Project Summary/Abstract
Therapies against immunologic checkpoint proteins, such as PD-L1, have revolutionized the treatment of multiple
malignancies, such as malignant melanoma, lung cancer, bladder cancer, or gastroesophageal cancer. All of these
therapies depend on predictive biomarker assays for PD-L1 protein expression in tumor tissue to identify patients who
will most likely respond. PD-L1 assays face two significant challenges: (i) they often need to be performed on small tissue
biopsies where tumor tissue is limited and (ii) the threshold for positivity by immunohistochemistry (1%) is very small,
highlighting crucial dependence on utmost precision and pre-analytical sample validity.
To address the issue of small and limited samples in biopsies, we have developed a chromogenic multiplexed
immunohistochemical assay that combines PD-L1 assessment with tissue and cell type-specific markers to provide a
combined diagnostic and predictive assay on a single histologic slide. To move multiplexed assays around PD-L1 into the
clinic, there exists an important knowledge gap: what are the patient-specific factors and specimen-related pre-analytical
variables that can influence the readout of PD-L1 positivity? Very limited knowledge is available about these variables.
Importantly, because the PD-L1 positivity threshold is so low and requires the reliable separation of two very small
numbers (<1% vs. ≥1%), even minute pre-analytical variabilities would be expected to have significant negative impact on
assay validity. In that respect, PD-L1 testing in tissue is particularly in need of extensive characterization and control of
pre-analytical variability, even more so than other assays whose cut-off points lie in more favorable ranges.
Our proposal is based on the hypothesis that both patient-specific factors (such as molecular features of the cancer,
current immune status, prior drug therapy, etc.) and specimen-related factors (such as timing of biopsy, size of tissue,
ischemic time, fixation protocol, etc.) can significantly influence subsequent biomarker measurements. We further
hypothesize that solid knowledge about these influences will allow controlling for and mitigating patient-specific and
specimen-related effects and will lead to more accurate and valid biomarker assessment. Aim 1: We will create a cohort
to test the influence of patient-specific context factors. We will make use of our extensive immuno-oncology database
and biobank of >5,000 patients. Aim 2: We will test how specimen-related pre-analytical variables affect the assay using
a wide variety of fresh, frozen, and formalin fixed tissue types and sizes. Aim 3: Once we have defined optimal patient-
specific and specimen-related procedures, we will validate our multiplex assay in a prospective cohort of immunotherapy
patients at MSKCC.
Significance: This project will yield abundant data about the pre-analytical variables that influence a PD-L1 multiplex
immunohistochemistry assay. These data will inform optimal specimen acquisition and handling and strategies for
avoiding or mitigating inaccurate assay results. Innovation: This will be the first study to systematically explore both
patient context factors and specimen pre-analytics in multiplexed immunohistochemical testing for immuno-oncology.
Our close collaboration with commercial test developers as part of our research team will accelerate the translation of
our scientific findings into optimized assays for the direct benefit of patients.
