Figure 2 - uploaded by Shamim Mushtaq
Content may be subject to copyright.
Illustration of various hotspot metrics. Various methods of calculating IC hotspots have been described in the literature, and include methods based upon rank-ordering of IC density across subregions, or based upon inferential testing. (A) Rank-order-based approaches, which define hotspots as either 'top 3' (the three most densely infiltrated subregions) or 'top 30%' (the top 30% most densely infiltrated subregions). (B) The Getis-Ord Gi* method, which uses inferential statistical testing to estimate p values indicating the likelihood of each subregion being a hotspot or a coldspot. The Getis-Ord test statistic follows a normal distribution and can be thought of as a measure of local IC density in neighboring subregions, relative to overall IC density. IC, immune cell; TC, tumor cell.
Source publication
Modern histologic imaging platforms coupled with machine learning methods have provided new opportunities to map the spatial distribution of immune cells in the tumor microenvironment. However, there exists no standardized method for describing or analyzing spatial immune cell data, and most reported spatial analyses are rudimentary. In this review...
Contexts in source publication
Context 1
... characterization of H&E breast cancer TIL hotspots and coldspots (as binary 'absent or present' variables) were shown to predict progression-free survival, independent of clinical risk factors, in a cohort from the Cancer Genome Atlas and Carolina Breast Cancer Study [44]. A commonly adopted quantitative method for analyzing hotspots is to rasterize the tumor area, rank-order the subregions by IC density, and define hotspots as either a fixed number, fixed percentile, or a fixed SD cut-off of highest-density subregions (Figure 2A) [13,29,34]. In a series of 53 melanoma patients treated with pembrolizumab, it was concluded that a hotspot definition of IC densities of 30% of the highest-ranking HPFs resulted in the best prediction of the clinical response to pembrolizumab and outperformed conventional assessments of mean IC density [13]. ...
Context 2
... limitation is that it rests upon the presumption that all tumors have hotspots, whereas it is biologically possible that IC hotspots occur at different frequencies across tumors. As an alternative, the Getis-Ord Gi* is a probabilistic approach for defining hotspots and coldspots within a tumor ( Figure 2B). The approach involves defining neighbors for each tumor subregion (i.e. usually defined as immediately adjacent subregions) and conducting a formal hypothesis test using these data to estimate a likelihood that any given neighborhood is a hotspot, given the totality of the spatial data. ...
Context 3
... metric was only weakly correlated with lymphocyte density, indicating that the two metrics may be useful in conjunction for stratifying tumors by immune response. Figure 2 illustrates the top 3, top 30%, and Getis-Ord Gi* approaches for defining hotspots in a breast cancer specimen. In this example, the mean IC hotspot density varies substantially depending on the method (top 3: 3.6 Â 10 À3 /μm 2 ; top 30%: 1.6 Â 10 À4 /μm 2 ; Getis-Ord p < 0.1: 2.2 Â 10 À4 /μm 2 ; Getis-Ord p < 0.05: 1.8 Â 10 À4 /μm 2 ). ...
Context 4
... characterization of H&E breast cancer TIL hotspots and coldspots (as binary 'absent or present' variables) were shown to predict progression-free survival, independent of clinical risk factors, in a cohort from the Cancer Genome Atlas and Carolina Breast Cancer Study [44]. A commonly adopted quantitative method for analyzing hotspots is to rasterize the tumor area, rank-order the subregions by IC density, and define hotspots as either a fixed number, fixed percentile, or a fixed SD cut-off of highest-density subregions (Figure 2A) [13,29,34]. In a series of 53 melanoma patients treated with pembrolizumab, it was concluded that a hotspot definition of IC densities of 30% of the highest-ranking HPFs resulted in the best prediction of the clinical response to pembrolizumab and outperformed conventional assessments of mean IC density [13]. ...
Context 5
... limitation is that it rests upon the presumption that all tumors have hotspots, whereas it is biologically possible that IC hotspots occur at different frequencies across tumors. As an alternative, the Getis-Ord Gi* is a probabilistic approach for defining hotspots and coldspots within a tumor ( Figure 2B). The approach involves defining neighbors for each tumor subregion (i.e. usually defined as immediately adjacent subregions) and conducting a formal hypothesis test using these data to estimate a likelihood that any given neighborhood is a hotspot, given the totality of the spatial data. ...
Context 6
... metric was only weakly correlated with lymphocyte density, indicating that the two metrics may be useful in conjunction for stratifying tumors by immune response. Figure 2 illustrates the top 3, top 30%, and Getis-Ord Gi* approaches for defining hotspots in a breast cancer specimen. In this example, the mean IC hotspot density varies substantially depending on the method (top 3: 3.6 Â 10 À3 /μm 2 ; top 30%: 1.6 Â 10 À4 /μm 2 ; Getis-Ord p < 0.1: 2.2 Â 10 À4 /μm 2 ; Getis-Ord p < 0.05: 1.8 Â 10 À4 /μm 2 ). ...
Similar publications
Spatial transcriptomics allows for the analysis of a cell's gene expression in the context of its physical location. With spatial transcriptomics data, investigators often want to find genes of interest whose spatial patterns are biologically relevant in multiple samples. However, due to confounding factors in spatial data that produce noise across...
Citations
... Conflicting data have been reported regarding the association of single immune markers with pCR in early BC, highlighting thus the challenges of multiplex assays 9,25 . However, these multidimensional methods could provide further directions beyond the immune cell abundance, hence a growing body of literature is focusing on the spatial interactions among the different cells within the TME 6 . In the ARTEMIS trial on neoadjuvant chemotherapy for TNBC, a closer spatial proximity of T cells to cancer cells was associated with increased pCR rates 25 . ...
Breast cancer (BC) represents a heterogeneous ecosystem and elucidation of tumor microenvironment components remains essential. Our study aimed to depict the composition and prognostic correlates of immune infiltrate in early BC, at a multiplex and spatial resolution. Pretreatment tumor biopsies from patients enrolled in the EORTC 10994/BIG 1-00 randomized phase III neoadjuvant trial (NCT00017095) were used; the CNN11 classifier for H&E-based digital TILs (dTILs) quantification and multiplex immunofluorescence were applied, coupled with machine learning (ML)-based spatial features. dTILs were higher in the triple-negative (TN) subtype, and associated with pathological complete response (pCR) in the whole cohort. Total CD4+ and intra-tumoral CD8+ T-cells expression was associated with pCR. Higher immune-tumor cell colocalization was observed in TN tumors of patients achieving pCR. Immune cell subsets were enriched in TP53 -mutated tumors. Our results indicate the feasibility of ML-based algorithms for immune infiltrate characterization and the prognostic implications of its abundance and tumor-host interactions.
... This is a time consuming process requiring pooling or separation of groups of spectra, or intense dimensionality reduction. The central hypothesis of this study is that the use of deep learning approaches on chemical image data, where spectral information is not separated from its spatial context, will lead to improved performance in predicting disease recurrence, where spatial context within the tumor microenvironment is known to be a contributory factor [49][50][51] . ...
... Within the field of digital pathology substantial advances have been made towards the use of digital images acquired from standard H&E specimens within automated pathological support systems 40,54 . Within the breast cancer field the approach has successfully demonstrated the capability to classify disease grade 55,56 , subtype 56,57 and for automated image segmentation, including the identification of the location and distribution of tumor infiltrating lymphocytes, which is itself an important prognostic indicator 49,[58][59][60] . More recently the field has demonstrated potential to classify survival probability 61,62 and risk of recurrence using OncoType-DX genetic risk scores as ground truth 63 . ...
Predicting long-term recurrence of disease in breast cancer (BC) patients remains a significant challenge for patients with early stage disease who are at low to intermediate risk of relapse as determined using current clinical tools. Prognostic assays which utilize bulk transcriptomics ignore the spatial context of the cellular material and are, therefore, of limited value in the development of mechanistic models. In this study, Fourier-transform infrared (FTIR) chemical images of BC tissue were used to train deep learning models to predict future disease recurrence. A number of deep learning models were employed, with champion models employing two-dimensional and two-dimensional-separable convolutional networks found to have predictive performance of a ROC AUC of approximately 0.64, which compares well to other clinically used prognostic assays in this space. All-digital chemical imaging may therefore provide a label-free platform for histopathological prognosis in breast cancer, opening new horizons for future deployment of these technologies.
... 28 As mentioned, understanding the spatial measures of TILs is critical; however, there is no standardized approach. 29 A report from the IIOBWG provides recommendations for spatial characterization of TILs in breast cancer using computational methods. 29 This report suggests that novel imaging platforms could be used to improve the quantification and spatial characterization of sTILs over H&E slides using segmentation, subregion sampling, density estimation, and calculation of the sTIL score. ...
... 29 A report from the IIOBWG provides recommendations for spatial characterization of TILs in breast cancer using computational methods. 29 This report suggests that novel imaging platforms could be used to improve the quantification and spatial characterization of sTILs over H&E slides using segmentation, subregion sampling, density estimation, and calculation of the sTIL score. However, a major caveat is that different computational methods on TILs measure different spatial metrics, causing discrepancies. ...
Introduction
In this review, we evaluate the role of stromal tumor-infiltrating lymphocytes (sTILs) as a biomarker in human epidermal growth factor receptor 2 (HER2)-positive breast cancer, exploring the prognostic and predictive potential in various treatment settings.
Methods
Data from multiple clinical trials in the early and metastatic settings, focusing on TILs’ correlation with pathologic complete response (pCR), progression-free survival (PFS), and overall survival across early and metastatic HER2-positive breast cancer were summarized. This review also discusses TILs’ assessment methods, interobserver variability, and emerging technologies to assess TILs.
Results
TILs have been identified as a highly reproducible biomarker that predicts pCR in patients receiving neoadjuvant therapy and serves as a prognostic indicator for long-term outcomes in several breast cancer subtypes, including HER2-positive. Studies indicate that higher TIL levels correlate with better recurrence-free survival rates. Despite these findings, there is no consensus on the optimal TIL threshold for clinical decision making, and further research is required on how to incorporate TILs into routine clinical practice.
Conclusions
TILs represent a promising biomarker in HER2-positive breast cancer, particularly in early disease settings. This assessment could guide treatment de-escalation or intensification, tailoring therapies to individual patient profiles. Due to their prognostic importance, TILs can be added to pathology reports. However, further validation in clinical trials is essential for the widespread adoption of TILs in clinical practice.
... Thus, numerous histological foundation models [6,31,32,55,58] have been proposed, pre-trained on large-scale datasets for general-purpose representations. Unlike natural images, the analysis of cell spatial distribution within WSIs has been verified as clinically important, associated with the molecular profile [44], tumor progression [9], prognostic biomarkers [37], etc. Heavy reliance on the foun-dation models, combined with an oversight of cell spatial distribution, results in high computational costs and suboptimal performance. Fundamentally, this limitation stems from the absence of cell-level annotations in existing WSI datasets [1,2,28,43] due to the extremely high cost 1 . ...
... Cell-level annotations and statistics hold significant clinical importance [37]. To illustrate it, we construct survival risk evaluation metrics based on the WSI-Cell5B and conduct Kaplan-Meier analyses on three types of cancer. ...
... To illustrate it, we construct survival risk evaluation metrics based on the WSI-Cell5B and conduct Kaplan-Meier analyses on three types of cancer. Based on clinical experience, the survival risk is highly influenced by the proportion and distribution of neoplastic and inflammatory cells [37]. Therefore, We construct two metrics as shown in the left of Figure 3: Cell Proportion Score (CPS) and Multi-scale Cell Proportion Score (MCPS). ...
It is clinically crucial and potentially very beneficial to be able to analyze and model directly the spatial distributions of cells in histopathology whole slide images (WSI). However, most existing WSI datasets lack cell-level annotations, owing to the extremely high cost over giga-pixel images. Thus, it remains an open question whether deep learning models can directly and effectively analyze WSIs from the semantic aspect of cell distributions. In this work, we construct a large-scale WSI dataset with more than 5 billion cell-level annotations, termed WSI-Cell5B, and a novel hierarchical Cell Cloud Transformer (CCFormer) to tackle these challenges. WSI-Cell5B is based on 6,998 WSIs of 11 cancers from The Cancer Genome Atlas Program, and all WSIs are annotated per cell by coordinates and types. To the best of our knowledge, WSI-Cell5B is the first WSI-level large-scale dataset integrating cell-level annotations. On the other hand, CCFormer formulates the collection of cells in each WSI as a cell cloud and models cell spatial distribution. Specifically, Neighboring Information Embedding (NIE) is proposed to characterize the distribution of cells within the neighborhood of each cell, and a novel Hierarchical Spatial Perception (HSP) module is proposed to learn the spatial relationship among cells in a bottom-up manner. The clinical analysis indicates that WSI-Cell5B can be used to design clinical evaluation metrics based on counting cells that effectively assess the survival risk of patients. Extensive experiments on survival prediction and cancer staging show that learning from cell spatial distribution alone can already achieve state-of-the-art (SOTA) performance, i.e., CCFormer strongly outperforms other competing methods.
... Users can customize the grid size using the scale argument, achieving a balance between resolution for detecting local-regional variations and the stability of estimates. Guidelines for optimizing grid size have been previously reported in studies such as Page, D. B. et al. 25 , Mi, H. et al. 26 , Francis, K. et al. 27 , Turkki, R. et al. 28 and Verghese, G. et al. 29 . ...
Understanding spatial dynamics within tissue microenvironments is crucial for deciphering cellular interactions and molecular signaling in living systems. These spatial characteristics govern cell distribution, extracellular matrix components, and signaling molecules, influencing local biochemical and biophysical conditions. Despite significant progress in analyzing digital pathology images, current methods for capturing spatial relationships are limited. They often rely on specific spatial features that only partially describe the complex spatial distributions of cells and are frequently tied to particular outcomes within predefined model frameworks. Furthermore, these methods are typically limited to field of view analysis, which restricts their capacity to capture spatial patterns across whole-slide images, thereby limiting their ability to fully address the complexities of tissue architecture. To address these limitations, we present SpatialQPFs (Spatial Quantitative Pathology Features), an R package designed to extract interpretable spatial features from cell imaging data using spatial statistical methodologies. Leveraging segmented cell information, our package offers a comprehensive toolkit for applying a range of spatial statistical methods within a stochastic process framework, including analyses of point process data, areal data, and geostatistical data. By decoupling feature extraction from specific outcome models, SpatialQPFs enables thorough large-scale spatial analyses applicable across diverse clinical and biological contexts. This approach enhances the depth and accuracy of spatial insights derived from tissue data, empowering researchers to conduct comprehensive spatial analyses efficiently and reproducibly. By providing a flexible and robust framework for spatial feature extraction, SpatialQPFs facilitates advanced spatial analyses, paving the way for new discoveries in tissue biology and pathology. SpatialQPFs code and documentation are publicly available at https://github.com/Genentech/SpatialQPFs.
... Secondly, hematoxylin and eosin staining oversimplifies the nature of the immune response, as it does not allow for the distinction between different TIL subtypes, each potentially playing different roles within the tumor microenvironment [42]. Thirdly, this method does not estimate the spatial aspects of the TIL distributions that could contain valuable clinical information [43][44][45]. Furthermore, the omission of intratumoral TILs from the assessment may miss the potential of comprehensive immune microenvironment evaluation [40,43], especially in cancers where the presence of intratumoral TILs indicates a strong immune response within the tumor core. ...
Introduction
Breast cancer (BC) presents diverse malignancies with varying biological and clinical behaviors, driven by an interplay between cancer cells and tumor microenvironment. Deciphering these interactions is crucial for personalized diagnostics and treatment. This study explores the prognostic impact of tumor proliferation and immune response patterns, assessed by computational pathology indicators, on breast cancer-specific survival (BCSS) models in estrogen receptor-positive HER2-negative (ER+HER2–) and triple-negative BC (TNBC) patients.
Materials and methods
Whole-slide images of tumor surgical excision samples from 252 ER+HER2– patients and 63 TNBC patients stained for estrogen and progesterone receptors, Ki67, HER2, and CD8 were analyzed. Digital image analysis (DIA) was performed for tumor tissue segmentation and quantification of immunohistochemistry (IHC) markers; the DIA outputs were subsampled by hexagonal grids to assess the spatial distributions of Ki67-positive tumor cells and CD8-positive (CD8+) cell infiltrates, expressed as Ki67-entropy and CD8-immunogradient indicators, respectively. Prognostic models for BCSS were generated using multivariable Cox regression analysis, integrating clinicopathological and computational IHC indicators.
Results
In the ER+HER2– BC, multivariable Cox regression revealed that high CD8+ density within the tumor interface zone (IZ) (HR: 0.26, p = 0.0056), low immunodrop indicator of CD8+ density (HR: 2.93, p = 0.0051), and low Ki67-entropy (HR: 5.95, p = 0.0.0061) were independent predictors of better BCSS, while lymph node involvement predicted worse BCSS (HR: 3.30, p = 0.0013). In TNBC, increased CD8+ density in the IZ stroma (HR: 0.19, p = 0.0119) and Ki67-entropy (HR: 3.31, p = 0.0250) were independent predictors of worse BCSS. Combining these independent indicators enhanced prognostic stratification in both BC subtypes.
Conclusions
Computational biomarkers, representing spatial properties of the tumor proliferation and immune cell infiltrates, provided independent prognostic information beyond conventional IHC markers in BC. Integrating Ki67-entropy and CD8-immunogradient indicators into prognostic models can improve patient stratification with regard to BCSS.
... Our study revealed that, CD4 + T cells and monocytes increased in the T1 group after treatment, while CD8 + T cells and B cells decreased in the T2 group after treatment. The use of immune cell in ltration as a novel biomarker for predicting the prognosis of patients with various types of cancer holds great promise [23,24] . While previous reports have implicated immune in ltration in affecting tumor patient prognosis [25][26][27] , the interaction mechanism between prognosis and the tumor microenvironment remains incompletely understood. ...
Background
Immunoneoadjuvant therapy has garnered considerable attention owing to significant strides in cancer treatment. We aimed to explore the molecular mechanisms underpinning immunoneoadjuvant therapy through a comprehensive multiomics analysis using samples from a registered clinical trial cohort.
Methods
Preoperative samples were collected from 16 patients, and postoperative samples were obtained from 12 among them. RNA-seq and Olink proteomics were employed to identify key genes before and after neoadjuvant treatment. The weighted coexpression network was constructed using Weighted gene co-expression network analysis (WGCNA). Furthermore, the proportion of infiltrated immune cells was calculated using xCell based on normalized expression data derived from RNA-seq.
Results
Patients were stratified into T1 (good efficacy) and T2 (poor efficacy) groups based on Tumor Regression Grade (TRG) to neoadjuvant immunotherapy. Compared to the T2 group (TRG2 and TRG3), the T1 group (TRG0 and TRG1) showed significant differences in pathways related to inflammatory response and myeloid leukocyte activation. Furthermore, the T1 group exhibited elevated levels of CD8 + T cells and B cells. The top two factors with the highest area under the Receiver Operating Characteristic (ROC) curve were CD8a molecule (CD8A) (1.000) and C-C motif chemokine ligand 20 (CCL20) (0.967). Additionally, the expression of Placenta Growth Factor (PGF) and TNF receptor superfamily member 21 (TNFRSF21) proteins significantly increased compared to the T2 group. High expression of CD8A and PGF were associated with favorable and poor prognosis in gastric cancer patients, respectively. Immunoinfiltration analysis revealed a positive correlation between CD8A and Dendritic Cell (DC) levels, while a negative correlation was observed with Myeloid-derived suppressor cell (MDSC) levels.
Conclusions
Through multiomics analysis, we discovered that CD8A is linked to enhanced treatment response and tumor regression. Conversely, PGF exhibited contrasting effects, hinting at a potential adverse influence on treatment outcomes.
... Rapid advances in digital histopathology have also allowed the extraction of clinically relevant information embedded in tumor slides by applying machine learning and artificial intelligence methods, capitalizing on recent advancements in image analysis through deep learning 4 . Key advances are already underway, as whole-slide images (WSIs) of tissue stained with hematoxylin and eosin (H&E) have been used to diagnose tumors computationally [5][6][7][8] , classify cancer types 7,[9][10][11][12][13] , distinguish tumors with low or high mutation burden 14 , identify genetic mutations 6,[15][16][17][18][19][20][21][22][23] , predict patient survival [24][25][26][27][28][29][30] , detect DNA methylation patterns 31 and mitosis 32 , and quantify tumor immune infiltration 33 , tumor-infiltrating lymphocytes (TILs) 34 and spatial immune cell infiltration 35 . ...
Advances in artificial intelligence have paved the way for leveraging hematoxylin and eosin-stained tumor slides for precision oncology. We present ENLIGHT–DeepPT, an indirect two-step approach consisting of (1) DeepPT, a deep-learning framework that predicts genome-wide tumor mRNA expression from slides, and (2) ENLIGHT, which predicts response to targeted and immune therapies from the inferred expression values. We show that DeepPT successfully predicts transcriptomics in all 16 The Cancer Genome Atlas cohorts tested and generalizes well to two independent datasets. ENLIGHT–DeepPT successfully predicts true responders in five independent patient cohorts involving four different treatments spanning six cancer types, with an overall odds ratio of 2.28 and a 39.5% increased response rate among predicted responders versus the baseline rate. Notably, its prediction accuracy, obtained without any training on the treatment data, is comparable to that achieved by directly predicting the response from the images, which requires specific training on the treatment evaluation cohorts.
... Users can customize the grid size using the scale argument, achieving a balance between resolution for detecting localregional variations and the stability of estimates. Guidelines for optimizing grid size have been previously reported in studies such as Page, D. B. et al. 22 , Mi, H. et al. 23 , Francis, K. et al. 24 , Turkki, R. et al. 25 and Verghese, G. et al. 26 . ...
Understanding the spatial dynamics within tissue microenvironments is crucial for deciphering cellular interactions and molecular signaling in living systems. These spatial characteristics govern cell distribution, extracellular matrix components, and signaling molecules, influencing local biochemical and biophysical conditions. Decoding these features offers insights into physiological processes, disease progression, and clinical outcomes. By elucidating spatial relationships between cell types, researchers uncover tissue architecture, cell communication networks, and microenvironment dynamics, aiding in the identification of biomarkers and therapeutic targets. Digital pathology imaging, including Hematoxylin and Eosin (H&E) staining, provides high-resolution histological information that offer intricate insights into cell-cell spatial relationships with greater details. However, current methods for capturing cell-cell spatial interactions are constrained by either methodological scopes or implementations restricted to script-level access. This limitation undermines generalizability and standardization, crucial for ensuring reproducibility. To address these limitations, we introduce SpatialQPFs , an extendable R package designed for extraction of interpretable spatial features from digital pathology images. By leveraging segmented cell information, our package provides researchers with a comprehensive toolkit for applying a range of spatial statistical methods within a stochastic process framework which includes analysis of point pattern data, areal data, and geostatistical data. This allows for a thorough analysis of cell spatial relationships, enhancing the depth and accuracy of spatial insights derived from the tissue, thereby empowering researchers to conduct comprehensive spatial analyses efficiently and reproducibly.
... It is of paramount importance to tackle such analyses using a multidisciplinary approach, whereby immunologists, pathologists, and statisticians work together to ensure proper hypothesis generation, histological annotation, and statistical modelling/testing. For detailed information on analysing and reporting spatial immune profiling data, please see the companion manuscript from The International Immuno-Oncology Biomarker Working Group [52]. ...
Recent advances in the field of immuno‐oncology have brought transformative changes in the management of cancer patients. The immune profile of tumours has been found to have key value in predicting disease prognosis and treatment response in various cancers. Multiplex immunohistochemistry and immunofluorescence have emerged as potent tools for the simultaneous detection of multiple protein biomarkers in a single tissue section, thereby expanding opportunities for molecular and immune profiling while preserving tissue samples. By establishing the phenotype of individual tumour cells when distributed within a mixed cell population, the identification of clinically relevant biomarkers with high‐throughput multiplex immunophenotyping of tumour samples has great potential to guide appropriate treatment choices. Moreover, the emergence of novel multi‐marker imaging approaches can now provide unprecedented insights into the tumour microenvironment, including the potential interplay between various cell types. However, there are significant challenges to widespread integration of these technologies in daily research and clinical practice. This review addresses the challenges and potential solutions within a structured framework of action from a regulatory and clinical trial perspective. New developments within the field of immunophenotyping using multiplexed tissue imaging platforms and associated digital pathology are also described, with a specific focus on translational implications across different subtypes of cancer. © 2024 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.
