Small Molecule, Multimodal [¹⁸F]-PET and Fluorescence Imaging Agent Targeting Prostate Specific Membrane Antigen: First-in-Human Study

Abstract

Purpose:

To conduct a first-in-human study of [¹⁸F]-BF3-Cy3-ACUPA by determining its safety, biodistribution, radiation dosimetry, and feasibility of tumor detection by preoperative positron emission tomography (PET) as well as intraoperative fluorescence imaging in patients with prostate-specific membrane antigen (PSMA) positive tumors.

Methods:

Ten patients aged 66 ± 7 years received a 6.5 ± 3.2 mCi intravenous injection of [¹⁸F]-BF3-Cy3-ACUPA and underwent PET/computed tomography (PET/CT) imaging. Radiation dosimetry of [¹⁸F]-BF3-Cy3-ACUPA, normal organ biodistribution, and tumor uptakes were examined. Two patients were pre-scheduled for radical prostatectomy (RP) with extended pelvic lymphadenectomy ~ 24 hours following [¹⁸F]-BF3-Cy3-ACUPA injection and imaging. Without reinjection, intraoperative fluorescence imaging was performed on freshly excised tissue during RP. Frozen sections of excised tissue during RP underwent confirmatory histopathology and multiphoton microscopy.

Results:

Absorbed doses by organs including the kidneys and salivary glands were similar to ⁶⁸Ga-PSMA-11 imaging. [¹⁸F]-BF3-Cy3-ACUPA physiologic radiotracer accumulation and urinary/biliary excretion closely resembled the distribution of other published PSMA tracers including [¹⁸F]-JK-PSMA-7, [¹⁸F]-PSMA-1007, [¹⁸F]-DCFPyL, [¹⁸F]-DCFBC. ¹⁹F-BF3-Cy3-ACUPA was retained in PSMA⁺ cancer in patients for at least 24 hours, allowing for intraoperative fluorescence assessment of the prostate and of embedded prostate cancer without contrast reinjection. After 24 hours, the majority had decayed or cleared from the blood pool. Preoperative PET and fluorescence imaging findings were confirmed with final histopathology and multiphoton microscopy.

Conclusions:

Our first-in human results demonstrate that [¹⁸F]-BF3-Cy3-ACUPA is both safe and useful in humans. Larger trials with this PET tracer are expected to further define its capabilities and its clinical role in the management of PSMA⁺ tumors, especially in prostate cancer.

Microabstract

A number of radioactive tracers are currently being actively examined to image prostate-specific membrane antigen (PSMA) positive tumors, especially primary and recurrent prostate cancer. We showed a first-in-human study of [¹⁸F]-BF3-Cy3-ACUPA by determining its safety, biodistribution, radiation dosimetry, and feasibility of tumor detection by preoperative positron emission tomography (PET) as well as intraoperative fluorescence imaging in patients with PSMA positive tumors.

INTRODUCTION

Prostate cancer (PCa)-specific positron emission tomography (PET) agents are increasingly used to indicate patient candidacy for radical prostatectomy (RP), distinguish localized from disseminated disease, and monitor treatment response. In recent years, many new PET agents targeting prostate-specific membrane antigen (PSMA) have been developed and clinically translated in light of the frequent overexpression of PSMA in PCa. However, while such PET agents, notably ⁶⁸Ga-PSMA-11, have become invaluable in the diagnosis, staging, and treatment planning of PCa, they are of less value for intraoperative guidance during RP, the most commonly used curative intervention in PCa. One important aspect during surgical resection of PCa is to determine lymph node status. As lymph node metastases can be only a few millimeters in diameter, they can escape detection on PET using the currently available agents due to limited spatial resolution of the modality.

Contrast agents that can demarcate small, early-stage metastases during surgery are critically needed to ensure curative RP. It has been reported that 4–48% of patients will have recurrent disease after RP, bringing along considerable health and cost implications ^1-6. To address the aforementioned need, preoperative planning using PET imaging followed by intraoperative fluorescence imaging during surgery have been explored to improve the clinical success of RP, reduce biochemical recurrence, and reduce costs associated with disease recurrence ⁷. However, most efforts in this area have focused on using single modality agents for preoperative PET imaging and intraoperative fluorescence imaging, respectively; this requires multiple injections, and moreover, since the same receptor is targeted, this can lead to PSMA receptor saturation which may cause false negatives and result in inadequate intervention ^8-14.

In contrast to a single modality agent approach, a multimodality agent approach combining PET and fluorescence into a single molecule have been alternatively proposed ^14,15. In this vein, our group has developed [¹⁸F]-BF3-Cy3-ACUPA, a PSMA⁺ PCa-specific multimodal positron-emitting (¹⁸F) and fluorescence imaging agent ^16,17. Our probe uniquely employs an alkylammonio methyl trifluoroborate (AMBF₃) fluoride isotopic exchange (IEX) trap to allow rapid and direct ¹⁸F labeling to produce [¹⁸F]-BF3-Cy3-ACUPA ^18-23. Since the clinical potential of the AMBF₃ IEX ¹⁸F-PET prosthetic group has never been evaluated in patients, in this study, we explored the first-in-human use of [¹⁸F]-BF3-Cy3-ACUPA by determining its safety, biodistribution, radiation dosimetry, and feasibility of tumor detection by preoperative PET and intraoperative fluorescence imaging.

MATERIALS AND METHODS

Patients

This study was approved by the local Institutional Ethical Committee (Cerrahpasa University School of Medicine protocol #-29281604.01.01-139206). Written informed consent was obtained from all patients to receive [¹⁸F]-BF3-Cy3-ACUPA. All procedures were performed in accordance with the ethical standards of the institution, national research committee and with the 1964 Helsinki declaration and its later amendments. Patient studies were conducted between October 2018 and December 2019 at Istanbul University- Cerrahpasa, Cerrahpasa Medical Faculty. The inclusion criteria were suspected PCa (including a prior digital rectal examination, elevated prostate-specific antigen (PSA) serum level, and confirmed PCa (existing TNM and/or Gleason-scored biopsy) as well prior PSMA⁺ non-prostate cancers.

To evaluate the safety of [¹⁸F]-BF3-Cy3-ACUPA administration, vital parameters such as body temperature, blood pressure, and heart rate were measured at multiple time points up to 4 h after injection.

Radiosynthesis and Quality Control of [¹⁸F]-BF3-Cy3-ACUPA

¹⁹F-BF3-Cy3-ACUPA (Fig. 1) was synthesized and radiolabeled in 140 μg (100 nmol) aliquots as previously described ^16,17. Radiolabeling, purification, and preparation of the agent was performed in a certified Good Laboratory Practice (GLP) facility (Istanbul University-Cerrahpasa, Cerrahpasa Medical Faculty). Pharmaceutical grade [¹⁸F]-fluoride ion (for bone imaging) was used for isotopic exchange. On average, IEX was conducted with 152 ± 45 mCi of [¹⁸F]-fluoride ion and resulted in 11.6 ± 7.7 mCi of [¹⁸F]-BF3-Cy3-ACUPA (decay uncorrected). The obtained specific activity was greater than 0.116 ± 0.077 mCi/nmol. Radiolabeling yield and purity was confirmed by radio-high performance liquid chromatography (HPLC), where an average HPLC purity of 94 ± 11 % was observed. Products were tested for sterility, endotoxins, pH, fluorescence, and absence of precipitates. All materials for injection were passed through sterile 0.2 μm syringe filters prior to patient injection.

Fig.1. Structure and radiosynthesis.

A) Structure of [¹⁸F]-BF3-Cy3-ACUPA, containing AMBF3, Cy3, and PSMA targeting moieties, that was radiolabeled through IEX radiochemistry. B) Radio-HPLC showing a pure [¹⁸F]-BF3-Cy3-ACUPA radiosynthesis (Patient No. 1). C) Chemical structures of other contemporary standalone PET-only [¹⁸F]-PSMA tracers including [¹⁸F]-JK-PSMA-7, [¹⁸F]-DCFPyL, [¹⁸F]-PSMA-1007, [¹⁸F]-DCFBC, and [⁶⁸Ga]-PSMA-11.

[¹⁸F]-BF3-Cy3-ACUPA PET/CT and PET/MRI Acquisition

All ten patients underwent PET/computed tomography (CT) using a Discovery 710 PET/CT scanner (GE Healthcare). Additionally, pelvic PET/magnetic resonance imaging (MRI) was performed in six patients using a Signa PET/MRI scanner (GE Healthcare). Whole-body PET images were acquired from the top of the skull to the mid-thigh. Patients received an average of 6.5 ± 3.2 mCi of [¹⁸F]-BF3-Cy3-ACUPA through the cephalic vein by simple injection using a shielded syringe in 3–5 mL saline. An average time of 73 ± 27 min was allowed to pass between injection and imaging acquisition. PET data was acquired using QuantWorks (GE Healthcare).

Emission data was corrected for random scatter and decay, and reconstruction was performed using the ordered subset expectation maximization (OSEM) algorithm. Image co-registration was performed for PET and CT datasets. Attenuation correction was performed using unenhanced low-dose CT data (120 kV; 59 mAs) for PET/CT acquisitions, whereby CT images were reconstructed to a slice thickness of 3.75 mm, increment of 0.12 mm, and noise index of 15; for PET/MRI acquisitions, Zero Echo Time (ZTE) MRI sequences focusing on the bone were used for soft tissue attenuation correction of PET images.

Regions of Interest on Preoperative PET Imaging

All PET images were assessed by qualified nuclear radiologists. Regions of interest (ROIs) were drawn around PSMA⁺ lesions (areas of focally increased uptake). Additionally, ROIs with 20 mm diameter spherical volumes (4.19 mL) were manually drawn around other organ tissues of interest including the right/left (R/L) lacrimal, R/L parotid, R/L submandibular, urinary bladder, red bone marrow, gallbladder, adrenal, brain, breast, upper/lower large intestine wall, stomach wall, kidney, liver, lungs, muscle, pancreas, osteogenic cell, skin, spleen, testes, and thyroid.

Radiation Dosimetry and Tracer Biodistribution

Radiation dosimetry analysis including absorbed and effective dose calculations were performed using OLINDA/EXM 1.0 software (Vanderbilt University). Total counts were obtained by adding all counts from manually drawn ROIs on individual CT slices.

Tracer biodistribution to PSMA⁺ lesions and other organ tissues of interest was determined by calculating the mean standard uptake value (SUV_mean) within the aforementioned ROIs across sequentially acquired PET images using the OLINDA/EXM 1.0 or Amide 1.0.4 software. Blood radioactivity concentration was measured in the left ventricle of the heart.

Intraoperative Fluorescence Imaging

Of the ten patients included in this study, two patients were pre-scheduled for RP with extended pelvic lymphadenectomy without reinjection ~ 24 hours following [¹⁸F]-BF3-Cy3-ACUPA injection and preoperative imaging. There is no commercially available fluorescence imaging device for wide-field fluorescence imaging of Cy3 fluorophore-containing tissue in the operating room (OR). Therefore, a first-generation custom fluorescence interference filter block device was assembled and placed on the OR back table. This device contained a Solis 525C LED illuminator (Thor Labs) with a dominant wavelength of 525 nm and a typical collimated optical power of 3.1W. The illuminating light was focused through a plano-convex lens through a fluorescence interference filter block containing a dichroic mirror with a 533–580 nm reflected band and a 595–800 nm transmission band. A 600 ± 40 nm bandpass barrier filter was used to control emitted light that was transmitted to an attached color Complementary Metal Oxide Semiconductor (CMOS) camera. An exciter filter was not used in this setup because of potential concern that important anatomical landmark data would be lost if optical filtration was too stringent. This device was used to collect fluorescence images and videos of whole, intact, fresh, resected tissue and associated margins immediately following surgical resection to qualitatively ascertain the fluorescence ¹⁹F-BF3-Cy3-ACUPA signal. Imaging was performed ex vivo and did not interfere with standard-of-care RP. TIFF files were captured using ThorCam Software v3.2.0 and converted to .AVI using Image J v.1.51j8. ThorCam Software v3.2.0 was used to transform red channel CMOS data into black and white images.

Histopathological Confirmation

Following RP, excised tissue was submitted to pathology where frozen sections were first prepared from excised ^18/19F-BF3-Cy3-ACUPA-containing tissue and then stained for immunohistochemistry to confirm preoperative PET and intraoperative fluorescence colocalization within PSMA⁺ tissue under a confocal microscope. Histopathological evaluation was performed by dedicated uropathologists, using International Society of Urological Pathology nomenclature, blinded to preoperative PET/CT, preoperative PET/MRI (Patient 6 shown in Fig. S5), and intraoperative fluorescence imaging results. Human prostate tissue sections were stained with DAPI (blue, nuclear stain). Control, neighboring sections were stained with hematoxylin (violet, nuclear stain), eosin (light blue, cytoplasm stain), and horse-radish-peroxidase (HRP) conjugated PSMA⁺ antibody, and then treated with 3,3'-Diaminobenzidine (brown, oxidized polybenzimidazole product).

Confirmation by Post-Surgical Multiphoton Microscopy

Post-surgical Cy3 fluorescence imaging was also performed using multiphoton microscopy (MPM) on excised, frozen tissue sections with an Olympus FV1000 multiphoton microscope at the Microscopy CoRE and Advanced Bioimaging Center at the Icahn School of Medicine at Mount Sinai, at room temperature and at 780 nm excitation. First, intact biopsy cores were examined at low resolution for fast tissue orientation and gross PSMA+ area assessment (z-resolution = 1.1 μm/pixel, NA objective = 4×/0.4). Next, intact biopsy cores were examined at high resolution tissue to yield a tiled image (2D optical section at a focal plane approximately 20 μm below tissue’s surface) (z-resolution = 0.48, 4, and 1.2 μm/pixel; NA objective = 25×/1.0).

Statistical Analysis

All errors were reported as ± standard deviation. SigmaPlot 10.0 software was used for statistical analysis of dosimetry data including the calculation of median, mean, minimum, maximum, 25^th percentile, and 75^th percentile values. Dosimetry data of [¹⁸F]-BF3-Cy3-ACUPA was compared to that of contemporary [¹⁸F]-PET tracers including [¹⁸F]-JK-PSMA-7, [¹⁸F]-PSMA-1007, [¹⁸F]-DCFPyL, and [¹⁸F]-DCFBC. To determine whether differences in dosimetry data were significant between [¹⁸F]-BF3-Cy3-ACUPA and the other tracers, P-values were calculated using the following equation with a two-tailed t distribution:

$$t=\frac{\mid {\chi }_1-{\chi }_2\mid }{\sqrt{(\frac{s_1^2}{n_1}+\frac{s_2^2}{n_2})}},$$

where t is the statistical value used to calculate p; χ₁, s₁, and n₁ are the published average dosimetry value, standard deviation, and sample size, respectively, for each tested tissue using [¹⁸F]-JK-PSMA-7, [¹⁸F]-PSMA-1007, [¹⁸F]-DCFPyL, or [¹⁸F]-DCFBC; and χ₂, s₂, and n₂ are the average dosimetry value, standard deviation, and sample size (n = 10) for [¹⁸F]-BF3-Cy3-ACUPA measured in this study. P-values < 0.05 were considered statistically significant.

Tumor-to-blood ratios (TBR) were calculated by the dividing SUV_mean values of lesions by the SUV_mean value in the red bone marrow for individual patients.

RESULTS

Patient Characteristics

A total of ten patients were included in this study (mean age = 66 ± 7 years). Detailed patient characteristics are provided in Table 1. Nine patients (Patients 1–3, 5–10) were candidates for [¹⁸F]-BF3-Cy3-ACUPA PET imaging because of clinically suspected PCa; an additional patient (Patient 4) with PSMA⁺ colorectal cancer and was included in the study to achieve an adequate sample size for biodistribution measurements. Of the nine patients with clinically suspected PCa, PSA levels were measured immediately prior to [¹⁸F]-BF3-Cy3-ACUPA PET imaging, and the mean elevated serum PSA level was 31 ± 63 μg/L. Eight of these patients had undergone prostate biopsies yielding Gleason scores > 4+3 (or 3+4). The same eight patients had also undergone CT and/or MRI a minimum of 3 weeks prior to the study and received TNM classifications based on CT and/or MRI. As PET studies were not available at the time of TNM classification, TNM classifications as reported in Table 1 may not corroborate with the data obtained from [¹⁸F]-BF3-Cy3-ACUPA PET/CT and/or PET/MRI.

Table 1.

Patient Characteristics

Patient No.	Age	Diagnosis (post biopsy) *	Gleason	TNM (Clinical) ┼	PSA (time of scan, ug/L)	Mets Imaged	Dose (mCi)	Imaging Modality	Imaging Time Point(s) (min post injection)	HPLC (%)	Starting Activity [18F]- fluoride ion (mCi)	Recovered Labeled Product (mCi)
1	72	Metastatic Prostate Cancer	4+3	T2bN1M1	199	Lung, adrenal, bone, LN	3.25	GE 710 PETCT	Multiple, 9-119 min	99.1	120	3.5
2	60	Prostatectomize d LN met	5+3	T2aN0M0	12	LN	2.79	GE 710 PETCT+MR	Multiple, 6-121 min	74	182	3.18
3	62	No Biopsy	n.a.	n.a	5.14	N	7.28	GE 710 PETCT+MR	46 min	99.8	101	17
4	57	Histology confirmed PSMA+ Cancer	n.a.	n.a.	n.a.	LN	6.53	GE 710 PETCT+MR	60 min	99.8	101	17
5	65	Prostatectomize d LN mets	4+3	T3bN0M0	0.25	LN	3.20	GE 710 PETCT	64 min	98.3	224	13.2
6	74	Prostate Cancer	4+4	T3bN1M0	8	LN	9.84	GE 710 PETCT+MR	132 min	72	160	22
7	65	Metastatic Prostate Cancer	4+3	T3N1M0	12	Prostate ECE	5.00	GE 710 PETCT+MR	85 min	98.3	224	13.2
8	73	Prostate Cancer	3+4	T2aN0M0	10.5	N	5.66	GE 710 PETCT	78 min	99.5	134	11
9	74	Prostate Cancer	4+4	T3aN1M0	6.46	LN	10.99	GE 710 PETCT	60 min	99.8	152	32.5
10	57	Prostate Cancer	4+3	T3bN0M0	32	LN, EPI	10.89	GE 710 PETCT+MR	59 min	99	128	29.2
Average (SD)	66 (± 7)				31 (± 63 ug/L)◊		6.5 (± 3.2 mCi)		Δ Injection to First scan╪ 73 (± 27 min)	94 (± 11%)	152 (± 45 mCi)	11.6 (± 7.7 mCi)

^aDiagnosis made by urologists, independent of [¹⁸F]-PET imaging.

^bSD-variation due to high PSA value in Patient 1.

^cPatients receiving multiple scans were not included in the calculation.

Adverse Events

All patients tolerated the drug product well. No signs or symptoms of adverse effects from injection of [¹⁸F]-BF3-Cy3-ACUPA tracer were reported.

Biodistribution

Maximum intensity projection (MIP) images of [¹⁸F]-BF3-Cy3-ACUPA PET/CT in Patient 1 with widely disseminated PCa and Patients 6 and 8 with localized PCa are depicted in Fig. 2, 3, and 4, respectively. In Patient 1 who had multiorgan metastatic disease involving the bones, lymph nodes, and adrenals, the pattern of physiologic biodistribution was not visually affected by the presence of metastatic disease.

Fig. 2. [¹⁸F]-BF3-Cy3-ACUPA PET/CT imaging of Patient 1 with disseminated PCa.

Patient 1 had undergone RP and was receiving adjuvant luteinizing hormone-releasing hormone (LHRH) analog therapy prior to this study. Patient 1 also had a hip replacement surgery prior to this study. The patient’s serum PSA was 199 μg/L. A) Coronal and B) Sagittal whole-body PET (red)/CT (grey) MIP images show general and focal [¹⁸F]-BF3-Cy3-ACUPA distribution. Specific PCa lesions (“a–d”) are indicated by magenta arrows. C) CT only (inverted, grey) and D) [¹⁸F]-BF3-Cy3-ACUPA PET (red)/CT (inverted, grey) transverse cross-sections show specific PCa lesions (indicated by magenta markings).

Fig 3. [¹⁸F]-BF3-Cy3-ACUPA and [⁶⁸Ga]-PSMA-11 PET/CT imaging in Patient 6.

Patient 6 received [⁶⁸Ga]-PSMA-11 PET/CT one month prior to [¹⁸F]-BF3-Cy3-ACUPA PET/CT. Serum PSA was measured at 8 μg/L prior to receiving [¹⁸F]-BF3-Cy3-ACUPA. A) Coronal PET (grey) MIP image (left) shows general and focal [¹⁸F]-BF3-Cy3-ACUPA distribution (patient with urinary catheter). B) Fused PET (red)/CT (grey, top) or CT only (grey, bottom) transverse cross-sections with [¹⁸F]-BF3-Cy3-ACUPA and C) fused PET (red)/CT (grey, top) or CT only (grey, bottom) transverse cross-sections with [⁶⁸Ga]-PSMA-11 show the same lesions (magenta arrows). D) Coronal PET (grey) MIP image with [⁶⁸Ga]-PSMA-11 shows similar general distribution and focal PCa to A).

Fig. 4. [¹⁸F]-BF3-Cy3-ACUPA PET/CT imaging of Patient 8.

Patient 8 was a candidate for RP based on prior biopsy and [⁶⁸Ga]-PSMA-11 imaging. As part of this study, this patient underwent [¹⁸F]-BF3-Cy3-ACUPA injection and PET/CT imaging, followed by RP ~24 hours later. A) [⁶⁸Ga]-PSMA-11 PET MIP image. B) [¹⁸F]-BF3-Cy3-ACUPA PET MIP image. C) [⁶⁸Ga]-PSMA-11 PET/CT image shows localized PSMA+ disease (arrow) but no nodal or distant disease. D) Similar PET delineations (magenta arrowheads) are observed on the 90-min post-injection [¹⁸F]-BF3-Cy3-ACUPA PET/CT image. Without reinjection, the patient underwent RP 24 hours later (after 18F decayed). E) Fluorescence images of excised prostate tissue using a first-generation back-table Cy3 fluorescence imaging device. F) Post-surgical prostate co-registration shows PCa in an MRI apparent diffusion coefficient image, PET image, and fluorescence image; post-surgical fluorescence image shows a positive margin in the posterior urethra/bladder neck (indicated by a dashed magenta border).

Patients 1 and 2 were repeatedly imaged 6, 31, 61, and 121 min post injection. Fig. S6 shows decay uncorrected, relative percentile changes in organ activity in Patient 2.

Radiation Dosimetry

The total body dose was 7.07 ± 1.49 E-03 mSv/MBq. The effective dose equivalent was 1.30 ± 0.54 E-02 mSv/MBq and the effective dose was 8.87 ± 2.17 E-03 mSv/MBq. The highest organ uptakes and radiation doses were observed in the salivary glands, kidneys, and lacrimal glands Partial organ absorbed dose distribution of [¹⁸F]-BF3-Cy3-ACUPA to different organ tissues are reported in Table S1, Fig. S1, and Fig. S2.

Comparison to Contemporary ¹⁸F-PSMA Tracers

In the entire patient cohort, major physiologic radiotracer accumulation was observed in the salivary and lacrimal glands, liver, spleen and intestines in a pattern resembling the distribution of other [¹⁸F]-PSMA tracers (Table S1). Box plots illustrate the closeness of dosimetry between [¹⁸F]-BF3-Cy3-ACUPA and other published PSMA⁺ tracers in Fig. S1, wherein the dose of [¹⁸F]-BF3-Cy3-ACUPA was statistically similar or lower than that of other published PSMA⁺ tracers. Tracer biodistribution in an expanded organ panel is reported in Fig. S2.

Tumor-to-Blood Ratio of PSMA⁺ Lesions and Lymph Nodes

TBRs are presented in Table 2. One patient had distant metastatic PCa, six patients had PCa nodal metastasis, and one patient had PSMA⁺ colorectal cancer. TBRs were quantified for five lesions within the prostate gland, one bone lesion, one lesion in the adrenal gland, and four lymph nodes. PSMA⁺ lesions within the prostate fossa showed a TBR of 13 ± 11 (SUV_mean). Lymph nodes showed a TBR of 19 ± 21 (SUV_mean).

Table 2.

Tumor-to-blood ratios (TBR) of primary and metastatic lesions

Patient No.	Age	Gleason	TNM (Clinical)┼	PSA (time of scan, ug/L)	TBR - Primary Prostate Lesion/Marrow*	TBR - Metastasis/Marrow*
1	72	4+3	T2bN1M1	199	2.8	34.5 (Bone) 34.6 (Adrenal)
2	60	5+3	T2aN0M0	12	n.a.	n.a.
3	62	n.a.	n.a	5.14	n.a.	n.a.
4	57	n.a.	n.a.	n.a.	n.a.	n.a.
5	65	4+3	T3bN0M0	0.25	n.a.	3.6 (LN)
6	74	4+4	T3bN1M0	8	10.7	3.9 (LN)
7	65	4+3	T3N1M0	12	32.5	50.2 (LN)
8	73	3+4	T2aN0M0	10.5	12.5	n.a.
9	74	4+4	T3aN1M0	6.46	n.a.	31.0 (LN)
10	57	4+3	T3bN0M0	32	7.4	4.8 (LN)

LN = lymph node.

Intraoperative Fluorescence Imaging of Resected PSMA⁺ Lesions and Lymph Nodes

Of the total ten patients, only five patients (Patients 1, 2, 5, 6, and 8) were candidates for RP based on [⁶⁸Ga]-PSMA-11 PET imaging at 10 months, 3 months, 2 months, 1 month, and 1 month prior (Patient 6 shown in Fig. S4). Three patients (Patients 1, 2, and 5) had undergone RP prior to this study. The two remaining patients (Patients 6 and 8) were pre-scheduled for RP with extended pelvic lymphadenectomy ~ 24 hours following [¹⁸F]-BF3-Cy3-ACUPA injection and PET imaging.

In Patients 6 and 8, the superficial presence of PSMA⁺ tissue was suggestive of positive PCa margins (see Movie S1). In Patient 6, four resected lymph nodes were imaged following pelvic lymph node dissection (see Fig. S3). To best observe a visible difference in fluorescence intensity between PCa containing and non-PCa containing nodes, all four nodes were imaged together using same-intensity fluorescence imaging settings. Fluorescence imaging intensity was greater in two of the four removed nodes. All nodes were sent to histopathology for analysis. The two nodes with greater fluorescence imaging were confirmed to contain PCa foci.

Post-Surgical Confirmation by Histopathology and MPM Fluorescence Imaging of Resected PSMA⁺ Lesions

[¹⁸F]-BF3-Cy3-ACUPA left a permanent fluorescence record in frozen sections of excised tissue, allowing confirmation of PSMA⁺ status as determined by preoperative PET imaging and intraoperative fluorescence imaging. In MPM, the emitted two-photon excited signal was detected by three photomultiplier tube (PMT) channels: 1) second harmonic generation (SHG) at 390 nm for collagen specific contrast (Blue); 2) intrinsic fluorescence at 420–460 nm (granular and stromal tissue; green); and 3) Cy3 fluorescence at 500–550 nm for capturing PSMA+ cells throughout the imaged tissue (red). Fig. 5 shows a representative MPM image of a biopsy core, where the red signal indicates Cy3 fluorescence originating from glandular tissue within the prostate. Additional morphological contrast from stromal collagen enhanced the visualization of PSMA⁺ glands in the imaged tissue and provided additional morphological tissue information, not seen by confocal imaging (histopathology) alone.

Fig 5. Corroborative ¹⁸F-BF3-Cy3-ACUPA fluorescence imaging, diaminobenzidine histopathology, and multiphoton microscopy.

Patient 6 underwent [¹⁸F]-BF3-Cy3-ACUPA injection and PET/CT imaging, followed by RP ~24 hours later. A prostate specimen containing BF3-Cy3-ACUPA was excised during RP and sent to histopathology. A) At histopathology, after staining with DAPI (purple, nuclear stain). The specimen containing BF3-Cy3-ACUPA appears pink and is localized to glandular tissue under a confocal microscope. B) A neighboring frozen section was created from the prostate specimen and stained with hematoxylin (violet, nuclear stain), eosin (light blue, cytoplasm stain), horse-radish-peroxidase (HRP) conjugated PSMA⁺ antibody and then treated with 3,3'-Diaminobenzidine (DAB brown, oxidized polybenzimidazole product). DAB produced by the PSMA⁺ antibody is brown and is also localized to glandular tissue. Note the colocalization of ACUPA-Cy3-BF3 (pink signal) with PSMA⁺ immunohistochemistry (right, brown) in glandular prostate tissue. C–F) MPM imaging of unprocessed frozen sections containing ACUPA-Cy3-BF3 capturing endogenous two-photon excited PSMA⁺ fluorescence from Cy3 (red), cellular and elastin intrinsic fluorescence (green), and second harmonic from stromal collagen (Blue). Examples of C–E) PSMA⁺ fluorescence glandular tissue and a F) PSMA⁻ specimen. Note that MPM enhances the visualization of collagen fibers surrounding the prostate glands. Scale bars: A–B) 100 μm, C) 1000 μm, D–F) 100 μm.

DISCUSSION

This was a first-in-human study to investigate the safety, biodistribution, radiation dosimetry, and tumor-detecting ability of multimodal PET and fluorescence [¹⁸F]-BF3-Cy3-ACUPA in patients with PSMA⁺ malignancy. Altogether, our findings demonstrate that [¹⁸F]-BF3-Cy3-ACUPA is safe and useful in the clinical setting.

In our study, we determined that [¹⁸F]-BF3-Cy3-ACUPA dosimetry to normal, healthy tissue was generally lower than that of contemporary [¹⁸F]-PET tracers including [¹⁸F]-JK-PSMA-7, [¹⁸F]-PSMA-1007, [¹⁸F]-DCFPyL, and [¹⁸F]-DCFBC. Specifically, its distribution to the kidney, lung, liver, and splenic tissue was 1.4 to 6-fold lower than that of the published tracers (Table S1). An exception was its distribution to salivary and lacrimal tissue, which was similar (0.5–1.6×) to that of the published tracers; this may be due to the specific targeting of natural levels of PSMA expression in lacrimal and salivary gland tissue. [¹⁸F]-BF3-Cy3-ACUPA distribution to the bone and marrow was minimal, indicating that it does not defluoridate to produce significant concentrations of [¹⁸F]-fluoride ions in humans. Dosimetry findings, taken together, suggest that [¹⁸F]-BF3-Cy3-ACUPA may have superior in-patient PSMA⁺ affinity compared with contemporary [¹⁸F]-PSMA tracers. A larger patient sample is needed to confirm this hypothesis.

In this study of ten patients, nine had clinically suspected PCa. We also showed that [¹⁸F]-BF3-Cy3-ACUPA can be used to accurately delineate PCa in a diverse group of nine patients including those with metastatic disease, those undergoing RP, and those with suspected PCa, with varying PSA serum levels, Gleason scores, and TNM staging. PSMA⁺ PCa was visible in the prostate gland (n = 5) and lymph nodes (n = 6) on [¹⁸F]-BF3-Cy3-ACUPA PET imaging. Like other [¹⁸F]-PSMA tracers, [¹⁸F]-BF3-Cy3-ACUPA PET was useful in distinguishing between disseminated and localized PSMA⁺ cancer. The TBR of [¹⁸F]-BF3-Cy3-ACUPA for small lymph node metastases ranged from 3.6 to 50.2, suggesting that [¹⁸F]-BF3-Cy3-ACUPA may be exceptionally sensitive for identifying tiny PSMA⁺ tumor deposits in the body.

¹⁹F-BF3-Cy3-ACUPA is also useful for fluorescence imaging within ~24 h of injection to confirm PSMA⁺ cancer in the intraoperative (back-table) setting. In light of this, the [¹⁸F]-BF3-Cy3-ACUPA tracer stands apart from other contemporary [¹⁸F]-PSMA tracers which lack fluorescence utility and are therefore not useful in the intraoperative setting. The ~24 h temporal delay between injection and preoperative PET imaging and intraoperative fluorescence imaging is advantageous for three reasons: 1) The delay gives urologists more time to assess the PET image in determining the extent of lymph node dissection. 2) The delay removes ionizing radiation (from ¹⁸F decay) from the OR procedure (¹⁸F has a half-life of 108 min). 3) The delay allows for quantitative clearance of unbound ¹⁹F-BF3-Cy3-ACUPA from bloodstream and the urine. With the temporal delay, a patient can also travel and undergo delayed fluorescence-guided RP with the urologist of his or her choice without reinjection.

This first-in-human study demonstrated first time, to our knowledge, using [¹⁸F]-PET/FL multimodality imaging probe. There is currently no FDA-approved PSMA⁺ PET probe or [¹⁸F]-PET/FL multimodality imaging probe. Combined PET/FL contrast preclinical development is current and popular ^14,15. It is possible to compete with a multimodality probe by employing a mixture of standalone PET and FL probes. However, a multimodal probe would obviate the well-recognized complications associated with co-injected mixtures of standalone PET or FL contrast, including differences in blood clearance and non-specific tissue accumulation. In cases of limited ligand expression, using standalone probes would lead to PSMA⁺ receptors binding sites’ being blocked by a competing PET or FL probe, reducing the absolute PET or FL signal.

We used a unique IEX AMBF₃ radiochemistry to yield ¹⁹F-BF3-Cy3-ACUPA. IEX AMBF₃ radiochemistry is less technically challenging compared to many current strategies for [¹⁸F]-PET radio labeling. No immediate or long-term adverse effects were reported or observed in patients receiving [¹⁸F]-BF3-Cy3-ACUPA, suggesting that the AMBF₃ prosthetic group can be safely translated to other agents to yield contrast that is safe for human use.

Intraoperative fluorescence imaging of ex vivo prostate or lymph node tissue in the OR did not interfere with standard-of-care prostate removal. However, our Cy3 fluorescence imaging system has a few limitations. Firstly, given the lack of a standard device for Cy3 fluorescence wavelength (554 nm excitation, 565 nm emission) imaging in the OR ²⁴, standardized thresholds or approved parameters for molecularly targeted fluorescence PCa margin imaging do not exist and it was necessary to build a custom fluorescence imaging camera for real-time Cy3 wavelength imaging and perform qualitative evaluation. Secondly, our Cy3 fluorescence imaging system would benefit from future equipment modifications. From this study, our current wish list is: stereotaxic visualization, photomultiplier incorporation, lens autofocusing, and interchangeable exciter filters. Thirdly, it would be ideal if Cy3 fluorescence imaging could be used to image margins within the sterile field and not on a back-table. Fourthly, clear mathematical parameters to define certain signal-to-noise ratios as the positive margin are needed. These fluorescence parameters and improvements can only be developed for use in the OR with input from experienced urological surgeons.

CONCLUSION

In summary, this study demonstrates that [¹⁸F]-BF3-Cy3-ACUPA is both safe and useful in humans. Developing a multimodal, single molecule PET and FL contrast agent will be cheaper and faster than developing separate standalone PET and FL contrast agents in parallel. Larger trials with this PET tracer are expected to further define its capabilities and its clinical role in the management of PSMA⁺ malignancy, especially in PCa.

CLINICAL PRACTICE POINTS.

• Prostate-specific membrane antigen (PSMA), which is highly expressed in both localized and metastatic prostate cancer (PCa) as well as non-prostate origin cancer, is an ideal target for imaging and therapy of PCa.

• We showed that [¹⁸F]-BF3-Cy3-ACUPA is both safe and useful in humans. ¹⁸F]-BF3-Cy3-ACUPA demonstrated high PSMA-targeting efficiency and a favorable pharmacokinetic profile, allowing for sensitive and high-contrast detection of PSMA positive lesions by PET and optical imaging methods (intraoperative imaging/in vivo and ex vivo fluorescence microscopy)

• Our findings also suggested that [¹⁸F]-BF3-Cy3-ACUPA is versatile hybrid imaging agents.