Patient Selection

Three patients (Pt 001, Pt 002, and Pt 003) with pancreatic cancer scheduled for surgical resection were enrolled in this exploratory phase 0 trial, designed to assess the feasibility, biodistribution, and pharmacokinetics of PTP-01 in a first-in-human setting. Inclusion criteria required normal kidney function, ECOG performance status 0–2, and informed consent. Exclusion criteria included impaired renal function, participation in another investigational drug study within 30 days, known allergies to study components, and serious medical conditions such as uncontrolled infections or cardiovascular disease. Pregnant or breastfeeding women were also excluded. This study was conducted over 10 years ago, during which comprehensive clinical data—including patient demographics, tumor staging, and preoperative imaging—were collected. However, due to the passage of time, much of this information is no longer accessible. All patients met clinical eligibility criteria for surgical resection, confirming the absence of distant metastases via standard preoperative imaging (CT and/or MRI). Diagnostic biopsies were performed as part of the standard clinical workup prior to surgery, and CSP expression was evaluated during the study. Currently, histological validation data are available for only one patient. The study was IRB-approved, and all patients provided written informed consent. Safety was monitored through clinical labs, electrocardiograms, vital signs, and physical exams up to seven days post-PTP-01 administration.

PTP-01 Radiosynthesis and Quality Control

PTP-01 tetrameric peptide was synthesized using a multivalent scaffold-based conjugation approach, ensuring high binding affinity to CSP. The tetramer was generated by conjugating four CSP-targeting peptide units to a branched polyethylene glycol (PEG) linker, which enhances solubility and stability. This approach was chosen over the classical streptavidin–biotin method to improve pharmacokinetics and reduce potential immunogenicity. The synthesis was conducted in a GMP grade facility (CS Bio Company).

For radiolabeling, PTP-01 was labeled with 111In using 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) as the chelator at Cardinal Health (Charlottesville, VA). DOTA was selected due to its well-characterized stability with 111In and its established use in clinical radiopharmaceuticals. Briefly, 444 MBq of 111In (Cardinal Health) was mixed with 120 µg of peptide at 40ºC for 60 min, then passed through a 0.2 µm syringe filter (PVDF) into a new sterile evacuated vial.

Radiochemical purity was determined using instant thin-layer chromatography (ITLC) with a mobile phase of 0.1 M sodium citrate buffer (pH 5.0) to ensure effective chelation. Additionally, a Sep-Pak membrane filtration system was used with gamma-well counting to measure the activity of the flow-through solution and Sep-Pak residue. Each dose was also tested for endotoxins to ensure safety. Radiochemical purity consistently exceeded 95%, confirming the stability and suitability of the radiolabeled agent for clinical use.

PTP-01 Imaging Acquisition In Vivo Planar Scintigraphy

Each patient received a single intravenous 100 µg bolus dose of PTP-01 labeled with 370 MBq 111In. Whole-body planar scintigraphy was performed at 10 min post-injection, followed by scans at 1, 2, 3, 4, 20, and 50 h. SPECT/CT imaging was performed at 4, 20, and 50 h post-injection using a Siemens Symbia™ T6 dual-head SPECT/CT system with low-energy high-resolution (LEHR) collimators optimized for 111In imaging. However, most of these data are no longer accessible due to storage limitations from the original study period. Quantitative analysis was conducted at the time using fused SPECT/CT images, and planar imaging remains available for biodistribution assessment. Tumor ROIs were initially defined using fused SPECT/CT images and cross-referenced with planar data for consistency. The CT component was acquired using a low-dose protocol without IV or oral contrast, primarily for anatomical localization and attenuation correction. However, attenuation correction was not applied to the SPECT images, as full quantitative SPECT reconstruction was not supported by the scanner at the time. Quantitative analysis (%ID) was derived from planar scintigraphy data using calibration factors from a standard source (syringe) imaged under identical conditions. While SPECT/CT facilitated ROI placement, quantitative measurements were based on planar scintigraphy due to the absence of attenuation-corrected SPECT data. After the final pre-surgical imaging session, surgery occurred 24 to 28 h post-injection. In some cases, post-surgical imaging—including planar scintigraphy and SPECT/CT (when available)—was performed at 50 h post-injection after patients recovered from anesthesia.

Ex Vivo Planar Scintigraphy

Following surgery, ex vivo planar images of the resected tumor and adjacent normal tissue (e.g., pancreas and duodenum) were obtained at 28 h post-injection. This timing aligns with surgical resection, allowing immediate imaging of fresh specimens. A white light image was collected for one patient to confirm tumor localization, but these images were not systematically obtained.

ROI Selection and Quantification

Invicro® Imaging Services conducted the clinical trial image analysis, following the methods described by Siegel et al. [15] and Grimes et al. [16].

For in vivo images, we generated eight ROIs for each patient, including bladder, heart, left kidney, right kidney, liver, syringe, whole body, and tumor. Tumor ROIs were defined using fused SPECT/CT images to ensure precise anatomical localization, with adjustments made based on planar scintigraphy data for consistency across time points. Organ ROIs were manually created at the first time point and registered for all subsequent time points. Results were expressed as percent injected dose (%ID). Tumor-to-background ratios were calculated based on the concentration of PTP-01 in each region at each time point. In this study, quantitative measurements were not performed on normal pancreatic tissue. The primary focus was to assess tumor-specific uptake of PTP-01, and including normal pancreatic tissue was not part of the original ROI analysis protocol. We acknowledge this limitation, as tumor-to-normal tissue ratios could have provided additional context.

For ex vivo images, tumor ROIs were drawn using white light imaging, with the surrounding tissue considered background. Since the ex vivo data did not include a syringe, we applied syringe sensitivity values from the planar data analysis. The activity was calculated using planar data methods, ignoring the self-attenuation of the tissue because of tissue thickness.

CSP Immunohistochemistry

Immunohistochemical staining for CSP was performed on a Discovery Ultra Staining Module (Roche Diagnostics, USA) by the Biorepository and Tissue Research Facility at the University of Virginia. Tissue Sects. (4 μm) were deparaffinized, and heat-induced antigen retrieval was conducted using Cell Conditioner 1 (Roche Diagnostics). Endogenous peroxidases were blocked before incubation with anti-plectin antibody (Abcam, ab32528). Detection was performed using OmniMap anti-rabbit multimer with DISCOVERY ChromoMap DAB Kit (Roche Diagnostics). Slides were counterstained with hematoxylin, dehydrated, cleared, mounted, and scanned using the Aperio ScanScope (Leica Biosystems, USA). However, due to the passage of time, only a limited number of stained tissue sections and corresponding analyses were available.

CSP Density Estimation

The in vivo CSP antigen density was estimated using the Krogh cylinder model, which evaluates tissue microenvironmental factors such as vascularity from the quantitative imaging data [17, 18]. Capillary permeability (P), interstitial diffusivity (D), and void fraction were estimated based on molecular weight using empirical relationships derived from the literature [19].

The following input parameters were applied to all patients:

Association rate constant (kon) = 1 × 105 M−1 s−1

Dissociation constant (KD) = 90 nM

Molecular weight (MW) = 27.647 kDa

Mass dose = 100 µg

Internalization half-life = 13 h

Residualization half-life = 120 h

Bodyweight = 70 kg.

Blood clearance parameters were calculated individually for each patient based on a biexponential fit to blood %ID/g data. The parameters for patient Pt 001 were A = 0.492, t1/2,α = 0.33 h, t1/2,β = 6.4 h; for patient Pt 002, A = 0.338, t1/2,α = 0.36 h, t1/2,β = 11.5 h; for patient Pt 003, A = 0.423, t1/2,α = 0.65 h, t1/2,β = 9.6 h. CSP Antigen density was assumed to remain constant throughout the course of the simulation.

To determine CSP antigen density and vascularity multiple simulations were run, and the mean squared error (MSE) between the simulated and quantitative imaging data was calculated. The simulation with the lowest MSE was considered the best fit and the corresponding parameters were considered as the parameters of best fit.

Pharmacokinetic Analysis

Blood and urine samples were collected at multiple time points post-injection to assess the pharmacokinetics of PTP-01. Blood samples were obtained at approximately 10 min, 1, 2, 4, 20, and 50 h post-injection to measure circulating activity and determine clearance rates. Additionally, urine samples were collected at corresponding time points to evaluate renal clearance and urinary excretion of PTP-01. The primary pharmacokinetic data were derived from direct blood and urine sample measurements.

Blood and urine data were converted to %ID/mL for blood and %ID for urine using counts per minute (CPM), decay-corrected injected dose converted to CPM, and sample weights (with 1 g = 1 mL). For urine, total sample weights were used, and a fixed hematocrit fraction (0.45 for men, 0.40 for women) and a gamma counter efficiency fraction of 0.278 were applied.

The total accumulated %ID of urine was computed under the assumption that all activity in the blood was located in the plasma. Blood pharmacokinetics were analyzed using two approaches:

1.

Non-compartmental analysis: Data were fit to a single-exponential model, including at least three points based on goodness of fit, with the exclusion of Cmax.

2.

Biexponential model fitting: Data were also fit to a biexponential distribution to evaluate early and late clearance phases.

All pharmacokinetic analyses were conducted using MATLAB software, which provided the best-fit curves and corresponding parameters.