Artificial Intelligence–Powered Assessment of Pathologic Response to Neoadjuvant Atezolizumab in Patients With NSCLC: Results From the LCMC3 Study

Sanja Dacic, William D. Travis, Jennifer M. Giltnane, Filip Kos, John Abel, Stephanie Hilz, Junya Fujimoto, Lynette Sholl, Jon Ritter, Farah Khalil, Yi Liu, Amaro Taylor-Weiner, Murray Resnick, Hui Yu, Fred R. Hirsch, Paul A. Bunn, David P. Carbone, Valerie Rusch, David J. Kwiatkowski, Bruce E. JohnsonJay M. Lee, Stephanie R. Hennek, Ilan Wapinski, Alan Nicholas, Ann Johnson, Katja Schulze, Mark G. Kris, Ignacio I. Wistuba

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Introduction: Pathologic response (PathR) by histopathologic assessment of resected specimens may be an early clinical end point associated with long-term outcomes with neoadjuvant therapy. Digital pathology may improve the efficiency and precision of PathR assessment. LCMC3 (NCT02927301) evaluated neoadjuvant atezolizumab in patients with resectable NSCLC and reported a 20% major PathR rate. Methods: We determined PathR in primary tumor resection specimens using guidelines-based visual techniques and developed a convolutional neural network model using the same criteria to digitally measure the percent viable tumor on whole-slide images. Concordance was evaluated between visual determination of percent viable tumor (n = 151) performed by one of the 47 local pathologists and three central pathologists. Results: For concordance among visual determination of percent viable tumor, the interclass correlation coefficient was 0.87 (95% confidence interval [CI]: 0.84–0.90). Agreement for visually assessed 10% or less viable tumor (major PathR [MPR]) in the primary tumor was 92.1% (Fleiss kappa = 0.83). Digitally assessed percent viable tumor (n = 136) correlated with visual assessment (Pearson r = 0.73; digital/visual slope = 0.28). Digitally assessed MPR predicted visually assessed MPR with outstanding discrimination (area under receiver operating characteristic curve, 0.98) and was associated with longer disease-free survival (hazard ratio [HR] = 0.30; 95% CI: 0.09–0.97, p = 0.033) and overall survival (HR = 0.14, 95% CI: 0.02–1.06, p = 0.027) versus no MPR. Digitally assessed PathR strongly correlated with visual measurements. Conclusions: Artificial intelligence–powered digital pathology exhibits promise in assisting pathologic assessments in neoadjuvant NSCLC clinical trials. The development of artificial intelligence–powered approaches in clinical settings may aid pathologists in clinical operations, including routine PathR assessments, and subsequently support improved patient care and long-term outcomes.

Original languageEnglish
Pages (from-to)719-731
Number of pages13
JournalJournal of Thoracic Oncology
Volume19
Issue number5
DOIs
StatePublished - May 2024

Keywords

  • Artificial intelligence
  • Convolutional neural network
  • Digital pathology
  • NSCLC
  • Neoadjuvant

Fingerprint

Dive into the research topics of 'Artificial Intelligence–Powered Assessment of Pathologic Response to Neoadjuvant Atezolizumab in Patients With NSCLC: Results From the LCMC3 Study'. Together they form a unique fingerprint.

Cite this