Introduction

Prostate cancer (PC) is the most common cancer in men and ranks second in terms of mortality (1). Although PC initially responds to androgen deprivation therapy (ADT), it eventually progresses to metastatic castration-resistant prostate cancer (mCRPC) (2, 3).

Prostate-specific membrane antigen (PSMA), a type II transmembrane glycoprotein, is overexpressed in PC cells compared to normal prostate cells. The intensity of PSMA expression on PC cells is further increased depending on the Gleason score (GS) of PC, tumor aggressiveness, androgen independence, recurrence, and metastatic disease (4).

Gallium-68(68 Ga)-PSMA positron emission tomography/computed tomography (PET/CT) is a widely used imaging modality for primary tumor detection, staging, and treatment response assessment in patients with PC (5). However, 5–10% of the lesions in patients with PC may not express PSMA glycoprotein due to some alterations related to DNA damage repair, showing inter- or intra-lesion heterogeneity (4, 6). While PC initially does not show typical glucose uptake, it starts to show high glucose uptake, which is called the Warburg effect, after a series of mutations that occur in later periods (7). However, it has been shown that inhibition of PSMA in patients with mCRPC is positively correlated with a higher uptake of 2-Deoxy-2-[fluorine-18] fluoro-d-glucose (18F-FDG) and this is associated with the upregulation of glucose transporter 1 (8). It has also been reported that patients with mCRPC lesions with 18F-FDG affinity are less differentiated and express lower levels of androgen receptors (AR) (9).

Docetaxel is an agent that has been used in patients with mCRPC for many years and its contribution to survival has been confirmed (10). Cabazitaxel is another anticancer agent used in patients with docetaxel-resistant mCRPC and has been shown to improve overall survival (OS) (11). Although the action mechanism of taxanes is not fully understood, some of their activity is seen in their interaction with androgen signalling at both cytoplasmic and nuclear levels in PC cells, affecting androgen-sensitive regulatory mechanisms (12). However, some of the antitumor activity of taxanes may be due to an androgen-independent mechanism. In recent years, advances in molecular knowledge have led to the understanding that PC is a heterogeneous disease in which androgen-dependent and independent clones coexist (13).

In our study, we aimed to determine the prognostic role of volumetric parameters and lesion-based Pro-PET scores obtained from 68 Ga-PSMA PET/CT and 18F-FDG PET/CT in predicting OS in patients with mCRPC receiving taxane therapy.

Materials and methods

Study design and patient selection

A total of 391 patients who underwent 18F-FDG PET/CT and 68 Ga-PSMA PET/CT at our clinic between January 2019 and January 2022 were retrospectively reviewed. The imaging protocols were performed with a median interval of 3 days (1–7 days) between 68 Ga-PSMA and 18F-FDG imaging. All patients underwent 68 Ga-PSMA PET/CT followed by 18F-FDG PET/CT.

Patients without 68 Ga-PSMA PET/CT or 18F-FDG PET/CT imaging, patients with a Pro-PET score between 0 and 2, patients who did not receive docetaxel and cabazitaxel or whose treatment was incomplete, patients receiving combination therapy (177Lu-PSMA plus taxane), patients with additional malignancies, patients who died of other causes, and patients whose data were unavailable were excluded.

A total of 71 patients diagnosed with mCRPC who had a Pro-PET score of 3–5 and who were referred to taxan and taxan plus treatments by the oncology board and who had completed docetaxel or cabazitaxel treatment were included in the study (Fig. 1). Docetaxel 75 mg/m2 or cabazitaxel 25 mg/m2, in combination with prednisone 5 mg 2 × 1, was administered every 21 days for a median of six cycles. ADT treatment was continued if castration-resistant. Serum prostate-specific antigen (PSA) levels, GS, and date of death were recorded. Whole-body volumetric parameters were obtained from both imaging studies and recorded. This retrospective study was initiated after approval by the local ethics committee (approval number: 368–2023) and was conducted in accordance with good clinical practice guidelines and current local legislation.

Fig. 1
figure 1

Patient selection flowchart

Full size image

PET/CT protocol

At least 6 h of fasting was recommended prior to 18F-FDG PET/CT imaging and 18F-FDG injection was performed at a dose of 3.5–5.5 MBq/kg with a blood glucose level < 140 mg/dL measured by the fingerstick method. For 68 Ga-PSMA PET/CT imaging, patients received an injection of 2 MBq/kg. All images were acquired using a Discovery IQ 4 ring 20 cm axial field of view (FOV) PET/CT (GE Healthcare, Milwaukee, WI, USA). After injection, whole-body images from the vertex to the mid-thigh were taken at hour 1. After CT images (CT parameters: 120 kV, 80 mAs/slice, 700 mm transaxial FOV, 64 × 0.625 mm collimation, pitch 1.4. 0.5 s rotation time, 3.3 mm slice thickness, 512 × 512 matrix), PET images (PET parameters: 3D FOV 20 cm, ordered subset expectation–maximisation algorithm 5 iterations/12 subset, full width at half maximum 3 mm) were acquired at the bed side over 2.5 min in the same position so as to include the same regions. All patients were asked to drink water before PET/CT imaging and urinate immediately before imaging.

Evaluation of images

Images were evaluated consensually by two nuclear medicine physicians with at least 10 years of experience in PET/CT interpretation according to current guidelines (14, 15). After excluding physiologic uptake and false positives, uptakes higher than the mediastinal blood pool were defined as positive for 68 Ga-PSMA PET/CT or 18F-FDG PET/CT. Primary tumors and metastatic lesions were counted, measured, and recorded for both imaging modalities. In addition, patients were visually scored according to lesion involvement. According to the Pro-PET scoring system, patients were scored as follows: score 0: PSMA(-) FDG(-), score 1: PSMA( +) FDG( – ), score 2: PSMA( +) FDG( +) (FDG < PSMA), score 3: PSMA( +) FDG( +) (FDG = PSMA), score 4: PSMA( +) FDG( +) (FDG > PSMA), score 5: PSMA( – ) FDG( +) and recorded accordingly (16). The presence of at least one lesion for any score in organ-based and patient-based evaluation was considered positive for that score. Those with a patient-based score of 3–5 were included in the study.

In both 68 Ga-PSMA PET/CT and 18F-FDG PET/CT images of patients with scores of 3–5, semiautomatic VOIs (volume of interest) were drawn from each of the primary lesion and metastatic areas using a 40% SUV threshold, with the lesion within the image area in all three planes.

For primary tumor (P), lymph nodes (L), bone (B), and visceral (V) metastatic lesions, 18F-FDG tumor volume (TV-F) and PSMA tumor volume (TV-P) values of each lesion were automatically provided by the device and these data were summed and recorded as total tumor volume for FDG (TTV-F) and for PSMA (TTV-P). Total lesion glycolysis (TL-G) and total lesion PSMA (TL-P) values for each lesion were calculated and whole body total lesion glycolysis (TTL-G) and whole body total lesion PSMA (TTL-P) values were obtained and recorded. TTV-F and TTV-P values, which were obtained from 18F-FDG PET/CT and 68 Ga-PSMA PET/CT, were added to obtain the whole body total tumor volume FDG + PSMA (TTV-F + P). In addition, TTL-G and TTL-P values were added to obtain the whole body total lesion glycolysis + PSMA (TTL-G + P) value.

Statistical methods

The SPSS 26.0 (IBM Corporation, Armonk, New York, United States) program was used to analyse the variables. The sensitivity and specificity ratios for the relationship between the actual classification and the classification separated by the cut-off value calculated for each variable were examined and expressed by receiver operating characteristic (ROC) curve analysis. Kaplan–Meier (product limit method) log-rank analysis was used to examine the effect of factors on mortality and survival. The univariate and multivariate Cox regression analysis enter method was used to measure the effects of prognostic variables on mortality and survival. Quantitative variables were expressed as mean ± SD (standard deviation) and median (percentile 25%/percentile 75%), while categorical variables were expressed as n (%). Variables were analysed with 95% confidence interval (CI) and a p-value less than 0.05 was considered significant.

Results

The median age of the patients included in the study was 71 (56–89) years. The median serum PSA value was 16.4 (0.01–1852 ng/dl). After imaging, 48% of patients received first-line docetaxel, 11.2% received rechallenged docetaxel and 40.8% received cabazitaxel. Patients receiving rechallenged docetaxel had previously received at least one new generation anti-androgen-axis agent. All patients receiving cabazitaxel had previously received docetaxel and at least one new generation anti-androgen-axis agent. During the study period, 19 patients (26.7%) died, with a median survival of 10.6 months (95% Cl: 0.8–47.3).

Pro-PET evaluation showed a score of ≥ 3 in 28 patients (39.5%) for primary prostate lesions (Pscore), 29 patients (40.8%) for lymph node metastases (Lscore), 56 patients (78.9%) for bone metastases (Bscore), and 56 patients (78.9%) for visceral metastases (Vscore). The evaluation of lesions based on Pro-PET scoring is shown in Table 1.

Table 1 Evaluation of lesions on imaging based on Pro-PET scoring with 68 Ga-PSMA PET/CT and 18F-FDG PET/CT
Full size table

Prognostic factors

Based on the ROC curve analysis, when the TTV-P cut-off value was taken as ≥ 78.5 cm3, its sensitivity and specificity in predicting mortality were 68% and 67.4%, respectively, which was statistically significant (area under the curve [AUC] = 0.685 ± 0.069, P = 0.011). TTL-P cut-off value ≥ 278.15, TTV-F cut-off value ≥ 94.95 cm3, and TTL-G cut-off value ≥ 485.3 were statistically significant in predicting mortality. The cut-off values, AUC, sensitivity and specificity of all parameters are shown in Table 2.

Table 2 ROC curve cut-off, sensitivity, and specificity values in predicting mortality
Full size table

When comparing 1 and 2-year OS rates in Kaplan–Meier survival analysis, Pro-PET Lscore ≥ 3 and Vscore ≥ 3 were found to be prognostic factors predicting short OS (p = 0.015 and p < 0.001). As quantitative and volumetric assessments, TTV-P (≥ 78.5 cm3), TTL-P (≥ 278.8), TTV-F (≥ 94.98 cm3), TTL-G (≥ 458. 3), TTV-P + F (≥ 195.45 cm3), TTL-G + P (≥ 855.78), LTV-FDG (≥ 3.4 cm3), LFDG-SUVmax (≥ 3.2), LFDG-SUVmean (≥ 2. 25), LFDG-SUVpeak (≥ 2.55) and BTV-F (≥ 51.15 cm3) parameters were found to be prognostic factors predictive of short OS (p < 0.001, p = 0.002, p = 0.006, p = 0.007, p = 0.002, p < 0.001, p = 0.026, p = 0.007, p = 0.003, p = 0.003 and p = 0.006, respectively) (Table 3).

Table 3 Kaplan meier 1 and 2 year survey analysis
Full size table

Univariate Cox regression analysis revealed no statistically significant correlation between age, serum PSA levels, GS, and OS (Table 4). Finally, the multivariate Cox regression analysis showed that a Vscore ≥ 3 (95% CI: 7.069–98.251, p < 0.001) and TTL-G + P ≥ 855.78 (95% CI: 4.878–1037.860, p:0.006) were found to be independent prognostic factors predicting short OS (Table 5), (Figs. 2 and 3).

Table 4 Univariate analyses of factors for predicting overall survival
Full size table
Table 5 Multivariant cox regression analysis for OS
Full size table
Fig. 2
figure 2

Kaplan–Meier curves of overall survival according to (A) Vscore and (B) TTL-G + P

Full size image
Fig. 3
figure 3

A 69-year-old mCRPC (Gleason skor:4 + 5, TPSA:1129) patient with visceral score:5 and TTL-G + P: 9603.9 g/ml x cm3; 68 Ga-PSMA PET/CT A TTV-P: 610.2 cm3, TTL-P: 1415.4 g/ml x cm3 (MIP, PET and fusion image with and without VOIs) and 18F-FDG PET/CT B TTV-F: 1990.2 cm3, TTL-G: 8188.5 g/ml x cm3 (MIP, PET and fusion image with and without VOIs). The patient was given 6 cycles of cabazitaxel after imaging and the survival time was 5 months

Full size image

Discussion

In the current study where the measurements are based on the evaluation of both 68 Ga-PSMA PET/CT and 18F-FDG PET/CT images, a Vscore of ≥ 3 and the volumetric parameter TTL-G + P of ≥ 855.78 were independent prognostic factors predicting OS in patients with mCRPC receiving taxane therapy.

In the literature, terms such as anaplastic PC, neuroendocrine PC, and intermediate-grade atypical PC, have previously been used for PC with poor prognosis (17). Aggressive variant (AV) PC has been defined in the literature and this newly defined term has been suggested instead of the previous ones (18). It has been reported that AVPC should include at least one of seven features, including the presence of visceral metastasis (19, 20).

There are several studies in the literature on the correlation between volumetric parameters derived from 68 Ga-PSMA PET/CT and prognosis. Schmuck et al. reported that volumetric parameters may be a quantitative imaging biomarker especially in the evaluation of response to treatment (21). In the study of Grubmüller et al. in which patients with mCRPC receiving different treatments were included, the researchers found that the use of SUVmean, SUVmax, SUVpeak, and TTV-P parameters derived from 68 Ga-PSMA-11 PET/CT were suitable for assessing treatment response; however, they found no association with OS (22). In another study that included patients with mCRPC receiving second-generation ADT, pretreatment TTV-P ≥ 124.85 (OR: 11.4, 95% CI 1.11–116, 6, P = 0.04) was found to be an independent prognostic factor for OS (23). In a study where Can et al. evaluated patients receiving docetaxel and abiraterone-enzalutamide, TTV-P ≥ 116.88 was reported to be an independent prognostic factor in predicting mortality (24). In another study where Telli et al. studied patients receiving docetaxel and abiraterone-enzalutamide treatments, the researchers found that TTV-P ≥ 40.1 was an independent prognostic factor for survival (HR: 1.003, 95% CI 1.001–1.004, p = 0.001) in multivariate analysis (25). Siefert et al. found that TTV-P level was a negative prognostic factor on OS in their multivariant cox analysis in the group receiving 177Lutetium (177Lu) PSMA ligand treatment (26). In our univariate analysis, TTV-P ≥ 78.5 (95% CI 9.354–15.325) p < 0.001) and TTL-P ≥ 278.15 (95% CI 10.055–15.806 p = 0.002) were found to be negative prognostic factors for OS. However, in the multivariate analysis, the volumetric parameters obtained from 68 Ga-PSMA PET/CT alone were not found to be independent prognostic factors.

Normal cells derive energy from glucose through oxidative phosphorylation via the mitochondrial citric acid cycle. Most cancer cells derive their energy from lactic acid fermentation, which synthesises less ATP, even in the presence of oxygen (27). The dense glucose required for these cancer cells is transported into the cell via glucose transporters (GLUTs) (28). 18F-FDG PET/CT imaging is based on this mechanism (29). However, the role of 18F-FDG PET/CT imaging in PC is limited (30). It has been shown by immunohistochemical and micro-PET studies that differentiated PC cells take up glucose via sodium-glucose co-transporter 2 (31). Previous studies have shown that androgen increases GLUT1 expression in androgen-dependent PC. This suggests that androgen signalling may play a role in the regulation of GLUT1 (8, 32). However, the mechanisms underlying GLUT1 regulation by androgen signalling remain unclear. GLUT1 is a key mediator of 18F-FDG uptake into the cell and is associated with the aggressive phenotype of androgen-independent PC (33).

Although mCRPC is a heterogeneous disease, approximately 87% of patients with mCRPC have at least one 18F-FDG PET/CT positive metastasis (9). Bauckneh et al. found that decreased MTV (MTV < 325.97 cm3) and TLG (TLG < 844.86) values were associated with increased OS. They also reported that MTV was an independent prognostic factor for OS in multivariate analysis (34). In a retrospective study, Wibmer et al. found that TTL-G was independently associated with OS in patients with mCRPC receiving first-line treatment with abiraterone or enzalutamide (p < 0.001) and reported that it may qualify as a quantitative prognostic imaging biomarker (35). In our univariate analysis, TTV-F ≥ 94.95, TTL-G ≥ 458.3, LTV-F ≥ 3.4, LTL-G ≥ 10.3, LFDG-SUVmax ≥ 3.2, LFDG-SUVmean ≥ 2.25, LFDG-SUVpeak ≥ 2.55, and BTV-F ≥ 51.15 were found to be negative prognostic factors for OS. However, in multivariate analysis, volumetric parameters obtained from 18F-FDG PET/CT alone were not found to be independent prognostic factors.

In the study by Ferdinandus et al. where they used 18F-FDG PET/CT and 68 Ga-PSMA PET/CT together in patients receiving 177Lu-PSMA therapy, FDG-positive MTV and PSMA SUVmean were identified as potential biomarkers of OS. However, TTV-P level was not found to be a prognostic factor for OS (36). In our study, the TTL-G + P ≥ 855.78 (95% CI: 4.878–1037.860, p = 0.006) value that we obtained by summing 18F-FDG PET/CT and 68 Ga-PSMA PET/CT parameters was found to be an independent prognostic factor for short OS. A cut-off value for TTL-G + P has not been found in the literature.

Whitney et al. suggested that age was a prognostic factor for OS in mCRPC, but did not define a threshold (37). However, Guijarro et al. showed in their study that there was no significant correlation between age and OS and PFS in patients with metastatic PC (38). Similarly, in our study, we did not find an association between age and OS in the univariate analysis.

There are several studies that have reported that GS and serum PSA level are associated with OS in patients with PC. Oruç et al. found that especially baseline PSA and ∆PSA levels were independent prognostic factors for OS in patients with mCRPC receiving second-generation AR axis targeted therapy (23). Although several studies reported that patients with lower levels of serum PSA level had better OS, this parameter was not associated with OS (24, 25, 34). However, Ferdinandus et al. found no correlation between PSA level and OS (36). In our study, univariate Cox regression analysis showed no significant correlation between GS and serum PSA levels and OS.

Several studies have shown that the presence of visceral organ metastases in patients with PC is associated with a poor prognosis (37,38,39,40,41). An immunohistochemical study showed that membranous PSMA expression in PC metastases without neuroendocrine differentiation was significantly lower in liver metastases than in bone and lymph node metastases (6).

Several studies have shown that volumetric parameters obtained from 18F-FDG PET/CT and 68 Ga-PSMA PET/CT and score-based assessments showing increased glucose uptake may indicate tumor differentiation and contribute to disease management (42,43,44). In our study, which was different in terms of design, Lscore ≥ 3 and Vscore ≥ 3 were found to be prognostic factors in predicting OS. In addition, in the multivariate analyses, a Vscore of ≥ 3 was found to be an independent prognostic factor for short OS. Summarizing our findings, increased FDG uptake in lymph node and especially visceral organ metastases were associated with poor prognosis.

In light of the existing literature and the findings of our study, 18F-FDG PET/CT and 68 Ga-PSMA PET/CT findings show that increased glucose uptake may predict a prognosis where the disease transforms into aggressive PC. Therefore, we recommend the use of dual tracer PET/CT in routine clinical practice in order to determine which group of patients will or will not benefit from treatment, and consequently, the approach to combined therapies.

Strengths and limitations

The main limitation of our study is its retrospective and single-centre design. The second limitation is the lack of secondary biopsy in patients after progression to mCRPC. Another limitation is that the association of OS within the already small group of patients with visceral metastases was not compared between organs. A further limitation is that it was not possible to compare the two different taxane treatments due to a lack of patients. On the other hand, the homogeneous group of patients receiving only taxane treatment is the strength of our study.

Conclusion

We observed strong associations between OS and volumetric biomarkers derived from 68 Ga-PSMA PET/CT and 18F-FDG PET/CT imaging and Pro-PET scores in patients with mCRPC receiving taxane group therapies. 18F-FDG PET/CT provides prognostic information when used in combination with 68 Ga-PSMA PET/CT in patients with mCRPC. This finding may open a new window into the characterisation of mCRPC heterogeneity, allowing for individualised therapy. Prospective studies involving a larger number of patients treated with both taxanes and non-taxanes are needed.