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

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 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 ⁶⁸Ga-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, 9‐11]. Due to several advantages of the positron emitter fluorine-18, more and more frequently ¹⁸F-labelled PSMA ligands ([¹⁸F]F-PSMA-1007, [¹⁸F]F-DCFPyL) are used for PET/CT imaging [12‐15].

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 [20‐24]. 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 [¹¹¹In]In-PSMA-I&T RGS, the use of the γ-probe during salvage LA provided a specimen-based 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 [^99mTc]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 [^99mTc]Tc-PSMA-I&S in patients who underwent RGS and in whom we did a meticulous manual separation of the LN.

Conditions for optimal RGS (tracers, protocols) are still in development and under investigation [27]. Available reports on the application of ^99mTc-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 ^99mTc-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 [^99mTc]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 non-affected 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 (Table 2). This contrast is 2.5-fold lower than we found for [¹¹¹In]In-PSMA-617 [23]. The reduced contrast for ^99mTc]Tc-PSMA-I&S is caused by a 34% higher average tracer uptake in unaffected LN and a 66% lower average tracer uptake in LNM. A reason for this could be the different treatment protocol. While RGS was done with a sufficient count rate for ¹¹¹In-labelled PSMA ligands (2.8 days half-life) 48 h after injection of the tracer, RGS with ^99mTc-labeled PSMA ligands (6 h half-life) must be done earlier, in our study 24 h after tracer injection. Nevertheless, both tracers demonstrate the power to distinct between unaffected LN and LNM and therefore for the use for RGS. Another finding was a relatively high variation of the tracer uptake in the samples and that the range of uptake between LN and LNM overlap. Using the Youden Index as SUL cutoff, more false-positive (23.4% vs 7.9%) and false-negative (10.6% vs 5.4%) single samples were observed with technetium-99m than with the indium-111. All results based on ex situ measurements under optimal conditions like a high sensitivity germanium detector and separated single lymph nodes without any disturbing background signal from other tissue. Measurements of mixed tissue samples (LNM, LN lymph nodes with fibro fatty tissue) with a γ-probe during surgery, especially in situ, may be still challenging due to additional unavoidable physiological and unspecific background signal.

An important study from Maurer et al. [22] showed for ^99mTc 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 [^99mTc]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.

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 ¹¹¹In-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].

Ex situ analysis of [^99mTc]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.