Prospective Evaluation of ¹⁸F-DCFPyL PET/CT in Detection of High-Risk Localized Prostate Cancer: Comparison With mpMRI

Abstract

OBJECTIVE.

The purpose of this study was to assess the utility of PET with (2S)-2-[[(1S)-1-carboxy-5-[(6-(¹⁸F)fluoranylpyridine-3-carbonyl)amino]pentyl]carbamoylamino]pentanedioic acid (¹⁸F-DCFPyL), a prostate-specific membrane antigen (PSMA)-targeted radiotracer, in the detection of high-risk localized prostate cancer as compared with multiparametric MRI (mpMRI).

SUBJECTS AND METHODS.

This HIPAA-compliant prospective study included 26 consecutive patients with localized high-risk prostate cancer (median age, 69.5 years [range, 53–81 years]; median prostate-specific antigen [PSA] level, 18.88 ng/mL [range, 1.03–20.00 ng/mL]) imaged with ¹⁸F-DCFPyL PET/CT and mpMRI. Images from PET/CT and mpMRI were evaluated separately, and suspicious areas underwent targeted biopsy. Lesion-based sensitivity and tumor detection rate were compared for PSMA PET and mpMRI. Standardized uptake value (SUV) and PSMA PET parameters were correlated with histopathology score, and uptake in tumor was compared with that in nonmalignant tissue. On a patient level, SUV and PSMA tumor volume were correlated with PSA density.

RESULTS.

Forty-four tumors (one in Gleason grade [GG] group 1, 12 in GG group 2, seven in GG group 3, nine in GG group 4, and 15 in GG group 5) were identified at histopathology. Sensitivity and tumor detection rate of ¹⁸F-DCFPyL PET/CT and mpMRI were similar (PET/CT, 90.9% and 80%; mpMRI, 86.4% and 88.4%; p = 0.58/0.17). Total lesion PSMA and PSMA tumor volume showed a relationship with GG (τ = 0.27 and p = 0.08, τ = 0.30 and p = 0.06, respectively). Maximum SUV in tumor was significantly higher than that in nonmalignant tissue (p < 0.05). Tumor burden density moderately correlated with PSA density (r = 0.47, p = 0.01). Five true-positive tumors identified on ¹⁸F-DCFPyL PET/CT were not identified on mpMRI.

CONCLUSION.

In patients with high-risk prostate cancer, ¹⁸F-DCFPyL PET/CT is highly sensitive in detecting intraprostatic tumors and can detect tumors missed on mpMRI. Measured uptake is significantly higher in tumor tissue, and PSMA-derived tumor burden is associated with severity of disease.

Prostate cancer is the second leading cause of cancer death in men [1]. To date, diagnosis and subsequent management of primary prostate cancer are reliant on biopsy, which may be obtained via systematic cores or with image-guided biopsies using multiparametric MRI (mpMRI)-ultrasound fusion [2]. However, mpMRI has some limitations including in fully differentiating clinically significant disease, reliably staging primary tumors, and monitoring response after treatment [3–5]. Small-molecule radiotracers targeting prostate-specific membrane antigen (PSMA) offer a possible solution to this limitation. PSMA is a transmembrane protein that is expressed in most prostate cancer cells but shows low expression in normal tissue [6–8]. Higher PSMA expression correlates with more aggressive disease, suggesting that PSMA uptake may be prognostic [9]. Additionally, by their nature, PSMA-targeted PET/CT and PET/MRI offer whole-body staging in a single scan.

A number of small-molecule PSMA-targeted radiotracers have been developed [10, 11]. Recently, ¹⁸F-labeled ligands have been gaining popularity because of the longer half-life of their label compared with ⁶⁸Ga (110 vs 68 minutes), improved spatial resolution, and opportunity for mass-production [12, 13]. One of the latest in this line of radiotracers is (2S)-2-[[(1S)-1-carboxy-5-[(6-(¹⁸F) fluoranylpyridine-3-carbonyl)amino]pentyl]carbamoylamino]pentanedioic acid (¹⁸F-DCFPyL). This second-generation ¹⁸F-labeled PSMA targeting agent improves on its predecessor with improved tumor-to-background ratio and benefits from optimization of the ¹⁸F label [14–20]. High-risk disease constitutes an important form of prostate cancer with a wide range of aggressiveness, optimal treatment of which depends on accurate estimation of the disease burden and its proper and timely management [21]. Unfortunately, available conventional imaging methods are limited in their ability to deliver accurate information about disease burden and staging, yet the impact of novel imaging strategies such as targeted PET is yet to be explored for these indications [22].

In this study, we prospectively compared ¹⁸F-DCFPyL PSMA PET/CT and mpMRI for detection of high-risk localized prostate cancer.

Subjects and Methods

Study Population

This prospective single-institution HIPAA-compliant study was approved by the ethics committee of the National Cancer Institute (Protocol #17-C-0109, NCT03181867). Between August 2017 and June 2019, 26 consecutive patients with high-risk primary prostate cancer were imaged with ¹⁸F-DCFPyL PSMA tracer. Inclusion criteria were patients with known localized high-risk prostate cancer (i.e., serum prostate-specific antigen [PSA] level > 10 ng/mL, Gleason score of 8–10 at preimaging biopsy, or clinical stage > T2c) with evidence of disease on standard imaging (CT and bone scanning). Exclusion criteria included contraindications to PET/CT or MRI and androgen deprivation therapy before imaging. All 26 imaged patients were included in the analysis. Patient characteristics are reported in Table 1.

TABLE 1:

Patient and Lesion Characteristics

Characteristic	Value
Patient
Total no.	26
Median age (y)	69.5 (53–81)
Median PSA level (ng/mL)	18.88 (1.03–203)
PSA density (ng/mL2)	0.36 (0.09–2.03)
Lesion
Identified on MRI	43
Identified on PET/CT	50
Identified on both MRI and PET/CT	37
Pathology positive on biopsy	44
GG group 1	1 (2.3)
GG group 2	12 (27.3)
GG group 3	7 (15.9)
GG group 4	9 (20.5)
GG group 5	15 (34.1)
Pathology positive by zone
Peripheral zone	31
Transition zone	21
Spanning both zones	4
Extraprostatic uptake on PET
Lymph node uptake
No. of lesions	38
No. of patients	6
Distant metastasis
No. of sites	2
No. of patients	2

Note—Values in parentheses are ranges or percentages. PSA = prostate-specific antigen, GG = Gleason grade.

PET/CT and Multiparametric MRI Acquisition

All patients underwent ¹⁸F-DCFPyL PET/CT and mpMRI within a mean of 57 days (median, 30 days; range, 1–270 days) of each other. PET/CT was performed on a 3D time-of-flight (TOF) mode GE Discovery 710 camera (GE Healthcare), with a 20-cm coronal and a 70-cm axial FOV. Data were reconstructed with a maximum-likelihood expectation-maximization TOF-based algorithm. The scanner uses CT-based attenuation correction along with random, normalization, dead time, and model-based scatter corrections for attenuation correction and anatomic coregistration purposes. The ¹⁸F-DCFPyL tracer was synthesized under previously described good manufacturing practices [14]. Quality control testing was performed before injection to ensure proper doses and specific activity: radiochemical purity of 99.2% ± 0.4% and specific activity of 1842 ± 489 mCi/μmol (68,154 ± 18,093 MBq/μmol). Each patient received an IV bolus injection of ¹⁸F-DCFPyL with a mean dose of 295.45 MBq (range, 202.72–323.9 MBq), followed by a static whole-body PET/CT scan (from the top of the skull to the feet) approximately 2 hours after injection (2 minutes in each bed position). Dose-modulated transmission CT scans (120 kV, 80 mAs) were acquired before each PET scan for anatomic correlation and coregistration purposes. Before PET/CT image acquisition, patients were asked to void to decrease bladder activity. Readers carefully reviewed coregistered images in all three planes and at varying window levels to evaluate tissue adjacent to the bladder.

A 3-T mpMRI scanner (Achieva 3 T-TX, Philips Healthcare) was used with an endorectal coil (BPX-30, Medrad), filled with 45-mL Fluorinert (3M), and either the anterior half of a 32-channel cardiac SENSE coil (InVivo) or only the cardiac SENSE coil. The mpMRI acquisition fulfilled the minimum technical requirements of the Prostate Imaging Reporting and Data System version 2 (PI-RADSv2); and sequences obtained included T2-weighted images (axial, sagittal, and coronal), DW images (b values up to 1500–2000 mm/s²) available as apparent diffusion coefficient maps and high b-value sequences, and axial 3D T1-weighted fast field-echo dynamic contrast-enhanced images (temporal resolution of 5.6 seconds with 54 phases) [23].

Analysis of ¹⁸F-DCFPyL PET/CT Scans

All PET/CT scans were prospectively interpreted by two board-certified nuclear medicine physicians independently, resolving any disagreements by consensus. A MIM workstation (version 6.5.6, MIM Software) was used to review and annotate axial, coronal, and sagittal PET, CT, and PET/CT images. Areas of abnormal uptake of ¹⁸F-DCFPyL relative to surrounding background that were not associated with physiologic uptake were classified as intraprostatic, lymph node, or distant metastasis. The automated semiquantitative PET parameters included maximum standardized uptake value (SUV_max), mean SUV (SUV_mean), total lesion PSMA (TL-PSMA), and PSMA-derived tumor volume (PSMA-TV), which were derived from an automated ROI generated with a gradient-based segmentation method.

Multiparametric MRI Interpretation

One genitourinary radiologist with over 13 years of experience read the mpMRI scans. Areas suspicious for tumor were assigned a PI-RADSv2 score at a commercial PACS workstation [24]. At the time of mpMRI interpretation, whole prostate volume was measured using automated segmentation software (Dynacad, InVivo). Computer-aided diagnosis capabilities of this platform were not used.

Reference Standard

Patients underwent biopsy in accordance with the relevant institutional protocol. Both systematic and image-guided biopsies were obtained using a standard extended template or a transrectal ultrasound/MRI fusion biopsy platform (Uronav, InVivo). Biopsy procedures were done by a urologist or an interventional radiologist with more than 13 years of experience on the transrectal ultrasound/MRI fusion–guided biopsy platform. All biopsies performed in the study institution were reviewed by a highly experienced genitourinary pathologist with more than 30 years of experience. The mean time between PSMA PET and biopsy was 65 days (median, 61.5 days; range, 6–180 days); the mean time between mpMRI and biopsy was 52 days (median, 33 days; range, 1–179 days). For pathology results, the International Society of Urological Pathology criteria for Gleason grade (GG) groups were used [25]. To confirm that the same region was identified on PSMA PET and mpMRI, T2-weighted MRI was fused using MIM software with TOF attenuation correction PET using pelvic bones as fiducial markers.

Retrospective Quantitative Analysis

ROIs designating lesions with PI-RADSv2 scores of 1 or 2 were delineated on mpMRI by the genitourinary radiologist. This method was used because low-risk imaging features have a high negative predictive value and low false-negative rate for clinically significant disease [3, 26]. ROIs were used to extract SUVs from the TOF attenuation correction PET/CT after fusion with T2-weighted MRI in MIM software.

Statistical Analysis

For all lesions identified on either imaging modality, sensitivity and tumor detection rate of ¹⁸F-DCFPyL and mpMRI were compared using the Wald test on 2000 bootstrap samples at the patient level. Tumor detection rate was defined as the number of true-positives divided by the total number of lesions identified on imaging. For true-positive tumors identified on PET/CT, SUV measurements as well as TL-PSMA and PSMA-TV were correlated with GG using a Kendall τ-b for clustered data [27]. SUV measurements of tumor-positive (defined as GG ≥ 1) areas were also compared with SUV measurements from normal prostate, benign prostatic hyperplasia (BPH), and biopsy-negative intraprostatic regions using a paired Wilcoxon test for clustered data [28].

On a patient level, cumulative intraprostatic tumor burden was calculated for each patient by dividing TL-PSMA by prostate volume to give tumor burden density. PSA density (PSAD) was calculated by dividing PSA by prostate volume. PSAD was correlated with SUV_max and tumor burden density using the Spearman correlation test. All statistical tests were two-sided; a p value less than 0.05 was considered significant.

As a secondary evaluation, to assess whether ¹⁸F-DCFPyL PET has potential as second-line primary imaging, any tumor-positive areas identified on PET but not on MRI were considered an incremental detection contribution of ¹⁸F-DCFPyL PET. PET-detected lymph nodes and distant metastasis suggesting possible upstaging that would change management were also included in incremental detection contribution. Conversely, MRI lesions that were biopsy-positive but not detected on PET were included in tumor detection contribution of mpMRI.

Results

Patient and Lesion Characteristics

Patient and lesion characteristics are given in Table 1. The final study population consisted of 26 patients, with a total of 44 pathology-proven tumors. Of the 44 tumors, the most represented group was GG group 5 (15 tumors, 34.1%). One focus of GG group 1 disease was seen. Extraprostatic ¹⁸F-DCFPyL uptake was seen in eight patients. Six of the eight showed 38 possibly metastatic lymph nodes; the other two patients showed possible bone metastases.

Twelve lesions were identified on imaging that were biopsy negative; they included both benign tissue and high-grade prostatic intraepithelial neoplasia. Of them, seven were identified at PET/CT, two were identified at mpMRI, and three were identified on both PET/CT and mpMRI.

Lesion-Based ¹⁸F-DCFPyL PET Performance

Sensitivities and tumor detection rates for intraprostatic tumor detection on ¹⁸F-DCFPyL PET/CT and mpMRI are given in Table 2. Performance of both modalities was similar, with ¹⁸FDCFPyL PET achieving a sensitivity of 90.9% and a tumor detection rate of 80% (p = 0.58 and 0.17, respectively).

TABLE 2:

Performance of ¹⁸F-DCFPyL PET Versus Multiparametric MRI

Performance Measurement	18F-DCFPyL PET/CT	Multiparametric MRI	p a
Lesion level
Sensitivity (%)
Overall (n = 56)	90.9	86.4	0.58
Peripheral zone (n = 31)	90.9	81.8
Transition zone (n = 21)	88.9	88.9
Tumor detection rate (%)
Overall (n = 56)	80.0	88.4	0.17
Peripheral zone (n = 31)	74.1	81.8
Transition zone (n = 21)	84.2	94.1
Patient level
Incremental tumor detection	5	3
Incremental node detection	6	0
Incremental metastasis detection	2	0
No. (%) of patients with potential detection contribution	11 (42.3)b	3 (11.5)

Note—¹⁸F-DCFPyL = (2S)-2-[[(1S)-1-carboxy-5-[(6-(^tF)fluoranylpyridine-3-carbonyl)amino]pentyl]carbamoylamino]pentanedioic acid.

^aDerived from the Wald test on bootstrapped samples and summary statistics for zonal based analysis.

^bOne patient had both additional lymph nodes and additional true-positive tumor detected on PET/CT, reflected in the sum of patients who benefited from ¹⁸F-DCFPyL PET/CT.

For tumor-positive lesions, SUV_max (mean ± SD, 15.95 ± 12.26; range, 3.6–48.62), TL-PSMA (mean, 59.55 ± 144.43 [g/mL × cm³]; range, 0.16–757.30 [g/mL × cm³]), and PSMA-TV (mean, 4.56 ± 8.52 cm³; range, 0.05–38.9 cm³) were correlated with GG groups, with values by GG group given in Table 3. All three variables trended upward with increasing GG, with the highest mean SUV_max and TL-PSMA seen in GG group 4 and highest PSMA-TV in GG group 5. Overall, SUV_max showed weak correlation with GG (τ = 0.14 [95% CI, –0.11 to 0.40], p = 0.30), whereas PSMA-derived parameters correlated moderately (TL-PSMA: τ = 0.27 [95% CI, –0.01 to 0.56], p = 0.08; PSMA-TV: τ = 0.30 [95% CI, 0.01–0.59], p = 0.06) (Table 3 and Fig. 1).

TABLE 3:

SUV_max and PSMA-Derived Parameters in True-Positive Tumor Cases

Parameter	Overall	Gleason Grade Group					t(95% CI)a	p a
		1	2	3	4	5
SUVmax
Meanb	15.95 ± 12.26	6.40	9.34 ± 6.56	17.04 ± 10.67	19.76 ± 16.94	18.70 ± 12.86	0.14	0.30
95% CI	3.60–48.62		3.60–24.90	4.90–30.00	5.90–48.62	3.80–45.20	−0.11 to 0.40
TL-PSMA
Meanb	59.55 ± 144.43	1.56	6.11 ± 8.77	28.62 ± 32.17	124.20 ± 279.51	83.48 ± 136.60	0.27	0.08
95% CI	0.16–757.30		0.16–28.60	2.00–88.80	6.12–757.30	1.38–435.68	−0.01 to 0.56
PSMA-TV
Meanb	4.56 ± 8.52	0.30	0.89 ± 0.83	2.37 ± 2.76	6.71 ± 11.81	7.33 ± 10.71	0.30	0.06
95% CI	0.05–38.90		0.05–2.30	0.40–8.30	1.20–33.42	0.30–38.90	0.01–0.59

Note—SUV_max = maximum standardized uptake value, PSMA = prostate-specific membrane antigen, TL = total lesion, TV = tumor volume.

^aCorrelation is derived using the Kendall t-b method for clustered data as described by Shih et al. [27].

^bValues provided are mean or mean ± SD.

Fig. 1—

Total lesion prostate-specific membrane antigen (TL-PSMA) level and PSMA tumor volume (PSMA-TV) correlation with Gleason grade group.

A and B, Box-and-whisker plots show measurements of TL-PSMA (A) and PSMA-TV (B) for each lesion (solid dots). Horizontal lines within boxes denote medians; top and bottom lines denote mean + SD and mean – SD, respectively; whiskers denote upper and lower values of 95% CIs. Circles denote outliers. Correlation values are reported in Table 3.

Characterization of Tumor Tissue

Figure 2 presents a comparison of tumor-positive tissue to normal prostate, BPH, and biopsy-negative lesions. SUV measurements in BPH (mean SUV_max, 3.38 ± 0.70; SUV_mean, 1.98 ± 0.50) and normal prostate (mean SUV_max, 2.6 ± 0.51; SUV_mean, 1.43 ± 0.43) were significantly lower than the uptake in tumors (mean SUV_max, 15.95 ± 12.26; SUV_mean, 8.91 ± 5.96) (p < 0.001). BPH tissue showed no visually perceivable uptake on ¹⁸F-DCFPyL PET (Fig. 3). Biopsy-negative tissue showed significantly lower maximal uptake than tumor tissue (mean SUV_max, 5.69 ± 1.17; p = 0.04) (Table 4).

Fig. 2—

Comparison of standardized uptake value (SUV) in tumor (defined as tissue with Gleason grade of 1 or more) and nonmalignant tissue. Double asterisk denotes p < 0.001. Single asterisk denotes p < 0.05. BPH = benign prostatic hyperplasia.

A and B, Box-and-whisker plots show comparisons of maximum SUV (A) and mean SUV (B) for each lesion (solid dots). Horizontal lines within boxes denote medians; top and bottom lines denote mean + SD and mean – SD, respectively; whiskers denote upper and lower values of 95% CIs. Circles denote outliers.

Fig. 3—

Examples of intraprostatic cancer localization with (2S)-2-[[(1S)-1-carboxy-5-[(6-(¹⁸F)fluoranylpyridine-3-carbonyl)amino]pentyl]carbamoylamino]pentanedioic acid (¹⁸F-DCFPyL) PET/CT and with multiparametric MRI (mpMRI).

A, 70-year-old man (prostate-specific antigen [PSA] level, 5.18 ng/mL) who underwent mpMRI followed by ¹⁸F-DCFPyL PET/CT 8 days later. Lesion (Prostate Imaging Reporting and Data System version 2 score, 5) was reported in right mid base peripheral zone (yellow arrows) on mpMR images (middle left, axial T2-weighted image; middle right, apparent diffusion coefficient map; bottom left, DW image [b value, 2000 mm/s²]; bottom right, dynamic contrast-enhanced image). On PET (top left) and PET/CT (top right) images, area of high uptake in right mid base peripheral zone is identified independently (white and black arrows) with maximum standardized uptake value (SUV_max) of 24.9, total lesion (TL) prostate-specific membrane antigen (PSMA) level of 25.11, and molecular tumor volume (MTV) of 1.75 cm³. Benign prostatic hyperplasia in transition zone and left-sided benign peripheral zone tissue do not show ¹⁸F-DCFPyL uptake. Biopsy yielded Gleason 4 + 5 (Gleason grade [GG] group 5) prostate adenocarcinoma.

B, 76-year-old man (PSA level, 15.07 ng/mL) with area of relatively high uptake (white and black arrows) seen in right mid base peripheral zone on ¹⁸F-DCFPyL PET/CT images (top row) (SUV_max = 3.8, TL-PSMA = 4.16, MTV = 1.6 cm³). No obvious lesion was seen on any sequence in corresponding prostate slices from mpMRI examination (middle and bottom rows) performed 5 months later, which was interpreted as negative. Biopsy of area identified on PET/CT images yielded Gleason 4 + 5 (GG group 5) prostate adenocarcinoma.

TABLE 4:

SUV Measurements for Tumor, Biopsy-Negative Tissue, BPH, and Normal Prostate Tissue

Value	Tumor	Biopsy-Negative Tissue	BPH	Normal Prostate Tissue
No. of lesions	44	12	15	13
SUVmax
Median	9.60	5.80	3.50	2.70
Mean ± SD	15.95 ± 12.26	5.69 ± 1.17	3.38 ± 0.70	2.60 ± 0.51
pa		0.04	< 0.001	< 0.001
SUVmean
Median	6.45	4.50	1.93	1.45
Mean ± SD	8.91 ± 5.96	4.30 ± 0.93	1.98 ± 0.50	1.43 ± 0.43
pa		0.126	< 0.001	< 0.001

Note—Tumor was defined as tissue with a Gleason grade group of 1 or more. SUV = standardized uptake value, BPH = benign prostatic hyperplasia, SUV_max = maximum SUV, SUV_mean = mean SUV.

^aReported as a comparison of tumor to nonmalignant tissue.

Patient-Based ¹⁸F-DCFPyL PET Performance

On a patient level, ¹⁸F-DCFPyL PET had a sensitivity of 100%, whereas mpMRI had a sensitivity of 96% for detection of tumor in patients with clinically significant cancer. PSAD (mean, 0.45 ± 0.39 ng/mL² [range, 0.09–2.03 ng/mL²]) correlated moderately with patient-level SUV_max (mean, 19.11 ± 12.26 [range, 3.8–48.62], r = 0.37, p = 0.06) and tumor burden density (mean, 1.59 ± 2.41 [range, 0.06–8.96], r = 0.48, p = 0.01) (Fig. 4).

Fig. 4—

Correlation of tumor burden density and maximum standardized uptake value (SUV_max) with prostate-specific antigen (PSA) density.

A, Scatterplot shows correlation of tumor burden density (per-patient intraprostatic total lesion prostate-specific membrane antigen divided by prostate volume) with PSA density (r = 0.47, p = 0.01).

B, Scatterplot shows correlation of per-patient SUV_max (highest intraprostatic maximum uptake) with PSA density (r = 0.37, p = 0.06).

Potential Incremental Contribution of ¹⁸F-DCFPyL PET Over Multiparametric MRI

Detection differences between PET/CT and mpMRI are given in Table 2. PET/CT findings indicated possible detection contribution in 11 patients (42.3%). Of them, five patients had pathology-proven tumors that were not detected by mpMRI.

In one patient, mpMRI was equivocal and read as negative, but ¹⁸F-DCFPyL PET showed high uptake in an area that was ultimately positive for Gleason 4 + 5 prostate adenocarcinoma (Fig. 3).

Discussion

In a population with high-risk prostate cancer, ¹⁸F-DCFPyL PET provided a slight cancer detection benefit over mpMRI and was reliably sensitive in detecting intra-prostatic tumors with a remarkably lower level of background uptake in normal prostate and BPH tissue. Additionally, PSMA-derived parameters provided a relative estimate of tumor burden. A few other clinical investigations of ¹⁸F-DCFPyL PET have been conducted and have showed uptake in high-grade primary tumors and in biochemically recurrent disease [29–31]. Our results suggest a potential role for ¹⁸F-DCFPyL PET as a second-line method of assessing primary prostate cancer, especially in patients with negative mpMRI results but persistently elevated serum PSA levels, and opportunities to measure the total burden of disease for treatment planning.

Diagnosis of primary prostate cancer has greatly improved with the advent of mpMRI, which has led to better visualization of tumors and improved targeted biopsies and in turn to greater detection of clinically significant cancer [2, 32]. It has also been useful for focal therapy. However, mpMRI still suffers from variable image quality, interreader variability, and low specificity [3, 33–36]. Additionally, although mpMRI is generally an excellent modality for detection of intraprostatic tumors, some clinically significant cancers are still missed or under-estimated [22]. In our study, in one patient out of 26, mpMRI was read as negative, but ¹⁸F-DCFPyL PET/CT showed a tumor that proved to be GG group 5 at biopsy. Overall, ¹⁸F-DCFPyL PET/CT showed a potential incremental detection benefit that could change management in 42.3% of imaged patients, 45% of whom had at least one focus of clinically significant intraprostatic prostatic cancer seen on ¹⁸F-DCFPyL PET/CT but not on mpMRI. Thus, PSMA PET agents could be used to detect prostate cancers, assist in staging, and direct therapy.

In addition to patient-level benefits, ¹⁸F-DCFPyL PET showed a greater than 90% sensitivity for localization of intraprostatic tumors. This high sensitivity is a marked improvement from the first generation ¹⁸F compound, (N-[N-[(S)-1,3-dicarboxypropyl] carbamoyl]-4-F-fluorobenzyl-L-cysteine) (¹⁸F-DCFBC), which had a sensitivity of only 36% in a similar high-risk population [37]. Mean SUV_max with ¹⁸F-DCFBC PET was also lower, at 5.8 ± 4.4 versus 15.95 ± 12.26 for ¹⁸F-DCFPyL PET [37]. The new tracer’s high affinity for PSMA facilitates clear definition of tumor location. In a smaller cohort of patients than ours, Bauman et al. [31] also found that dominant intraprostatic lesions were well localized using this tracer. These results have implications for focal treatment targeting with radiation or ablative technologies and for posttreatment surveillance.

Multiparametric MRI has been used to generate prognostic biomarkers. Among them are calculated PSAD from MRI-derived prostate volume and PI-RADSv2 score; however, the latter can be relatively subjective [38–40]. PET-based biomarkers could be more quantitative. In this study, when normalized to prostate volume, cumulative tumor burden on PSMA PET correlated with serum PSA level in patients with high-risk disease. Schmuck et al. [41] found similar results using ⁶⁸Ga-PSMA with TL-PSMA and PSMA-TV as markers of whole-body tumor burden, recommending that this could assist in determining prognosis and facilitating therapy monitoring. Also, because PSMA expression is regulated independently of PSA, PSMA could serve as an additional biomarker in biochemically recurrent disease [9]. Of note, TL-PSMA and related PSMA-TV are derived from PET PSMA expression and mirror the commonly used FDG PET parameters of total lesion glycolysis and metabolic tumor volume in the setting of PSMA imaging. Both TL-PSMA and PSMA-TV showed correlation with increasing GG group, suggesting these measurements could correlate with severity of disease. In our limited patient population and accounting for clustered data, establishing statistical significance is difficult. However, with a larger dataset, more specific quantitative parameters could be identified.

PSMA is widely considered a promising target for prostate cancer, given consistent data showing that it is overexpressed in more aggressive prostate cancer. However, PSMA is also expressed in normal prostate tissue and BPH and is associated with neovascularity, which can be present in many types of tumors [7, 42]. In developing a PSMA binding ligand, there is concern that it will show confounding high expression in these regions. In normal prostate tissue and BPH, uptake could mask primary tumor, as occurs with other prostate PET agents. Our data show that BPH and normal tissue have low SUVmean with ¹⁸F-DCFPyL (both < 2). SUV_mean has long been used in liver imaging as a reliable measure of normal tissue, because it gives a better idea of overall background uptake rather than maximal uptake at a single point [43]. Additionally, ¹⁸F-DCFPyL as a ligand is effective at differentiating between prostate cancer, normal prostate, and BPH, with high significance. Biopsy-negative tissue also showed lower quantitative measurements than biopsy-positive tissue, but differences in SUV_mean were less significant. Of the 10 biopsy-negative lesions seen on PET/CT, three also showed abnormalities on mpMRI, and some showed high-grade prostatic intraepithelial neoplasia on pathology. The focus of our study was the ability to detect cancer, but elements of this “negative” tissue remain to be explored to determine prognostic significance.

This study had several limitations. First, imaging and analysis were performed only with patients who had high-risk disease, naturally skewing tumors toward higher grade disease and likely corresponding to areas of high PSMA expression. However, this study design reflects real-life application, because these patients are the ones who would likely be offered advanced PET. Second, biopsy is an inherently limited reference standard compared with whole-mount histopathology. The biopsy system used has been proven to have an accuracy of within 3 mm of lesion targets [44]. Our findings suggest that the benefits of this tracer in posttreatment monitoring can be explored in patients who are not recommended for radical prostatectomy. Additionally, without radical prostatectomy specimens, retrospective measurement of normal prostate and BPH is limited. For lesions with PI-RADSv2 scores of 1 and 2, our rate of clinically significant cancer detection is 0–10%, which we were able to leverage to provide comparisons to background tissue. Finally, ¹⁸F-DCFPyL was produced at the National Cancer Institute in this study, and it is currently not available outside the research setting. However, one or more PSMA PET agents are anticipated to be approved in the United States soon.

In conclusion, use of ¹⁸F-DCFPyL PET provided sensitivity and tumor detection rate comparable with that of mpMRI for intra-prostatic tumor localization and may provide additional cancer detection benefit. The tracer has much higher affinity for tumor than for surrounding normal tissue and BPH, and PSMA-derived measurements of tumor burden appear to correlate with severity of disease. These findings should inform imaging of patients with high-risk disease and should continue to be validated as more prospective ¹⁸F-DCFPyL PET trials accrue patients.