Computer-assisted stereology and automated image analysis for quantification of tumor infiltrating lymphocytes in colon cancer

Reduced Tumor Volume and Increased Necrosis of Human Breast Tumor Xenograft in Mice Pretreated by a Cocktail of Three Specific Anti-HER2 scFvs

PURPOSE

We aimed to assess the effects of a cocktail comprising three specific anti- HER2 scFvs on breast tumor formation in a xenograft mouse model and to evaluate quantitative changes in the tumor using stereological analysis.

METHODS

Three specific anti-HER2 phage antibodies were produced from a scFv-library using phage display technology. The cell binding capacities of the antibodies were assessed via FACS analysis. Soluble forms of the antibodies were prepared by infecting HB2151-E. coli cells and purified using a centrifugal ultrafiltration method. The purification process was evaluated by SDSPAGE analysis. Two forms of scFv cocktails were prepared, soluble scFv and phage-scFv cocktail, which contained an equal amount/phage of each of the three antibodies. Inbred female BALB/c mice were pretreated with 5 and 20 mg/kg of the soluble scFv cocktail and 1011 phage-scFv cocktail/ kg. The mice were then injected with 2×106 SKBR-3 human breast cancer cells. Total tumor, inflammatory and non-inflammatory volumes were estimated using the Cavalieri principle after preparing photomicrograph slides.

RESULTS

The anti-HER2 scFvs showed significantly higher binding to SKBR-3 cells compared to the isotype control. SDS-PAGE analysis confirmed the high purification of the scFvs. Stereological analysis revealed that the group pretreated with 20 mg/kg of the soluble scFv cocktail exhibited the highest reductions in total tumor volume, non-inflammatory volume, and inflammatory volume, with reductions of 73%, 78%, and 72%, respectively, compared to PBS-pretreated mice (P-value < 0.0001). The volumetric ratio of necrotic tissue to total tumor volume increased by 2.2-fold and 2- fold in the 20 mg/kg of soluble scFv cocktail and phage-scFv cocktail groups, respectively, compared to the PBS-treated mice (P-value < 0.05).

CONCLUSION

Pre-treatment with a 20 mg/kg anti-HER2 scFv cocktail resulted in a significant reduction in tumor volume and increased necrotic area in a human breast cancer xenograft model, indicating the remarkable anti-tumor effect of the cocktail in vivo.

Prognostic and predictive value of tumor infiltrating lymphocytes in combination with systemic inflammatory markers in colon cancer

OBJECTIVE

Systemic inflammatory indices and CD8(+) tumor infiltrating lymphocytes (TILs) in the tumor microenvironment are highly prognostic in colon cancer (CC) but combined assessment is less well studied. The purpose of this study was to investigate the prognostic and predictive value of CD8(+) TILs in combination with systemic inflammatory indices in patients with resected stage II-III colon cancer.

PATIENTS AND METHODS

Patients with stage II-III CC (n = 304) diagnosed between 2008 and 2016 were included. Pan-immune inflammation value (PIV) was used as a comprehensive inflammatory index and was calculated as: [neutrophil count × platelet count × monocyte count]/lymphocyte count. The mean density of CD8+ TILs in the periphery and center of the tumor was assessed and dichotomized at the 75th percentile. Combined inflammation score (CIS) was classified as "high" in patients with high PIV (>median) plus low mean CD8(+) TILs density, and CIS "low" in the remaining patients.

RESULTS

5-year DFS was 71% (78% in stage II, 63.4% in stage III). PIV was higher in right colon tumors, T4 tumors and in patients with obstruction / perforation. CD8(+) TIL density was lower in node positive tumors. High PIV and low CD8(+) TILs were associated with shorter disease-free survival (DFS). In multivariate analysis; age > 65 years, stage III disease and high CIS (PIVhigh / CD8low) were associated with shorter DFS. Among patients with stage II disease, patients with high CIS (PIVhigh / CD8low) derived significant benefit from adjuvant chemotherapy while those with low CIS derived no benefit.

CONCLUSION

Combined inflammation score may represent a new prognostic factor for localized colon cancer and predictor of chemotherapy response in patients with stage II disease.

Development and Validation of a Digital Analysis Method to Quantify CD3-immunostained T Lymphocytes in Whole Slide Images of Crohn's Disease Biopsies

The T-lymphocyte-mediated inflammation in Crohn's disease can be assessed by quantifying CD3-positive T-lymphocyte counts in colonic sections. We developed and validated a process to reliably quantify immunohistochemical marker-positive cells in a high-throughput setting using whole slide images (WSIs) of CD3-immunostained colonic and ileal tissue sections. In regions of interest (ROIs) and/or whole tissue sections of 40 WSIs from 36 patients with Crohn's disease, CD3-positive cells were quantified by an expert gastrointestinal pathologist (gold standard) and by image analysis algorithms developed with software from 3 independent vendors. Semiautomated quantification of CD3-positive cell counts estimated in 1 ROI per section were accurate when compared with manual analysis (Pearson correlation coefficient, 0.877 to 0.925). Biological variability was acceptable in digitally determined CD3-positive cell measures between 2 to 5 ROIs annotated on the same tissue section (coefficient of variation <25%). Results from computer-aided analysis of CD3-positive T lymphocytes in a whole tissue section and the average of results from 2 to 5 ROIs per tissue section lacked reliability (overestimation or underestimation and systematic bias), suggesting that absolute quantification of CD3-positive T lymphocytes in a whole tissue section may be more accurate. Semiautomated image analysis in WSIs demonstrated reproducible CD3-positive cell measures across 3 independent algorithms. A computer-aided digital image analysis method was developed and validated to quantify CD3-positive T lymphocytes in colonic and ileal biopsy sections from patients with Crohn's disease. Results support consideration of this digital analysis method for use in future Crohn's disease clinical studies.

Development of a Fully Automated Method to Obtain Reproducible Lymphocyte Counts in Patients With Colorectal Cancer

Colorectal cancer (CRC) is the third most common cancer worldwide. Although clinical outcome varies among patients diagnosed within the same TNM stage it is the cornerstone in treatment decisions as well as follow-up programmes. Tumor-infiltrating lymphocytes have added value when evaluating survival outcomes. The aim of this study was to develop a fully automated method for quantification of subsets of T lymphocytes in the invasive margin and central tumor in patients with CRC based on Deep Learning powered artificial intelligence. The study cohort consisted of 163 consecutive patients with a primary diagnosis of CRC followed by a surgical resection. Double-labeling immunohistochemical staining with cytokeratin in combination with CD3 or CD8, respectively, was performed on 1 representative slide from each patient. Visiopharm Quantitative Digital Pathology software was used to develop Application Protocol Packages for visualization of architectural details (background, normal epithelium, cancer epithelium, surrounding tissue), identification of central tumor and invasive margin as well as subsequent quantitative analysis of immune cells. Fully automated counts for CD3 and CD8 positive T cells were obtained in 93% and 92% of the cases, respectively. In the remaining cases, manual editing was required. In conclusion, the development of a fully automated method for counting CD3 + and CD8 + lymphocytes in a cohort of patients with CRC provided excellent results eliminating not only observer variability in lymphocyte counts but also in identifying the regions of interest for the quantitative analysis. Validation of the performance of the Application Protocol Packages including clinical correlation is needed.

Deep Learning-Based Mapping of Tumor Infiltrating Lymphocytes in Whole Slide Images of 23 Types of Cancer

The role of tumor infiltrating lymphocytes (TILs) as a biomarker to predict disease progression and clinical outcomes has generated tremendous interest in translational cancer research. We present an updated and enhanced deep learning workflow to classify 50x50 um tiled image patches (100x100 pixels at 20x magnification) as TIL positive or negative based on the presence of 2 or more TILs in gigapixel whole slide images (WSIs) from the Cancer Genome Atlas (TCGA). This workflow generates TIL maps to study the abundance and spatial distribution of TILs in 23 different types of cancer. We trained three state-of-the-art, popular convolutional neural network (CNN) architectures (namely VGG16, Inception-V4, and ResNet-34) with a large volume of training data, which combined manual annotations from pathologists (strong annotations) and computer-generated labels from our previously reported first-generation TIL model for 13 cancer types (model-generated annotations). Specifically, this training dataset contains TIL positive and negative patches from cancers in additional organ sites and curated data to help improve algorithmic performance by decreasing known false positives and false negatives. Our new TIL workflow also incorporates automated thresholding to convert model predictions into binary classifications to generate TIL maps. The new TIL models all achieve better performance with improvements of up to 13% in accuracy and 15% in F-score. We report these new TIL models and a curated dataset of TIL maps, referred to as TIL-Maps-23, for 7983 WSIs spanning 23 types of cancer with complex and diverse visual appearances, which will be publicly available along with the code to evaluate performance.

Code Available at:https://github.com/ShahiraAbousamra/til_classification.

A New Automated Histomorphometric MATLAB Algorithm for Immunohistochemistry Analysis Using Whole Slide Imaging

The use of animal models along with the employment of advanced and sophisticated stereological methods for assessing bone quality combined with the use of statistical methods to evaluate the effectiveness of bone therapies has made it possible to investigate the pathways that regulate bone responses to medical devices. Image analysis of histomorphometric measurements remains a time-consuming task, as the image analysis software currently available does not allow for automated image segmentation. Such a feature is usually obtained by machine learning and with software platforms that provide image-processing tools such as MATLAB. In this study, we introduce a new MATLAB algorithm to quantify immunohistochemically stained critical-sized bone defect samples and compare the results with the commonly available Aperio Image Scope Positive Pixel Count (PPC) algorithm. Bland and Altman analysis and Pearson correlation showed that the measurements acquired with the new MATLAB algorithm were in excellent agreement with the measurements obtained with the Aperio PPC algorithm, and no significant differences were found within the histomorphometric measurements. The ability to segment whole slide images, as well as defining the size and the number of regions of interest to be quantified, makes this MATLAB algorithm a potential histomorphometric tool for obtaining more objective, precise, and reproducible quantitative assessments of entire critical-sized bone defect image data sets in an efficient and manageable workflow.

Favorable prognostic impact of Natural Killer cells and T cells in high-grade serous ovarian carcinoma

Introduction: The aim of the present study was to investigate the prognostic impact of intratumoral cytotoxic T cells, Natural Killer (NK) cells, neutrophils and PD-L1 expression in patients with epithelial ovarian cancer.Methods: All patients diagnosed with high-grade serous carcinoma (HGSC) in Denmark in 2005 were included in the study. Immunohistochemical staining for PD-L1, CD8, CD66b and CD57 was performed on tumor tissue from 283 patients. Cell densities were analyzed using a digital image analysis method. The primary endpoint was overall survival (OS).Results: The median OS for HGSC patients was 30 months. It was 45 months in patients with high level of CD57+ NK cells (≥10 cells/mm²) compared with 29 month in patients with low level (<10 cells/mm²) (p = .0310). The median OS was 37 and 25 months in patients with high vs. low level of CD8+ T cells (cutoff 80 cells/mm²) (p = .0008). In multivariate analysis, high numbers of CD57+ NK cells and CD8+ T cells remained independent markers of favorable OS, adjusted hazard ratio (HR) 0.67; p = .041, and HR 0.72; p = .020, respectively. PD-L1 expression was associated with improved OS (37 months vs. 22 months, p = .0006), but was only borderline significant in the multivariate analysis (HR 0.77, p = .061). CD66b + neutrophils had no association with OS.Conclusions: In patients with HGSC tumor-infiltrating CD57+ NK cells and CD8+ T cells had favorable prognostic impact, while PD-L1 expression had borderline favorable prognostic significance. CD66b + neutrophils had no prognostic association. These findings may influence future immunotherapy development.

Immunogradient Indicators for Antitumor Response Assessment by Automated Tumor-Stroma Interface Zone Detection

The distribution of tumor-infiltrating lymphocytes (TILs) within the tumor microenvironment provides strong prognostic value, which is increasingly important with the arrival of new immunotherapy modalities. Both visual and image analysis-based assays are developed to assess the immune contexture of the tumors. We propose an automated method based on grid subsampling of microscopy image analysis data to extract the tumor-stroma interface zone (IZ) of controlled width. The IZ is a ranking of tissue areas by their distance to the tumor edge, which is determined by a set of explicit rules. TIL density profiles across the IZ are used to compute a set of novel immunogradient indicators that reflect TIL gradient towards the tumor. We applied this method on CD8 immunohistochemistry images of surgically excised hormone receptor-positive breast and colorectal cancers to predict overall patient survival. In both cohorts, the immunogradient indicators enabled strong and independent prognostic stratification, outperforming clinical and pathologic variables. Patients with breast cancer with low immunogradient levels had a prominent decrease in survival probability 5 years after surgery. Our study provides proof of concept that data-driven, automated, operator-independent IZ sampling enables spatial immune response measurement in the tumor-host interaction frontline for prediction of disease outcomes.

Quantitative digital image analysis of tumor-infiltrating lymphocytes in HER2-positive breast cancer

As visual quantification of the density of tumor-infiltrating lymphocytes (TILs) lacks in precision, digital image analysis (DIA) approach has been applied in order to improve. In several studies, TIL density has been examined on hematoxylin and eosin (HE)-stained sections using DIA. The aim of the present study was to quantify TIL density on HE sections of core needle biopsies using DIA and investigate its association with clinicopathological parameters and pathological response to neoadjuvant chemotherapy in human epidermal growth factor receptor 2 (HER2)-positive breast cancer. The study cohort comprised of patients with HER2-positive breast cancer, all treated with neoadjuvant anti-HER2 therapy. DIA software applying machine learning-based classification of epithelial and stromal elements was used to count TILs. TIL density was determined as the number of TILs per square millimeter of stromal tissue. Median TIL density was 1287/mm² (range, 123-8101/mm²). A high TIL density was associated with higher histological grade (P = 0.02), estrogen receptor negativity (P = 0.036), and pathological complete response (pCR) (P < 0.0001). In analyses using receiver operating characteristic curves, a threshold TIL density of 2420/mm² best discriminated pCR from non-pCR. In multivariate analysis, high TIL density (> 2420/mm²) was significantly associated with pCR (P < 0.0001). Our results indicate that DIA can assess TIL density quantitatively, machine learning-based classification algorithm allowing determination of TIL density as the number of TILs per unit area, and TIL density established by this method appears to be an independent predictor of pCR in HER2-positive breast cancer.

Budding, tumor-infiltrating lymphocytes, gland formation: scoring leads to new prognostic groups in World Health Organization low-grade colorectal cancer with impact on survival

Grading for colorectal carcinoma (CRC) is traditionally based on the percentage of gland formation. In recent years, high-grade CRC has become subject to more precise molecular grading strategies. Most, however, are low-grade cases according to the World Health Organization (WHO) with inhomogenous outcomes due to still insufficient characterization. On the other hand, budding and tumor-infiltrating lymphocytes have developed as interesting additive prognostic factors in CRC. Especially budding has been very well defined by the International Tumor Budding Consensus Conference recently. We analyzed a large collective of 576 WHO low-grade CRC cases, stages I to IV, diagnosed between 2005 and 2016 in terms of gland formation, budding, and tumor-infiltrating lymphocytes and developed a new, morphology-based risk score, taking into account each of the 3 parameters. For each parameter, 1 to 2 points were given, resulting in a sum score, dividing the CRC cases into a low-, an intermediate-, and a high-risk group. By our score, 179 (34.9%) of the cases were grouped as low risk, 241 (53.5) as intermediate risk, and 92 (35.5%) as high risk. The 3 groups differed significantly in pT, pN, and M as well as tumor stages, lymphatic vessel invasion, venous invasion, and overall survival (0.;P < .001 for low risk versus high risk, P = .038 for low versus intermediate risk, and P = .036 for intermediate versus high risk; log rank: median, 94.0 months [95% confidence interval {CI}, 74.9-113.1] for low risk; median, 63.0 months [95% CI, 44.0-82.0] for intermediate risk; and median, 40.0 months [95% CI, 23.4-56.7] for high risk) in Kaplan-Meier-analysis. Our proposed Bayreuth score enables separating the large group of WHO low-grade CRC cases into subgroups, which differ significantly in outcome.

Spatial Architecture and Arrangement of Tumor-Infiltrating Lymphocytes for Predicting Likelihood of Recurrence in Early-Stage Non-Small Cell Lung Cancer

PURPOSE

The presence of a high degree of tumor-infiltrating lymphocytes (TIL) has been proven to be associated with outcome in patients with non-small cell lung cancer (NSCLC). However, recent evidence indicates that tissue architecture is also prognostic of disease-specific survival and recurrence. We show a set of descriptors (spatial TIL, SpaTIL) that capture density, and spatial colocalization of TILs and tumor cells across digital images that can predict likelihood of recurrence in early-stage NSCLC.

EXPERIMENTAL DESIGN

The association between recurrence in early-stage NSCLC and SpaTIL features was explored on 301 patients across four different cohorts. Cohort D1 (n = 70) was used to identify the most prognostic SpaTIL features and to train a classifier to predict the likelihood of recurrence. The classifier performance was evaluated in cohorts D2 (n = 119), D3 (n = 112), and D4 (n = 112). Two pathologists graded each sample of D1 and D2; intraobserver agreement and association between manual grading and likelihood of recurrence were analyzed.

RESULTS

SpaTIL was associated with likelihood of recurrence in all test sets (log-rank P < 0.02). A multivariate Cox proportional hazards analysis revealed an HR of 3.08 (95% confidence interval, 2.1-4.5, P = 7.3 × 10-5). In contrast, agreement among expert pathologists using tumor grade was moderate (Kappa = 0.5), and the manual TIL grading was only prognostic for one reader in D2 (P = 8.0 × 10-3).

CONCLUSIONS

A set of features related to density and spatial architecture of TILs was found to be associated with a likelihood of recurrence of early-stage NSCLC. This information could potentially be used for helping in treatment planning and management of early-stage NSCLC.See related commentary by Peled et al., p. 1449.

The Prognostic Value of Tumor-Infiltrating lymphocytes in Stage II Colon Cancer. A Nationwide Population-Based Study

BACKGROUND: Additional prognostic markers are needed for better treatment stratification of stage II colon cancer (CC). We investigated the prognostic value of tumor-infiltrating lymphocytes (TILs) in a true population-based cohort of patients with stage II CC. MATERIAL AND METHODS: A total of 573 patients were included. Tumor blocks representing the deepest invasive part of the primary tumor were used for analysis. CD3+ and CD8+ TILs at the invasive front were evaluated by immunohistochemistry on whole tumor sections. The invasive area was manually outlined, and Visiopharm Integrator System software was used for quantification. Data were dichotomized for comparison with clinical data. The prognostic value was investigated in Cox proportional-hazard models for recurrence-free survival (RFS) and overall survival (OS). RESULTS: Low CD3+ or CD8+ TILs were significantly associated with poor RFS and OS (P = .0021 and P ≤ .0009, respectively, log-rank test). In multiple Cox regression analysis, low CD3+ and CD8+ TILs were associated with reduced RFS with hazard ratio (HR) = 1.386 (95% CI 1.039-1.850), P = .026, and HR = 1.394 (95% CI 1.029-1.890), P = .032, respectively, independent of age, T-stage, localization, perforation, and microsatellite instability (MSI). In the subgroups of patients with low CD3+ or CD8+ TILs, there was no difference in survival between patients with MSI and microsatellite-stable tumors, (P = .821 and P = .907, respectively). CONCLUSION: Low CD3+ and CD8+ TILs in the invasive area are both related to inferior prognosis of stage II CC, and we recommend either of these parameters to be considered as additional high-risk factor.

Budding and tumor-infiltrating lymphocytes - combination of both parameters predicts survival in colorectal cancer and leads to new prognostic subgroups

Tumor budding is an independent prognostic factor in colorectal cancer (CRC) and has recently been well defined by the International Concensus Conference on Tumor Budding (ITBCC). Tumor-infiltrating lymphocytes (TILs) are also an issue in different human cancers and correlate with prognosis in CRC. Here we evaluate the combination of budding and TILs in CRC with regard to prognosis. Hematoxylin and eosin (H&E)-stained slides of 501 CRC patients, diagnosed between 2005 and 2010, were reevaluated for tumor budding according to the ITBCC criteria. Low (n = 331) was compared to intermediate/high budding (n = 170). The percentage of TILs was also assessed, and the following four groups were established: low budding + TILs >5% (n = 162), low budding + TILS ≤5% (n = 169), high budding + TILS >5% (n = 68), high budding + TILs ≤5% (n = 93). The combination of both markers revealed highly significant differences in overall survival (OS) between the four groups (P = .001). The low budding/>5% TILs group showed longest OS, followed by high budding/>5% TILs cases, followed by tumors with low budding/≤5% TILs. OS was worst for the high budding/≤ 5% TILs group. The combined score also correlated with T, N, M, L, Vstaging, development of disease relapse and distant metastasis. Our study shows that - even in the age of molecular pathology - it is still important to pay special attention to tumor morphology for additional information on tumor behavior and prognosis. Combining different morphological parameters of tumor and tumor environment can help to further subdivide CRC into new prognostic groups.