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

Background The purpose of this study was to investigate the imaging properties of pulmonary metastases and benign opacities in 68Ga-PSMA positron emission tomography (PET) in patients with prostate cancer (PC). Methods 68Ga-PSMA-PET/CT scans of 739 PC patients available in our database were evaluated retrospectively for lung metastases and non-solid focal pulmonary opacities. Maximum standardized uptake values (SUVmax) were assessed by two- and three-dimensional regions of interest (2D/3D ROI). Additionally CT features of the lesions, such as location, morphology and size were identified. Results Ninety-one pulmonary metastases and fourteen opacities were identified in 34 PC patients. In total, 66 PSMA-positive (72.5%) and 25 PSMA-negative (27.5%) metastases were identified. The mean SUVmax of pulmonary opacities was 2.2±0.7 in 2D ROI and 2.4±0.8 in 3D ROI. The mean SUVmax of PSMA-positive pulmonary metastases was 4.5±2.7 in 2D ROI and in 4.7±2.9 in 3D ROI; this was significantly higher than the SUVmax of pulmonary opacities in both 2D and 3D ROI (p<0.001). The mean SUVmax of PSMA-negative metastases was 1.0±0.5 in 2D ROI and 1.0±0.4 in 3D ROI, and significantly lower than that of the pulmonary opacities (p<0.001). A significant (p<0.05) weak linear correlation between size and 3D SUVmax in lung metastases (ρSpearman=0.207) was found. Conclusion Based on the SUVmax in 68Ga-PSMA-PET alone, it was not possible to differentiate between pulmonary metastases and pulmonary opacities. The majority of lung metastases highly overexpressed PSMA, while a relevant number of metastases were PSMA-negative. Pulmonary opacities demonstrated a moderate tracer uptake, significantly lower than PSMA-positive lung metastases, yet significantly higher than PSMA-negative metastases.


Background
Worldwide, prostate cancer (PC) is considered the second most frequently diagnosed cancer in men and the fifth leading cause of cancer death [1]. Radiolabeled prostate-specific membrane antigen (PSMA) ligands such as 68 Ga-PSMA-HBED-CC have been introduced recently as promising radiotracers for the PET imaging of PC [2]. PSMA, or glutamate carboxypeptidase II, is a transmembrane protein expressed in the prostate as well as in many other tissues, and is significantly overexpressed in most prostate cancer cells [3]. Different studies have demonstrated the superiority of 68 Ga-PSMA-PET imaging regarding the detection of metastases in PC, both compared to current standard imaging (CT, MRI and bone scintigraphy) and other PET tracers, such as 18 F-Choline [4][5][6][7]. Especially at low serum PSA levels in biochemically recurrent prostate cancer, it improves the detection of metastatic lesions [8]. However, PSMA overexpression is not limited to prostate cancer; it is typically found in other malignant tumors, such as lung, colorectal, gastric, renal and thyroid cancer, particularly within the tumor neovasculature [9][10][11][12]. Furthermore, case studies have demonstrated elevated PSMA-expression in benign lesions such as sarcoidosis, Paget disease, meningioma and adrenal adenoma [13][14][15][16].
According to autopsy studies, pulmonary metastases are considered to be the second most common extranodal metastases in PC (46%), after bone metastases (90%) [17].
The clinical incidence of PC pulmonary metastases in a large retrospective review study has been 3.6% [18]. Although the data suggest that the presence of pulmonary metastases has no ominous impact on clinical course and disease outcome [19], early and reliable detection of lung metastases can be of high clinical importance for accurate staging and therapy planning.
Pyka et al. investigated the imaging and differentiation of pulmonary PC metastases, primary lung cancer and tuberculosis in 45 patients using 68 Ga-PSMA-PET. Within their study population, SUV analysis was not able to differentiate pulmonary metastases from lung cancer [20]. Other studies have confirmed PSMA-overexpression in primary lung cancer [9,21]. However, case reports have demonstrated PSMA-overexpression also in benign lung lesions such as pulmonary opacities and bronchiectasis [22], sarcoidosis [13] and tuberculosis [20].
Therefore, the aim of this study was to investigate the 68 Ga-PSMA-PET imaging properties of lung metastases and opacities in PC patients, and whether quantitative SUV analysis is able to differentiate benign from malignant lesions.

Study population
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 68 Ga-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.

Imaging protocol
PET/CT imaging was performed 67.0±33.1 min after intravenous injection of 125.9±26.9 MBq of 68 Ga-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 4mm 3 , 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, contrastenhanced 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.

Imaging analysis
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 68 Ga-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. 68 Ga-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.

Standard of reference
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.

Statistical analysis
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).

Lesion-based analysis of pulmonary metastases and opacities
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±6.3cm 2 (range 0.2 -29.5cm 2 ). 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 PSMAnegative 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, PSMApositive 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   Table 3 Comparison of size, 68 Ga-PSMA-HBED-CC uptake (SUV max ) and radiodensity (HU mean ) between pulmonary metastases (all, PSMA-positive and PSMA-negative) and opacities. PSMA-positive lung metastases demonstrated the highest tracer uptake, significantly higher than pulmonary opacities, whereas PSMA-negative metastases demonstrated the lowest tracer uptake, significantly lower than pulmonary opacities (both p<0.001). All pulmonary metastases taken together, the mean SUV max of pulmonary metastases and opacities did not significantly differ (p>0.05). SUV max Maximum standardized uptake value, HU mean Mean Hounsfield units.   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).

Patient-based analysis of pulmonary metastases
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.

Discussion
This study investigated the imaging characteristics of pulmonary metastases and opacities in 68 Ga-PSMA-PET. Based on the SUV max in 68 Ga-PSMA-PET alone, it was not possible to differentiate between pulmonary metastases and pulmonary opacities. The majority of lung metastases highly overexpressed PSMA, while a relevant number of metastases were PSMA-negative. Pulmonary opacities demonstrated a moderate tracer uptake, significantly lower than PSMA-positive lung metastases, yet significantly higher than PSMA-negative metastases.

68
Ga-PSMA-PET/CT for the assessment of lung metastases 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 68 Ga-PSMA-PET. Up to now, the largest study investigating the 68 Ga-PSMA-PET imaging of PC pulmonary metastases and primary lung cancer was conducted by Pyka et al., with 45 patients and 89 lesions Fig. 4 SUV max of PSMA-positive and PSMA-negative lung metastases, pulmonary opacities and the aorta. Bar chart comparing the mean SUV maxvalues of PSMA-positive and PSMA-negative pulmonary metastases, pulmonary opacities and the aorta. The mean SUV max of PSMA-positive lung metastases was significantly higher than that of pulmonary opacities (p<0.05). Contrary, the mean SUV max of PSMA-negative lung metastases was significantly lower than that of pulmonary opacities (p<0.05). Error bars represent the standard deviation. SUV max Maximum standardized uptake value Fig. 6 Correlation between size and 3D SUV max of metastases. Linear correlation according to a Spearman's correlation, including 95% confidence intervals. A weak significant association between the SUV max in pulmonary metastases in 3D regions of interest and their size was calculated (p<0.05, ρ Spearman =0.207, 95% CI [0.02, 0.396]). R 2 Coefficient of determination, r Spearman's rho, SUV max Maximum standardized uptake value Fig. 5 Mean SUV max of all lung metastases, pulmonary opacities and the aorta. Bar chart comparing the mean SUV max -values of all pulmonary metastases, pulmonary opacities and the aorta. There was no difference between the mean SUV max of all lung metastases compared to that of pulmonary opacities (p>0.05). Error bars represent the standard deviation. SUV max Maximum standardized uptake value [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 68 Ga-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 68 Ga-PSMA-PET is low [31,32]. Therefore, PSMA-positive lung metastases are clearly visible and detectable as a result of a high lesion-tobackground 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 68 Ga-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 68 Ga-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 PSMApositive metastases tended to be larger than PSMAnegative metastases.

Limitations
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.

Conclusions
Based on the SUV max in 68 Ga-PSMA-PET alone, it was not possible to differentiate between pulmonary metastases and pulmonary opacities. The majority of lung metastases highly overexpressed PSMA, while a relevant number of metastases were PSMA-negative. Pulmonary opacities demonstrated a moderate tracer uptake, significantly lower than PSMA-positive lung metastases, yet significantly higher than PSMA-negative metastases. Nevertheless, given the combined information of CT scans as well as follow-up scans and the patient's history, experienced readers should be able to diagnose pulmonary metastases in 68 Ga-PSMA-PET accurately in most cases. Ethical approval and consent to participate All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the principles of the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The present study was retrospective; for this type of study the local ethics committee waived formal consent.