Tumor-associated macrophages (TAM) play a controversial role in epithelial-mesenchymal transition (EMT) and prognosis of colorectal cancer (CRC). In particular, the microlocalization, polarization and prognostic impact of TAM in the immediate environment of invading CRC cells has not yet been established. To address this clinically relevant question, intraepithelial (iCD68) and stromal macrophages (sCD68), M1-macrophages (iNOS), M2-macrophages (CD163), cytokeratin-positive cancer cells (tumor buds) and expression of the anti-phagocytic marker CD47 were investigated in primary tumors of 205 well-characterized CRC patients. Cell-to-cell contacts between tumor buds and TAM were detected using high-resolution digital scans. The composition of the tumor microenvironment was analyzed with clinicopathological and molecular features. High CD68 counts predicted long term overall survival independent of microlocalization (iCD68 p=0.0016; sCD68 p=0.03), pT, pN, pM and post-operative therapy. CD68 infiltration correlated with significantly less tumor budding (iCD68 p=0.0066; sCD68 p=0.0091) and absence of lymph node metastasis (sCD68 p=0.0286). Cell-to-cell contact of sCD68 and invading cancer cells was frequent and ameliorated the detrimental prognostic effect of the tumor budding phenotype. Subgroup analysis identified long-term survival with CD47 loss and predominance of CD163 M2 macrophages ( = 0.0366). CD163 macrophages represented 40% of the total population, and positively correlated with total CD68 macrophage numbers (r[CD68/CD163] = 0.32; = 0.0001). Strong CD163 infiltration predicted lower tumor grade ( = 0.0026) and less lymph node metastasis ( = 0.0056). This study provides direct morphological evidence of an interaction between TAM and infiltrating cancer cells. The prognostic impact of TAM is modulated by phenotype, microlocalization and the expression of anti-phagocytic markers in CRC.

Background:In colorectal cancer (CRC), tumour budding at the invasion front is associated with lymph node (LN) and distant metastasis. Interestingly, tumour budding can also be detected in biopsies (intratumoural budding; ITB) and may have similar clinical importance. Here we investigate whether ITB in preoperative CRC biopsies can be translated into daily diagnostic practice.Methods:Preoperative biopsies from 133 CRC patients (no neoadjuvant therapy) underwent immunohistochemistry for pan-cytokeratin marker AE1/AE3. Across all biopsies for each patient, the densest region of buds at × 40 (high-power field; HPF) was identified and buds were counted.Results:A greater number of tumour buds in the biopsy was associated with pT stage (P=0.0143), LN metastasis (P=0.0007), lymphatic (P=0.0065) and venous vessel invasion (P=0.0318) and distant metastasis (cM1) (P=0.0013). Using logistic regression, a ‘scale' was developed to estimate the probability of LN and distant metastasis using the number of tumour buds (e.g. 10 buds per HPF: 64% chance of LN metastasis; 30 buds per HPF: 86% chance). Inter-observer agreement for ITB was excellent (intraclass correlation coefficient: 0.813).Conclusion:Tumour budding can be assessed in the preoperative biopsy of CRC patients. It is practical, reproducible and predictive of LN and distant metastasis. Intratumoural budding qualifies for further investigation in the prospective setting.

BackgroundTissue microarray (TMA) technology revolutionized the investigation of potential biomarkers from paraffin-embedded tissues. However, conventional TMA construction is laborious, time-consuming and imprecise. Next-generation tissue microarrays (ngTMA) combine histological expertise with digital pathology and automated tissue microarraying. The aim of this study was to test the feasibility of ngTMA for the investigation of biomarkers within the tumor microenvironment (tumor center and invasion front) of six tumor types, using CD3, CD8 and CD45RO as an example.MethodsTen cases each of malignant melanoma, lung, breast, gastric, prostate and colorectal cancers were reviewed. The most representative H&E slide was scanned and uploaded onto a digital slide management platform. Slides were viewed and seven TMA annotations of 1 mm in diameter were placed directly onto the digital slide. Different colors were used to identify the exact regions in normal tissue (n = 1), tumor center (n = 2), tumor front (n = 2), and tumor microenvironment at invasion front (n = 2) for subsequent punching. Donor blocks were loaded into an automated tissue microarrayer. Images of the donor block were superimposed with annotated digital slides. Exact annotated regions were punched out of each donor block and transferred into a TMA block. 420 tissue cores created two ngTMA blocks. H&E staining and immunohistochemistry for CD3, CD8 and CD45RO were performed.ResultsAll 60 slides were scanned automatically (total time < 10 hours), uploaded and viewed. Annotation time was 1 hour. The 60 donor blocks were loaded into the tissue microarrayer, simultaneously. Alignment of donor block images and digital slides was possible in less than 2 minutes/case. Automated punching of tissue cores and transfer took 12 seconds/core. Total ngTMA construction time was 1.4 hours. Stains for H&E and CD3, CD8 and CD45RO highlighted the precision with which ngTMA could capture regions of tumor-stroma interaction of each cancer and the T-lymphocytic immune reaction within the tumor microenvironment.ConclusionBased on a manual selection criteria, ngTMA is able to precisely capture histological zones or cell types of interest in a precise and accurate way, aiding the pathological study of the tumor microenvironment. This approach would be advantageous for visualizing proteins, DNA, mRNA and microRNAs in specific cell types using in situ hybridization techniques.

Tumor budding is a promising and cost-effective biomarker with strong prognostic value in colorectal cancer. However, challenges related to interobserver variability persist. Such variability may be reduced by immunohistochemistry and computer-aided tumor bud selection. Development of computer algorithms for this purpose requires unequivocal examples of individual tumor buds. As such, we undertook a large-scale, international, and digital observer study on individual tumor bud assessment. From a pool of 46 colorectal cancer cases with tumor budding, 3000 tumor bud candidates were selected, largely based on digital image analysis algorithms. For each candidate bud, an image patch (size 256 × 256 µm) was extracted from a pan cytokeratin-stained whole-slide image. Members of an International Tumor Budding Consortium (n = 7) were asked to categorize each candidate as either (1) tumor bud, (2) poorly differentiated cluster, or (3) neither, based on current definitions. Agreement was assessed with Cohen's and Fleiss Kappa statistics. Fleiss Kappa showed moderate overall agreement between observers (0.42 and 0.51), while Cohen's Kappas ranged from 0.25 to 0.63. Complete agreement by all seven observers was present for only 34% of the 3000 tumor bud candidates, while 59% of the candidates were agreed on by at least five of the seven observers. Despite reports of moderate-to-substantial agreement with respect to tumor budding grade, agreement with respect to individual pan cytokeratin-stained tumor buds is moderate at most. A machine learning approach may prove especially useful for a more robust assessment of individual tumor buds.

Ki67, caspase-3 and M30 staining is absent in most tumour buds, suggesting decreased proliferation and apoptosis. However, the fact that Ki67 and caspase-3 immunoreactivity was associated with unfavourable prognosis points to a heterogeneous population of tumour buds.