Ga68-PSMA PET/CT is an increasingly frequently used imaging modality in prostate cancer staging, treatment response evaluation and restaging.
PSA recurrence is commonly seen after radical prostatectomy and radiotherapy. Biochemical recurrence is defined as a serum PSA value above 0.2 ng/ml after RP and an increase of more than 2 ng/ml from the lowest value after RT. The main aim in these patients is to differentiate between local and systemic disease, as this may affect the treatment approach (8). Ga68-PSMA PET/CT can visualise recurrence areas that other conventional imaging techniques cannot find even at low PSA levels.
Published data suggest that there are a large number of BR patients with favourable prognostic features (ISUP ≤ 2, pT2 stage, negative surgical margins, PSAdt > 12 months; BR interval > 18 months) who may not experience a clinically significant recurrence (9). One of the unmet clinical needs is to identify the group of BR patients who may benefit from salvage radiation therapy and a surveillance-based approach.
In this study, we evaluated the efficacy of Ga68-PSMA PET/CT in the localisation and detection of recurrent disease in patients with biochemical recurrent prostate cancer. Our study group consisted of 42 post-RP and 16 post-RT patients with biochemical recurrence.
According to the results of biochemical recurrence patients evaluated in the study group (mean PSA 7.6 ng/ml, range 0.24–100 ng/ml), Ga68-PSMA PET/CT patient-based detection rate was found to be 57%. These results are similar to most publications in the literature (9–12); however, there are also publications in the literature with higher (7, 13–15) and lower detection rates (16).
In a meta-analysis including 3693 patients and 24 studies (primary and recurrent disease), it was concluded that PSA was a predictor of a positive PSMA PET scan, and increased pre-PET PSA levels were associated with Ga68-PSMA PET positivity (17). According to this meta-analysis, PSMA PET/CT positivity rates were 45%, 59%, 75% and 95% when PSA levels were 0.2–0.49 ng/ml, 0.5–0.99 ng/ml, 1.00-1.99 ng/ml and > 2.00 ng/ml, respectively. When the post-RP studies were evaluated separately, a non-statistically significant higher positivity rate was observed at 1-1.99 ng/ml PSA level (82% vs. 64%). In another study, PSMA PET/CT detection rates at PSA values of 0-0.19 ng/ml, 0.20–0.99 ng/ml, 1.00-1.99 ng/ml and ≥ 2 ng/ml were 21%, 41%, 91% and 92%, respectively (18). Statistically significant correlation between PSMA PET/CT positivity and PSA level has been reported in other studies in the literature (11, 12, 14, 18). In our study, PSA level was found to be significantly higher in PSMA PET positive patients. However, as the PSA level increased, PSMA PET/CT detection rate increased (23.8%, 68.8%, 81% for PSA levels < 1 ng/ml, 1–4 ng/ml and > 4 ng/ml, respectively). In this respect, the study results are compatible with the literature. Due to the heterogeneous distribution of the patient population in the study, PSA levels could not be categorised as 0.2–0.5 ng/ml, 0.5-1 ng/ml, 1–2 ng/ml and > 2 ng/ml as in most publications in the literature and PSMA PET/CT detection at very low PSA values could not be investigated.
In the study, the site of recurrence could be demonstrated at low PSA values (< 1 ng/ml) with a rate of 23.8%. This rate is lower than the rate of 44% obtained by Ceci et al (332 patients, 10) and Verburg et al (155 patients,19). The reason for this may be difference number of patients. Pelvic recurrence site was detected in 3 of 11 low PSA patients and pelvic radiotherapy was applied to 8 patients including these patients. There were no distant metastatic patients in low PSA patients in the study group. However, it should be kept in mind that these patients may have distant metastasis and pelvic salvage radiotherapy may not be an appropriate treatment option.
In a study by Ceci et al. with 70 patients, significant differences were found between 68Ga-PSMA PET/CT positive and negative patients in terms of PSA level (higher in positive patients) and PSAdt (faster kinetics in positive patients). In addition, PSAdt was shown to be a significant predictor of positive PSMA PET scan and detection of distant lesions. In ROC analysis, threshold values of 6.5 months for PSAdt (AUC: 0.868, 95% Confidence Interval: 0.767–0.969) and 0.83 ng/mL for PSA (AUC: 0.764, 95% Confidence Interval: 0.647–0.882) were found (13). In our study, a correlation was observed between PSA and PSMA PET/CT positivity, but no statistically significant difference was observed between PSAdt and PSA velocity. In addition, the PSA threshold value was found to be 1.005 with 86% sensitivity and 64% specificity (AUC: 0.779, 95% Confidence Interval: 0.614–0.943). PSAdt and PSAvel could be statistically analysed in only 35 patients. Threshold values were found to be 3.8 months with 86% sensitivity and 36% specificity for PSAdt (AUC: 0.509, 95% Confidence Interval: 0.302–0.715) and 1.25 ng/ml/year with 71% sensitivity and 79% specificity for PSA velocity (AUC: 0.723, 95% Confidence Interval: 0.547–0.899), but not statistically significant.
In a study by Verburg et al. local prostate tumour, extrapelvic lymph node and bone metastasis rates were higher in patients with higher PSA levels. Shorter PSAdt was associated with pelvic and extrapelvic lymph node, bone metastasis, visceral metastasis; higher Gleason score was associated with pelvic lymph node metastasis. Both PSA and PSAdt were found to be the only independent markers of scan positivity and extrapelvic lymph node metastasis, and PSAdt was found to be the only independent marker of bone metastases (19). In our study, PSA and Gleason score mean were found to be higher in the presence of bone involvement. In addition, PSA value was significantly higher in distant metastatic patients (mean 24.86 vs. 6.9). Although there was no difference in terms of PSAdt, it was observed to be shorter in distant metastatic patients and in positive uptake.
In the meta-analysis performed by Perera et al. in 2020, the overall positivity rates were 28% in the prostate bed, 38% in pelvic lymph nodes, 13% in extrapelvic lymph nodes, 22% in bone and 5% in visceral organs (17). In our study, the most common sites of involvement were the prostate gland (n = 15, 26%), pelvic lymph nodes (n = 12, 21%) and bone (n = 10, 18%) in parallel with this meta-analysis. In the same meta-analysis, statistically more positivity was found in the prostate bed in the post-radiotherapy group (52% vs. 22%). In this study, statistically more uptake was observed in patients with BR in the prostate lodge after RT, which is consistent with the literature. This may be due to the lack of surgical removal of the minimal residual source However, there was no significant difference between PSMA PET/CT positivity and Gleason score according to the publications including staging and restaging in the same meta-analysis, positivity was 72% in patients with Gleason score 7 and 80% in patients with Gleason score 8. In our study, the mean Gleason score was higher in PET positive patients. In the same time, more positivity was found in patients with high ISUP grade. Similarly, in the study published by Celli et al. in 2021, ISUP categories were defined as the only significant risk factor for clinical recurrence. More clinical relapses have been reported in ISUP grade 4–5 compared to grades 1–2 and 3 (9) and recurrence-free survival is longer in patients with lower ISUP grades (20.5 months in grade 1–2; 12.6 in grade 3 and 12.1 in grade 4–5). In our study, supradiaphragmatic lymph node and bone involvement leading to distant metastatic staging were more common in patients with high ISUP grade.
In the study by Ertürk et al. evaluating the relationship between PSMA PET/CT findings and ALP, LDH and PSA in 61 unoperated prostate cancer patients, no significant correlation was found between prostate SUVmax and PSA. Significant correlation was observed between high ALP levels and serum PSA and extensive bone metastasis (20). Another study reported a correlation between prostate SUVmax and ISUP grade (21). In our study, no significant correlation was found between prostate SUVmax and PSA and Gleson score. In the correlation analysis, only a positive correlation was observed between PSA and bone SUVmax.
The overall sensitivity and specificity of PSMA PET/CT were calculated as 75% and 99% on a lesion basis and 77% and 97% on a patient basis, respectively (17). The sensitivity value calculated in our study was 73%, which is at a similar level, but the specificity value was lower (78.6%). Other diagnostic performance parameters calculated in the study were positive predictive value (90.9%) and negative predictive value (50%). The reason for the low specificity and negative predictive value compared to the literature may be that the correlation is mostly made with clinical follow-up based on PSA and other imaging methods, not histopathological confirmation. However, due to its high sensitivity and positive predictive value, we can say that PSMA PET/CT provides a high probability of disease localisation in biochemical recurrence. In this study, 9 of the 11 patients who were evaluated as false negative had negative PSA after RT and/or ADT, and the other 2 patients had a decrease in PSA after salvage RT after PSA increase became apparent in the follow-up.
PSA value was below 1 ng/ml in 9 of 11 patients who were evaluated as false negative in the study. In 6 of them, the PSA value was between 0.2–0.5 ng/ml. In one patient (postoperative 5th month), the follow-up PSA value was 33.9 ng/ml, and since the control PSA value decreased to 16 ng/ml at the time of PSMA PET without treatment, it was thought that this may be secondary to laboratory error or postoperative fluctuation. The PSA value of the last remaining patient was 4.77 ng/ml.
Unspecific bone involvement and celiac ganglion involvement were present in 2 of the 3 patients with suspicious uptake. In a multicentre retrospective cohort analysis of 298 patients in which biochemical recurrence was investigated in 2022 (53 patients eligible for final diagnostic analysis, 81 positive bone lesions), Ga68-PSMA PET results were evaluated (22). Accordingly, most bone involvement was due to benign lesions. Bone involvement was observed in 32.9% of patients and metastasis rate was 34.6% (28/81 lesions). Initial Gleason score ≥ 8, age at recurrence ≤ 71 years and SUVmax > 6.21 were found to be independent predictors of prostate cancer metastasis. In a recent retrospective analysis of 348 patients, unspecific bone involvement was observed in 51.4% and the most common localisation was the costa and pelvis (23). In another study with F18-PSMA and Ga68-PSMA (24), similar metastasis detection rates were found with both radiopharmaceuticals (n = 72 vs 64). Unspecific bone uptake was more common with F18-PSMA PET/CT than with Ga68-PSMA PET/CT (34.2% vs. 16.7%). There was no significant difference between PSA groups (< 1, 1–5, > 5 ng/ml) in the detection of metastasis and unspecific bone involvement. It has been emphasised that focal bone involvement in biochemical recurrent patients with PSA ≤ 5 ng/ml and SUV > 4 is likely to be false positive. In our study, unspecific bone involvement was found in the left femoral neck and SUVmax was 2.9 (PSA: 0.45).
Sympathetic ganglia are among the tissues where Ga68-PSMA expression is increased and is one of the potential causes of false positives that may lead to direct treatment regimen change. In histological studies, glutamate carboxypeptidase II was found to be highly expressed in non-myelinated schwann cells in the ganglia of the sympathetic nervous system. This explains the increased Ga68-PSMA-ligand uptake in sympathetic ganglia. Sympathetic ganglia are located along the vertebral column and can be easily confused with lymph node metastases located near the cervical, celiac or sacral ganglia. In the study by Rischpler et al. (25), Ga68-PSMA uptake was highest in the celiac ganglion (SUVmax: 2.9 ± 0.8) and SUVmax values in the cervical and sacral ganglia were 2.4 ± 0.6 and 1.7 ± 0.5, respectively. In addition, PSMA expression was significantly higher in lymph node metastases (cervical SUVmax: 18.0 ± 16.2, celiac 13.5 ± 12.3, ;sacral SUVmax: 13.4 ± 11.6). In addition, ganglia were predominantly band or teardrop shaped and rarely nodular in shape resembling lymph node metastasis. Therefore, PSMA ligand uptake and lesion configuration should be carefully examined in the differentiation of lymph node metastases and sympathetic ganglia. In our study, one patient (PSA: 0.46 ng/ml, Gleason score: 3 + 3) had a fusiform nodular lesion in the upper abdomen showing low level PSMA receptor activity. When it was re-evaluated due to the absence of morphological change in the follow-up, it was found to be celiac ganglion and SUVmax value was higher compared to the literature (SUVmax:5,6).
Increased PSMA expression due to neovascularisation in hepatocellular carcinoma has been reported in the literature (26). However, reports of increased PSMA uptake due to hepatitis are rare. In a case report published by Adediran et al. increased uptake on Ga68-PSMA PET/CT due to radiation hepatitis in the liver dome after radiotherapy in a patient diagnosed with invasive lobular carcinoma of the right breast (27). In our study, one patient with false positive liver uptake had histopathological confirmed cholestatic hepatitis.
The study has some limitations. One of these is that the study was retrospective. In addition, histopathological correlation of most of the positive uptake could not be performed and correlation was made with clinical follow-up/laboratory/other radiological methods. In addition, some patients were followed up in an external centre and referred to our centre only for imaging. Therefore, detailed clinical, laboratory and pathological data of each patient could not be obtained. Another limitation is that the total number of patients included in the study was small and therefore the number of patients in some subgroups was insufficient in statistical analysis. Because of these limitations, the promising performance and efficacy of Ga-68 PSMA PET/CT on disease management in prostate cancer restaging should be confirmed in a larger prospective study.