSMA-positive lesions were included in this study.
Finally, it remains to be examined how the application can be optimized for ex-situ and in-situ measurements of sample mixtures with γ-probes and how dual labelled PSMA-targeting agents for radio-and fluorescence guides surgery can improve this approach [27].

Conclusion
Ex situ analysis of [ 99m Tc]Tc-PSMA-I&S uptake at single LN level showed good performance for ex situ distinction of tumor-bearing vs tumor-free LN during RGS.