The Pulse
AI-Designed PARP-1 PET Tracer Reaches Four Cancer Patients
Researchers evaluated the AI-designed gallium-68 PET tracer [68Ga]Ga-DOTA-FZPF in four patients with breast or ovarian cancer. Early scans showed PARP-1-related uptake, but limited contrast in a liver metastasis points to a need for larger
AI.info Team · Written with AI ·
On October 11, researchers reported that an AI-designed PET tracer had been used in an exploratory scan study involving four people with cancer. The gallium-68 compound, [68Ga]Ga-DOTA-FZPF, targets PARP-1, a DNA-repair enzyme, and was evaluated in three breast cancer patients and one ovarian cancer patient.
The small study marks a move from computer-assisted molecular design and animal testing to human imaging. It does not establish the tracer as a diagnostic test: the first scans offer limited evidence about how reliably it can show PARP-1 in different tumors or distinguish lesions from surrounding tissue.
From fuzuloparib to a PET tracer
The researchers started with fuzuloparib, a PARP inhibitor, and used a staged computational workflow to prioritize chemical variants for testing. The process combined a recurrent neural network to generate candidate structures with docking, synthetic-complexity scoring, target prediction and molecular-dynamics simulations. The workflow helped narrow the candidates; researchers still had to synthesize and test the compounds.
Two gallium-68 tracers advanced through laboratory evaluation: DOTA-FZPF and NOTA-FZPF. In cell experiments, both showed greater uptake in cancer cells with higher PARP-1 expression than in a low-expression cell line, and excess fuzuloparib reduced uptake. In mice, the DOTA version accumulated more strongly in tumors than in the low-PARP-1 model and showed lower uptake in several organs than its NOTA counterpart, so the team selected it for human scans.
Four patients, with uneven signal
The exploratory human evaluation reported no adverse events or abnormalities in vital signs following injection. In one patient with left breast cancer, the lesion’s maximum standardized uptake value (SUVmax) was 3.7, compared with 1.8 in the opposite breast; pathology also showed high PARP-1 expression in the lesion.
A scan in a patient with liver metastases showed a much narrower difference: SUVmax was 1.3 in the liver lesion and 1.1 in normal liver tissue. The biopsy indicated low PARP-1 expression in that lesion. The result illustrates a practical challenge for imaging targets inside organs that also take up the tracer: the signal may not separate a lesion clearly from background tissue.
Early evidence, not a validated test
The study’s human findings come from just four patients, and the reported tumor uptake was modest. The researchers point to the small cohort and low PARP-1 expression in some lesions as possible explanations; the DOTA component’s hydrophilicity may also limit how well the compound enters cells. The authors also report a modest tumor-to-liver ratio, a potential constraint when assessing liver metastases.
PET imaging that measures PARP-1 could eventually help researchers examine target expression across multiple sites without relying on a biopsy from each one, or assess whether a PARP inhibitor is engaging its target. But those uses require evidence that the scans correspond consistently to tumor biology and perform well across a larger patient group. The initial study offers a human test of the design approach, not proof of clinical accuracy or patient benefit.