Related Experiment Video
Updated: May 1, 2026
Cancer-Associated Fibroblasts from Mouse Mammary Tumors as Tools for Molecular and Computational Studies
Published on: July 3, 2025
A contamination-adjusted tissue thrombopoietic framework identifies a clinically relevant CAF-myeloid niche in
Guocai Mao1,2, Jing Li3, Xiaofei Song1,4
1National Clinical Research Center for Hematologic Diseases, Jiangsu Institute of Hematology, The First Affiliated Hospital of Soochow University, Suzhou, China.
Background:
Platelet-related cancer biology is often studied through circulating markers, yet tissue-centered thrombopoietic programs in lung cancer remain poorly defined and are vulnerable to blood-contamination artifacts. We developed a contamination-adjusted megakaryopoiesis-platelet tumor axis (MPTA) framework to test whether thrombopoietic activity in lung cancer is embedded in a stromal-myeloid niche with clinical relevance.
Methods:
We used paired bulk RNA sequencing from a local cohort of 99 tumor-adjacent pairs as the discovery engine. After paired differential expression and explicit adjustment for blood-contamination structure, a locked tissue-centered MPTA score was evaluated using public single-cell, bulk, spatial, and platelet-transcript reference layers together with de-identified local clinicopathologic and survival data.
Results:
Paired analysis identified a tumor-enriched contamination-adjusted MPTA, which remained significantly higher in tumors than in adjacent tissues after correction (0.34 vs -0.26; p = 1.4 × 10^-5). Direct public single-cell scoring together with bulk back-projection and supportive spatial analyses localized the composite signal primarily to a CAF-myeloid niche, while tumor epithelial cells contributed more selectively to the tumor-thrombopoietic component rather than dominating the composite signal. The strongest translational signal was observed locally, where tumor_MPTA was retained as the primary prognostic readout and predicted worse overall survival (hazard ratio 1.43, 95% confidence interval 1.15-1.78; p = 0.00144), whereas delta_MPTA was retained only as a supportive within-patient metric because of proportional-hazards sensitivity in fully adjusted models. External TCGA analyses preserved tumor-normal directionality but provided only limited, non-confirmatory prognostic support, and platelet transcriptomes did not recapitulate the tissue-derived program.
Conclusions:
A contamination-adjusted, tissue-centered MPTA framework supports a clinically relevant CAF-myeloid thrombopoietic niche in lung cancer. For clinical interpretation, tumor_MPTA-not delta_MPTA-should be regarded as the primary prognostic readout. The strongest evidence comes from paired human tissue and local survival analyses, whereas the external layers are supportive rather than confirmatory, and platelet transcriptomes define a biological boundary rather than a validation layer.
