PET/CT Images: Measuring PD-L1 Status and Predicting Response to Immunotherapy in NSCLC
Posted: Tuesday, August 10, 2021
To preclude ineffectively treating the approximately half of patients with non–small cell lung cancer (NSCLC) who may not respond to immune checkpoint inhibitor therapy, an important element of targeted immunotherapy is predicting prognosis using PD-L1 biomarker analysis. In a recent study in the Journal for ImmunoTherapy of Cancer, Matthew B. Schabath PhD, of the Moffitt Cancer Center, Tampa, and colleagues reported on the development of a noninvasive approach to measuring biomarker levels; it avoids the risks of biopsies and inadequate samples and uses, instead, deep learning of PET/CT images.
“Because images are routinely obtained and are not subject to sampling bias per se, we propose that the individualized risk assessment information provided by these analyses may be useful as a future clinical decision support tool pending larger prospective trials,” said study coauthor Robert Gilles, PhD, also of Moffitt Cancer Center, in an institutional press release.
Utilizing machine learning with 18F-FDG PET/CT images and clinical data from 697 patients with NSCLC, the study investigators focused on six cohorts from three institutions to train, validate, and assess the prognostic value of a deeply learned score (DLS). It investigated the association of the DLS and clinical characteristics on clinical outcomes. According to the study findings, the PD-L1 DLS was able to discriminate between patients with PD-L1–positive disease and those with PD-L1–negative disease in the training, validation, and two external cohorts. In addition, the authors noted that the DLS may prove to be a useful surrogate for immunohistochemistry (IHC), given the novel score was found to be indistinguishable from the IHC-derived PD-L1 status in predicting both progression-free and overall survival.
This study is reportedly the first to develop a PD-L1 radiomic signature and then use it for response prediction. Thus, it may prove to be an effective clinical decision-making tool.
Disclosure: For full disclosures of the study authors, visit jitc.bmj.com.
