Skip to main content
Premium Trial:

Request an Annual Quote

Intra-Tumor Heterogeneity Patterns Correlate With Survival in Lung Cancer Patients

NEW YORK — Researchers used both proteomic and genomic approaches to tease out patterns of intra-tumor heterogeneity within lung adenocarcinoma patients that correlated with survival.

Lung adenocarcinoma is known to be heterogeneous, and while intra-tumor heterogeneity is thought to contribute to treatment failure and drug resistance, more remains to be learned about its spatial distribution. Using multi-region mass spectrometry, single-cell copy number sequencing, and cyclic immunofluorescence, researchers from Dana-Farber Cancer Institute and elsewhere examined the intra-tumor heterogeneity of lung adenocarcinomas. As they reported in the journal Cell Genomics on Wednesday, they found two distinct spatial patterns, which they called clustered or random geographic diversification. Further analyses tied the random geographic diversification pattern to decreased cell adhesion as well as to decreased survival.

"We have presented a comprehensive dataset that illustrates the extent of spatial intra-tumor proteomic heterogeneity across tens of regions in single tissue sections, depicts spatial patterns of tumor-infiltrating immune cells, and elucidates spatial intra-tumor genomic heterogeneity of single tumor cells in lung adenocarcinoma," wrote senior author Franziska Michor from Dana-Farber and colleagues in their paper.

For their analysis, the researchers performed multi-region MALDI-TOF profiling of thousands of regions of interest within frozen tumor tissue samples from a discovery cohort of 95 patients, followed by a validation cohort of 52 patients in which about 1,000 regions of interest were examined.

With this, they examined the geographic diversification of samples in the proteomic space using a novel analysis that focuses on the association between molecular features and geographic locations. This placed samples into either clustered or random geographic diversification patterns. Within this cohort, tumors falling into the random geographic diversification were further associated with an increased risk of death.

This association further held in the validation cohort.

Through additional bulk RNA-sequencing of 53 tumors from the discovery cohort, the researchers uncovered certain transcriptional programs that were less active in the different geographic diversification cohorts. For instance, among the random geographic diversification samples, there was decreased expression of genes in cell-cell adhesion-related pathways like cadherin binding, transmembrane transport, and extracellular matrix genes. Decreased expression of these programs could lead to a decrease in cell-cell adhesion, boosting cellular migration and invasion and possibly contributing to the random geographic diversification pattern.

Meanwhile, the researchers also used cyclic immunofluorescence in more than 10 million cells from FFPE tumor samples from a dozen patients to examine the tumors' cellular composition. While they found no correlation between the proteomic-derived geographic diversification and the immune infiltration geographic diversification, the researchers did notice that clustered geographic diversification tumors had higher tumor cell content overall. This could stem from decreased cell motility leading to higher tumor cell density, the researchers said.

They additionally conducted single-cell analyses of nearly 2,000 cells isolated from different regions of frozen tumor tissue samples from seven patients from the discovery cohort. Clustering analysis of some tumors found that single cells from the same region tended to cluster together and shared a common ancestral lineage, reflecting a clustered geographic diversification pattern, while cells from other tumors did not show such clustering, reflecting a random geographic diversification pattern. This suggests that both spatial patterns could be detected by both proteomic and genomic analyses.

The findings suggested to the researchers that "higher-order tumor structural features, such as spatial" intra-tumor heterogeneity, could also be used as cancer biomarkers. "This strategy will provide new insights for the future development of prognostic biomarkers in tissue sections," they added.