Supplementary Figure S9. The combined T-cell status of precursor and carcinoma tissue categories in relation to anatomic location. (A-B) CD3+CD4+ and CD3+CD8+ T-cell densities in overall tissue regions stratified by anatomic location. (C-D) CD3+CD4+FOXP3+ and CD3+CD8+FOXP3+ T-cell densities in overall tissue regions stratified by anatomic location. P values were calculated with the Mann-Whitney U test (Wilcoxon rank-sum test) as compared to colorectal normal mucosa near precursor lesions (Normal mucosa) and carcinoma with proficient mismatch repair (pMMR). ***: P <0.0005, **: P <0.005, *: P <0.05. Abbreviations: CA, colorectal invasive carcinoma; dMMR, deficient mismatch repair; Non-serrated, non-serrated adenomas, Normal, colorectal normal mucosa near the precursor lesions; pMMR, proficient mismatch repair; Serrated without SSL, serrated lesions including hyperplastic polyp and traditional serrated adenoma; SSL, sessile serrated lesions.

Antibodies and Staining Conditions for Multiplex Identification of T-cell Subsets and Epithelial Cells

The immune microenvironment is a crucial component of colorectal carcinoma that has been well characterized, but much less is known about the immune microenvironment of colorectal carcinoma precursors. We hypothesized that T-cell infiltrates might differ across the colorectal neoplastic spectrum. We leveraged the prospective cohort incident-tumor biobank method, which provided formalin-fixed, paraffin-embedded tumor tissue specimens (N = 1,825) from 790 colorectal carcinoma precursors (including hyperplastic polyps, sessile serrated adenomas, traditional serrated adenomas, tubular adenomas, tubulovillous adenomas, and villous adenomas) and 1,035 colorectal carcinomas. We performed an in situ multispectral immunofluorescence assay for CD3, CD4, CD8, FOXP3 (negative, low, or high expression), PTPRC (CD45RO and CD45RA), MKI67 (Ki-67), and KRT (keratin) combined with supervised machine learning. CD3+CD4+ cells were more abundant than CD3+CD8+ cells in most precursors. In conventional adenomas, greater villous component correlated with fewer intraepithelial CD3+CD8+ cells. Serrated lesions, including hyperplastic polyps and sessile serrated lesions, exhibited higher densities of intraepithelial CD3+CD8+ cells compared with other precursors and carcinomas. Age strata of patients with precursors (including early-onset precursors) were not associated with differential T-cell infiltration patterns. Compared with invasive colorectal carcinoma, precursors generally showed higher densities of CD3+CD4+ cells and CD3+CD8+ cells with phenotypes of naive (CD45RA+CD45RO-), memory (CD45RA-CD45RO+), and regulatory (FOXP3+Low and FOXP3+High) in intraepithelial and lamina propria/stromal regions. In conclusion, T-cell infiltration patterns vary across different histopathologic types of the colorectal neoplastic spectrum from precursors to invasive carcinomas. Our findings shed light on how the tumor-immune microenvironment evolves during precursor development and progression to colorectal carcinoma.

Supplementary Figure S1. Flow diagram for T-cell population and distribution analyses. Abbreviations: HPFS, Health Professionals Follow-up Study; NHS, Nurses’ Health Study; NHS II, Nurses’ Health Study II.

Supplementary Figure S3. Identification of distinct T-cell subsets via the combinational expression of T-cell markers (membrane CD3, membrane CD4, membrane CD8, membrane CD45RA, membrane CD45RO, nucleus FOXP3), cell cycle marker (nucleus MKI67), epithelial marker (cytoplasm KRT), and DNA marker (nucleus DAPI) at single-cell resolution. Scale bar: 2.5 (μm).

Objective:To test the hypothesis that the association of smoking with long-term colorectal cancer incidence may be stronger for tumours with higher mutational and neoantigen loads. Methods and analysis:In the Nurses' Health Study (1980-2012) and the Health Professionals Follow-up Study (1986-2012), our novel prospective cohort incident-tumour biobank method (PCIBM) used 3053 incident colorectal carcinoma cases including 752 cases with whole-exome sequencing data. Using the multivariable duplication-method Cox regression model with the inverse probability weighting to adjust for the selection bias due to tissue availability, we assessed a differential association of cigarette smoking with colorectal carcinoma incidence by an exome-wide tumour mutational burden (e-TMB) or neoantigen load. Results:The association of pack-years smoked with colorectal cancer incidence differed by e-TMB (Pheterogeneity<0.001). Multivariable-adjusted HRs for e-TMB-high (≥10 mutations/megabase) tumours were 1.28 (95% CI 0.72 to 2.28) and 2.56 (95% CI 1.61 to 4.07) for 1-19 and ≥20 pack-years (vs 0 pack-years; Ptrend<0.001), respectively. In contrast, pack-years smoked were not associated with e-TMB-low tumour incidence (Ptrend=0.67). A similar differential association was observed for the neoantigen load (Pheterogeneity=0.017). The differential association by e-TMB appeared consistent in the strata of CpG island methylator phenotype status, BRAF mutation or lymphocytic infiltrates. Conclusions:Smoking is more strongly associated with the long-term incidence of colorectal carcinoma harbouring higher mutational and neoantigen loads. Our PCIBM-based evidence supports the immunosuppressive effect of smoking and the potential of smoking cessation in improving antitumour immunity for cancer prevention and treatment.

Supplementary Table S5 shows results of univariate analysis for 77 chronically elevated serum alanine aminotransferase (cALT)-defined nonalcoholic fatty liver disease in the Pancreatic Cancer Case-Control Consortium (PanC4) data

Table S3. Sensitivity analysis of the association between plasma 25-hydroxyvitamin level and disease-free survival, overall survival, and time to recurrence, excluding patients who recurred or died within three months of blood collection

Supplementary Table S3 shows results of univariate analysis for 36 biopsy-confirmed nonalcoholic fatty liver disease SNPs in the PanScan sample

SNPs in adiponectin and leptin receptor genes and their association with receptor expression

Supplementary Table S4 shows results of univariate analysis for 17 imaging and biopsy validated SNPs in the PanScan sample

BackgroundGrowing evidence supports the importance of characterizing the organizational patterns of various cellular constituents in the tumor microenvironment in precision oncology. Most existing data on immune cell infiltrates in tumors, which are based on immune cell counts or nearest neighbor-type analyses, have failed to fully capture the cellular organization and heterogeneity.MethodsWe introduce a computational algorithm, termed Tumor-Immune Partitioning and Clustering (TIPC), that jointly measures immune cell partitioning between tumor epithelial and stromal areas and immune cell clustering versus dispersion. As proof-of-principle, we applied TIPC to a prospective cohort incident tumor biobank containing 931 colorectal carcinoma cases. TIPC identified tumor subtypes with unique spatial patterns between tumor cells and T lymphocytes linked to certain molecular pathologic and prognostic features. T lymphocyte identification and phenotyping were achieved using multiplexed (multispectral) immunofluorescence. In a separate hepatocellular carcinoma cohort, we replaced the stromal component with specific immune cell types-CXCR3+CD68+ or CD8+-to profile their spatial relationships with CXCL9+CD68+ cells.ResultsSix unsupervised TIPC subtypes based on T lymphocyte distribution patterns were identified, comprising two cold and four hot subtypes. Three of the four hot subtypes were associated with significantly longer colorectal cancer (CRC)-specific survival compared to a reference cold subtype. Our analysis showed that variations in T-cell densities among the TIPC subtypes did not strictly correlate with prognostic benefits, underscoring the prognostic significance of immune cell spatial patterns. Additionally, TIPC revealed two spatially distinct and cell density-specific subtypes among microsatellite instability-high colorectal cancers, indicating its potential to upgrade tumor subtyping. TIPC was also applied to additional immune cell types, eosinophils and neutrophils, identified using morphology and supervised machine learning; here two tumor subtypes with similarly low densities, namely 'cold, tumor-rich' and 'cold, stroma-rich', exhibited differential prognostic associations. Lastly, we validated our methods and results using The Cancer Genome Atlas colon and rectal adenocarcinoma data (n = 570). Moreover, applying TIPC to hepatocellular carcinoma cases (n = 27) highlighted critical cell interactions like CXCL9-CXCR3 and CXCL9-CD8.ConclusionsUnsupervised discoveries of microgeometric tissue organizational patterns and novel tumor subtypes using the TIPC algorithm can deepen our understanding of the tumor immune microenvironment and likely inform precision cancer immunotherapy.

Figure S2. Dose-response relationship and test of linearity for the hazard ratio (HR) of disease-free survival (A), overall survival (B), and time to recurrence (C) by continuous predicted vitamin D scores, with reference of plasma 25-hydroxyvitamin D set at 12 ng/ml

Supplementary Table S9 shows descriptive Statistics for PanScan sample

Table S4. Hazard ratio of disease-free survival, overall survival, time to recurrence by quartiles of predicted vitamin D score

Association between SNPs in the leptin receptor gene and pancreatic cancer mortality by sex

Supplementary Table S10 shows descriptive Statistics for PanC4 sample

Supplementary Table S2 shows results of univariate analysis for 22 imaging-defined nonalcoholic fatty liver disease SNPs in the PanScan sample

Supplementary Table S6 shows results of univariate analysis for 22 imaging defined nonalcoholic fatty liver disease SNPs in the PanC4 sample

Association between leptin levels and pancreatic cancer mortality using spline regression.