⁶⁸ Ga-PSMA-PET/CT for the evaluation of pulmonary metastases and opacities in patients with prostate cancer

For this retrospective study, we obtained approval from our institutional ethics review board. We extracted 739 consecutive patients with confirmed prostate cancer from our local database who underwent at least one ⁶⁸Ga-PSMA-PET/CT between September 2013 and April 2017. By manual review of all reports and scans, we identified twenty patients with lung metastases and fourteen patients with pulmonary opacities, according to the criteria described below. Prostate cancer was histologically proven in all patients. Only patients with no other known type of cancer but PC were included. All available additional information from clinical records was considered.

Using a conventional ⁶⁸Ge/⁶⁸Ga radionuclide generator (Eckert & Ziegler Radiopharma GmbH, Berlin, Germany), ⁶⁸Ga was eluted and then compounded with PSMA-HBED-CC (ABX GmbH, Radeberg, Germany) according to the method described previously [23, 24].

PET/CT imaging was performed 67.0±33.1 min after intravenous injection of 125.9±26.9 MBq of ⁶⁸Ga-PSMA-HBED-CC. PET scans were acquired using a Gemini Astonish TF 16 PET/CT scanner (Phillips Medical Systems) in 3D acquisition mode [25]. Axial, sagittal and coronal slices were reconstructed (144 voxels with 4mm³, isotropic). Prior to each PET scan, a CT was performed for anatomical mapping and attenuation correction (30 mAs, 120 kVp). Each bed position was acquired for 1.5 min with a 50% overlap. In nineteen patients, contrast-enhanced CT (CE-CT, 162 - 215 mAs, 120 kVp, slide thickness 3.0 mm) was performed using 70-120 ml of contrast agent (Ultravist® 370, Bayer Schering Pharma, Berlin, Germany), which was injected intravenously with a delay of 70 seconds for the venous phase. In fifteen patients, only unenhanced CT was available.

Two experienced observers analyzed the PET/CT scans using Visage 7.1 (Visage Imaging GmbH, Berlin, Germany). All scans were reviewed for suggestive pulmonary lesions, and lesions were classified to either “metastases” or “opacities”. All the following criteria had to be fulfilled for the diagnosis of lung metastasis: (I) CT imaging with a singular or multiple masses. (II) New appearance or change of size of lesions compared to previous studies. (III) Lesions need to be round or oval. (IV) No signs of benignity such as fat or calcification. (V) Synchronous other distant metastases. For the diagnosis of pulmonary opacity, the following criteria had to be fulfilled: (I) CT-imaging of an irregular-shaped or confluent focal opacity. (II) Lesion must not be nodular or a solid mass. (III) Lesion must not be classified as metastasis. Only intrapulmonary lesions >5mm were considered. Patients with a history of or signs for a malignancy other than PC were excluded. Overall, 20 patients with lung metastases and fourteen patients with pulmonary opacities were identified. Up to ten lesions per patient were analyzed. In case a patient was imaged more than once, only the most recent ⁶⁸Ga-PSMA-PET scan was included in this study. As a result, ninety-one lung metastases and fourteen pulmonary opacities were analyzed. For all lesions, location and morphology were described. The sizes of metastases were measured based on the CT scan.

Standardized uptake values (SUV) were normalized for body weight by the software using the equation SUV = C _tis /Q _inj /BW, where C _tis is the lesion activity concentration in MBq per milliliter, Q _inj is the activity injected in MBq, and BW is the bodyweight in kilograms. For PET data quantification, a two-dimensional region of interest (2D ROI) and a three-dimensional region of interest (3D ROI) were defined. ⁶⁸Ga-PSMA-HBED-CC uptake of all lesions was quantified using maximum standardized uptake values (SUV_max). To differentiate PSMA-positive from PSMA-negative lesions, the SUV_max of the blood pool measured within the descending thoracic aorta was set as a reference; lesions with SUV_max-values 20% or more above blood pool were considered PSMA-positive. All values were recorded in the transaxial, attenuation-corrected PET-slice representing the greatest extent of the respective lesion. Regions of interest were defined avoiding the periphery of lesions to minimize partial volume effects. The readers were blinded to the results of other diagnostic procedures and the clinical history of the patients.

A reference standard was created by presenting each case to an adjudication panel consisting of experts in the fields of nuclear medicine, radiology and urology. All available data (clinical records and follow-up data, radionuclide imaging, radiographs, CT, MRI, histology, and intraoperative findings) were taken into consideration for the standard of reference and a final diagnosis for every lesion was documented.

The descriptive statistics are reported as mean, median and/or range when applicable. The Mann-Whitney U test was used for the comparison of SUV_max values of pulmonary opacities and lung metastases. SUV_max values in 2D and 3D ROI were compared using the Wilcoxon signed-rank test. To determine the relationship between SUV_max and size of metastases, a Spearman’s rank correlation was used. The significance level was set to α < 0.05. Statistical analyses were conducted with SPSS 23 for Mac (IBM Corp, Armonk, NY).

In total, 91 lung metastases were detected in 20 of 739 (2.7%) patients and 14 pulmonary opacities were detected in 14 of 739 (1.9%) patients. Patients’ characteristics are summarized in Table 1. Mean patients’ age was 70.6±8.1 years. Median GS was 9 (range 6 – 10). Mean PSA level was 123.6±300.2 ng/ml.

The lesions’ characteristics such as location and morphology are summarized in Table 2, all detailed results in Table 3. The mean size of metastases was 11.0±6.3cm² (range 0.2 – 29.5cm²). The mean SUV_max of all lung metastases was 3.5±2.8 in 2D and 3.7±3.0 in 3D ROI. In total, 66 PSMA-positive (72.5%) and 25 PSMA-negative (27.5%) metastases were identified. No significant difference regarding the size of metastases was measured between both groups. Examples of pulmonary opacities are illustrated in Fig. 1, examples of PSMA-positive and PSMA-negative metastases in Figs. 2 and 3. The mean SUV_max of PSMA-positive metastases was 4.5±2.7 in 2D and 4.7±2.9 in 3D ROI. The mean SUV_max of PSMA-negative metastases was 1.0±0.5 in 2D and 1.0±0.4 in 3D ROI. In pulmonary opacities, the mean SUV_max was 2.2±0.7 in 2D ROI and 2.4±0.8 in 3D ROI. Overall, PSMA-positive lung metastases demonstrated the highest tracer uptake, significantly higher than pulmonary opacities (p<0.001). PSMA-negative metastases demonstrated the lowest tracer uptake, significantly lower than pulmonary opacities (p<0.001). All pulmonary metastases taken together, there was no difference in mean SUV_max between pulmonary metastases and opacities (p>0.05). The mean SUV_max of PSMA-positive, PSMA-negative and all metastases compared to pulmonary opacities and the aorta is illustrated in Figs. 4 and 5.

The mean SUV_max obtained by 3D ROI was significantly higher than that obtained by 2D ROI in pulmonary opacities (p<0.05) as well as in PSMA-positive lung metastases (p<0.001). There was no difference in SUV_max of PSMA-negative metastases between 2D and 3D ROI (p>0.05).

In all pulmonary metastases taken together, HU_mean was -129.8±133.7 in CE-CT and -211.9±164.0 in unenhanced CT. In PSMA-negative metastases, HU_mean was -127.4±74.6 in CE-CT and -237.2±138.9 in unenhanced CT; in PSMA-positive metastases, HU_mean was -130.4±145.7 in CE-CT and -199.3±175.8 in unenhanced CT.

We calculated a weak, significant positive linear relationship between size and 3D SUV_max of metastases (Fig. 6, ρ_Spearman=0.207, 95% CI [0.002; 0.396], p<0.05).

Of 20 patients with lung metastases, three patients (15%) had ten or more metastases, twelve patients (60%) had two to ten metastases, and five patients (25%) had a single metastasis. Regarding the tracer uptake, twelve patients (60%) had PSMA-positive lung metastases only, seven patients (35%) had mixed metastases, and one patient (5%) had PSMA-negative metastases only.

Our study demonstrates that, although a majority of PC lung metastases were PSMA-positive, a considerable share of metastases was PSMA-negative and could therefore not be detected directly by ⁶⁸Ga-PSMA-PET. Up to now, the largest study investigating the ⁶⁸Ga-PSMA-PET imaging of PC pulmonary metastases and primary lung cancer was conducted by Pyka et al., with 45 patients and 89 lesions [20]. Pyka et al. demonstrated that, due to the high tracer uptake in lung cancer, differentiation between primary lung cancer and lung metastases by SUV analysis was not possible [20]. Within their cohort, mean SUV_max of lung metastases was 4.4±3.9, which is consistent with the SUV_max calculated in our cohort. Although Pyka et al. did not explicitly differentiate between PSMA-positive and PSMA-negative metastases, they observed a great heterogeneity of tracer uptake in pulmonary metastases; many metastases showed only a faint tracer uptake [20]. This finding is also consistent with our study, in which 27.5% of metastases were PSMA-negative. A possible explanation for the difference of ⁶⁸Ga-PSMA-HBED-CC uptake in lung metastases is the diversity of phenotypes in metastases; especially neuroendocrine trans-differentiation has been identified as a main factor for the loss of PSMA-expression in visceral metastases [26]. This can be explained by the fact that a large part of neuroendocrine prostate cancer cells does not express generic PC biomarkers such as P501S, PSMA, and PSA [27]. Neuroendocrine trans-differentiation was proven histologically in the single patient of our study who had only PSMA-negative pulmonary metastases. A case report by Shetty et al. of a non-PSMA-avid PC lung metastasis suggests that an uncommon variant of the primary PC, in this case ductal adenocarcinoma, can be another cause for missing PSMA-expression [28]. However, the detection rate of the more common lymph node and bone metastases in PSMA-PET appear to be much higher. Schwenck et al. reported detection rates of 94% for lymph node and 98% for bone metastases [29]. Regarding the overall incidence of PSMA-negativity in prostate cancer, Mannweiler et al. found 5% of primary prostate cancer and 15% of prostate cancer metastases to be PSMA-negative in immunohistochemistry [30].

Compared to most other tissues, background tracer uptake of the lungs in ⁶⁸Ga-PSMA-PET is low [31, 32]. Therefore, PSMA-positive lung metastases are clearly visible and detectable as a result of a high lesion-to-background contrast. The low PSMA-expression in the normal lung parenchyma has been confirmed immunohisto-chemically, since bronchioles and terminal bronchioles of normal lung did not stain [33].

Regarding the prevalence in our cohort, lung metastases were present in 2.7% of patients who underwent ⁶⁸Ga-PSMA-PET. This is consistent with the prevalence in imaging of 3.6% reported by Fabozzi et al. [18].

Pulmonary opacities in our study demonstrated a moderate PSMA-expression, not significantly different from pulmonary metastases. There have been some case reports of elevated radiotracer uptake in pulmonary opacities in ⁶⁸Ga-PSMA-PET [22, 31].

It is thought that tracer uptake in pulmonary opacities might not be associated with an increased avidity of the lesion, but due to increased capillary penetration caused by inflammation, which results in a higher tracer activity in the interstitial space [22]. However, another explanation could be the PSMA-expression in the neovasculature of physiologic regenerative and reparative conditions as shown by Gordon et al. [34]. This question could be a subject of investigation for future studies.

A significant positive linear correlation between size and 3D SUV_max of PSMA-positive lung metastases was observed. This is likely due to the fact that PSMA-positive metastases tended to be larger than PSMA-negative metastases.

This retrospective study is limited by the fact that diagnoses of pulmonary lesions were not confirmed histopathologically since no biopsies of most of the lesions were performed. Although, to our best knowledge, no patients with secondary malignancy were included, we cannot rule out in each case that solitary lesions considered as metastases were in fact primary pulmonary neoplasms. The study included only lesions > 5 mm in diameter. PSMA-avidity in smaller lesions close to the spatial resolution of the scanner (4.7 mm) could be underestimated due to partial volume effect. However, there was no significant difference between the mean size of PSMA-negative and PSMA-positive lung metastases. Therefore, the variant PSMA-avidity cannot be explained by partial volume effect only.