Utilizing AI for Quantitative Assessment of Prostate Cancer Recurrence

Manual assessments of prostate specific membrane antigen (PSMA) scoring and measures such as standardized uptake value (SUV) mean and maximum SUV (SUVmax) on positron emission tomography (PET) scans can be tedious and are not always accurate, noted Hong Song, M.D., Ph.D., in a recent interview.

With this in mind, Dr. Song and colleagues recently evaluated the combination of aPROMISE software (PYLARIFY AI, Exini Diagnostics AB/Lantheus Holdings), an FDA-cleared deep learning platform for quantitative assessment of PSMA PET/CT images, and piflufostat F 18 (PYLARIFY®, Lantheus Holdings) to assess 69 patients with prostate cancer recurrence and the impact of quantitative measures upon subsequent biochemical progression-free survival.

In some of the findings from the study, presented at the recent Society for Nuclear Medicine and Molecular Imaging (SNMMI) conference, the researchers noted that higher PSMA-avid total tumor volume (PSMAAttv) and a higher aPSMA score for bone metastases were both associated with shorter biochemical progression-free survival in patients with prostate cancer.

(Editor’s note: For related content, see “Recurrent Prostate Cancer and Low PSA Levels: Can an Emerging PSMA PET Agent Have an Impact?,” “Emerging PET Radiotracer May Offer Multiple Advantages in Detecting Prostate Cancer” and “Can Pre-Op MRI Staging Help Predict Prostate Cancer Recurrence After a Prostatectomy?”)

“There are quantitative tools now that are readily available to help us have this prognostic evaluation of (patients) who may be at high risk for subsequent progression (of prostate cancer and) should be followed more closely. … There is more information than meets the eye in the scan that we can now quantify and extract,” emphasized Dr. Song, an assistant professor of radiology (nuclear medicine) at Stanford University.

For more insights from Dr. Song, watch the video below.