sing a new test that unites cell-imaging technology with machine learning, researchers have found that patients whose tumors contain high concentrations of specific immune cells are likely to survive longer than patients with tumors carrying lower densities of these cells. The test, called ImmunoPROFILE, uses fluorescent staining to mark different types of immune cells within a tumor, then analyzes a digital image of the tumor with machine learning technology to calculate the density of those cell types. Researchers used the test to analyze 2,023 tumor specimens collected over three years from 27 tumor types. They found that high densities of CD8+ T cells (which attack diseased and cancerous cells), PD-1+ T cells (so-called exhausted T cells), and FOXP3+ T cells (which regulate the T cell response) were associated with longer overall survival across tumor types regardless of therapy. The findings suggest that tests that quantify multiple immune cell populations may have greater predictive power than those that count just one, and that clinical use of multiplexed imaging assays provides previously unrealized prognostic information across multiple tumor types.
