apjcc.waocp.

com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

DOI:10.31557/APJCC.2024.9.4.793 REVIEW

Dissecting the Tumor Microenvironment (TME) to Decipher

New Immunotherapy Targets by Using Artificial Intelligence
Mohd Umar Rehmani1, Soham Aggarwal2, Chandra Prakash Prasad1, Mayank
Singh1

Department of Medical Oncology (Lab) BRAIRCH AIIMS, New Delhi, India. 2Harvard T H Chan Indian Research Centre (HRC)
1

Mumbai India.

Abstract
Cancer remains one of the leading causes of death worldwide, second only to cardiovascular diseases. Standard
cancer treatments, such as chemotherapy, radiation, and surgery, are effective for primary tumors but often fail
to eliminate disseminated tumor cells responsible for metastasis. This limitation underscores the necessity for
advanced therapeutic strategies, leading to the rise of immunotherapy, which leverages the immune system to
combat cancer. Despite its promise, immunotherapy faces challenges, including variable patient responses and
immune-related toxicities, complicating the prediction of treatment efficacy. Here, artificial intelligence (AI)
emerges as a vital tool that can enhance the precision and effectiveness of immunotherapy by analyzing the intricate
tumor microenvironment (TME). This paper explores the limitations of current immunotherapies and examines
how AI can address these challenges. It discusses the TME’s role in shaping immune responses, highlighting how
understanding its complexities can improve predictive power and treatment outcomes. Furthermore, we address
the limitations of AI in cancer research and propose future directions for its integration into clinical practice,
with the potential to revolutionize personalized cancer therapy and improve overall patient care.

Keywords: Artificial Intelligence- Immunotherapy- Tumor Microenvironment

Asian Pac J Cancer Care, 9 (4), 793-799 Submission Date: 09/23/2024 Acceptance Date: 11/03/2024

Introduction
Cancer continues to be one of the leading causes of gap has led to the rise of immunotherapy, a promising
mortality. According to recent statistics from the American treatment modality that harnesses the immune system
Cancer Society, cancer remains the second most common to fight cancer [2]. While immunotherapy has shown
cause of mortality in the USA, particularly among those significant success in some cancers, challenges such as
under 85 years old. The COVID-19 pandemic exacerbated variable patient responses, immune-related toxicities, and
this crisis by delaying cancer diagnosis and treatment due the complexity of predicting treatment efficacy remain.
to healthcare facility closures, economic uncertainties, This is where artificial intelligence (AI) emerges as a
and patients’ fear of exposure to the virus. These delays critical tool, offering the potential to enhance the precision
have raised concerns about an increase in late-stage and effectiveness of immunotherapy. By leveraging
cancer diagnoses, potentially contributing to higher AI to analyze the intricate tumor microenvironment
mortality rates at the community level [1]. The standard (TME), researchers can gain insights into the dynamic
cancer treatments, such as surgery, chemotherapy, and interactions between cancer cells, immune cells, and other
radiotherapy are often effective at treating the primary components. Understanding these interactions can reveal
tumor. However, their inability to eliminate dispersed novel immunotherapy targets and improve treatment
tumor cells responsible for metastasis highlights the outcomes. AI can help predict responses, minimize
need for more advanced therapeutic strategies. This toxicities, and guide more personalized cancer therapies

Corresponding Author:
Dr. Mayank Singh
Department of Medical Oncology (Lab) BRAIRCH AIIMS New Delhi, India.
Email: [email protected]

Asian Pacific Journal of Cancer Care• Vol 9• Issue 4 793

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

by unravelling the TME landscape. We will explore the To address these gaps, there is a pressing need for
limitations of current immunotherapies and examine more accurate, reproducible, and cost-effective predictive
how AI can address these challenges. We will discuss the biomarkers to inform clinical decisions. This is where
role of the TME in shaping immune responses and how artificial intelligence (AI) comes into play. By analyzing
deciphering its complexities could enhance the predictive vast amounts of data from tumor microenvironments, AI
power and efficacy of cancer immunotherapy. Finally, we has the potential to uncover new immunotherapy targets
will explore the limitations of AI in cancer research and the and improve patient outcomes in this evolving landscape
future directions for its integration into clinical practice. of cancer treatment.

2. Cancer immunotherapy 3. Artificial Intelligence

Cancer immunotherapy represents a revolutionary Artificial intelligence (AI) is a branch of computer
approach that utilizes the body’s immune system to combat science that enables machines to perform tasks typically
cancer. Recently, this strategy has gained significant reserved for humans, such as learning, thinking, and
attention due to its promising results, with innovations problem-solving [15, 16]. It includes subsets like machine
ranging from immune checkpoint inhibitors to adoptive learning (ML) and deep learning (DL), which have become
cell therapies [3, 4]. Today, cancer immunotherapy is widely used, including in cancer research. Researchers
employed across a various cancers, from hematological can utilize off-the-shelf AI products or develop custom
malignancies to solid tumors. This shift has been software pipelines to enhance productivity, uncover
largely inspired by the remarkable successes of immune hidden insights, and improve cancer immunotherapy
checkpoint inhibitors (ICIs) in melanoma patients and by better understanding the tumor microenvironment
CAR-T cell therapies in blood cancers like leukemia (TME) [11, 17]. Machine learning focuses on pattern
and multiple myeloma [5, 6]. At the heart of this recognition and is often called a learning machine due
immunotherapeutic approach are immune checkpoints to its ability to learn from data. ML has been utilized in
key inhibitory receptors that tumors exploit to evade T cancer research for quite some time and there are various
cell activity, a phenomenon known as immune escape. The tools available. A component or method of ML is artificial
most notable checkpoint inhibitors currently in use include neural network or ANN which is kind of resembles brain
programmed cell death 1 (PD-1), programmed death neural networks and consists of components or units
ligand 1 (PD-L1), and cytotoxic T lymphocyte-associated known as neurons organized into multiple layers. These
antigen 4 (CTLA-4). Immunotherapy drugs such as layers include an input layer receiving the input data, an
nivolumab (anti-PD-1), atezolizumab (anti-PD-L1), and output layer producing the final output, and a few hidden
ipilimumab (anti-CTLA-4) have shown considerable layers involved in computation and abstraction [18, 19].
anti-tumor effects, heralding a new era in cancer treatment Deep learning is a kind of ML that utilizes multilayered
[5, 7]. While CAR-T cell therapy has demonstrated ANNs, which makes it quite impressive. DL has shown
impressive outcomes in hematological cancers, it is now major contribution or development in the field of image
making strides in addressing the challenges posed by processing and the branch or method used is known as
solid tumors. Advances in loco-regional delivery and the computer vision. These deep ANNs may use a component
identification of new biomarkers are paving the way for known as convolutions which reduces the raw pixel to
targeted therapies that can tackle tumor heterogeneity relevant information [11]. Because of the expanded design
and immune suppression—hallmarks of cancer [8-10]. of deep ANNs, higher degrees of computation and data
However, despite these promising developments, the representation are supported which enables deep neural
effectiveness of immunotherapy varies significantly networks to learn complicated patterns and abstract more
among patients. High treatment costs and unpredictable information. Usually, large datasets are employed to train
responses remain major challenges. Some individuals Deep neural networks. Today, in cancer research the
experience substantial benefits, while others may show terms AI and DL are used interchangeably quite a lot, the
little to no response, sometimes accompanied by serious concept of AI overlaps with deep learning. During deep
side effects and toxicities [11, 12]. Identifying patients learning, the network learns to perform tasks from inputs.
who are most likely to benefit from these therapies is These inputs may be images such as histopathological
crucial for enhancing diagnostic accuracy and minimizing or radiology images as in the case of oncology [20].
unnecessary toxicities. This could also alleviate the Additionally, Multimodal deep learning models integrate
financial burden of treatment, making it more accessible, diverse data types, enhancing data interpretation by
especially in developing countries. A few biomarkers, considering various sources and host factors [21]. By
such as the expression of PD-L1, microsatellite instability harnessing these advanced AI techniques, we can deepen
(MSI), tumor mutational burden (TMB), and the number our understanding of the tumor microenvironment (TME)
of tumor-infiltrating lymphocytes (TILs), have been and its implications for immunotherapy, particularly in
identified to predict responses to immune checkpoint response prediction and efficacy.
inhibitors [13, 14]. Yet, the predictive performance of these
biomarkers whether used individually or in combination 3.1 Unraveling the Tumor Microenvironment: Insights
remains suboptimal. Some tumors exhibit resistance and Innovations Through AI
despite the presence of these biomarkers, while others The tumor microenvironment (TME) refers to the
may respond favorably without them [15]. complex cellular landscape in which tumors and cancer

794 Asian Pacific Journal of Cancer Care• Vol 9• Issue 4

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

stem cells exist. It significantly influences the growth, like gene expression or images, and as such are unable
behavior, and intercellular communication of cancer cells to completely appreciate and express the depth, diversity
[22, 23]. Comprising a variety of elements, the TME and dynamism of TME [37-40]. Multimodal models are
includes immune cells such as neutrophils, macrophages, useful for combining multiple and diverse data entities
and lymphocytes, as well as non-immune cells like and as such can combine both the spatial and non-spatial
fibroblasts and vascular endothelial cells. This intricate data together [41]. The emergence of such models, which
environment is not just a mere aggregation of cancer converge diverse data types—such as genomic, clinical,
cells; rather, it is a heterogeneous mix of resident and and imaging data enhances our ability to identify relevant
invading host cells, extracellular matrix components, and patterns and improve diagnostic accuracy than ever before
secreted factors. From the onset of tumor development, [42, 43]. These sophisticated models can analyze complex
cancer cells and the constituents of the TME establish a relationships among various TME components, offering a
dynamic and bidirectional relationship that fosters cancer comprehensive understanding of how these interactions
cell survival, metastatic spread, and regional invasion. affect patient prognosis and response to immunotherapy
Immune cells play a crucial role in this interplay; they can ultimately paving the way for enhanced therapeutic
both support and hinder tumor development, influencing strategies.
carcinogenesis, tumor progression, metastasis, and
recurrence [24, 25]. Traditional assessment methods, 3.2 How AI can help in Immunotherapy: response
including western blotting, coimmunoprecipitation, prediction and efficacy
and real-time quantitative polymerase chain reaction, AI is transforming the landscape of cancer
have provided insights into the interactions between immunotherapy by enhancing response prediction and
tumors and their microenvironments. However, newer efficacy. One of the key ways AI contributes is through its
high-throughput technologies, such as genomics, ability to identify new biomarkers and quantitatively assess
proteomics, and single-cell sequencing, have revealed the existing ones, such as tumor mutational burden (TMB),
TME’s complexity, indicating that current approaches are microsatellite instability (MSI), and PD-1 expression
insufficient [26-28]. Here, artificial intelligence (AI) can [44]. By analyzing image data from oncology, including
bridge the gap, utilizing deep learning to synthesize vast radiology and histopathology slides, AI can extract critical
amounts of data from multiple sources and uncover novel information that has often been overlooked in traditional
insights. AI has the potential to transform TME analysis by medical settings. The core principle of image-based
managing extensive datasets and conducting sophisticated biomarkers lies in the recognition that routinely acquired
image analyses [29]. By evaluating quantitative and images contain much more data than is currently utilized
spatial characteristics of tumor and immune cells within [11]. With the aid of deep learning algorithms, particularly
the TME, AI can reveal the predictive prognostic value of those based on artificial neural networks, AI can abstract
the environment and provide new avenues for therapeutic meaningful insights from these images. This capability is
intervention [30, 31]. Its deep learning algorithms are particularly relevant in oncology, where radiology images
adept at extracting information from histopathological confirm malignancy and histopathology provides insights
images, such as H & E-stained slides, allowing for accurate into tumor characteristics and staging. A significant
quantification of immune cells and tumor-associated advantage of AI is its ability to leverage the vast amount of
structures [32]. Moreover, AI can predict biomarkers like imaging data available in cancer diagnosis and treatment.
TMB, MSI, and PD-L1 expression directly from imaging These images can serve as raw material for training AI
data, enhancing our understanding of tumor phenotypes models that predict immunotherapy responses and assess
and their evolution throughout treatment [30, 33, 34]. This treatment efficacy. Deep learning systems can extract
capability is invaluable, as it enables real-time tracking far more information from radiological and histological
of molecular changes that may influence therapeutic images than what is typically harnessed in healthcare
outcomes. Also, by integrating AI with advancements settings.
in spatial transcriptomics and other high-throughput Deep Radiomics: The emergence of deep radiomics a
technologies, researchers can uncover detailed insights refined version of classical radiomics has revolutionized
into cellular interactions and positional relationships the extraction of features from medical images. While
within the TME. Currently available research, however, traditional radiomics software focused on a limited set of
has shown that the complex interactions with cancer cells features such as shape, intensity, and texture, deep radiomics
can cause the precise activity of immune cells to change employs convolutional neural networks (CNNs) to access
and even reverse. Therefore, to comprehend this dynamic a broader spectrum of characteristics. This allows for the
condition of TME, more research is required [27, 35, 36]. direct forecasting of target categories from radiology
The issue with TME as it exists today is sample bias in image data, enhancing flexibility and providing valuable
pathology or expensive, unilateral information-based insights relevant to immunotherapy. Additionally, AI
high-end approaches. Through cell quantification and can predict key biomarkers like TMB, MSI, and PD-L1
localisation, pathologists use histological research to expression from radiology images, capturing how
identify the TME. However, this approach is susceptible these markers change in response to treatment [32, 45].
to sample bias, whereas methods like single-cell genomics Although the performance of these predictions may not yet
and spatial transcriptomics are costly, time-consuming, be optimal, the ability to track molecular and phenotypic
and usually rely on information from a single source, changes during treatment is invaluable for personalized

Asian Pacific Journal of Cancer Care• Vol 9• Issue 4 795

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

cancer therapy [46, 47]. treatment [51]. Immunotherapy is typically administered

Computational Pathology: AI has also given rise to following two or three therapeutic modalities. Therefore,
the field of computational pathology, which improves there is a few-month lag between initial treatments or
the analysis of histopathology [30, 48]. Traditionally, regimens such as surgery, radiotherapy, and chemotherapy
pathologists examined hematoxylin and eosin (H&E) to immunotherapeutic treatment. The issue stems
slides to assess tumor characteristics. Although from data training, which is often carried out on initial
conventional histopathology allows us to look for various histopathology samples, even if the tumor niche may have
types of cells and get a basic understanding of their changed by the time immunotherapy is administered.
quantitative and spatial features but is invasive and can One possible course of action is to collect samples at a
be subject to sample bias. AI can quantify biomarkers like later stage as well and switch immunotherapy to earlier
PD-L1 expression from H&E slides and provide scoring treatment modalities.
that aids in identifying patients sensitive to immune In conclusion, immunotherapy has demonstrated
checkpoint inhibitors (ICIs) [32, 49, 50]. Furthermore, remarkable promise in treating cancer, particularly
computational pathology utilizes deep learning to count in patients who have undergone two or more lines of
immune cells, such as tumor-infiltrating lymphocytes conventional therapies and are at risk of recurrence.
(TILs). By abstracting complex visual patterns and Immune checkpoint inhibitors and adoptive cell therapies
analyzing cell morphology, deep learning can determine have the potential not only to address primary tumors but
tumor sensitivity to immunotherapy, assessing features also to target dispersed or disseminated malignant cells,
such as cell shape, phenotype, and spatial relationships leading to complete remission. However, challenges
from raw histopathology images [51, 52]. remain, including limited applicability to certain patient
The integration of these advanced AI techniques not populations and high costs. As the saying goes, “Every
only quantifies immunotherapy-related biomarkers but also problem has the seed of its own solution hidden within
enhances the prediction of therapeutic responses through it.” The vast data generated from routine procedures
scores like the immunoscore and immunophenoscore such as radiology and histopathology, combined with the
[53-55]. By effectively forecasting the probability of complexity of cancer, presents a significant opportunity
successful immunotherapy outcomes, AI is facilitating for training artificial intelligence models. These models
the development of more targeted and personalized can predict treatment responses and efficacy, allowing for
approaches to cancer treatment. more tailored therapies that benefit only those patients
most likely to respond. This approach could significantly
4. Limitations of AI and what needs to be done reduce the burden on healthcare systems and lower
There are a few limitations to the use of AI, among treatment costs.
them first one is related to trustworthiness. It has to do with This interdisciplinary collaboration has deepened our
the human mind’s capacity to comprehend its intricate understanding of the tumor microenvironment, an area
algorithms. The traditional or earlier ML models could that has remained elusive despite extensive research.
be explained, but the more recent Deep ANNs-based DL With advancements in immunomics and technologies
models are quite intricate and the presence of multilayered like next-generation sequencing (NGS), omics, single-cell
neural networks with a substantial number of hidden sequencing, and spatial transcriptomics, we are finally
layers for computation makes their algorithm hard to able to gain insights into this complex environment.
be explained by the human mind and have thus attained AI plays a crucial role in integrating diverse data sources
the moniker of “Black box models”. Another limitation through its multilayered deep neural networks, enabling
is related to the training of AI models such as DL-based the extraction of relevant patterns and information.
models. Other than requiring a large amount of data, By combining these AI capabilities with spatial
the quality of data should also be high. If the data is transcriptomics, we stand to enhance our understanding of
unclear, artefactual, or noisy, then it would be difficult to the cancer ecosystem, paving the way for the discovery of
get accurate results. To compensate for this even larger new therapeutic targets and improving predictions of
amount of data is needed to reach a decent result [56]. immunotherapy responses and overall treatment efficacy
Another issue is related to generalization and is quite in clinical settings.
prominent in a field where data varies a lot from location As medical, analytical, and research technologies
to location, between different healthcare centres and continue to advance alongside data collection methods
nations. The data should represent a real-world scenario and computer science, it is likely that AI and its deep
so as to provide a better generalized result. Another learning subset will fundamentally transform cancer
problem related is data bias. The data may be bias based diagnosis and treatment.
on ethnicity, age, and gender. For this large amount
of data is needed, so as to make the model bias-free Acknowledgments
and generalized. Further proper validation is needed
to utilize these models in the real-world scenario on a Statement of Transparency and Principals
routinely basis [48, 57]. A major conceptual limitation
of the AI model is based on histopathology and is related Funding Statement
to computational pathology. As we know the H and E No funding resources
slides of IHC samples are acquired at the initial stage of

796 Asian Pacific Journal of Cancer Care• Vol 9• Issue 4

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

Approval org/10.1158/1078-0432.CCR-22-0390
This is a review article so it does not require any 12. Wang DY, Salem J, Cohen JV, Chandra S, Menzer C, Ye F,
Approval. Zhao S, et al. Fatal Toxic Effects Associated With Immune
Checkpoint Inhibitors: A Systematic Review and Meta-
analysis. JAMA oncology. 2018 Dec 01;4(12):1721-1728.
Conflict of Interest
https://doi.org/10.1001/jamaoncol.2018.3923
None to declare 13. Research CFDE. FDA approves pembrolizumab for adults
and children with TMB-H solid tumors. FDA. 2024 08 09;.
Ethical Declaration 14. Planchard D, Popat S, Kerr K, Novello S, Smit EF, Faivre-
This is a review article so it does not require any Finn C, Mok TS, et al. Metastatic non-small cell lung
Approval. cancer: ESMO Clinical Practice Guidelines for diagnosis,
treatment and follow-up. Annals of Oncology: Official
Journal of the European Society for Medical Oncology. 2018
References Oct 01;29(Suppl 4):iv192-iv237. https://doi.org/10.1093/
1. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. annonc/mdy275
CA: a cancer journal for clinicians. 2024;74(1):12-49. 15. Lee JS, Ruppin E. Multiomics Prediction of Response
https://doi.org/10.3322/caac.21820 Rates to Therapies to Inhibit Programmed Cell Death 1
2. Farkona S, Diamandis EP, Blasutig IM. Cancer immunotherapy: and Programmed Cell Death 1 Ligand 1. JAMA oncology.
the beginning of the end of cancer?. BMC medicine. 2016 2019 Nov 01;5(11):1614-1618. https://doi.org/10.1001/
05 05;14:73. https://doi.org/10.1186/s12916-016-0623-5 jamaoncol.2019.2311
3. Looi C, Chung FF, Leong C, Wong S, Rosli R, Mai C. 16. Alowais SA, Alghamdi SS, Alsuhebany N, Alqahtani
Therapeutic challenges and current immunomodulatory T, Alshaya AI, Almohareb SN, Aldairem A, et al.
strategies in targeting the immunosuppressive pancreatic Revolutionizing healthcare: the role of artificial intelligence
tumor microenvironment. Journal of experimental & clinical in clinical practice. BMC Medical Education. 2023 09
cancer research: CR. 2019 04 15;38(1):162. https://doi. 22;23(1):689. https://doi.org/10.1186/s12909-023-04698-z
org/10.1186/s13046-019-1153-8 17. Elemento O, Leslie C, Lundin J, Tourassi G. Artificial
4. Fu Y, Liu S, Zeng S, Shen H. From bench to bed: the tumor intelligence in cancer research, diagnosis and therapy. Nature
immune microenvironment and current immunotherapeutic Reviews. Cancer. 2021 Dec;21(12):747-752. https://doi.
strategies for hepatocellular carcinoma. Journal of org/10.1038/s41568-021-00399-1
experimental & clinical cancer research: CR. 2019 09 18. Shao F, Shen Z. How can artificial neural networks
09;38(1):396. https://doi.org/10.1186/s13046-019-1396-4 approximate the brain?. Frontiers in Psychology. 2023 01
5. Havel JJ, Chowell D, Chan TA. The evolving landscape 09;13:970214. https://doi.org/10.3389/fpsyg.2022.970214
of biomarkers for checkpoint inhibitor immunotherapy. 19. Prelaj A, Boeri M, Robuschi A, Ferrara R, Proto C, Lo
Nature Reviews. Cancer. 2019 03;19(3):133-150. https:// Russo G, Galli G, et al. Machine Learning Using Real-
doi.org/10.1038/s41568-019-0116-x World and Translational Data to Improve Treatment
6. Mishra AK, Gupta A, Dagar G, Das D, Chakraborty A, Selection for NSCLC Patients Treated with Immunotherapy.
Haque S, Prasad CP, et al. CAR-T-Cell Therapy in Multiple Cancers. 2022 01 16;14(2):435. https://doi.org/10.3390/
Myeloma: B-Cell Maturation Antigen (BCMA) and Beyond. cancers14020435
Vaccines. 2023 Nov 16;11(11):1721. https://doi.org/10.3390/ 20. Goldenberg SL, Nir G, Salcudean SE. A new era: artificial
vaccines11111721 intelligence and machine learning in prostate cancer. Nature
7. Nishino M, Ramaiya NH, Hatabu H, Hodi FS. Monitoring Reviews. Urology. 2019 07;16(7):391-403. https://doi.
immune-checkpoint blockade: response evaluation and org/10.1038/s41585-019-0193-3
biomarker development. Nature Reviews. Clinical Oncology. 21. Perez-Lopez R, Ghaffari Laleh N, Mahmood F, Kather JN.
2017 Nov;14(11):655-668. https://doi.org/10.1038/ A guide to artificial intelligence for cancer researchers.
nrclinonc.2017.88 Nature Reviews. Cancer. 2024 06;24(6):427-441. https://
8. Uslu U, Castelli S, June CH. CAR T cell combination therapies doi.org/10.1038/s41568-024-00694-7
to treat cancer. Cancer Cell. 2024 08 12;42(8):1319-1325. 22. Anderson NM, Simon MC. The tumor microenvironment.
https://doi.org/10.1016/j.ccell.2024.07.002 Current biology: CB. 2020 08 17;30(16):R921-R925. https://
9. Dagar G, Gupta A, Masoodi T, Nisar S, Merhi M, Hashem doi.org/10.1016/j.cub.2020.06.081
S, Chauhan R, et al. Harnessing the potential of CAR-T 23. Visser KE, Joyce JA. The evolving tumor microenvironment:
cell therapy: progress, challenges, and future directions From cancer initiation to metastatic outgrowth. Cancer
in hematological and solid tumor treatments. Journal of Cell. 2023 03 13;41(3):374-403. https://doi.org/10.1016/j.
Translational Medicine. 2023 07 07;21(1):449. https://doi. ccell.2023.02.016
org/10.1186/s12967-023-04292-3 24. Joyce JA, Pollard JW. Microenvironmental regulation of
10. Utkarsh K, Srivastava N, Kumar S, Khan A, Dagar G, Kumar metastasis. Nature Reviews. Cancer. 2009 04;9(4):239-252.
M, Singh M, Haque S. CAR-T cell therapy: a game-changer https://doi.org/10.1038/nrc2618
in cancer treatment and beyond. Clinical & Translational 25. Lei X, Lei Y, Li J, Du W, Li R, Yang J, Li J, Li F, Tan
Oncology: Official Publication of the Federation of Spanish H. Immune cells within the tumor microenvironment:
Oncology Societies and of the National Cancer Institute of Biological functions and roles in cancer immunotherapy.
Mexico. 2024 06;26(6):1300-1318. https://doi.org/10.1007/ Cancer Letters. 2020 02 01;470:126-133. https://doi.
s12094-023-03368-2 org/10.1016/j.canlet.2019.11.009
11. Ghaffari Laleh N, Ligero M, Perez-Lopez R, Kather JN. Facts 26. Rangamuwa K, Aloe C, Christie M, Asselin-Labat M,
and Hopes on the Use of Artificial Intelligence for Predictive Batey D, Irving L, John T, et al. Methods for assessment
Immunotherapy Biomarkers in Cancer. Clinical Cancer of the tumour microenvironment and immune interactions
Research: An Official Journal of the American Association in non-small cell lung cancer. A narrative review. Frontiers
for Cancer Research. 2023 01 17;29(2):316-323. https://doi. in Oncology. 2023;13:1129195. https://doi.org/10.3389/

Asian Pacific Journal of Cancer Care• Vol 9• Issue 4 797

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

fonc.2023.1129195 C, et al. A Deep Learning Approach for Histology-Based

27. Walsh LA, Quail DF. Decoding the tumor microenvironment Nucleus Segmentation and Tumor Microenvironment
with spatial technologies. Nature Immunology. 2023 Characterization. Modern Pathology: An Official Journal
Dec;24(12):1982-1993. https://doi.org/10.1038/s41590- of the United States and Canadian Academy of Pathology,
023-01678-9 Inc. 2023 08;36(8):100196. https://doi.org/10.1016/j.
28. Peng H, Wu X, Liu S, He M, Xie C, Zhong R, Liu J, et al. modpat.2023.100196
Multiplex immunofluorescence and single-cell transcriptomic 41. Khader F, Müller-Franzes G, Wang T, Han T, Tayebi Arasteh
profiling reveal the spatial cell interaction networks in the S, Haarburger C, Stegmaier J, et al. Multimodal Deep
non-small cell lung cancer microenvironment. Clinical and Learning for Integrating Chest Radiographs and Clinical
Translational Medicine. 2023 01;13(1):e1155. https://doi. Parameters: A Case for Transformers. Radiology. 2023
org/10.1002/ctm2.1155 Oct;309(1):e230806. https://doi.org/10.1148/radiol.230806
29. Xie T, Huang A, Yan H, Ju X, Xiang L, Yuan J. Artificial 42. Niehues JM, Quirke P, West NP, Grabsch HI, Treeck M,
intelligence: illuminating the depths of the tumor Schirris Y, Veldhuizen GP, et al. Generalizable biomarker
microenvironment. Journal of Translational Medicine. prediction from cancer pathology slides with self-supervised
2024 08 29;22(1):799. https://doi.org/10.1186/s12967- deep learning: A retrospective multi-centric study. Cell
024-05609-6 Reports. Medicine. 2023 04 18;4(4):100980. https://doi.
30. Echle A, Rindtorff NT, Brinker TJ, Luedde T, Pearson org/10.1016/j.xcrm.2023.100980
AT, Kather JN. Deep learning in cancer pathology: a new 43. Chen RJ, Lu MY, Williamson DFK, Chen TY, Lipkova J,
generation of clinical biomarkers. British Journal of Cancer. Noor Z, Shaban M, et al. Pan-cancer integrative histology-
2021 02;124(4):686-696. https://doi.org/10.1038/s41416- genomic analysis via multimodal deep learning. Cancer
020-01122-x Cell. 2022 08 08;40(8):865-878.e6. https://doi.org/10.1016/j.
31. Laak J, Litjens G, Ciompi F. Deep learning in histopathology: ccell.2022.07.004
the path to the clinic. Nature Medicine. 2021 05;27(5):775- 44. Hoang D, Dinstag G, Shulman ED, Hermida LC, Ben-Zvi
784. https://doi.org/10.1038/s41591-021-01343-4 DS, Elis E, Caley K, et al. A deep-learning framework to
32. Hondelink LM, Hüyük M, Postmus PE, Smit VTHBM, Blom predict cancer treatment response from histopathology
S, Thüsen JH, Cohen D. Development and validation of a images through imputed transcriptomics. Nature Cancer.
supervised deep learning algorithm for automated whole- 2024 09;5(9):1305-1317. https://doi.org/10.1038/s43018-
slide programmed death-ligand 1 tumour proportion score 024-00793-2
assessment in non-small cell lung cancer. Histopathology. 45. Mu W, Jiang L, Shi Y, Tunali I, Gray JE, Katsoulakis E, Tian
2022 03;80(4):635-647. https://doi.org/10.1111/his.14571 J, Gillies RJ, Schabath MB. Non-invasive measurement of
33. Jain MS, Massoud TF. Predicting tumour mutational PD-L1 status and prediction of immunotherapy response
burden from histopathological images using multiscale using deep learning of PET/CT images. Journal for
deep learning. Nat Mach Intell. 2020 Jun;2(6):356–62.. Immunotherapy of Cancer. 2021 06;9(6):e002118. https://
2020;2(3):356-62. doi.org/10.1136/jitc-2020-002118
34. Sha L, Osinski BL, Ho IY, Tan TL, Willis C, Weiss H, 46. Sverzellati N, Marrocchio C. Deep learning-based radiomics:
Beaubier N, et al. Multi-Field-of-View Deep Learning pacing immunotherapy in lung cancer. The Lancet. Digital
Model Predicts Nonsmall Cell Lung Cancer Programmed Health. 2023 07;5(7):e396-e397. https://doi.org/10.1016/
Death-Ligand 1 Status from Whole-Slide Hematoxylin and S2589-7500(23)00086-9
Eosin Images. Journal of Pathology Informatics. 2019;10:24. 47. Liao C, Chen Y, Wu Y, Chao H, Chiu H, Wang T, Chen J,
https://doi.org/10.4103/jpi.jpi_24_19 Shiao T, Lu C. Personalized prediction of immunotherapy
35. Liu W, Puri A, Fu D, Chen L, Wang C, Kellis M, Yang response in lung cancer patients using advanced radiomics
J. Dissecting the tumor microenvironment in response to and deep learning. Cancer Imaging: The Official Publication
immune checkpoint inhibitors via single-cell and spatial of the International Cancer Imaging Society. 2024 09
transcriptomics. Clinical & Experimental Metastasis. 2024 30;24(1):129. https://doi.org/10.1186/s40644-024-00779-4
08;41(4):313-332. https://doi.org/10.1007/s10585-023- 48. Kleppe A, Skrede O, De Raedt S, Liestøl K, Kerr DJ,
10246-2 Danielsen HE. Designing deep learning studies in cancer
36. Laury AR, Zheng S, Aho N, Fallegger R, Hänninen S, diagnostics. Nature Reviews. Cancer. 2021 03;21(3):199-
Saez-Rodriguez J, Tanevski J, et al. Opening the Black 211. https://doi.org/10.1038/s41568-020-00327-9
Box: Spatial Transcriptomics and the Relevance of Artificial 49. Kapil A, Meier A, Zuraw A, Steele KE, Rebelatto MC,
Intelligence-Detected Prognostic Regions in High-Grade Schmidt G, Brieu N. Deep Semi Supervised Generative
Serous Carcinoma. Modern Pathology: An Official Journal Learning for Automated Tumor Proportion Scoring on
of the United States and Canadian Academy of Pathology, NSCLC Tissue Needle Biopsies. Scientific Reports. 2018
Inc. 2024 07;37(7):100508. https://doi.org/10.1016/j. Nov 26;8(1):17343. https://doi.org/10.1038/s41598-018-
modpat.2024.100508 35501-5
37. Papalexi E, Satija R. Single-cell RNA sequencing to explore 50. Liu J, Zheng Q, Mu X, Zuo Y, Xu B, Jin Y, Wang Y, et
immune cell heterogeneity. Nature Reviews. Immunology. al. Automated tumor proportion score analysis for PD-
2018 01;18(1):35-45. https://doi.org/10.1038/nri.2017.76 L1 (22C3) expression in lung squamous cell carcinoma.
38. Adan A, Alizada G, Kiraz Y, Baran Y, Nalbant A. Flow Scientific Reports. 2021 08 05;11(1):15907. https://doi.
cytometry: basic principles and applications. Critical org/10.1038/s41598-021-95372-1
Reviews in Biotechnology. 2017 03;37(2):163-176. https:// 51. Cifci D, Foersch S, Kather JN. Artificial intelligence to
doi.org/10.3109/07388551.2015.1128876 identify genetic alterations in conventional histopathology.
39. Shi L, Zhang Y, Wang H. Prognostic prediction based on The Journal of Pathology. 2022 07;257(4):430-444. https://
histopathologic features of tumor microenvironment in doi.org/10.1002/path.5898
colorectal cancer. Frontiers in Medicine. 2023;10:1154077. 52. Woerl A, Eckstein M, Geiger J, Wagner DC, Daher T,
https://doi.org/10.3389/fmed.2023.1154077 Stenzel P, Fernandez A, et al. Deep Learning Predicts
40. Rong R, Sheng H, Jin KW, Wu F, Luo D, Wen Z, Tang Molecular Subtype of Muscle-invasive Bladder Cancer
from Conventional Histopathological Slides. European

798 Asian Pacific Journal of Cancer Care• Vol 9• Issue 4

 apjcc.waocp.com Mohd Umar Rehmani, et al: Dissecting the Tumor Microenvironment (TME) to Decipher New Immunotherapy Targets

Urology. 2020 08;78(2):256-264. https://doi.org/10.1016/j.

eururo.2020.04.023
53. Hu J, Cui C, Yang W, Huang L, Yu R, Liu S, Kong Y. Using
deep learning to predict anti-PD-1 response in melanoma
and lung cancer patients from histopathology images.
Translational Oncology. 2021 01;14(1):100921. https://doi.
org/10.1016/j.tranon.2020.100921
54. Xie C, Vanderbilt C, Feng C, Ho D, Campanella G, Egger J,
Plodkowski A, et al. Computational biomarker predicts lung
ICI response via deep learning-driven hierarchical spatial
modelling from H&E .
55. Angell H, Galon J. From the immune contexture to the
Immunoscore: the role of prognostic and predictive immune
markers in cancer. Current Opinion in Immunology. 2013
04;25(2):261-267. https://doi.org/10.1016/j.coi.2013.03.004
56. Schömig-Markiefka B, Pryalukhin A, Hulla W, Bychkov A,
Fukuoka J, Madabhushi A, Achter V, et al. Quality control
stress test for deep learning-based diagnostic model in
digital pathology. Modern Pathology: An Official Journal
of the United States and Canadian Academy of Pathology,
Inc. 2021 Dec;34(12):2098-2108. https://doi.org/10.1038/
s41379-021-00859-x
57. Obermeyer Z, Powers B, Vogeli C, Mullainathan S.
Dissecting racial bias in an algorithm used to manage the
health of populations. Science (New York, N.Y.). 2019 Oct
25;366(6464):447-453. https://doi.org/10.1126/science.
aax2342

This work is licensed under a Creative Commons Attribution-

Non Commercial 4.0 International License.

Asian Pacific Journal of Cancer Care• Vol 9• Issue 4 799