Measuring multiple parameters of CD8+ tumor-infiltrating lymphocytes in human cancers by image analysis

Response to ipilimumab therapy in metastatic melanoma patients: potential relevance of CTLA-4+ tumor infiltrating lymphocytes and their in situ localization

Immune checkpoint inhibitors, including ipilimumab (IPI), achieve a clinical benefit in a small proportion of melanoma patients highlighting the need to investigate predictive biomarkers. In this study, we characterized tumor infiltrating lymphocytes (TILs), focusing on the CTLA-4⁺ subset, and evaluated their possible predictive significance. We characterized TIL density, cell type, and localization in 40 melanoma lesions from 17 patients treated with IPI. Associations of TILs with IPI timing, tissue localization, and response to IPI were estimated using a linear mixed-effects modelling approach. We found that most of TIL subsets increased in situ upon IPI therapy, with particular reference to FoxP3⁺ cells. TILs and TIL subsets, such as CD3⁺, CD45RO⁺, CTLA-4⁺, CD4⁺, CD8⁺ T cells, CD20⁺ B cells, and NKp46⁺ NK cells, showed significantly different spatial distributions in the tumor microenvironment being higher at the invasive margin (IM) as compared to the tumor center (TC) (P value < 0.001 for TIL score and P value < 0.05 for all subsets). Remarkably, high TIL score and density of CD3⁺, CD8⁺ T cells, and CTLA-4⁺ immune cells were significantly associated with a better response to IPI (P values = 0.002, 0.023, 0.007, and 0.001, respectively, for responders vs non-responders). In conclusion, we provide a detailed analysis of CTLA-4⁺ TIL distribution in melanoma tissues taking into account localization, relationship with CD3⁺/CD8⁺ TILs, and changes in response to IPI treatment. We identified that CTLA-4⁺ TILs may represent a marker of IPI response, alone or with CD3⁺/CD8⁺ subsets, although this requires confirmation in larger studies.

Advances in theranostic biomarkers for tumor immunotherapy

Cancer treatment has known a revolution with the emergence of immune checkpoint inhibitors. However, accurate theranostic biomarkers are lacking. In this review, we discuss different types of biomarkers currently under investigation. First, we focus on tissue biomarkers including PD-L1 expression by immunohistochemistry-the first Food and Drug Administration-approved biomarker-despite conflicting results. In addition, we report on novel biomarkers, including protein-based, molecular (tumor mutational load, immune signature…), circulating (neutrophil-to-lymphocyte ratio, serum cytokines…), and imaging-based biomarkers (radiomic signatures and positron-emission tomography using radiolabeled antibodies). We highlight the limitations of each candidate biomarker and finally discuss combinatorial approaches for their use and the opportunity to switch from a predictive strategy of biomarker research to an adaptive one in the field of cancer immunotherapy.

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.

Characterisation of tumour microenvironment and immune checkpoints in primary central nervous system diffuse large B cell lymphomas

Primary central nervous system diffuse large B cell lymphoma (PCNS-DLBCL) is a rare and aggressive entity of diffuse large B cell lymphoma (DLBCL). Elements of the tumour microenvironment (TME) including tumour-infiltrating lymphocytes (TILs) and tumour-associated macrophages (TAMs) have been associated with survival in DLBCL but their composition and prognostic impact in PCNS-DLBCL are unknown. Programmed cell death-1 (PD1)/programmed death-ligand 1 (PD-L1) immune checkpoint may represent a therapeutic option. Here, we aimed to characterise PD1/PDL1 immune checkpoints and the composition of the TME in PCNS-DLBCL. We collected tumour tissue and clinical data from 57 PCNS-DLBCL and used immunohistochemistry to examine TAMs (CD68, CD163), TILs (CD3, CD4, CD8, PD1) and tumour B cells (PAX5/PDL1 double stains, PDL1). The PDL1 gene was evaluated by fluorescence in situ hybridization (FISH). PAX5/PDL1 identified PDL1 expression by tumour B cells in 10/57 cases (17.5%). PDL1 gene translocation was a recurrent cytogenetic alteration in PNCS-DLBCL (8/47.17%) and was correlated with PDL1 positive expression in tumour B cells. The TME consisted predominantly of CD163 (+) M2 TAMs and CD8 (+) TILs. Most TAMs expressed PDL1 and most TILs expressed PD1. The density of TAMs and TILs did not associate with outcome. We showed that expression of PD1 on TILs and PDL1 on TAMs, but not the expression of PDL1 on tumour B cells was correlated with better prognosis. These findings support a significant role of TME composition and PD1/PDL1 crosstalk in PCNS-DLBCL pathogenesis and bring new insights to the targeted therapy of this aggressive lymphoma.

CTLA-4 Immunohistochemistry and Quantitative Image Analysis for Profiling of Human Cancers

There is an important need in immuno-oncology to develop reliable immunohistochemistry (IHC) to assess the expression of CTLA-4⁺ tumor-infiltrating lymphocytes in human cancers and quantify them with image analysis (IA). We used commercial polyclonal and monoclonal antibodies and characterized three chromogenic cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) assays with suitable specificity and sensitivity for use in formalin-fixed, paraffin-embedded (FFPE) tissues. We found variable numbers of CTLA-4⁺ lymphocytes in multiple types of cancer and secondary lymphoid organs (SLOs) and other normal human tissues. Combining CTLA-4 with CD3, CD4, or CD8 by immunofluorescence showed that CTLA-4⁺ lymphocytes in SLOs and tumors were typically CD3⁺ and CD4⁺, but not CD8⁺. Individual lymphocytes expressed CTLA-4 either as primarily granular cytoplasmic staining or as excentric globular deposits. The CTLA-4/FoxP3 (forkhead box P3 protein) duplex IHC demonstrated that CTLA-4⁺/FoxP3^- lymphocytes predominated in the germinal centers of SLOs and tumor tertiary lymphoid structures (TLSs), whereas CTLA-4⁺/FoxP3⁺ lymphocytes populated the T-cell zone of SLOs and TLSs, plus tumor stroma. IA scoring was highly comparable with pathologist scoring for CTLA-4 and CTLA-4/FoxP3 assays and a FoxP3 single IHC. Our findings show that CTLA-4 IHC can be used to reliably label lymphocytes in FFPE human tissues, making it possible to investigate the role of CTLA-4 in the tumor microenvironment.

Integrative Molecular Characterization of Resistance to Neoadjuvant Chemoradiation in Rectal Cancer

PURPOSE

Molecular properties associated with complete response or acquired resistance to concurrent chemotherapy and radiotherapy (CRT) are incompletely characterized.Experimental Design: We performed integrated whole-exome/transcriptome sequencing and immune infiltrate analysis on rectal adenocarcinoma tumors prior to neoadjuvant CRT (pre-CRT) and at time of resection (post-CRT) in 17 patients [8 complete/partial responders, 9 nonresponders (NR)].

RESULTS

CRT was not associated with increased tumor mutational burden or neoantigen load and did not alter the distribution of established somatic tumor mutations in rectal cancer. Concurrent KRAS/TP53 mutations (KP) associated with NR tumors and were enriched for an epithelial-mesenchymal transition transcriptional program. Furthermore, NR was associated with reduced CD4/CD8 T-cell infiltrates and a post-CRT M2 macrophage phenotype. Absence of any local tumor recurrences, KP/NR status predicted worse progression-free survival, suggesting that local immune escape during or after CRT with specific genomic features contributes to distant progression.

CONCLUSIONS

Overall, while CRT did not impact genomic profiles, CRT impacted the tumor immune microenvironment, particularly in resistant cases.

Looking beyond the hype: Applied AI and machine learning in translational medicine

Big data problems are becoming more prevalent for laboratory scientists who look to make clinical impact. A large part of this is due to increased computing power, in parallel with new technologies for high quality data generation. Both new and old techniques of artificial intelligence (AI) and machine learning (ML) can now help increase the success of translational studies in three areas: drug discovery, imaging, and genomic medicine. However, ML technologies do not come without their limitations and shortcomings. Current technical limitations and other limitations including governance, reproducibility, and interpretation will be discussed in this article. Overcoming these limitations will enable ML methods to be more powerful for discovery and reduce ambiguity within translational medicine, allowing data-informed decision-making to deliver the next generation of diagnostics and therapeutics to patients quicker, at lowered costs, and at scale.

Diffuse distribution of tumor-infiltrating lymphocytes is a marker for better prognosis and chemotherapeutic effect in triple-negative breast cancer

PURPOSE

High-density tumor-infiltrating lymphocytes (TILs) are a prognostic marker for triple-negative breast cancer (TNBC). However, lymphocytic infiltration is heterogeneous in its pattern. We aimed to explore the utility of TIL distribution patterns against TIL density for predicting TNBC prognosis and chemotherapeutic effects.

METHODS

Primary invasive TNBC cases were retrieved from a single institutional cohort, and archived samples were reviewed by two board-certificated pathologists. We used 154 consecutive surgical specimens from patients with standard adjuvant therapy, and 80 biopsies taken before primary systemic chemotherapy. The average density of stromal TILs was scored at 10% intervals, while the distribution pattern of TILs was evaluated as diffuse or non-diffuse. The association between TILs and prognosis or pathological complete response (pCR) was statistically analyzed.

RESULTS

A diffuse pattern of TILs at primary surgery correlated with better prognosis (relapse-free survival [RFS], hazard ratio [HR] 3.71, 95% confidence interval [CI] 1.60-8.57; overall survival [OS], HR 3.87, 95% CI 1.46-10.27), as well as high TIL density (≥ 50%; RFS, HR 4.51, 95% CI 2.06-9.90; OS, HR 3.28, 95% CI 1.32-8.14). Diffuse TIL pattern and nodal status were independent prognostic factors in multivariate analysis. Diffuse TIL pattern upon biopsy was associated with higher pCR rate (diffuse, 46%; non-diffuse, 21%; P = 0.032). All high TIL cases had diffuse patterns and the best outcome. Interobserver concordance was moderate (k = 0.53-0.55; distribution pattern) to good (weighted k = 0.67-0.69; density), and it was faster to assess the distribution pattern than to assess the density of TIL.

CONCLUSIONS

Showing similar clinical impacts to the TIL density, diffuse TILs could be a predictive marker for better prognosis and higher pCR. The assessment of TIL distribution pattern is simple, faster, and practical. Heterogeneous tumor immunity may contribute to further stratification of TNBC treatment.

Characterization of the immune microenvironment of NSCLC by multispectral analysis of multiplex immunofluorescence images

Multiplex immunofluorescence (MIF) staining of tumor sections combined with computational pathology quantifies phenotypic variants of tumor and immune cells and assesses their spatial relationships. Here, we discuss a MIF panel composed of cytokeratin, PD-L1, PD1, CD8, CD68, and Ki67 applied to non-small cell lung cancer (NSCLC) to demonstrate key components of the immune response to this cancer. We also describe a method of whole-slide multiplex imaging and digital multispectral image analysis. Key aspects of marker labeling and digital tissue and cellular classification are highlighted. We then illustrate how digital analysis can measure the spatial relationships among important cell types. This approach is presented in the context of a multidisciplinary team of scientists who together can optimize the combined methods to increase the impact of the study findings. Recommendations are provided to assist others to apply similar methods to further understand the immune response to NSCLC.

Immune cell score in pancreatic cancer-comparison of hotspot and whole-section techniques

An immune cell score (ICS) was introduced for predicting survival in pancreatic ductal adenocarcinoma (PDAC). Few studies have compared different methods of evaluating immune infiltrate. This study compared ICSs determined in whole sections or tissue microarray-like hotspots for predicting survival after PDAC surgery. We included in 79 consecutive patients from a single geographical area that underwent surgery for PDAC (R0/R1, stages I-III). We performed digital image analyses to evaluate CD3 and CD8 staining. ICSs were classified as low, moderate, or high, based on the numbers of immune cells in the tumour core and invasive margin. We compared ICS groups determined with the hotspot and whole-section techniques. Associations between ICS and survival were analysed with Cox regression models, adjusted for sex, age, tumour stage, differentiation grade, perineural invasion, and resection radicality. In hotspot ICS analysis, 5-year overall survival rates for low, moderate, and high groups were 12.1%, 26.3%, and 26.8%, respectively (p = 0.193). In whole-section analyses, overall survival rates were 5.3%, 26.4%, and 43.8%, respectively (p = 0.030). In the adjusted Cox model, whole-section ICS groups were inversely associated with the overall mortality hazard ratio (HR): low, moderate, and high ICS groups had HRs of 1.00, 0.42 (95% CI 0.20-0.88), and 0.27 (95% CI 0.11-0.67), respectively. The number of immune cells per square millimetre in the tumour core and the invasive margin were significantly higher and had a wider range in hotspots than in whole-tissue sections. Accordingly, ICS could predict survival in patients with PDAC after surgery. Whole tissue section ICSs exhibited better prognostic value than hotspot ICSs.

Automated image analysis of NSCLC biopsies to predict response to anti-PD-L1 therapy

Background

Immune checkpoint therapies (ICTs) targeting the programmed cell death-1 (PD1)/programmed cell death ligand-1 (PD-L1) pathway have improved outcomes for patients with non-small cell lung cancer (NSCLC), particularly those with high PD-L1 expression. However, the predictive value of manual PD-L1 scoring is imperfect and alternative measures are needed. We report an automated image analysis solution to determine the predictive and prognostic values of the product of PD-L1+ cell and CD8+ tumor infiltrating lymphocyte (TIL) densities (CD8xPD-L1 signature) in baseline tumor biopsies.

Methods

Archival or fresh tumor biopsies were analyzed for PD-L1 and CD8 expression by immunohistochemistry. Samples were collected from 163 patients in Study 1108/NCT01693562, a Phase 1/2 trial to evaluate durvalumab across multiple tumor types, including NSCLC, and a separate cohort of 199 non-ICT- patients. Digital images were automatically scored for PD-L1+ and CD8+ cell densities using customized algorithms applied with Developer XD™ 2.7 software.

Results

For patients who received durvalumab, median overall survival (OS) was 21.0 months for CD8xPD-L1 signature-positive patients and 7.8 months for signature-negative patients (p = 0.00002). The CD8xPD-L1 signature provided greater stratification of OS than high densities of CD8+ cells, high densities of PD-L1+ cells, or manually assessed tumor cell PD-L1 expression ≥25%. The CD8xPD-L1 signature did not stratify OS in non-ICT patients, although a high density of CD8+ cells was associated with higher median OS (high: 67 months; low: 39.5 months, p = 0.0009) in this group.

Conclusions

An automated CD8xPD-L1 signature may help to identify NSCLC patients with improved response to durvalumab therapy. Our data also support the prognostic value of CD8+ TILS in NSCLC patients who do not receive ICT.

Trial registration

ClinicalTrials.gov identifier: NCT01693562.

Study code: CD-ON-MEDI4736-1108.

Interventional study (ongoing but not currently recruiting).

Actual study start date: August 29, 2012.

Primary completion date: June 23, 2017 (final data collection date for primary outcome measure).

Electronic supplementary material

The online version of this article (10.1186/s40425-019-0589-x) contains supplementary material, which is available to authorized users.

Immune Exclusion Is Frequent in Small-Cell Carcinoma of the Bladder

Small-cell cancer of the urinary bladder is a rare but highly aggressive disease. It is currently unclear whether immune checkpoint therapies that have been approved for urothelial carcinomas will also be efficient in small-cell carcinomas. In this study, we analyzed potential predictors of response including PD-L1 expression and the quantity and location of tumor-infiltrating lymphocytes (TILs) in 12 small-cell and 69 “classical” urothelial cancers by immunohistochemistry. The analysis revealed that small-cell carcinomas were characterized by the virtual absence of PD-L1 expression and an “immune-excluded” phenotype with only a few TILs in the center of the tumor (CT). In small-cell carcinomas, the average immune cell density in the CT (CD3: 159 ± 206, CD8: 87 ± 169 cells/mm²) was more than 3 times lower than that in the urothelial carcinomas (CD3: 625 ± 800, p < 0.001; CD8: 362 ± 626 cells/mm², p = 0.004) while there was no significant difference in the immune cell density at the invasive margin (IM) (small-cell carcinomas CD3: 899 ± 733, CD8: 404 ± 433 cells/mm²; urothelial carcinomas CD3: 1167 ± 1206, p = 0.31; CD8: 582 ± 864 cells/mm², p = 0.27). Positive PD-L1 staining was found in 39% of urothelial cancers, but in only 8% of small-cell bladder cancer cases (p = 0.04). Concordant with these data, a sharp decrease of PD-L1 positivity from >80% to 0% positive cells and of TILS in the CT from 466-1063 CD3-positive cells/mm² to 50-109 CD3-positive cells/mm² was observed in two cancers with clear-cut progression from “classical” urothelial to small-cell carcinoma. In conclusion, these data demonstrate that small-cell bladder cancer commonly exhibits an immune-excluded phenotype.

Introduction to Digital Image Analysis in Whole-slide Imaging: A White Paper from the Digital Pathology Association

The advent of whole-slide imaging in digital pathology has brought about the advancement of computer-aided examination of tissue via digital image analysis. Digitized slides can now be easily annotated and analyzed via a variety of algorithms. This study reviews the fundamentals of tissue image analysis and aims to provide pathologists with basic information regarding the features, applications, and general workflow of these new tools. The review gives an overview of the basic categories of software solutions available, potential analysis strategies, technical considerations, and general algorithm readouts. Advantages and limitations of tissue image analysis are discussed, and emerging concepts, such as artificial intelligence and machine learning, are introduced. Finally, examples of how digital image analysis tools are currently being used in diagnostic laboratories, translational research, and drug development are discussed.

Multiplexed Immunohistochemistry for Molecular and Immune Profiling in Lung Cancer-Just About Ready for Prime-Time?

As targeted molecular therapies and immuno-oncology have become pivotal in the management of patients with lung cancer, the essential requirement for high throughput analyses and clinical validation of biomarkers has become even more intense, with response rates maintained in the 20%–30% range. Moreover, the list of treatment alternatives, including combination therapies, is rapidly evolving. The molecular profiling and specific tumor-associated immune contexture may be predictive of response or resistance to these therapeutic strategies. Multiplexed immunohistochemistry is an effective and proficient approach to simultaneously identify specific proteins or molecular abnormalities, to determine the spatial distribution and activation state of immune cells, as well as the presence of immunoactive molecular expression. This method is highly advantageous for investigating immune evasion mechanisms and discovering potential biomarkers to assess mechanisms of action and to predict response to a given treatment. This review provides views on the current technological status and evidence for clinical applications of multiplexing and how it could be applied to optimize clinical management of patients with lung cancer.

Quantitative Characterization of CD8+ T Cell Clustering and Spatial Heterogeneity in Solid Tumors

Quantitative characterization of the tumor microenvironment, including its immuno-architecture, is important for developing quantitative diagnostic and predictive biomarkers, matching patients to the most appropriate treatments for precision medicine, and for providing quantitative data for building systems biology computational models able to predict tumor dynamics in the context of immune checkpoint blockade therapies. The intra- and inter-tumoral spatial heterogeneities are potentially key to the understanding of the dose-response relationships, but they also bring challenges to properly parameterizing and validating such models. In this study, we developed a workflow to detect CD8+ T cells from whole slide imaging data, and quantify the spatial heterogeneity using multiple metrics by applying spatial point pattern analysis and morphometric analysis. The results indicate a higher intra-tumoral heterogeneity compared with the heterogeneity across patients. By comparing the baseline metrics with PD-1 blockade treatment outcome, our results indicate that the number of high-density T cell clusters of both circular and elongated shapes are higher in patients who responded to the treatment. This methodology can be applied to quantitatively characterize the tumor microenvironment, including immuno-architecture, and its heterogeneity for different cancer types.

Checkpoint blockade-based immunotherapy in the context of tumor microenvironment: Opportunities and challenges

A dynamic and mutualistic interaction between tumor cells and tumor microenvironment (TME) promotes the progression and metastasis of solid tumors. Cancer immunotherapy is becoming a major treatment paradigm for a variety of cancers. Although immunotherapy, especially the use of immune checkpoint inhibitors, has achieved clinical success, only a minority of patients exhibits durable responses. Clinical studies directed at identifying appropriate biomarkers and immune profiles that can be used to predict immunotherapy responses are presently being conducted. Combining treatment strategies tailored to cancer-immune interactions are designed to increase the rate of durable clinical response in patients. It is essential to establish a reasonable tumor classification strategy according to TME to improve cancer immunotherapy. In the current review, a modified classification of TME is proposed, and optimization of TME classification is needed through detailed and integrated molecular characterization of large patient cohorts in the future.