2. UroToday Home
  3. Conference Highlights
  4. ESMO 2023
  5. ESMO 2023 Prostate Cancer

ESMO 2023

ESMO 2023: Application of Novel Machine Learning Model in 68Ga-PSMA-11 PET/CT – Predicting Survival in Oligometastatic Prostate Cancer Patients

(UroToday.com) The 2023 ESMO annual meeting included a session on prostate cancer, featuring a presentation by Dr. Mikaela Dell'Oro discussing the application of novel machine learning model in 68Ga-PSMA-11 PET/CT. Biochemical recurrence is estimated to occur in ≥ 25% of patients with prostate cancer following primary curative therapy.


Machine learning models are being developed for lesion detection and tracking to provide a comprehensive view of disease burden, allowing clinicians to quantify and predict effectiveness of treatment for individual lesions.1 This study, presented at the ESMO 2023 congress, applied novel AI-assisted technology to automatically extract features from 68Ga-PSMA-11 PET/CT images that correlate with treatment intervention and survival data to create a scoring system.

 Between 2015 and 2016, 185 men with oligometastatic prostate cancer had a baseline and follow-up PSMA PET/CT scan (at ∼6-months) while being treated per standard clinical care. Inclusion criteria was low-disease burden as defined as negative/oligometastatic disease (>3 lesions) on bone scintigraphy and abdominal CT scans. Lesions were quantified and matched between time points using AIQ Solutions technology. Imaging features were extracted from each patient, including change in basic features (SUVmax, SUVmean, and number of lesions at baseline), and heterogeneity features (intra-patient heterogeneity of disease and response). Univariate predictive power of overall survival (OS) prediction of each measure was determined using Cox regression models. An AI approach was evaluated to predict OS using 5-fold cross-validation of a random survival forest. Model performance was evaluated using the c-index.

Overall, there were 1,233 lesions identified at baseline and 1,605 identified during follow-up. The top univariate predictors of survival were all heterogeneity features:

  • Proportion of lesions increasing (c-index = 0.62)
  • Number of stable lesions (c-index = 0.62)
  • Number of decreasing lesions (c-index = 0.60)
  • Number of new lesions (c-index = 0.59) 

In an individual scan, the proportion of increasing lesions >29% correlated with poorer progression:
The AI model was able to predict responders vs suboptimal responders based on whether they had a treatment intervention or observation alone (35%) (c-index = 0.83 in both cases). The following shows a response assessment map of patients demonstrating a higher TRAQinform Profile score for patients (PS136 and PS016) who were predicted to do favorably compared to patients (PS170 and PS019) who did not:
Dr. Dell'Oro concluded her presentation discussing the application of novel machine learning model in 68Ga-PSMA-11 PET/CT with the following take-home points:

  • This study demonstrates that an AI-assisted lesional response analysis can help predict response and prognosis of oligometastatic prostate cancer patients using 68Ga-PSMA-11 PET/CT images
  • These results support further studies to validate these findings in a prospective cohort

Presented by: Mikaela Dell'Oro, Research Fellow, The University of Western Australia, Perth, Australia

Written by: Zachary Klaassen, MD, MSc – Urologic Oncologist, Associate Professor of Urology, Georgia Cancer Center, Wellstar MCG Health, @zklaassen_md on Twitter during the 2023 European Society of Medical Oncology (ESMO) Annual Meeting, Madrid, Spain, Fri, Oct 20 – Tues, Oct 24, 2023.

Reference:

  1. Lindgren BS, Frantz S, Minarik D, et al. Applications of artificial intelligence in PSMA PET/CT for prostate cancer imaging. Semin Nucl Med. 2023 Jun 23.