AUA 2022: Artificial Intelligence PET Interpretation vs Human Specialists: The LATAM Study

(UroToday.com) The 2022 American Urological Association (AUA) Annual Meeting included a late-breaking abstract session and a presentation by Dr. Diego Barreiro discussing the LATAM study assessing artificial intelligence PET interpretation versus human specialists. The interpretation of the total volume of disease in PET may help with decision-making in mCRPC. Furthermore, artificial intelligence could help quantify all lesions, measuring the total tumor volume and the percentage of increase or response. This may be important given that human specialists have a subjective interpretation. The objective of this study was to compare the PET interpretation by artificial intelligence vs human specialists.


This study included a survey on QUESTIONPRO platform, which was released through various digital channels. The survey gathered the following parameters: Specialists’ year of experience (<5, 5-10, 10-20, >20 years) and interpretation of 3 clinical cases of real mCRPC patients (basal PET and follow-up PET after a treatment). The human expert had to identify the response percentage (0-100%). The participants were consulted on the following:

  1. If the quantification of basal and follow-up tumor volume, would impact their decision-making?
  2. Would they consider referring the patient to an organization which is able to prepare their patient’s tumor response?

The platform MIM Software Encore was used for the quantification of tumor volume through artificial intelligence. The programming staff developed a workflow that divides the skeletal system, recognizing the structures to be analyzed in PET choline studies:

 

image-0.jpg 

 

 The virtual survey was completed by 134 to specialists of multiple LATAM countries (Argentina, Mexico, Brazil, Guatemala, Honduras, Nicaragua, Panama, Cuba, Dominican Republic, Venezuela, Colombia, Ecuador, Peru, Bolivia, Chile, Paraguay, Uruguay) USA and Spain. Regarding years of experience, the participants were divided as follows:

  • <5 years: 13.97%
  • 5-10 years: 24.58%
  • 10-20 years: 30.73%
  • >20 years: 30.73%

 

image-1.jpg 

 

The analysis of the statistical significance shows the superiority of artificial intelligence (Case 1: p <0.001 Case 2: p <0.001 Case 3: p <0.005) and 83% of the participants considered that artificial intelligence does contribute to their decision-making.


Dr. Barreiro concluded this presentation by discussing the LATAM study assessing artificial intelligence PET interpretation versus human specialists with the following concluding statements:

  • Measuring the total tumor volume with the artificial intelligence model generated significantly greater results than the ones provided by expert interpretation, regardless of their years of experience
  • Specialists consider that the implementation of the artificial intelligence model in the PET/CT is essential and generates an impact on their decision-making

 


Presented By: Diego Barreiro, MD, Universidad de Buenos Aires, Buenos Aires, Argentina
Co-Authors: Danny Mena, Sonora, Mexico; Federico Losco, Buenos Aires, Argentina; Andres Lopez-Cortes, Madrid, Spain; Juan Pablo Sade, Buenos Aires, Argentina

Written By: Zachary Klaassen, MD, MSc – Urologic Oncologist, Assistant Professor of Urology, Georgia Cancer Center, Augusta University/Medical College of Georgia, @zklaassen_md on Twitter during the 2022 American Urological Association (AUA) Annual Meeting, New Orleans, LA, Fri, May 13 – Mon, May 16, 2022.

email news signup