Purpose:

Although high T-cell density is a well-established favorable prognostic factor in colorectal cancer, the prognostic significance of tumor-associated plasma cells, neutrophils, and eosinophils is less well-defined.

Experimental Design:

We computationally processed digital images of hematoxylin and eosin (H&E)–stained sections to identify lymphocytes, plasma cells, neutrophils, and eosinophils in tumor intraepithelial and stromal areas of 934 colorectal cancers in two prospective cohort studies. Multivariable Cox proportional hazards regression was used to compute mortality HR according to cell density quartiles. The spatial patterns of immune cell infiltration were studied using the GTumor:Immune cell function, which estimates the likelihood of any tumor cell in a sample having at least one neighboring immune cell of the specified type within a certain radius. Validation studies were performed on an independent cohort of 570 colorectal cancers.

Results:

Immune cell densities measured by the automated classifier demonstrated high correlation with densities both from manual counts and those obtained from an independently trained automated classifier (Spearman's ρ 0.71–0.96). High densities of stromal lymphocytes and eosinophils were associated with better cancer-specific survival [Ptrend < 0.001; multivariable HR (4th vs 1st quartile of eosinophils), 0.49; 95% confidence interval, 0.34–0.71]. High GTumor:Lymphocyte area under the curve (AUC0,20μm; Ptrend = 0.002) and high GTumor:Eosinophil AUC0,20μm (Ptrend < 0.001) also showed associations with better cancer-specific survival. High stromal eosinophil density was also associated with better cancer-specific survival in the validation cohort (Ptrend < 0.001).

Conclusions:

These findings highlight the potential for machine learning assessment of H&E-stained sections to provide robust, quantitative tumor-immune biomarkers for precision medicine.

Translational Relevance

Drawing from a large database of two prospective U.S. cohort studies, we identified lymphocytes, plasma cells, neutrophils, and eosinophils in digital images of hematoxylin and eosin (H&E)–stained sections from more than 930 colorectal cancers using pathologist-supervised machine learning algorithms and studied their associations with cancer-specific mortality, while extensively adjusting for potential confounders, including microsatellite instability status, CpG island methylator phenotype, long-interspersed nucleotide element-1 methylation, and KRAS, BRAF, and PIK3CA mutation status. We found that high densities not only of stromal lymphocytes but also eosinophils were associated with better cancer-specific survival, and greater proximity of lymphocytes and eosinophils to tumor cells was also associated with better cancer-specific survival. These findings highlight the potential for machine learning assessment of H&E-stained sections to provide robust, quantitative tumor-immune biomarkers for precision medicine and identify previously underappreciated immune cell subsets as harboring prognostic relevance.

Colorectal cancer is the third most common malignancy and the second most common cause of cancer deaths worldwide (1). The prognostic classification of colorectal cancer has been mainly based on disease stage (2, 3). However, each tumor is unique and driven by complex pathologic processes, many of which are mediated by interactions between the neoplastic cells and the host (4). Additional prognostic parameters that more fully capture these associations may therefore help us classify patients into more homogenous, therapeutically relevant subgroups.

Colorectal carcinoma is composed of a mixture of cell types, including tumor cells, fibroblasts, endothelial cells, and immune cells, that cumulatively comprise the tumor microenvironment (5). Accumulating evidence indicates that tumors may evoke an adaptive antitumor immune response (6). Indeed, higher densities of tumor-infiltrating T lymphocytes have been associated with improved clinical outcome in colorectal cancer (7). The prognostic value of Immunoscore, generated by measuring CD3+ and CD8+ T-cell densities in the tumor center and invasive margin, has been validated in an international multi-institutional study involving several thousand patients with colon cancer from 13 countries (8). Although innate immune cells, such as neutrophils and eosinophils, also represent a major cell population in colorectal tumors, their potential prognostic significance has not been as well-defined as that of T cells (7, 9). Although the potential prognostic value of plasma cells in colorectal cancer has attracted relatively little attention, B cells and tertiary lymphoid structures have been reported to be associated with prognosis in colorectal cancer (7) and immunotherapy response in melanoma (10). We therefore sought to examine plasma cells, which represent the only B-cell subset that can be reliably identified in hematoxylin and eosin (H&E)–stained tissue sections. Analyses of different types of immune cells involved in adaptive and innate immunity could help develop improved prognostic biomarkers and may lead to a better understanding of colorectal cancer biology.

In this study, we identified and quantified lymphocytes, plasma cells, neutrophils, and eosinophils in tumor intraepithelial and stromal areas, using supervised machine learning on digital images of H&E-stained tissue microarrays (TMAs) containing tumor tissue from patients with colorectal cancer in two large U.S.-based prospective cohort studies. For our primary aim, we tested the hypothesis that higher densities of these immune cells might be associated with better prognosis. In exploratory analyses, we investigated the relationships of the densities of these cell types with tumor and patient characteristics as well as the prognostic significance of spatial characteristics of the immune cell infiltrates in relation to tumor cells. To validate the findings, we analyzed an independent cohort of 570 colorectal cancers.

Study population and data collection

We utilized two U.S.-nationwide prospective cohort studies, the Nurses' Health Study (NHS, 121,701 women followed since 1976) and the Health Professionals Follow-up Study (HPFS, 51,529 men followed since 1986). In these populations, we documented 4,420 incident colorectal carcinoma cases during follow-up until 2014. We analyzed immune cell densities in 934 adenocarcinomas, based on the availability of follow-up data and adequate tissue specimens in TMAs (Table 1). We included both colon and rectal carcinoma based on the colorectal continuum model (11). We utilized the inverse probability weighting (IPW) method and covariate data from the 4,420 cases to adjust for selection bias due to tumor tissue availability in the 934 cases.

Table 1.

Clinical, pathologic, and molecular characteristics of colorectal cancer cases according to tumor stromal immune cell densities.

Stromal immune cell density (cells/mm2)
Median (25th–75th percentiles)
CharacteristicaNLymphocytePlasma cellNeutrophilEosinophil
All cases 934 413 (212–785) 15 (4.5–41) 46 (20–93) 15 (4.8–40) 
Sex 
 Female (NHS) 521 (56%) 446 (234–804) 17 (5.2–43) 51 (22–106) 13 (4.5–37) 
 Male (HPFS) 413 (44%) 366 (196–726) 13 (3.6–39) 39 (18–78) 16 (5.2–42) 
Age (years) 
 <65 289 (31%) 412 (221–766) 17 (5.3–39) 49 (23–93) 17 (5.2–42) 
 ≥65 645 (69%) 414 (210–790) 15 (4.4–42) 45 (19–93) 14 (4.6–38) 
Year of diagnosis 
 1995 or before 308 (33%) 435 (189–869) 14 (3.7–41) 45 (21–99) 15 (4.8–41) 
 1996–2000 306 (33%) 398 (206–745) 15 (5.6–45) 46 (20–95) 14 (4.9–34) 
 2001–2008 320 (34%) 423 (241–763) 16 (4.6–37) 46 (19–88) 15 (4.6–43) 
Family history of colorectal cancer in first-degree relative(s) 
 Absent 735 (79%) 394 (212–765) 14 (4.1–38) 45 (20–92) 14 (4.9–38) 
 Present 194 (21%) 499 (234–916) 19 (6.4–62) 49 (20–97) 16 (4.7–44) 
History of inflammatory bowel disease 
 Absent 920 (98%) 413 (212–782) 15 (4.5–41) 46 (20–93) 15 (4.7–40) 
 Present 13 (1.4%) 400 (270–982) 14 (5.0–35) 38 (22–65) 18 (10–59) 
Tumor location 
 Cecum 163 (18%) 435 (224–846) 16 (5.4–43) 45 (20–99) 15 (5.6–39) 
 Ascending to transverse colon 306 (33%) 443 (212–820) 17 (4.6–44) 46 (20–93) 12 (4.6–35) 
 Splenic flexure to sigmoid colon 280 (30%) 395 (198–770) 15 (4.3–38) 45 (20–92) 17 (5.6–43) 
 Rectum 181 (19%) 407 (212–726) 14 (3.6–40) 46 (20–90) 15 (3.6–41) 
Tumor differentiation 
 Well to moderate 843 (90%) 407 (212–758) 15 (4.4–40) 45 (20–91) 15 (4.9–42) 
 Poor 89 (9.6%) 569 (223–1019) 20 (7.3–45) 55 (24–123) 11 (4.7–33) 
Extent of signet-ring cells (%) 
 0 814 (87%) 413 (212–785) 15 (4.5–42) 46 (20–92) 15 (4.8–40) 
 1–50 108 (12%) 456 (224–779) 16 (4.4–38) 44 (20–100) 12 (4.1–34) 
 ≥51 12 (1.3%) 281 (162–980) 13 (4.4–48) 46 (25–79) 34 (12–81) 
Extent of extracellular mucin (%) 
 0 550 (59%) 413 (215–767) 15 (5.2–41) 46 (21–91) 14 (4.6–38) 
 1–50 274 (29%) 436 (217–850) 16 (3.7–43) 47 (20–91) 16 (5.4–43) 
 ≥51 110 (12%) 364 (192–777) 14 (3.7–37) 40 (19–112) 13 (4.0–41) 
AJCC disease stage 
 I 198 (23%) 558 (306–986) 23 (8.5–62) 49 (26–97) 27 (8.5–65) 
 II 285 (33%) 411 (224–798) 13 (4.1–39) 51 (23–101) 13 (4.6–33) 
 III 249 (29%) 355 (189–668) 14 (4.1–33) 44 (19–83) 11 (3.4–33) 
 IV 135 (16%) 280 (160–590) 10 (4.1–25) 32 (12–75) 10 (3.7–29) 
MSI status 
 Non–MSI-high 754 (83%) 385 (196–732) 14 (4.3–39) 44 (19–89) 15 (4.5–40) 
 MSI-high 153 (17%) 596 (311–966) 21 (7.1–48) 59 (30–117) 13 (5.2–31) 
CIMP status 
 Low/negative 707 (82%) 397 (198–736) 15 (4.3–39) 44 (19–90) 15 (4.5–40) 
 High 157 (18%) 559 (282–976) 20 (7.6–48) 55 (29–127) 12 (4.8–30) 
Mean LINE-1 methylation level 
 ≥60% 571 (63%) 436 (225–814) 15 (4.3–41) 46 (21–98) 15 (4.7–39) 
 <60% 335 (37%) 376 (184–717) 15 (4.9–41) 44 (18–84) 14 (4.6–40) 
KRAS mutation 
 Wild-type 538 (59%) 433 (217–820) 17 (5.2–43) 47 (22–94) 13 (4.8–37) 
 Mutant 368 (41%) 381 (206–738) 14 (3.3–38) 44 (18–93) 16 (4.5–43) 
BRAF mutation 
 Wild-type 775 (85%) 406 (208–773) 15 (4.3–41) 45 (20–90) 15 (4.5–41) 
 Mutant 138 (15%) 471 (223–882) 17 (6.0–43) 52 (23–126) 11 (5.2–33) 
PIK3CA mutation 
 Wild-type 709 (83%) 394 (202–743) 14 (4.2–39) 44 (20–94) 14 (4.4–38) 
 Mutant 141 (17%) 522 (264–945) 20 (5.6–51) 54 (23–104) 15 (6.1–40) 
Neoantigen load 
 Q1 (lowest) 105 (25%) 394 (234–743) 19 (4.6–48) 48 (21–78) 18 (6.2–48) 
 Q2 104 (25%) 335 (186–602) 15 (4.8–32) 43 (18–88) 13 (4.7–42) 
 Q3 105 (25%) 421 (185–783) 14 (3.3–39) 45 (20–101) 14 (6.1–38) 
 Q4 (highest) 103 (25%) 463 (278–886) 18 (7.6–45) 53 (20–113) 13 (5.1–34) 
Stromal immune cell density (cells/mm2)
Median (25th–75th percentiles)
CharacteristicaNLymphocytePlasma cellNeutrophilEosinophil
All cases 934 413 (212–785) 15 (4.5–41) 46 (20–93) 15 (4.8–40) 
Sex 
 Female (NHS) 521 (56%) 446 (234–804) 17 (5.2–43) 51 (22–106) 13 (4.5–37) 
 Male (HPFS) 413 (44%) 366 (196–726) 13 (3.6–39) 39 (18–78) 16 (5.2–42) 
Age (years) 
 <65 289 (31%) 412 (221–766) 17 (5.3–39) 49 (23–93) 17 (5.2–42) 
 ≥65 645 (69%) 414 (210–790) 15 (4.4–42) 45 (19–93) 14 (4.6–38) 
Year of diagnosis 
 1995 or before 308 (33%) 435 (189–869) 14 (3.7–41) 45 (21–99) 15 (4.8–41) 
 1996–2000 306 (33%) 398 (206–745) 15 (5.6–45) 46 (20–95) 14 (4.9–34) 
 2001–2008 320 (34%) 423 (241–763) 16 (4.6–37) 46 (19–88) 15 (4.6–43) 
Family history of colorectal cancer in first-degree relative(s) 
 Absent 735 (79%) 394 (212–765) 14 (4.1–38) 45 (20–92) 14 (4.9–38) 
 Present 194 (21%) 499 (234–916) 19 (6.4–62) 49 (20–97) 16 (4.7–44) 
History of inflammatory bowel disease 
 Absent 920 (98%) 413 (212–782) 15 (4.5–41) 46 (20–93) 15 (4.7–40) 
 Present 13 (1.4%) 400 (270–982) 14 (5.0–35) 38 (22–65) 18 (10–59) 
Tumor location 
 Cecum 163 (18%) 435 (224–846) 16 (5.4–43) 45 (20–99) 15 (5.6–39) 
 Ascending to transverse colon 306 (33%) 443 (212–820) 17 (4.6–44) 46 (20–93) 12 (4.6–35) 
 Splenic flexure to sigmoid colon 280 (30%) 395 (198–770) 15 (4.3–38) 45 (20–92) 17 (5.6–43) 
 Rectum 181 (19%) 407 (212–726) 14 (3.6–40) 46 (20–90) 15 (3.6–41) 
Tumor differentiation 
 Well to moderate 843 (90%) 407 (212–758) 15 (4.4–40) 45 (20–91) 15 (4.9–42) 
 Poor 89 (9.6%) 569 (223–1019) 20 (7.3–45) 55 (24–123) 11 (4.7–33) 
Extent of signet-ring cells (%) 
 0 814 (87%) 413 (212–785) 15 (4.5–42) 46 (20–92) 15 (4.8–40) 
 1–50 108 (12%) 456 (224–779) 16 (4.4–38) 44 (20–100) 12 (4.1–34) 
 ≥51 12 (1.3%) 281 (162–980) 13 (4.4–48) 46 (25–79) 34 (12–81) 
Extent of extracellular mucin (%) 
 0 550 (59%) 413 (215–767) 15 (5.2–41) 46 (21–91) 14 (4.6–38) 
 1–50 274 (29%) 436 (217–850) 16 (3.7–43) 47 (20–91) 16 (5.4–43) 
 ≥51 110 (12%) 364 (192–777) 14 (3.7–37) 40 (19–112) 13 (4.0–41) 
AJCC disease stage 
 I 198 (23%) 558 (306–986) 23 (8.5–62) 49 (26–97) 27 (8.5–65) 
 II 285 (33%) 411 (224–798) 13 (4.1–39) 51 (23–101) 13 (4.6–33) 
 III 249 (29%) 355 (189–668) 14 (4.1–33) 44 (19–83) 11 (3.4–33) 
 IV 135 (16%) 280 (160–590) 10 (4.1–25) 32 (12–75) 10 (3.7–29) 
MSI status 
 Non–MSI-high 754 (83%) 385 (196–732) 14 (4.3–39) 44 (19–89) 15 (4.5–40) 
 MSI-high 153 (17%) 596 (311–966) 21 (7.1–48) 59 (30–117) 13 (5.2–31) 
CIMP status 
 Low/negative 707 (82%) 397 (198–736) 15 (4.3–39) 44 (19–90) 15 (4.5–40) 
 High 157 (18%) 559 (282–976) 20 (7.6–48) 55 (29–127) 12 (4.8–30) 
Mean LINE-1 methylation level 
 ≥60% 571 (63%) 436 (225–814) 15 (4.3–41) 46 (21–98) 15 (4.7–39) 
 <60% 335 (37%) 376 (184–717) 15 (4.9–41) 44 (18–84) 14 (4.6–40) 
KRAS mutation 
 Wild-type 538 (59%) 433 (217–820) 17 (5.2–43) 47 (22–94) 13 (4.8–37) 
 Mutant 368 (41%) 381 (206–738) 14 (3.3–38) 44 (18–93) 16 (4.5–43) 
BRAF mutation 
 Wild-type 775 (85%) 406 (208–773) 15 (4.3–41) 45 (20–90) 15 (4.5–41) 
 Mutant 138 (15%) 471 (223–882) 17 (6.0–43) 52 (23–126) 11 (5.2–33) 
PIK3CA mutation 
 Wild-type 709 (83%) 394 (202–743) 14 (4.2–39) 44 (20–94) 14 (4.4–38) 
 Mutant 141 (17%) 522 (264–945) 20 (5.6–51) 54 (23–104) 15 (6.1–40) 
Neoantigen load 
 Q1 (lowest) 105 (25%) 394 (234–743) 19 (4.6–48) 48 (21–78) 18 (6.2–48) 
 Q2 104 (25%) 335 (186–602) 15 (4.8–32) 43 (18–88) 13 (4.7–42) 
 Q3 105 (25%) 421 (185–783) 14 (3.3–39) 45 (20–101) 14 (6.1–38) 
 Q4 (highest) 103 (25%) 463 (278–886) 18 (7.6–45) 53 (20–113) 13 (5.1–34) 

aPercentage indicates the proportion of patients with a specific clinical, pathologic, or molecular characteristic among all patients.

Study physicians, blinded to exposure data, reviewed the medical records related to colorectal cancer and recorded clinical information including the American Joint Committee on Cancer (AJCC) tumor, node, metastases stage (the 5th or 6th edition) and tumor location. The National Death Index was used to ascertain deaths of study participants and identify unreported lethal colorectal cancer cases. Patients were followed until death or end of follow-up (January 1, 2014, for HPFS; May 31, 2014, for NHS), whichever came first. Survival time was defined as the period from the date of colorectal cancer diagnosis to death or the end of follow-up for those who had not died.

Formalin-fixed paraffin-embedded tissue was collected from hospitals where the patients underwent resections of primary tumors. Tissue sections from all colorectal cancer cases were reviewed, and the diagnosis was confirmed by a single-study pathologist (S. Ogino). Histopathologic features including tumor differentiation (well/moderate vs poor), extent of signet-ring-cell morphology, and extent of extracellular mucin were recorded. DNA was extracted from formalin-fixed paraffin-embedded tumor blocks to evaluate microsatellite instability (MSI) and CpG island methylator phenotype (CIMP) status (12), KRAS (13), BRAF, and PIK3CA mutations (14), long-interspersed nucleotide element-1 (LINE-1) methylation level (15), and neoantigen load calculated from whole-exome sequencing of tumor and normal DNA pairs (ref. 16; Supplementary Table S1). TMAs were constructed using tissue cores from formalin-fixed paraffin-embedded tissue from colorectal cancer surgical resection specimens, as previously described (17). The core diameter was 0.6 mm, and cores were selected from nonperipheral regions of primary tumors to best represent overall tumor morphology. The study was conducted in accordance with the U.S. Common Rule. Written informed consent was obtained from all participants. The study protocol was approved by the institutional review boards of the Brigham and Women's Hospital and Harvard T.H. Chan School of Public Health (Boston, MA), and those of participating registries as required.

As an independent validation cohort, we included The Cancer Genome Atlas (TCGA) colorectal adenocarcinoma study (18). We used the clinical elements and survival outcome data included in the integrated TCGA Pan-Cancer Clinical Data Resource (19). Digitized H&E-stained histologic slides were available for 616 cases in TCGA data portal. We excluded cases with unrepresentative images (image not showing primary colorectal cancer; image obscured by slide markings or out of focus; or image scanned below 20× magnification), or cases with no follow-up data, resulting in 570 patients in the final analyses (Supplementary Table S2).

Immune cell detection and quantification

Two 4-μm sections, separated by a vertical depth of at least 50 μm, were cut from TMA blocks and H&E-stained in a single batch. The H&E-stained sections were scanned using the Vectra 3.0 Automated Quantitative Pathology Imaging System (Akoya Biosciences) equipped with a 20× objective. Only cores containing more than >10% tumor across the entire imaged core area were included in the subsequent analyses. In total, there were one to eight images per tumor [median 3, interquartile range (IQR) 2–4].

QuPath v0.1.2, an open source software for digital pathology image analysis (20), was used to detect and count intraepithelial and stromal lymphocytes, plasma cells, eosinophils, and neutrophils. These cell types were chosen for the analysis based on their distinctive morphologic characteristics in the H&E-stained sections. The evaluation was performed blinded to the study end points. The method began with two parallel, independent pathologist-supervised algorithmic processing steps: (i) cell detection and classification and (ii) tissue category classification (Fig. 1; Supplementary Table S3). These steps were based on earlier described functions included in QuPath (20).

Figure 1.

Immune cell detection and quantification and area segmentation in H&E-stained colorectal cancer TMAs using automated image analysis.

Figure 1.

Immune cell detection and quantification and area segmentation in H&E-stained colorectal cancer TMAs using automated image analysis.

Close modal

Cell identity data were exported from QuPath with cell coordinates, and tissue categories were exported as binary tissue category mask images. Individual cells were assigned to tissue categories through coordinate mapping, wherein the coordinates for each tissue category were identified from the tissue category mask images using the R statistical programming language (version 3.5.3; R Foundation for Statistical Computing) and the imager package. To calculate the immune cell density in each tissue compartment, we first counted the number of cells of immune cell type of interest within that compartment and then divided the counts by the tissue area in mm2. For cases with multiple tumor cores, the average (mean) densities of all available cores was calculated.

To validate the accuracy of the immune cell quantification method in relation to human assessment, immune cells in 80 tumor core images were manually annotated by a pathologist (J.P. Väyrynen), cumulatively yielding 13,713 classified immune cells. To evaluate interobserver reproducibility for automated counting, another study physician (S.A. Väyrynen) processed all tumor core images independently with a separately trained cell classifier.

We employed a nonlinear dimension reduction algorithm named Uniform Manifold Approximation and Projection (UMAP; using the umap R package; ref. 21), to project the high dimensional data, i.e., the intraepithelial and stromal densities of lymphocytes, plasma cells, neutrophils, and eosinophils, into two-dimensional space for visual inspection, in order to examine potential TMA or year-of-diagnosis–related effects on immune cell densities.

Evaluation of tumor-immune spatial relationships

The multitype G-function “G-cross” is a statistical analysis method based on the theory of point processes (22). Tumor:immune cell G-cross (GTumor:Immune cell) represents a nearest neighbor distance distribution function, which estimates the probability of finding at least one immune cell within an r radius of any tumor cell. We calculated the empiric distribution functions based on observed nearest neighbor distances with the spatstat R package (23). Both intraepithelial and stromal immune cells were included in this analysis. Estimation of the G-cross function is impeded by edge effects due to the unobservable points outside the analysis window. Therefore, we applied edge correction via the Kaplan–Meier method. To quantify the level of immune cell infiltration likely capable of effective cell-to-cell interaction with tumor cells, we computed the area under the curve of the G-cross function between 0 and 20 μm (AUC0,20μm). The distance range of 0 to 20 μm was chosen prior to analysis based upon a previously described method (22). For downstream survival analysis, the patients were grouped into ordinal AUC quartile categories (C1 to C4 from low to high AUC values).

Statistical analysis

The statistical analyses were conducted using SAS software (version 9.4, SAS Institute), and all P values were two-sided. Our primary hypothesis tested associations between the densities of intraepithelial and stromal lymphocytes, plasma cells, neutrophils, and eosinophils and cancer-specific survival using multivariable adjusted Cox proportional hazards regression models. We used the stringent two-sided α level of 0.005 for null hypothesis testing (24). All other analyses represented secondary analyses, where we interpreted our data cautiously, in addition to using the α level of 0.005.

To assess the relationships between immune cell densities and clinicopathologic features, we used Spearman's correlation test for continuous or ordinal variables, and the Wilcoxon rank-sum test or Kruskal–Wallis test for categorical variables, as appropriate. We estimated cumulative survival probabilities using the Kaplan–Meier method and compared the differences between categories using the log-rank test. For our primary analyses of colorectal cancer–specific mortality, deaths resulting from other causes were censored. We analyzed overall mortality (the NHS/HPFS and TCGA) and progression-free interval (TCGA) as secondary outcome measures. Univariable and multivariable Cox proportional hazards regression models were used to calculate the HR and 95% confidence interval (CI) for colorectal cancer–specific and overall mortality, and progression-free interval. In the NHS/HPFS cohorts, using 4,420 incident colorectal cancer cases, we applied the IPW method (25–27) to reduce selection bias due to the availability of tumor tissue. Detailed descriptions of the methods used in the survival analyses are presented in Supplementary Table S4.

Evaluation of immune infiltrate with computational pathology

To identify neutrophils, eosinophils, plasma cells, and lymphocytes in H&E-stained TMA cores from colorectal carcinomas, we conducted automated slide scanning and digital image analyses. Automated immune cell detection and classification demonstrated high concordance in relation to both a pathologist and an independently trained automated classifier. The Spearman's rank correlation coefficients (ρ) between automated method and manual counting in a subset of 80 tumor cores were 0.95 for lymphocytes, 0.74 for plasma cells, 0.71 for neutrophils, and 0.84 for eosinophils (Supplementary Fig. S1A), whereas the Spearman ρ values for core-level results between two independently trained classifiers were 0.96 for lymphocytes, 0.74 for plasma cells, 0.71 for neutrophils, and 0.83 for eosinophils (Supplementary Fig. S1B). These results suggested that these immune cells could be reproducibly identified via pathologist-supervised machine learning algorithms. The Spearman ρ values for immune cell densities between two randomly chosen cores of tumors with two or more available cores were 0.56 for lymphocytes, 0.43 for plasma cells, 0.35 for neutrophils, and 0.43 for eosinophils (Supplementary Fig. S1C), indicating moderate core-to-core correlation.

Characteristics of immune infiltrate in relation to clinicopathologic features

We analyzed immune cell densities in 934 colorectal cancer cases in the NHS/HPFS cohorts (Table 1). Of the four immune cell types under study, lymphocytes had the highest overall density, followed by neutrophils, plasma cells, and eosinophils (Fig. 2A). The cell densities were consistent across TMAs (Supplementary Fig. S2), and visualization of colorectal cancer cases using UMAP according to immune cell densities showed no clear TMA-related or year-of-diagnosis–related effects (Supplementary Fig. S3). There was predominantly a low-to-moderate correlation between the densities of different immune cell types; the highest was observed between stromal lymphocytes and stromal plasma cells (Spearman ρ = 0.73; Fig. 2B).

Figure 2.

Relationships between densities of intraepithelial and stromal immune cells and clinicopathologic features. A, Boxplots of the distribution of intraepithelial (IEL) and stromal (S) immune cell densities. B, Correlation matrix of Spearman correlation coefficients between the densities of intraepithelial and stromal immune cells. C, Heatmap of the relationships between clinicopathologic features and the densities of lymphocytes, plasma cells, neutrophils, and eosinophils. P values are based on the correlation analysis of immune cell densities and continuous or ordinal variables (AJCC stage, Neoantigen load) by Spearman's rank correlation test or the comparison of immune cell densities across categorical variable categories (tumor location, tumor differentiation, MSI status, CIMP status) by the Kruskal–Wallis test or Wilcoxon rank-sum test.

Figure 2.

Relationships between densities of intraepithelial and stromal immune cells and clinicopathologic features. A, Boxplots of the distribution of intraepithelial (IEL) and stromal (S) immune cell densities. B, Correlation matrix of Spearman correlation coefficients between the densities of intraepithelial and stromal immune cells. C, Heatmap of the relationships between clinicopathologic features and the densities of lymphocytes, plasma cells, neutrophils, and eosinophils. P values are based on the correlation analysis of immune cell densities and continuous or ordinal variables (AJCC stage, Neoantigen load) by Spearman's rank correlation test or the comparison of immune cell densities across categorical variable categories (tumor location, tumor differentiation, MSI status, CIMP status) by the Kruskal–Wallis test or Wilcoxon rank-sum test.

Close modal

The relationships between immune cell densities and main tumor characteristics are summarized in Fig. 2C. Advanced tumor stage was associated with lower densities of stromal lymphocytes, plasma cells, neutrophils, and eosinophils (P < 0.001 for all) and intraepithelial neutrophils (P = 0.001), whereas poor differentiation was associated with higher densities of intraepithelial lymphocytes (P < 0.001) and plasma cells (P = 0.002).

Given that mismatch repair deficiency is common in poorly differentiated tumors (28), we hypothesized that the increased burden of immunogenic neopeptides (29) in these tumors might be associated with higher immune cell densities. Supporting our hypothesis, mismatch repair deficiency, as measured by an MSI-high phenotype, was strongly associated with higher densities of both intraepithelial and stromal lymphocytes (P < 0.001). Unexpectedly, intraepithelial and stromal neutrophils were also more frequent in MSI-high tumors (P < 0.001), whereas there was no evidence of a strong association between MSI status and eosinophil or plasma cell density. Intraepithelial densities of lymphocytes and neutrophils were positively correlated with estimated neoantigen load (P < 0.001). Of other tumor molecular features, LINE-1 hypomethylation was associated with lower intraepithelial lymphocyte density (P < 0.001), BRAF mutation was associated with higher intraepithelial lymphocyte density (P < 0.001), and PIK3CA mutation was associated with higher stromal lymphocyte density (P = 0.009; Supplementary Fig. S4).

Survival analyses

Our primary aim was to evaluate the prognostic significance of intraepithelial and stromal immune cells. During the median follow-up time of 12.3 years (IQR, 8.7–16.3 years) for censored cases, there were 572 all-cause deaths, including 290 colorectal cancer–specific deaths.

In univariable survival analyses (Fig. 3 and Table 2; Supplementary Fig. S5), stromal lymphocytes, plasma cells, neutrophils, and eosinophils were significantly associated with better cancer-specific survival (all Ptrend < 0.001), whereas intraepithelial lymphocytes (Ptrend = 0.002) and neutrophils (Ptrend = 0.003) were also significantly associated with cancer-specific survival (Supplementary Fig. S4).

Figure 3.

IPW-adjusted Kaplan–Meier curves of colorectal cancer–specific survival according to ordinal quartile categories (C1–C4) of stromal lymphocyte (A), plasma cell (B), neutrophil (C), and eosinophil (D), densities.

Figure 3.

IPW-adjusted Kaplan–Meier curves of colorectal cancer–specific survival according to ordinal quartile categories (C1–C4) of stromal lymphocyte (A), plasma cell (B), neutrophil (C), and eosinophil (D), densities.

Close modal
Table 2.

Densities of intraepithelial and stromal immune cells and patient survival with IPW.

Colorectal cancer–specific survivalOverall survival
Number of casesNumber of eventsUnivariableMultivariablebNumber of eventsUnivariableMultivariable
HR (95% CI)aHR (95% CI)aHR (95% CI)aHR (95% CI)a
Tumor intraepithelial region 
Lymphocyte density 
 C1 234 94 1 (referent) 1 (referent) 162 1 (referent) 1 (referent) 
 C2 233 72 0.72 (0.52–0.99) 0.77 (0.56–1.07) 139 0.83 (0.64–1.08) 0.83 (0.65–1.07) 
 C3 234 64 0.68 (0.49–0.96) 0.73 (0.52–1.02) 130 0.79 (0.61–1.03) 0.77 (0.59–1.01) 
 C4 233 60 0.58 (0.41–0.82) 0.68 (0.48–0.98) 141 0.82 (0.63–1.06) 0.73 (0.56–0.94) 
Ptrendc   0.002 0.029  0.11 0.013 
Plasma cell density 
 C1 343 123 1 (referent) 1 (referent) 227 1 (referent) 1 (referent) 
 C2 196 42 0.54 (0.37–0.77) 0.53 (0.36–0.76) 105 0.65 (0.50–0.83) 0.67 (0.52–0.86) 
 C3 199 64 0.92 (0.66–1.26) 0.96 (0.69–1.33) 115 0.88 (0.69–1.12) 0.93 (0.73–1.20) 
 C4 196 61 0.82 (0.59–1.14) 0.77 (0.55–1.08) 125 0.88 (0.68–1.13) 0.87 (0.68–1.12) 
Ptrendc   0.44 0.36  0.44 0.48 
Neutrophil density 
 C1 372 137 1 (referent) 1 (referent) 246 1 (referent) 1 (referent) 
 C2 188 57 0.79 (0.57–1.09) 1.10 (0.81–1.51) 111 0.77 (0.59–0.99) 0.96 (0.76–1.21) 
 C3 187 49 0.64 (0.46–0.90) 0.74 (0.51–1.06) 107 0.67 (0.52–0.87) 0.70 (0.54–0.91) 
 C4 187 47 0.63 (0.44–0.90) 0.67 (0.46–0.97) 108 0.73 (0.57–0.93) 0.67 (0.52–0.87) 
Ptrendc   0.003 0.012  0.002 <0.001 
Eosinophil density 
 C1 482 171 1 (referent) 1 (referent) 316 1 (referent) 1 (referent) 
 C2 150 40 0.62 (0.43–0.88) 0.64 (0.42–0.98) 84 0.72 (0.54–0.95) 0.80 (0.61–1.07) 
 C3 151 40 0.73 (0.51–1.05) 0.73 (0.50–1.06) 87 0.87 (0.68–1.11) 0.80 (0.61–1.04) 
 C4 151 39 0.71 (0.49–1.03) 0.82 (0.58–1.16) 85 0.75 (0.57–0.99) 0.81 (0.62–1.06) 
Ptrendc   0.027 0.084  0.030 0.044 
Tumor stromal region 
Lymphocyte density 
 C1 233 110 1 (referent) 1 (referent) 173 1 (referent) 1 (referent) 
 C2 234 66 0.54 (0.39–0.74) 0.64 (0.46–0.88) 132 0.64 (0.49–0.83) 0.64 (0.49–0.83) 
 C3 233 60 0.49 (0.35–0.68) 0.59 (0.42–0.83) 134 0.64 (0.50–0.83) 0.66 (0.51–0.85) 
 C4 234 54 0.43 (0.31–0.61) 0.51 (0.36–0.71) 133 0.59 (0.45–0.77) 0.56 (0.43–0.72) 
Ptrendc   <0.001 <0.001  <0.001 <0.001 
Plasma cell density 
 C1 233 88 1 (referent) 1 (referent) 156 1 (referent) 1 (referent) 
 C2 234 86 1.01 (0.74–1.38) 0.95 (0.70–1.30) 150 1.03 (0.79–1.34) 1.01 (0.78–1.31) 
 C3 233 55 0.62 (0.44–0.89) 0.56 (0.39–0.78) 127 0.79 (0.61–1.03) 0.72 (0.55–0.93) 
 C4 234 61 0.61 (0.43–0.86) 0.61 (0.43–0.86) 139 0.78 (0.61–1.01) 0.73 (0.57–0.93) 
Ptrendc   <0.001 <0.001  0.016 0.002 
Neutrophil density 
 C1 234 96 1 (referent) 1 (referent) 169 1 (referent) 1 (referent) 
 C2 233 68 0.71 (0.51–0.99) 0.88 (0.64–1.23) 135 0.67 (0.52–0.88) 0.81 (0.62–1.06) 
 C3 234 67 0.66 (0.48–0.91) 0.88 (0.64–1.22) 130 0.67 (0.52–0.86) 0.79 (0.61–1.01) 
 C4 233 59 0.55 (0.39–0.78) 0.67 (0.47–0.96) 138 0.64 (0.51–0.82) 0.67 (0.51–0.87) 
Ptrendc   <0.001 0.034  <0.001 0.003 
Eosinophil density 
 C1 233 95 1 (referent) 1 (referent) 161 1 (referent) 1 (referent) 
 C2 234 79 0.90 (0.66–1.23) 0.88 (0.65–1.19) 147 0.89 (0.69–1.14) 0.87 (0.67–1.11) 
 C3 234 68 0.73 (0.52–1.01) 0.73 (0.52–1.02) 146 0.84 (0.65–1.07) 0.92 (0.71–1.19) 
 C4 233 48 0.47 (0.32–0.67) 0.49 (0.34–0.71) 118 0.59 (0.45–0.76) 0.63 (0.48–0.82) 
Ptrendc   <0.001 <0.001  <0.001 0.002 
Colorectal cancer–specific survivalOverall survival
Number of casesNumber of eventsUnivariableMultivariablebNumber of eventsUnivariableMultivariable
HR (95% CI)aHR (95% CI)aHR (95% CI)aHR (95% CI)a
Tumor intraepithelial region 
Lymphocyte density 
 C1 234 94 1 (referent) 1 (referent) 162 1 (referent) 1 (referent) 
 C2 233 72 0.72 (0.52–0.99) 0.77 (0.56–1.07) 139 0.83 (0.64–1.08) 0.83 (0.65–1.07) 
 C3 234 64 0.68 (0.49–0.96) 0.73 (0.52–1.02) 130 0.79 (0.61–1.03) 0.77 (0.59–1.01) 
 C4 233 60 0.58 (0.41–0.82) 0.68 (0.48–0.98) 141 0.82 (0.63–1.06) 0.73 (0.56–0.94) 
Ptrendc   0.002 0.029  0.11 0.013 
Plasma cell density 
 C1 343 123 1 (referent) 1 (referent) 227 1 (referent) 1 (referent) 
 C2 196 42 0.54 (0.37–0.77) 0.53 (0.36–0.76) 105 0.65 (0.50–0.83) 0.67 (0.52–0.86) 
 C3 199 64 0.92 (0.66–1.26) 0.96 (0.69–1.33) 115 0.88 (0.69–1.12) 0.93 (0.73–1.20) 
 C4 196 61 0.82 (0.59–1.14) 0.77 (0.55–1.08) 125 0.88 (0.68–1.13) 0.87 (0.68–1.12) 
Ptrendc   0.44 0.36  0.44 0.48 
Neutrophil density 
 C1 372 137 1 (referent) 1 (referent) 246 1 (referent) 1 (referent) 
 C2 188 57 0.79 (0.57–1.09) 1.10 (0.81–1.51) 111 0.77 (0.59–0.99) 0.96 (0.76–1.21) 
 C3 187 49 0.64 (0.46–0.90) 0.74 (0.51–1.06) 107 0.67 (0.52–0.87) 0.70 (0.54–0.91) 
 C4 187 47 0.63 (0.44–0.90) 0.67 (0.46–0.97) 108 0.73 (0.57–0.93) 0.67 (0.52–0.87) 
Ptrendc   0.003 0.012  0.002 <0.001 
Eosinophil density 
 C1 482 171 1 (referent) 1 (referent) 316 1 (referent) 1 (referent) 
 C2 150 40 0.62 (0.43–0.88) 0.64 (0.42–0.98) 84 0.72 (0.54–0.95) 0.80 (0.61–1.07) 
 C3 151 40 0.73 (0.51–1.05) 0.73 (0.50–1.06) 87 0.87 (0.68–1.11) 0.80 (0.61–1.04) 
 C4 151 39 0.71 (0.49–1.03) 0.82 (0.58–1.16) 85 0.75 (0.57–0.99) 0.81 (0.62–1.06) 
Ptrendc   0.027 0.084  0.030 0.044 
Tumor stromal region 
Lymphocyte density 
 C1 233 110 1 (referent) 1 (referent) 173 1 (referent) 1 (referent) 
 C2 234 66 0.54 (0.39–0.74) 0.64 (0.46–0.88) 132 0.64 (0.49–0.83) 0.64 (0.49–0.83) 
 C3 233 60 0.49 (0.35–0.68) 0.59 (0.42–0.83) 134 0.64 (0.50–0.83) 0.66 (0.51–0.85) 
 C4 234 54 0.43 (0.31–0.61) 0.51 (0.36–0.71) 133 0.59 (0.45–0.77) 0.56 (0.43–0.72) 
Ptrendc   <0.001 <0.001  <0.001 <0.001 
Plasma cell density 
 C1 233 88 1 (referent) 1 (referent) 156 1 (referent) 1 (referent) 
 C2 234 86 1.01 (0.74–1.38) 0.95 (0.70–1.30) 150 1.03 (0.79–1.34) 1.01 (0.78–1.31) 
 C3 233 55 0.62 (0.44–0.89) 0.56 (0.39–0.78) 127 0.79 (0.61–1.03) 0.72 (0.55–0.93) 
 C4 234 61 0.61 (0.43–0.86) 0.61 (0.43–0.86) 139 0.78 (0.61–1.01) 0.73 (0.57–0.93) 
Ptrendc   <0.001 <0.001  0.016 0.002 
Neutrophil density 
 C1 234 96 1 (referent) 1 (referent) 169 1 (referent) 1 (referent) 
 C2 233 68 0.71 (0.51–0.99) 0.88 (0.64–1.23) 135 0.67 (0.52–0.88) 0.81 (0.62–1.06) 
 C3 234 67 0.66 (0.48–0.91) 0.88 (0.64–1.22) 130 0.67 (0.52–0.86) 0.79 (0.61–1.01) 
 C4 233 59 0.55 (0.39–0.78) 0.67 (0.47–0.96) 138 0.64 (0.51–0.82) 0.67 (0.51–0.87) 
Ptrendc   <0.001 0.034  <0.001 0.003 
Eosinophil density 
 C1 233 95 1 (referent) 1 (referent) 161 1 (referent) 1 (referent) 
 C2 234 79 0.90 (0.66–1.23) 0.88 (0.65–1.19) 147 0.89 (0.69–1.14) 0.87 (0.67–1.11) 
 C3 234 68 0.73 (0.52–1.01) 0.73 (0.52–1.02) 146 0.84 (0.65–1.07) 0.92 (0.71–1.19) 
 C4 233 48 0.47 (0.32–0.67) 0.49 (0.34–0.71) 118 0.59 (0.45–0.76) 0.63 (0.48–0.82) 
Ptrendc   <0.001 <0.001  <0.001 0.002 

aIPW was applied to reduce a bias due to the availability of tumor tissue after cancer diagnosis (see “Statistical Analysis” subsection for details).

bThe multivariable Cox regression model initially included sex, age, year of diagnosis, family history of colorectal cancer, tumor location, tumor differentiation, disease stage, MSI, CIMP, KRAS, BRAF, and PIK3CA mutations, and LINE-1 methylation level. A backward elimination with a threshold P of 0.05 was used to select variables for the final models.

cPtrend value was calculated across the four ordinal categories of the density of each immune cell within tumor intraepithelial and stromal regions in the IPW-adjusted Cox regression model.

In multivariable Cox proportional hazards regression models (Table 2; Supplementary Table S5), higher densities of stromal lymphocytes, plasma cells, and eosinophils were associated with longer cancer-specific survival (all Ptrend < 0.001) independent of potential confounders, including MSI, CIMP, BRAF mutation, LINE-1 methylation, tumor stage, and tumor grade. For eosinophil density C4 versus C1, the HR for colorectal cancer–specific mortality was 0.49 (95% CI, 0.34–0.71).

As secondary analyses, we examined the survival association of stromal lymphocytes, plasma cells, and eosinophils with colorectal cancer mortality in strata of tumor MSI status. The trends between higher densities of these cell types and better cancer-specific survival did not significantly differ by MSI status (Pinteraction > 0.3, Supplementary Table S6).

To directly compare the relative prognostic value of stromal lymphocyte, plasma cell, and eosinophil densities, we included these three variables in one Cox regression model, adjusting for each other, and used backward elimination with a threshold P value of 0.05 to select significant variables (Supplementary Table S4). This analysis resulted in stromal lymphocyte density (P < 0.001) and stromal eosinophil density (P = 0.011) remaining in the final model. Multivariable-adjusted HRs for colorectal cancer mortality according to the densities of stromal lymphocytes and eosinophils are shown in Supplementary Table S7.

Spatial analysis

Given our ability to precisely identify each immune cell's location, we explored whether spatial characteristics of immune infiltrates in relation to tumor cells would be associated with patient survival. We employed the GTumor:Immune cell (r) function (22) to estimate the likelihood of any tumor cell in the sample having at least one immune cell of the specified type within an r μm radius. The AUC of the function within the specified radius r is influenced by both the density of immune cells (higher density results in a higher AUC) and the location of the immune cells (immune cells located closer to tumor cells result in a higher AUC). Examples of different patterns of immune cell infiltration and corresponding GTumor:Immune cell plots are presented in Fig. 4A–D.

Figure 4.

Spatial analysis of tumor-immune infiltrates with Tumor:Immune cell G-cross function (GTumor:Immune cell). A–D, Example lymphocyte infiltration patterns and corresponding GTumor:Lymphocyte(r) plots, estimating the probability of any tumor cell having at least one neighboring lymphocyte within an r μm radius. High immune cell infiltrate localizing near tumor cells (D) results in a higher AUC than low immune cell infiltrate mainly locating away from tumor cells (A). The G-cross function was summarized as AUC within a 20 μm radius (AUC0,20μm). E–H, GTumor:Immune cell AUC0,20μm quartiles (C1–C4) in relation to cancer-specific survival. The multivariable Cox regression models initially included sex, age, year of diagnosis, family history of colorectal cancer, tumor location, tumor differentiation, disease stage, MSI, CIMP, KRAS, BRAF, and PIK3CA mutations, and LINE-1 methylation level. A backward elimination with a threshold P of 0.05 was used to select variables for the final models.

Figure 4.

Spatial analysis of tumor-immune infiltrates with Tumor:Immune cell G-cross function (GTumor:Immune cell). A–D, Example lymphocyte infiltration patterns and corresponding GTumor:Lymphocyte(r) plots, estimating the probability of any tumor cell having at least one neighboring lymphocyte within an r μm radius. High immune cell infiltrate localizing near tumor cells (D) results in a higher AUC than low immune cell infiltrate mainly locating away from tumor cells (A). The G-cross function was summarized as AUC within a 20 μm radius (AUC0,20μm). E–H, GTumor:Immune cell AUC0,20μm quartiles (C1–C4) in relation to cancer-specific survival. The multivariable Cox regression models initially included sex, age, year of diagnosis, family history of colorectal cancer, tumor location, tumor differentiation, disease stage, MSI, CIMP, KRAS, BRAF, and PIK3CA mutations, and LINE-1 methylation level. A backward elimination with a threshold P of 0.05 was used to select variables for the final models.

Close modal

We hypothesized that higher AUC of GTumor:Immune cell, reflecting a high density of the specified immune cell type clustering near tumor cells, would be associated with better survival. We conducted univariable and multivariable Cox regression analyses, using quartiles of GTumor:Immune cell AUC0,20μm as the input (Fig. 4EH). We restricted the radius to 20 μm to model close and plausibly direct interactions between tumor cells and specified immune cells. The analysis indicated that both high GTumor:Lymphocyte AUC0,20μm (Ptrend = 0.002) and high GTumor:Eosinophil AUC0,20μm (Ptrend < 0.001) were associated with better cancer-specific survival, whereas GTumor:Plasma cell and GTumor:Neutrophil AUC0,20μm quartiles were not statistically significant in the multivariable Cox regression models with α = 0.005. Cox proportional hazards models with composite variables jointly classifying tumors according to GTumor:immune cell and immune cell density uncovered differential significance for tumor-immune cell proximity and immune cell density for different immune cells (Supplementary Table S8). For example, both lymphocyte density and proximity to tumor cells contributed to better cancer-specific survival, whereas, for plasma cells, only density contributed to better cancer-specific survival.

Validation cohort

We analyzed an independent validation cohort of 570 colorectal cancer cases in TCGA (Supplementary Table S2). Compared with the NHS/HPFS cohorts, the survival data of this cohort were limited by short follow-up and low numbers of events, particularly for the cancer-specific survival analysis (19). For cancer-specific survival analysis, during the median follow-up time of 1.8 years (IQR, 1.1–3.0 years) for censored cases, there were 75 events; for overall survival analysis, during the median follow-up time of 1.8 years (IQR, 1.1–2.9 years) for censored cases, there were 120 events; and for progression-free interval analysis, during the median follow-up time of 1.7 years (IQR, 1.0–2.8 years) for censored cases, there were 151 events.

In agreement with the findings of the NHS/HPFS cohorts, high stromal eosinophil density was associated with longer cancer-specific survival, overall survival, and progression-free interval (all Ptrend < 0.001) in multivariable Cox proportional hazards regression models (Supplementary Table S9). For stromal eosinophil density C4 versus C1, the HR for colorectal cancer–specific mortality was 0.21 (95% CI, 0.10–0.47). The point estimates for stromal lymphocyte density C4 versus C1 were close to those seen in the NHS/HPFS cohorts (cancer-specific survival: HR 0.55, 95% CI, 0.28–1.09; overall survival: HR 0.74, 95% CI, 0.44–1.25; progression-free interval: HR 0.61, 95% CI, 0.37–1.00), although statistical significance was not reached at α = 0.005. High intraepithelial eosinophil density was also associated with longer overall survival (Ptrend < 0.001) and progression-free interval (Ptrend < 0.001), whereas higher intraepithelial lymphocyte density was associated with longer progression-free interval (Ptrend = 0.009). Intraepithelial or stromal densities of plasma cells or neutrophils were not significantly associated with survival at α = 0.005.

Consistent with findings in the NHS/HPFS cohorts, high GTumor:Eosinophil AUC0,20μm was associated with longer overall survival (Ptrend < 0.001) and progression-free interval (Ptrend < 0.001), and high GTumor:Lymphocyte AUC0,20μm was also associated with longer progression-free interval (Ptrend = 0.002; Supplementary Table S10).

In this study, we evaluated the prognostic significance of computationally phenotyped immune cells in the colorectal cancer tumor microenvironment utilizing two large U.S.-based prospective cohort studies, as well as an independent cohort of 570 colorectal cancers from TCGA. Our main findings indicate that high densities of lymphocytes and eosinophils in tumor stroma are associated with better survival independent of potential confounders. These results support the potential of machine learning–based evaluation of immune cell infiltrate utilizing H&E-stained sections as a prognostic tool for colorectal cancer and identify previously underappreciated immune cell subsets as harboring prognostic relevance in colorectal cancer.

Lymphocytes are a heterogeneous group of cells with roles in both adaptive immunity (T cells, B cells) and innate immunity [natural killer (NK) cells] and which contribute to a wide variety of immune regulatory and effector functions, including cytokine production (T cells, B cells, NK cells), antigen presentation (B cells), cytotoxicity (cytotoxic T cells and NK cells), and immunologic memory (memory T and B cells). High densities of CD3+ and CD8+ T cells are considered promising favorable prognostic markers in colorectal cancer (8), and there is some evidence that a high density of MS4A1+ (CD20+) B cells is also associated with longer survival (30). H&E staining–based evaluation is not able to distinguish different types of lymphocytes. However, we found that high stromal lymphocyte density was strongly associated with lower cancer mortality independent of MSI status and other potential confounders in the NHS/HPFS cohorts. In those cohorts, the prognostic significance of intraepithelial lymphocytes appeared weaker than that of stromal lymphocytes. Although the reason for this is unclear, it may be related to generally higher lymphocyte densities in tumor stroma, providing for more robust outcomes analysis for stromal rather than intraepithelial regions.

Plasma cells are terminally differentiated B cells specialized in antibody production (31). Using a cohort of 557 patients with stage I–IV colorectal cancer, Berntsson and colleagues found that high SDC1+ (CD138+) plasma cell density was associated with better survival in colorectal cancer in univariable but not multivariable Cox regression models (30). However, there are few other studies assessing their relationship with survival in colorectal cancer (32). In our study, results from the NHS/HPFS cohorts but not the TCGA validation cohort support the association between higher stromal plasma cell infiltration and better cancer-specific survival.

Neutrophils are primary effector cells of innate immunity, representing a first-line defense against microbial infection (33). Emerging evidence suggests that neutrophils, recruited to human tumors, may enhance or inhibit tumor progression by multiple mechanisms including cytotoxicity and the release of inflammatory mediators, growth factors, and proteases (33). In colorectal cancer, several studies, utilizing CEACAM8 (CD66b; refs. 34–36) or MPO (37), to detect neutrophils have reported an association between higher neutrophil density and better survival. Notably, however, these markers are not specific to neutrophils, as CEACAM8 is also expressed by eosinophils (34, 38) and MPO is expressed by some macrophages. To overcome this limitation, in a recent study, neutrophils were recognized by morphology and manually counted in H&E-stained sections (39). That study could not demonstrate a statistically significant association between neutrophil density and overall survival, although the study's power was limited by sample size (221 patients with stage I–IV colorectal cancer). In our study, neutrophils were also identified by their morphology in H&E-stained images. The analyses of the NHS/HPFS cohorts support the association between high intraepithelial and stromal neutrophil density and better overall survival (Ptrend < 0.001 and Ptrend = 0.003, respectively), although statistical significance was not achieved with cancer-specific analysis at the α level of 0.005 (Intraepithelial, Ptrend = 0.012; stromal, Ptrend = 0.034), or in the TCGA validation cohort.

Eosinophils are innate immune cells that play a central role in defense against parasitic infection and are also involved in the pathogenesis of allergy and asthma (40). Like neutrophils, eosinophils can be reliably identified by their morphology in H&E-stained sections, as was done in our study. Several earlier studies found an association between higher eosinophil densities in colorectal cancer and better survival (41, 42). Our study, utilizing a large sample of two U.S. prospective cohort studies, as well as TCGA, supports the association between higher density of stromal eosinophils and better cancer-specific survival and overall survival. This finding is consistent with recent experimental evidence for an antitumorigenic role for eosinophils in a mouse colorectal cancer model (43). In that model, the tumor-inhibiting role of eosinophils was independent of CD8+ T cells (43). In our study, there was only moderate-or-weak correlation between eosinophil density and lymphocyte density (stromal: Spearman ρ = 0.47; intraepithelial: Spearman ρ = 0.17), suggesting that the factors modulating the density of these immune cells may be at least partially independent. Indeed, our study did not support an association between the density of stromal eosinophils and MSI or neoantigen load and, to our knowledge, such relationships have not been previously reported.

We utilized machine learning–based image analysis in the detection of four immune cell types with distinctive morphology in H&E-stained sections. During the past few years, such methods have been increasingly applied for the analysis of histopathology images (44). Many widely used deep learning methods use pixel patches (such as 100 × 100 pixels) as an input (44, 45), whereas the method we used included separate explicit steps for cell detection (based on high hematoxylin optical density in nuclei) and classification (based on morphologic features). We used QuPath, an open source software, which has been widely adapted and validated for digital pathology image analysis (20, 46, 47). Our method achieved good to excellent accuracy in relation to laborious manual counting (Spearman ρ = 0.71–0.95), and good to excellent reproducibility between two independently trained classifiers (Spearman ρ = 0.71–0.96), dependent on cell type. Better performance in detecting lymphocytes and eosinophils compared with plasma cells and neutrophils may be related to higher variability of morphology among plasma cells relative to other cell types and issues related to distinguishing neutrophils with nuclear lobules that are only partially visible, as well as differentiating neutrophils from apoptotic nuclear fragments and mitoses. A particular benefit of machine learning–based quantification is that all the images have been evaluated with the same criteria, improving data consistency.

The format of our data also enabled analysis of spatial features of immune cell infiltration in relation to tumor cells. We utilized the GTumor:Immune cell (r) function to approximate the likelihood of tumor cells having an immune cell neighbor of the specified type within an r μm radius. We hypothesized that high GTumor:Immune cell AUC0,20μm, correlating with a high density of immune cells clustering near tumor cells, would be associated with better outcome, and that this association would be more pronounced in analyses involving cell types that may show cytotoxicity toward tumor cells, including lymphocytes, neutrophils, and eosinophils, although such an association may not hold for the GTumor:Plasma cell analysis, as potential antitumor antibody production would not require close contact to tumor cells. Supporting our hypothesis, we observed that high AUC0,20μm for lymphocytes and eosinophils had a statistically significant multivariable-adjusted association with longer survival. This association was also observed for eosinophils in the TCGA validation cohort. However, in the NHS/HPFS cohorts, the GTumor:Eosinophil C3 category did not follow the trend toward better prognosis, the reasons for which are not clear. There are many potential approaches to characterizing tumor-immune spatial interactions (48, 49), and optimal methods to improve prognostic classification require further investigation.

Several limitations need to be considered in the interpretation of our results. First, information on cancer treatment was lacking. Nevertheless, treatments had been based mainly on disease stage rather than tumor-immune infiltrate, and we adjusted the survival analyses for disease stage. Second, the study evaluated multiple hypotheses. However, we used a stringent α level of 0.005 to improve reproducibility and decrease false positive findings (24). Third, the NHS/HPFS cohort analysis was performed using TMAs, and the immune cell infiltrates present in the small area of each tumor in TMA format may not fully represent immune cell infiltrates in the whole tumor area. However, numerous studies have generated reproducible results using TMAs (50). In addition, multiple TMA core images were examined for most tumors (mean of 3.2 cores per tumor). Because tumor regions for TMA coring were selected without specific regard to immune cell infiltrates, any measurement errors due to the use of the TMA format would likely drive our results toward the null hypotheses. Moreover, TMAs enabled us to investigate more than 900 tumors with uniform staining quality, improving the accuracy of the immune cell detection. Fourth, the TMAs only included cores from the tumor center, precluding examination of the immune cell infiltrate at the invasive margin and tertiary lymphoid structures, both of which have been reported to harbor prognostic significance (7, 8, 10).

Use of TCGA as a validation cohort imposed some additional limitations. The follow-up period in this cohort was short, and cancer-specific survival data had to be approximated based on “tumor_status” and “vital_status” variables, as described before (19). The event rate was low, resulting in poor statistical power, particularly for cancer-specific survival analysis. The H&E staining and image quality were highly variable, thereby hindering optimal recognition of immune cell types using nuclear morphology in a subset of images and resulting in lower lymphocyte, plasma cell, and neutrophil densities as compared with the NHS/HPFS cohorts. Most notably, identification of plasma cells, based predominantly on finely detailed nuclear morphology, was difficult in a large subset of cases. This may account for differences between the cohorts regarding the significance of this cell type. Conversely, eosinophil densities were similar in both cohorts, and stromal eosinophil density consistently had strong prognostic value.

The advantages of the study include the availability of a comprehensive data set of potential confounding factors in the NHS/HPFS cohorts, such as family history of colorectal cancer, MSI status, CIMP, LINE-1 methylation, and BRAF mutation status, which were included in the survival analyses. Nevertheless, several clinically relevant data elements were unavailable, such as tumor budding, NRAS status, and history of Lynch syndrome. In addition, the study population was based on a large number of hospitals across the United States, facilitating the generalizability of our results. Moreover, examination of multiple immune cell types at the same time enabled a more granular view of the colorectal cancer tissue microenvironment than analyses based on a single-cell type.

In conclusion, our results support the potential of machine learning–based evaluation of the immune cell infiltrates utilizing H&E-stained sections as a prognostic parameter in colorectal cancer. In particular, high density of eosinophils in tumor stroma was associated with favorable outcome. The results also suggest that the spatial patterns of immune cell infiltrate in relation to tumor cells harbor biologically and prognostically relevant information.

R. Nishihara is an employee of Pfizer. C.S. Fuchs is a paid consultant for Agios, Bain Capital, CytomX Therapeutics, Daiichi-Sankyo, Eli Lilly, Entrinsic Health, Evolveimmune Therapeutics, Genentech, Merck, Taiho, and Unum Therapeutics; holds ownership interest (including patents) in CytomX Therapeutics, Entrinsic Health, and Evolveimmune Therapeutics; and reports receiving other remuneration from Amylin Pharmaceuticals. J.A. Meyerhardt reports receiving other commercial research support from Taiho. M. Giannakis reports receiving commercial research grants from Bristol-Myers Squibb and Merck. J. Borowsky, M.C. Lau, J.A. Nowak and S. Ogino are listed as co-inventors on a provisional application for a patent titled “System for and Method of Discovering Spatially-Derived Signatures of Tumor-Immune Cell Interactions through Tumor-Immune Partitioning and Clustering” regarding novel methods for characterizing immune cell distributions in solid tumors that has been filed through Partners Healthcare. No potential conflicts of interest were disclosed by the other authors.

The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.

Conception and design: J.P. Väyrynen, M. Zhao, C.S. Fuchs, S. Ogino, J.A. Nowak

Development of methodology: J.P. Väyrynen, M.C. Lau, K. Haruki, S.A. Väyrynen, J. Borowsky, R. Nishihara

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.P. Väyrynen, M.C. Lau, K. Haruki, S.A. Väyrynen, J. Borowsky, K. Fujiyoshi, K. Arima, T.S. Twombly, S. Aminmozaffari, M. Song, A.T. Chan, C.S. Fuchs, S. Ogino

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.P. Väyrynen, M.C. Lau, K. Haruki, J. Borowsky, Y. Baba, M. Song, K. Wu, J.A. Meyerhardt, M. Giannakis, S. Ogino, J.A. Nowak

Writing, review, and/or revision of the manuscript: J.P. Väyrynen, M.C. Lau, K. Haruki, S.A. Väyrynen, J. Borowsky, M. Zhao, K. Fujiyoshi, K. Arima, T.S. Twombly, J. Kishikawa, N. Akimoto, T. Ugai, A. Da Silva, M. Song, K. Wu, A.T. Chan, C.S. Fuchs, J.A. Meyerhardt, M. Giannakis, S. Ogino, J.A. Nowak

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Arima, T.S. Twombly, S. Gu, S. Aminmozaffari, S. Shi, S. Ogino

Study supervision: C.S. Fuchs, S. Ogino, J.A. Nowak

We would like to thank the participants and staff of the NHS and the HPFS for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, and WY. The authors assume full responsibility for analyses and interpretation of these data. Portions of this research were conducted on the O2 High Performance Compute Cluster, supported by the Research Computing Group, at Harvard Medical School. See http://rc.hms.harvard.edu for more information. The results shown here are in part based upon data generated by the TCGA Research Network: https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga. This work was supported by U.S. NIH grants (P01 CA87969 to M.J. Stampfer; UM1 CA186107 to M.J. Stampfer; P01 CA55075 to W.C. Willett; UM1 CA167552 to W.C. Willett; U01 CA167552 to W.C. Willett and L.A. Mucci; P50 CA127003 to C.S. Fuchs; R01 CA118553 to C.S. Fuchs; R01 CA169141 to C.S. Fuchs; R01 CA137178 to A.T. Chan; K24 DK098311 to A.T. Chan; R35 CA197735 to S. Ogino; R01 CA151993 to S. Ogino; K07 CA190673 to R. Nishihara; R03 CA197879 to K. Wu; R21 CA222940 to K. Wu and M. Giannakis; and R21 CA230873 to K. Wu and S. Ogino); by Cancer Research UK Grand Challenge Award (OPTIMISTICC, UK C10674/A27140 to M. Giannakis and S. Ogino); by Nodal Award (2016-02) from the Dana-Farber Harvard Cancer Center (to S. Ogino); by the Stand Up to Cancer Colorectal Cancer Dream Team Translational Research Grant (SU2C-AACR-DT22-17 to C.S. Fuchs and M. Giannakis), administered by the American Association for Cancer Research, a scientific partner of SU2C; and by grants from the Project P Fund, The Friends of the Dana-Farber Cancer Institute, Bennett Family Fund, and the Entertainment Industry Foundation through National Colorectal Cancer Research Alliance. K. Haruki was supported by fellowship grants from the Uehara Memorial Foundation and the Mitsukoshi Health and Welfare Foundation. S.A. Väyrynen was supported by grants from the Finnish Cultural Foundation and Orion Research Foundation sr. J. Borowsky was supported by a grant from the Australia Awards-Endeavour Scholarships and Fellowships Program. K. Fujiyoshi was supported by a fellowship grant from the Uehara Memorial Foundation. K. Arima was supported by grants from Overseas Research Fellowship from Japan Society for the Promotion of Science (JP201860083). K. Wu was supported by an Investigator Initiated Grant from the American Institute for Cancer Research (AICR). M. Giannakis was supported by a Conquer Cancer Foundation of ASCO Career Development Award. A.T. Chan is a Stuart and Suzanne Steele MGH Research Scholar.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Bray
F
,
Ferlay
J
,
Soerjomataram
I
,
Siegel
RL
,
Torre
LA
,
Jemal
A
. 
Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J Clin
2018
;
68
:
394
424
.
2.
Benson
AB
,
Venook
AP
,
Al-Hawary
MM
,
Cederquist
L
,
Chen
Y-J
,
Ciombor
KK
, et al
NCCN guidelines insights: colon cancer, version 2.2018
.
J Natl Compr Canc Netw
2018
;
16
:
359
69
.
3.
Benson
AB
,
Venook
AP
,
Al-Hawary
MM
,
Cederquist
L
,
Chen
Y-J
,
Ciombor
KK
, et al
Rectal cancer, version 2.2018, NCCN Clinical practice guidelines in oncology
.
J Natl Compr Canc Netw
2018
;
16
:
874
901
.
4.
Ogino
S
,
Nowak
JA
,
Hamada
T
,
Milner
DA
,
Nishihara
R
. 
Insights into pathogenic interactions among environment, host, and tumor at the crossroads of molecular pathology and epidemiology
.
Annu Rev Pathol
2019
;
14
:
83
103
.
5.
Hanahan
D
,
Weinberg
RA.
Hallmarks of cancer: the next generation
.
Cell
2011
;
144
:
646
74
.
6.
O'Donnell
JS
,
Teng
MWL
,
Smyth
MJ
. 
Cancer immunoediting and resistance to T cell-based immunotherapy
.
Nat Rev Clin Oncol
2019
;
16
:
151
67
.
7.
Alexander
PG
,
McMillan
DC
,
Park
JH
. 
The local inflammatory response in colorectal cancer - type, location or density? A systematic review and meta-analysis
.
Cancer Treat Rev
2020
;
83
:
101949
.
8.
Pagès
F
,
Mlecnik
B
,
Marliot
F
,
Bindea
G
,
Ou
F
,
Bifulco
C
, et al
International validation of the consensus Immunoscore for the classification of colon cancer: a prognostic and accuracy study
.
Lancet
2018
;
391
:
2128
39
.
9.
Kather
JN
,
Halama
N.
Harnessing the innate immune system and local immunological microenvironment to treat colorectal cancer
.
Br J Cancer
2019
;
120
:
871
82
.
10.
Helmink
BA
,
Reddy
SM
,
Gao
J
,
Zhang
S
,
Basar
R
,
Thakur
R
, et al
B cells and tertiary lymphoid structures promote immunotherapy response
.
Nature
2020
;
577
:
549
55
.
11.
Yamauchi
M
,
Lochhead
P
,
Morikawa
T
,
Huttenhower
C
,
Chan
AT
,
Giovannucci
E
, et al
Colorectal cancer: a tale of two sides or a continuum?
Gut
2012
;
61
:
794
7
.
12.
Ogino
S
,
Kawasaki
T
,
Kirkner
GJ
,
Kraft
P
,
Loda
M
,
Fuchs
CS
. 
Evaluation of markers for CpG island methylator phenotype (CIMP) in colorectal cancer by a large population-based sample
.
J Mol Diagn
2007
;
9
:
305
14
.
13.
Ogino
S
,
Kawasaki
T
,
Brahmandam
M
,
Yan
L
,
Cantor
M
,
Namgyal
C
, et al
Sensitive sequencing method for KRAS mutation detection by Pyrosequencing
.
J Mol Diagn
2005
;
7
:
413
21
.
14.
Liao
X
,
Lochhead
P
,
Nishihara
R
,
Morikawa
T
,
Kuchiba
A
,
Yamauchi
M
, et al
Aspirin use, tumor PIK3CA mutation, and colorectal-cancer survival
.
N Engl J Med
2012
;
367
:
1596
606
.
15.
Irahara
N
,
Nosho
K
,
Baba
Y
,
Shima
K
,
Lindeman
NI
,
Hazra
A
, et al
Precision of pyrosequencing assay to measure LINE-1 methylation in colon cancer, normal colonic mucosa, and peripheral blood cells
.
J Mol Diagn
2010
;
12
:
177
83
.
16.
Giannakis
M
,
Mu
XJ
,
Shukla
SA
,
Qian
ZR
,
Cohen
O
,
Nishihara
R
, et al
Genomic correlates of immune-cell infiltrates in colorectal carcinoma
.
Cell Rep
2016
;
15
:
857
65
.
17.
Chan
AT
,
Ogino
S
,
Fuchs
CS
. 
Aspirin and the risk of colorectal cancer in relation to the expression of COX-2
.
N Engl J Med
2007
;
356
:
2131
42
.
18.
The Cancer Genome Atlas Network
. 
Comprehensive molecular characterization of human colon and rectal cancer
.
Nature
2012
;
487
:
330
7
.
19.
Liu
J
,
Lichtenberg
T
,
Hoadley
KA
,
Poisson
LM
,
Lazar
AJ
,
Cherniack
AD
, et al
An integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics
.
Cell
2018
;
173
:
400
6
.
20.
Bankhead
P
,
Loughrey
MB
,
Fernández
JA
,
Dombrowski
Y
,
McArt
DG
,
Dunne
PD
, et al
QuPath: open source software for digital pathology image analysis
.
Sci Rep
2017
;
7
:
16878
.
21.
McInnes
L
,
Healy
J
,
Melville
J
. 
UMAP: uniform manifold approximation and projection for dimension reduction
.
bioRxiv
2018
;
arXiv
:
1802
.
22.
Barua
S
,
Fang
P
,
Sharma
A
,
Fujimoto
J
,
Wistuba
I
,
Rao
AUK
, et al
Spatial interaction of tumor cells and regulatory T cells correlates with survival in non-small cell lung cancer
.
Lung Cancer
2018
;
117
:
73
9
.
23.
Baddeley
A
,
Turner
R
. 
spatstat: An R package for analyzing spatial point patterns
.
J Stat Softw
2005
;
12
:
282
90
.
24.
Benjamin
DJ
,
Berger
JO
,
Johannesson
M
,
Nosek
BA
,
Wagenmakers
E-J
,
Berk
R
, et al
Redefine statistical significance
.
Nat Hum Behav
2018
;
2
:
6
10
.
25.
Liu
L
,
Nevo
D
,
Nishihara
R
,
Cao
Y
,
Song
M
,
Twombly
TS
, et al
Utility of inverse probability weighting in molecular pathological epidemiology
.
Eur J Epidemiol
2018
;
33
:
381
92
.
26.
Seaman
SR
,
White
IR.
Review of inverse probability weighting for dealing with missing data
.
Stat Methods Med Res
2013
;
22
:
278
95
.
27.
Hamada
T
,
Cao
Y
,
Qian
ZR
,
Masugi
Y
,
Nowak
JA
,
Yang
J
, et al
Aspirin use and colorectal cancer survival according to tumor CD274 (programmed cell death 1 ligand 1) expression status
.
J Clin Oncol
2017
;
35
:
1836
44
.
28.
Ogino
S
,
Nosho
K
,
Kirkner
GJ
,
Kawasaki
T
,
Meyerhardt
JA
,
Loda
M
, et al
CpG island methylator phenotype, microsatellite instability, BRAF mutation and clinical outcome in colon cancer
.
Gut
2009
;
58
:
90
6
.
29.
Grasso
CS
,
Giannakis
M
,
Wells
DK
,
Hamada
T
,
Mu
XJ
,
Quist
M
, et al
Genetic mechanisms of immune evasion in colorectal cancer
.
Cancer Discov
2018
;
8
:
730
49
.
30.
Berntsson
J
,
Nodin
B
,
Eberhard
J
,
Micke
P
,
Jirström
K
. 
Prognostic impact of tumour-infiltrating B cells and plasma cells in colorectal cancer
.
Int J Cancer
2016
;
139
:
1129
39
.
31.
Nutt
SL
,
Hodgkin
PD
,
Tarlinton
DM
,
Corcoran
LM
. 
The generation of antibody-secreting plasma cells
.
Nat Rev Immunol
2015
;
15
:
160
71
.
32.
Wouters
MCA
,
Nelson
BH.
Prognostic significance of tumor-infiltrating B cells and plasma cells in human cancer
.
Clin Cancer Res
2018
;
24
:
6125
35
.
33.
Coffelt
SB
,
Wellenstein
MD
,
de Visser
KE
. 
Neutrophils in cancer: neutral no more
.
Nat Rev Cancer
2016
;
16
:
431
46
.
34.
Governa
V
,
Trella
E
,
Mele
V
,
Tornillo
L
,
Amicarella
F
,
Cremonesi
E
, et al
The interplay between neutrophils and CD8+T cells improves survival in human colorectal cancer
.
Clin Cancer Res
2017
;
23
:
3847
58
.
35.
Wikberg
ML
,
Ling
A
,
Li
X
,
Öberg
Å
,
Edin
S
,
Palmqvist
R
. 
Neutrophil infiltration is a favorable prognostic factor in early stages of colon cancer
.
Hum Pathol
2017
;
68
:
193
202
.
36.
Galdiero
MR
,
Bianchi
P
,
Grizzi
F
,
Di Caro
G
,
Basso
G
,
Ponzetta
A
, et al
Occurrence and significance of tumor-associated neutrophils in patients with colorectal cancer
.
Int J cancer
2016
;
139
:
446
56
.
37.
Droeser
RA
,
Hirt
C
,
Eppenberger-Castori
S
,
Zlobec
I
,
Viehl
CT
,
Frey
DM
, et al
High myeloperoxidase positive cell infiltration in colorectal cancer is an independent favorable prognostic factor
.
PLoS One
2013
;
8
:
e64814
.
38.
Yoon
J
,
Terada
A
,
Kita
H
. 
CD66b regulates adhesion and activation of human eosinophils
.
J Immunol
2007
;
179
:
8454
62
.
39.
Berry
RS
,
Xiong
M-J
,
Greenbaum
A
,
Mortaji
P
,
Nofchissey
RA
,
Schultz
F
, et al
High levels of tumor-associated neutrophils are associated with improved overall survival in patients with stage II colorectal cancer
.
PLoS One
2017
;
12
:
e0188799
.
40.
Ramirez
GA
,
Yacoub
M-R
,
Ripa
M
,
Mannina
D
,
Cariddi
A
,
Saporiti
N
, et al
Eosinophils from physiology to disease: a comprehensive review
.
Biomed Res Int
2018
;
2018
:
9095275
.
41.
Prizment
AE
,
Vierkant
RA
,
Smyrk
TC
,
Tillmans
LS
,
Lee
JJ
,
Sriramarao
P
, et al
Tumor eosinophil infiltration and improved survival of colorectal cancer patients: Iowa Women's Health Study
.
Mod Pathol
2016
;
29
:
516
27
.
42.
Nielsen
HJ
,
Hansen
U
,
Christensen
IJ
,
Reimert
CM
,
Brünner
N
,
Moesgaard
F
. 
Independent prognostic value of eosinophil and mast cell infiltration in colorectal cancer tissue
.
J Pathol
1999
;
189
:
487
95
.
43.
Reichman
H
,
Itan
M
,
Rozenberg
P
,
Yarmolovski
T
,
Brazowski
E
,
Varol
C
, et al
Activated eosinophils exert antitumorigenic activities in colorectal cancer
.
Cancer Immunol Res
2019
;
7
:
388
400
.
44.
Komura
D
,
Ishikawa
S.
Machine learning methods for histopathological image analysis
.
Comput Struct Biotechnol J
2018
;
16
:
34
42
.
45.
Saltz
J
,
Gupta
R
,
Hou
L
,
Kurc
T
,
Singh
P
,
Nguyen
V
, et al
Spatial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images
.
Cell Rep
2018
;
23
:
181
93
.
46.
Loughrey
MB
,
Bankhead
P
,
Coleman
HG
,
Hagan
RS
,
Craig
S
,
McCorry
AMB
, et al
Validation of the systematic scoring of immunohistochemically stained tumour tissue microarrays using QuPath digital image analysis
.
Histopathology
2018
;
73
:
327
38
.
47.
Bankhead
P
,
Fernández
JA
,
McArt
DG
,
Boyle
DP
,
Li
G
,
Loughrey
MB
, et al
Integrated tumor identification and automated scoring minimizes pathologist involvement and provides new insights to key biomarkers in breast cancer
.
Lab Invest
2018
;
98
:
15
26
.
48.
Maley
CC
,
Koelble
K
,
Natrajan
R
,
Aktipis
A
,
Yuan
Y
. 
An ecological measure of immune-cancer colocalization as a prognostic factor for breast cancer
.
Breast Cancer Res
2015
;
17
:
131
.
49.
Setiadi
AF
,
Ray
NC
,
Kohrt
HE
,
Kapelner
A
,
Carcamo-Cavazos
V
,
Levic
EB
, et al
Quantitative, architectural analysis of immune cell subsets in tumor-draining lymph nodes from breast cancer patients and healthy lymph nodes
.
PLoS One
2010
;
5
:
e12420
.
50.
Camp
RL
,
Neumeister
V
,
Rimm
DL
. 
A decade of tissue microarrays: progress in the discovery and validation of cancer biomarkers
.
J Clin Oncol
2008
;
26
:
5630
7
.

Supplementary data