Применение методов машинного обучения к данным геномики, транскриптомики и биомедицинской визуализации для решения медицинских задач /Application of Machine Learning to Genomic, Transcriptomic, and Imaging Data in Medical Problems тема диссертации и автореферата по ВАК РФ 00.00.00, кандидат наук Сарачаков Александр Евгеньевич

• Сарачаков Александр Евгеньевич
• кандидат наук
• 2026

Оглавление диссертации кандидат наук Сарачаков Александр Евгеньевич

Table of Contents

Introduction

Chapter 1. Literature review

1.1 Infrared imaging

1.2 Multiplexed immunofluorescence (MxIF)

1.2.1 Analytical Workflows of MxIF data

1.2.2 Machine Learning Methods for MxIF cell segmentation

1.2.3 Single-Cell Phenotyping and Clustering

1.2.4 Biological and Clinical Insights from MxIF

1.3 Integrating Imaging with Multi-Omics Approaches

1.3.1 Challenges in Data Integration

1.3.2 Computational Techniques for Multi-Modal Data analysis

1.3.3 Multi-modal integration for biomedical data analysis

1.3.4 Integrating Histopathology with Multi-Omics Data in Oncology

1.4 Applications of Deep Learning in Normal and Cancer Tissues

1.4.1 Lymphoid Organs and Tissue Architecture

1.4.2 Follicular Lymphoma Tumor Microenvironment and Architecture

1.4.3 Biomarker Discovery via Imaging and Multi-Omics Integration

1.4.4 High-Resolution Imaging of Novel Cellular Transformations

1.5 Clinical Trials and Translational Relevance

1.5.1 ML for Drug Response and Combination Therapy

1.5.2 Challenges in Deployment

1.5.3 Future Outlook

Chapter 2. Methods

2.1 Hardware Setup

2.2 Data Acquisition and Preprocessing

2.3 Neural Network Architecture and Training

2.4 Adaptive Denoising Agent

2.5 Alignment Using Reinforcement Learning

2.6 Evaluation Metrics

2.7 Tissue samples collection

2.8 Tissue Collection and Preparation

2.9 Multiplexed Immunofluorescence (MxIF) Staining and Imaging

2.10 Image Processing, Cell Segmentation, and Phenotyping

2.11 Single-Cell RNA Sequencing

2.12 Bulk RNA Sequencing

2.13 Genomic DNA Sequencing and Clonal Analysis

2.14 Spatial Analysis of Tissue Architecture

2.15 Quantification and Statistical Analysis

Chapter 3. Near-infrared light for enhanced vein visualization

3.1 Study 1: NIR-to-Visible Vein Imaging with CNN and RL

3.2 Study 2: Adaptive Denoising and Alignment Agents

Chapter 4. Spatial Mapping of Human Hematopoiesis in Bone Marrow

4.1 Bone Marrow Cellular Composition Across Ages

4.2 Marrow Adiposity Constrains Hematopoietic Output

4.3 Vascular architecture: preserved density, remodeled microvasculature

4.4 Age-Related Morphological Changes in Hematopoietic Cells

4.5 Spatial Localization of HSPCs and Other Cells: Conserved Niches in Humans

4.6 Evidence of a Megakaryocytic Niche and Its Attenuation with Age

4.7 Identification of Distinct Marrow Cell Communities

4.8 Integrated model of aging-associated topographic remodeling

Chapter 5. Multi-omic Profiling of Follicular Lymphoma Tumor Architecture

5.1 Cohort and Multi-Modal Atlas Construction

5.2 Genomic Alterations and Cellular Composition of FL

5.3 High-Plex Spatial Profiling of the Lymph Node Microenvironment

5.4 BCR Signaling and TME Interactions in High-Risk Patients

5.5 Distinct Histological Patterns Associated with Early Relapse

5.6 Stromal Remodeling and Fibrosis in the Tumor Microenvironment

5.7 Translational impact

Chapter 6. Duvelisib and Docetaxel in Anti-PD-1 Refractory Recurrent/Metastatic HNSCC

6.1 Clinical Efficacy and Survival Outcomes

6.2 Safety Profile and Tolerability

6.3 Immune and Genomic Correlates of Outcome

Chapter 7. A Novel Case of Glial Transdifferentiation in Renal Medullary Carcinoma Brain Metastasis

7.1 Histopathology and immunohistochemistry: evidence of glial mimicry

7.2 Molecular and epigenomic analyses: clonal origin and CNS-like shift

7.3 Spatial phenotyping by CODEX and tumor microenvironment remodeling

Conclusions

List of Symbols, Abbreviations

Bibliography

Введение диссертации (часть автореферата) на тему «Применение методов машинного обучения к данным геномики, транскриптомики и биомедицинской визуализации для решения медицинских задач /Application of Machine Learning to Genomic, Transcriptomic, and Imaging Data in Medical Problems»

Аннотация

В данной диссертации представлены вычислительные подходы, объединяющие современные методы биомедицинской визуализации, мультиомиксные данные и алгоритмы машинного обучения, для решения актуальных клинических проблем. В работе рассмотрены алгоритмы визуализации сосудов в ближнем инфракрасном диапазоне, которые повышают точность обнаружения и проекции подкожных вен при рутинных венозных доступах. Благодаря сверточным нейронным сетям и обучению с подкреплением эти системы адаптивно подавляют шум на изображениях в инфракрасном спектре и точно проецируют контуры вен на руку пациента, обеспечивая удобное отображение сосудистой сети в реальном времени.

Мультиплексная иммунофлуоресценция, позволяющая визуализировать экспрессии белков на клеточном разрешении, была использована для анализа нормальных и патологических тканей. Применение машинного обучения для сегментации, кластеризации и пространственного анализа позволило охарактеризовать архитектонику здорового костного мозга и выявило в нем изменения гемопоэтических ниш с возрастом. Эти алгоритмы были применены к анализу тканей фолликулярной лимфомы, совместно с интеграцией мультиплексных изображений и омиксных данных они выявили ключевые изменения микроокружения опухоли. В частности, у пациентов из группы высокого риска, была обнаружена специфическая пространственная архитектура фолликулов и ремоделирование стромы, что демонстрирует прогностический потенциал анализа изображений и транскриптомных профилей.

Кроме того, в данной диссертации описан редкий случай глиальной трансдифференциации в метастазе медуллярной карциномы почки, подчеркивающий ценность глубокой визуализации и молекулярного анализа при распознавании смешанных опухолей. Разработанные методики показали клиническую значимость в

исследовании комбинированной терапии дувелисибом и доцетакселом при рецидивирующей или метастатической карциноме головы и шеи, где анализ данных методами машинного обучения обеспечил более детальное понимание причин ответа пациентов на терапию и обнаружил потенциальные биомаркеры благоприятного прогноза.

В совокупности представленная работа демонстрирует, как надёжные алгоритмы, от обучения с подкреплением для инфракрасной визуализации до анализа тканевых образцов на клеточном уровне, развивают фундаментальные исследования и трансляционную медицину, открывая путь к более точной диагностике, эффективным методам лечения и глубокому пониманию биологии заболеваний.

Abstract

This dissertation focuses on computational frameworks to integrate advanced imaging, multi-omics data, and machine learning for addressing key challenges in modern biomedical research and clinical practice. The work begins by introducing near-infrared (NIR) imaging pipelines designed to improve both detection and projection of subcutaneous veins for routine venous access procedures. Through the use of convolutional neural networks and reinforcement learning, these pipelines adaptively denoise raw infrared data and align projected visible overlays, enabling more accurate and user-friendly visualization of vasculature in real time. This not only enhances clinical efficiency but also serves as a blueprint for generalizable computer vision approaches in medical imaging.

Next, the focus shifts to multiplexed immunofluorescence (MxIF), a high-dimensional imaging modality that captures spatial protein expression profiles at single-cell resolution. By leveraging machine learning-driven segmentation, clustering, and spatial analysis, MxIF reveals fundamental tissue architecture in both normal and diseased conditions. In particular, single-cell mapping of bone marrow highlights how hematopoietic niches evolve with aging, showcasing the utility of MxIF for dissecting complex biological processes. These workflows are further extended to cancer contexts, where multi-omic integration of imaging and genomic data uncovers critical alterations in tumor microenvironments. A study in follicular lymphoma identifies distinct tissue architectures and stromal remodeling unique to high-risk patient groups, illustrating the promise of combining imaging with transcriptomic profiles for enhanced prognostication.

Beyond hematologic malignancies, the dissertation also presents an unusual instance of glial transdifferentiation in a brain metastasis of renal medullary carcinoma, underscoring the value of deep imaging and molecular analyses for identifying rare or mixed-phenotype tumors. Finally, the methods developed herein are shown to inform clinical research, exemplified by a trial evaluating combined duvelisib and docetaxel

therapy in recurrent or metastatic head and neck carcinoma. Machine learning-based data evaluation enables more nuanced insights into patient responses and potential biomarkers.

The current work demonstrates how robust machine learning pipelines, ranging from reinforcement learning for infrared imaging to single-cell resolution analyses of tissue specimens, can advance both basic biological understanding and translational medicine. By uniting infrared imaging, MxIF techniques, and multi-omics data integration, these studies pave the way for more precise diagnostics, improved therapeutic interventions, and a deeper understanding of disease biology.

Заключение диссертации по теме «Другие cпециальности», Сарачаков Александр Евгеньевич

197 Conclusions

This dissertation demonstrates how uniting modern machine learning with multi-modal biomedical evidence can deliver concrete, clinically relevant insight - from photons to phenotypes to patients. It advances along a deliberately translational arc: beginning with the design of an intelligent near-infrared vein-visualization system that couples deep-learning segmentation with reinforcement-learning alignment, improving the reliability of a real-time medical imaging device.

It then establishes, at single-cell resolution, a spatial atlas of healthy human bone marrow across the adult lifespan, showing that canonical niches are preserved even as adiposity-linked constraints erode specific progenitor pools, thereby clarifying how tissue topology and composition shape hematopoietic output with age.

Building on those foundations, the work integrates high-plex tissue imaging with single-cell and bulk genomics to reveal that the architecture of the lymphoma microenvironment - fragmented follicles, stromal remodeling, and aligned transcriptional programs such as heightened B-cell receptor signaling, can distinguish indolent from high-risk follicular lymphoma and serve as early markers of aggressive disease.

In a prospective clinical setting, the dissertation evaluates the combination of the PI3K5/y inhibitor duvelisib with docetaxel for immunotherapy-refractory HNSCC, achieving meaningful activity (-19% objective responses and ~65% disease control) while using computational analyses to begin separating responder from non-responder biology, an illustration of how data-driven methods can interpret, and eventually guide, treatment decisions.

Finally, by resolving an enigmatic brain lesion in a patient with renal medullary carcinoma as a metastasis that underwent glial-like transdifferentiation, the thesis documents an extreme of tumor plasticity with multi-omic and histopathologic corroboration, setting a precedent for integrated diagnostics in atypical presentations.

Taken together, these studies supply both results and a reproducible strategy: rigorously link imaging, spatial context, and molecular state; validate with orthogonal modalities; and interpret findings in their clinical frame. The through-line is a practical vision of integrative precision medicine, in which heterogeneous patient data - slides, scans, sequences, and clinical covariates, are synthesized by robust models to support individualized decisions. The chapters collectively move this vision from concept to practice: a smarter visualization device ready for controlled clinical evaluation; a marrow reference map to benchmark pathology; risk-stratifying tissue features in lymphoma; a trial framework that couples therapeutics with computational correlative science; and an evidence-based workflow for rare, phenotypically deceptive tumors.

While earlier sections have detailed limitations and next steps, the broader implication is clear: when engineering, biology, and data science are woven together end-to-end, they yield insights neither discipline could achieve alone, and those insights are already close to the bedside. The work here provides stepping stones: methodological, conceptual, and translational, toward clinical systems that learn from multi-scale data to benefit individual patients. It closes with a simple proposition for the field: continue building from sensing to understanding to action, and the promise of algorithm-assisted discovery will translate into measurably better care.

Notwithstanding the successes and insights of these studies, it is important to acknowledge their limitations. These limitations temper the interpretation of results and point to areas where caution is needed or where further work is required.

A general limitation across all projects is scale. Many of the studies were necessarily limited in sample size - either by design (as in the single-case analysis) or by practical constraints (patients available for the trial or rare tissue samples for FL). For instance, while the bone marrow atlas included a respectable number of donors, it is still a relatively small cross-section of the human population; environmental factors, ethnic genetic diversity, or subtle health differences among donors could influence marrow composition, but our study may not have captured all such variability. Similarly, the FL study, though

comprehensive in data per patient, included only a subset of patients from a single institution's cohort - it may not represent the full heterogeneity of FL worldwide. The trial in HNSCC was a single-arm study with 26 patients; with such a cohort, statistical power to detect modest biomarker effects is low, and any correlations found must be deemed exploratory. The RMC case, being N=1, inherently cannot distinguish which findings are unique to this patient versus which are generalizable features of RMC biology under extreme conditions. Thus, one overarching limitation is that some conclusions (especially those of correlation and association) would benefit from validation in larger, independent cohorts. This is particularly relevant for the prognostic markers found in FL and the predictive markers suggested in HNSCC; these need testing in independent patient sets or prospective trials to ensure they hold true broadly and are not artifacts of our limited sampling.

Another limitation is the retrospective or observational nature of most of the studies. The bone marrow study was observational - we can note that fat correlates with fewer progenitors, but we cannot from that data alone prove causation (although mouse studies support it). The FL analysis was retrospective with outcomes already known, raising the possibility of selection bias (perhaps the most extreme cases were studied, exaggerating differences). The HNSCC trial lacked a control arm, so while outcomes seem better than historical controls, we cannot be absolutely certain how much benefit duvelisib added versus patient selection or other factors. For the RMC case, we are interpreting changes post-therapy without a pre-therapy brain sample; while the timeline and multi-modal evidence strongly support our interpretation, we did not experimentally manipulate the tumor (other than the therapy it received as part of care), so we rely on circumstantial evidence to infer causes (e.g. attributing H3K27me3 loss to the drug and phenotype changes to environment). These limitations mean that causality in many of our interpretations is inferred but not formally proven. Future functional experiments (like manipulating fat content in ex vivo marrow cultures, or testing BCR signaling inhibitors in

high-risk FL cells) would be needed to confirm mechanistic underpinnings of some correlations we observed.

Each individual study also has specific technical limitations. In the vein visualization project, one limitation was that the device was tested in a controlled environment on a limited set of volunteers; we did not conduct a full clinical trial to measure improvement in IV success rates. Therefore, while we improved technical metrics (image contrast, alignment error), the actual clinical impact in practice remains to be validated. There is also a limitation that our device might still struggle in certain scenarios (for example, extremely deep veins or very rapid movement might pose challenges beyond the range of our correction). So, further development and testing is needed before clinical adoption.

For the bone marrow atlas, a limitation was the resolution of cell phenotyping. We used a panel of markers that allowed broad classification (HSPC, progenitor, various mature cells), but within, say, the myeloid lineage or lymphoid lineage, there are finer subdivisions we did not distinguish (e.g. specific HSC subsets, or T-cell subtypes). The imaging could detect some but not all distinctions, and some overlap in markers means there is a degree of imperfection in classification (for example, some rare cell types might have been misclassified). Additionally, our age grouping was cross-sectional; ideally, one would want longitudinal samples to truly track how an individual's marrow changes with age, but that is hardly feasible in humans. There might also be unmeasured confounders -for example, undiagnosed health issues in older donors that could affect their marrow (we assumed all were "healthy"). We attempted to minimize such effects, but they remain considerations.

In the FL study, one limitation was the breadth of assays and potential batch effects. We performed WES, bulk RNA-seq, scRNA-seq, and multiplex imaging - each of these has its own technical noise and all were done on limited tissue from the same biopsy. In some cases, not all assays yielded high-quality data for every patient (some had too few cells for scRNA or variable tissue quality for imaging). We mitigated this with careful QC and computational batch correction where possible, but the possibility of technical artifacts

influencing results is real. For instance, differences in sample processing could potentially affect the detection of certain cell types in imaging. Also, while we identified intriguing patterns (like enriched BCR signaling or collagen deposition), the study wasn't set up to determine if those are simply correlated with aggressive disease or actively driving it. That distinction matters if one were to target these features therapeutically. Another limitation is that our analyses were partly retrospective fits to outcome; a prospective validation is needed to confirm that, say, our image-based risk clusters truly predict relapse in an unbiased fashion.

Regarding the HNSCC trial, beyond sample size and lack of control, another limitation is that the biomarker analysis was constrained by the tissue available and the timing of biopsies. We had pre-treatment biopsies on most patients, but very few on-treatment or post-progression biopsies. It would have been illuminating to see how the tumor microenvironment changed under duvelisib (did T-cells increase, etc.), but we lacked paired samples for many patients. Also, due to clinical considerations, some assays were done on subsets of patients (e.g., not all patients had fresh tissue for multiplex imaging, so some immune analyses were done by flow cytometry or IHC on fewer cases). This patchwork of data types can introduce bias - for example, maybe only patients with accessible biopsy sites got the full analysis, and they might have different outcomes. Furthermore, any suggestion of biomarker (like HPV status influence) is based on subgroup trends and needs cautious interpretation. Toxicity management also posed a limitation - with more patients or a refined protocol, we might have managed duvelisib dosing better to reduce discontinuations; our limited experience meant one responding patient stopped therapy early, which although insightful scientifically, is a suboptimal patient outcome.

In the RMC case, limitations include the reliance on archival tissue for some comparisons (e.g. we compared the brain metastasis to a prior liver metastasis biopsy -those were different sites and timepoints). The ideal scenario, which is almost impossible, would have been to have a biopsy of the brain lesion before it underwent changes, or serial

biopsies to watch the progression of phenotype. We had to reconstruct the timeline from one snapshot in time. Also, while we did extensive analyses, in a single case we could not perform every possible experiment - for example, we did not do a full proteomic analysis or metabolomic analysis of the tumor, which could have added further insights into how its metabolism or protein expression shifted. We did however cover the key genomic and phenotypic angles. Another limitation is that any therapeutic insights from one case are speculative; we suggested, for instance, that the tumor became ATRT-like, but that does not automatically mean ATRT therapies would work - that was not tested. Thus any clinical actionability from this finding (beyond diagnosis) remains to be determined.

Finally, a cross-cutting limitation of our approach is the complexity of data integration and potential for overfitting or spurious correlations. When one mines large datasets with many features (especially in small cohorts), one can always find patterns, but not all will be true associations - some may occur by chance. We mitigated this with knowledge-driven filtering (focusing on biologically plausible patterns and cross-validating findings with orthogonal data), but the risk remains. For example, in FL, we found a correlation between certain cytokines and fibroblast abundance; while intriguing, it's possible that's coincidental in our cohort and not a universal rule. The use of proper statistical corrections and validation in independent samples is needed to separate the truly robust patterns from incidental ones.

The work presented in this dissertation opens up numerous avenues for future research, building on both the specific findings and the methodologies developed. One broad future direction is scaling and generalizing the integrative approach demonstrated here. For example, in follicular lymphoma, a next step would be to validate and refine the prognostic markers identified. This could involve a multi-center study where biopsies from a larger FL patient cohort are analyzed with the same imaging and sequencing pipeline to see if the "fingerprints" of high risk (like fused follicles or high FRC content) hold true. Such validation would pave the way for incorporating these markers into clinical practice -perhaps developing an automated image analysis tool that can examine a routine biopsy

and output a risk score. Additionally, future studies could explore interventions suggested by our findings: in FL, testing whether targeting the microenvironment (e.g. with anti-fibrosis drugs or immune modulators) in patients with those high-risk features could improve outcomes. It would also be insightful to extend our multi-omic approach to related lymphomas or transformed FL cases to see how general the principles are.

For the bone marrow atlas, a compelling future direction is to apply similar spatial analytic techniques to diseased marrow. One could study how conditions like myelodysplastic syndromes (MDS) or aplastic anemia deviate from the "normal aging" baseline we established. The methods we developed for cell segmentation and niche identification can be directly applied to patient samples with those disorders to pinpoint exactly what goes wrong (for instance, does MDS exaggerate the fat-driven suppression of hematopoiesis? Do leukemic blasts disrupt the niches we mapped?). Another direction is to incorporate spatial transcriptomics or single-cell RNA-seq on the same tissues to add a molecular layer to the atlas. In our study, we inferred functionality (like aging bias) by proxies; having gene expression data for spatially located cells could validate, for example, that aged HSPCs in situ show senescence markers or that adipocytes produce certain inhibitory factors. Moreover, longitudinal studies in model organisms, informed by our human data, could test hypotheses like "reducing marrow fat in aged mice improves progenitor frequency," bridging our correlational findings to causation. On a technical front, improving the resolution of cell type identification (with more markers or advanced segmentation algorithms, potentially using newer deep learning models) could allow future atlases to distinguish even more subtle cell states and track even finer changes with age or disease.

For the vein visualization device, future work is clearly oriented towards translation and optimization. This means conducting clinical trials where nurses use the device in real IV insertion scenarios to formally measure improvement in success rates, time to cannulation, and patient satisfaction compared to the standard practice. The feedback from such trials will guide any needed modifications to the device or algorithm (for instance,

making the system more ergonomic, or tuning the display for various lighting conditions in hospitals). Integration with hospital workflows is another step - for example, ensuring the device is user-friendly for a non-engineer and can calibrate itself easily. On the algorithmic side, one might explore using even more advanced computer vision techniques, such as training the CNN on a much larger dataset (now that we have a working prototype, one could accumulate more training images from different skin tones and physiological conditions to improve robustness). One could also investigate 3D vein mapping by adding depth sensors, and use ML to reconstruct sub-surface vessel pathways - useful for more complex scenarios like identifying the best vein trajectory for long catheter insertions. Another future direction is extending the concept to artery or organ visualization during surgery (imagine a similar AR projection but for highlighting blood vessels or tumors during an operation, using ML to guide surgeons). The success of our vein project is a springboard to many such "AI-augmented reality" tools in medicine.

In the context of the HNSCC trial, a critical next step is to verify the efficacy of the duvelisib-docetaxel combination in a larger patient population, ideally through a randomized controlled trial (RCT). An RCT could compare docetaxel alone vs. docetaxel+duvelisib in PD-1-refractory HNSCC to quantify the added benefit in a statistically rigorous way. Within such a trial, the biomarkers we identified (HPV status, certain mutations, immune infiltration) can be pre-specified for validation - checking, for example, if HPV-negative patients indeed have a significantly higher response rate to the combo than HPV-positive patients. If validated, that could lead to a stratified treatment approach where patients are selected for PI3K-inhibitor therapy based on their tumor's molecular profile. Additionally, exploring combinations to mitigate toxicity or enhance efficacy is another future direction. For instance, perhaps using an alternate PI3K inhibitor with a different toxicity profile, or adding a PD-1 inhibitor back into the mix (once tumors are re-sensitized by duvelisib) could be tried. Another interesting route is to further analyze the rich data from this trial - for example, performing single-cell sequencing on pre- and post-treatment biopsies (should they be available in future studies) to see how the tumor

ecosystem changes under treatment. That can yield insights into mechanisms of resistance that develop, guiding next-line therapies. On the informatics side, one could use the trial data to train predictive models: given a patient's baseline multi-omic data, can we predict if they'll respond to the combo? This might involve machine learning models that incorporate variables like gene mutations, gene expression, and immunofluorescence metrics. With enough data, this could become a clinical decision support tool to identify ideal candidates for PI3K-based therapy.

For the RMC glial transdifferentiation case, future research could extend its findings in several ways. At the clinical level, it would be valuable to compile a registry of similar cases (if any) from multiple institutions - perhaps other RMC patients who had unusual metastases or other SMARCB1-deficient tumors that showed unexpected behavior. By comparing notes on even a handful of such cases, we could determine if the phenomenon we observed is extremely rare or if, for instance, prolonged EZH2 inhibitor therapy in SMARCB1-null cancers sometimes leads to that ATRT-like epigenetic shift. Another direct offshoot would be laboratory modeling: establishing cell lines or patient-derived organoids/xenografts from the patient's tumor (if possible) to experimentally probe its biology. One could test, for example, whether the brain metastasis cells respond to therapies used for ATRT or glioblastoma (like certain chemotherapy regimens or radiation sensitizers) differently than typical RMC cells. This might open therapeutic insights -perhaps a combination of kidney cancer therapy and brain tumor therapy could be relevant for such transdifferentiated metastases. Additionally, lab studies could explore the mechanism of transdifferentiation: using CRISPR or drug treatments on RMC cell models to see if inducing certain pathways (MYC overexpression, EZH2 inhibition, co-culture with neural cells) can trigger glial marker expression or other lineage changes. This would deepen our understanding of tumor plasticity - knowledge that could apply to other cancers with lineage switching (for instance, small cell lung cancer emerging from treated adenocarcinoma, etc.). On the technical side, the methods we used for the RMC case can be turned into a general pipeline for investigating any diagnostically challenging tumor.

The future could see pathologists routinely employing integrated genomic and epigenomic analysis for unusual cancers to avoid misdiagnosis and to guide personalized treatment; our work serves as a blueprint for that. One could imagine creating a machine learning classifier that, when given combined data (mutations, methylation, IHC results), can output the likelihood that a tumor is a metastasis with altered phenotype versus a new primary, aiding pathology in real time.

Beyond project-specific directions, there are overarching future directions stemming from this dissertation's theme. One is the further development of multi-modal machine learning models. We largely performed serial or parallel analyses of different data types and then combined interpretations. A future goal would be to train unified models that take all data types as input and directly predict outcomes or diagnoses. This would involve advanced deep learning architectures that can handle image data (pathology slides, radiology) alongside sequence data and even clinical features, learning complex patterns. With larger datasets, this becomes feasible and could yield highly accurate and holistic predictive tools. The results from our studies would serve as features or intermediate labels to guide such models (for example, a deep model for FL might implicitly learn to detect follicle shapes and fibrotic areas in images as well as BCR gene signatures in genomics when predicting relapse - essentially automating what we did manually).

Another direction is fostering better data integration and sharing. The richness of the datasets we generated (especially multi-omic ones) makes them valuable for the research community. A future step is to share these datasets (with appropriate de-identification) on public repositories and to develop interactive platforms where researchers or clinicians can query them. For instance, an interactive atlas for bone marrow where users can explore cell distributions by age, or a public repository of multiplex FL images and sequences for others to validate and build upon our findings. This would amplify the impact of the work and enable meta-analyses combining data from multiple studies, which is particularly useful in rare scenarios like RMC.

In terms of direct clinical translation, several steps could bring elements of this dissertation closer to routine care. The vein device could be commercialized and implemented in hospitals, improving everyday procedures. The prognostic markers in FL could be developed into a combined testing kit (imagine a pathology service that does an immunofluorescence panel and sequencing on an FL biopsy and provides a report on risk features). The HNSCC trial approach could integrate into precision oncology practice - for example, testing a patient's tumor for PI3K pathway activation to determine if adding a PI3K inhibitor to salvage therapy is warranted. And the integrative diagnostic method used for the RMC case could become part of molecular tumor boards for rare cancers (where difficult cases are analyzed by comprehensive sequencing and methylation analysis, not just standard pathology).

Finally, an important future direction is training the next generation of researchers and clinicians in this integrative approach. As implied throughout this dissertation, meaningful progress required understanding of biology, clinical context, and machine learning. One impact of this work could be educational - informing curriculum development that teaches interdisciplinary skills, so that future teams can replicate and improve upon studies like these more readily. Encouraging cross-talk between AI developers and medical experts will be crucial; this dissertation can serve as a case study in how such collaboration yields real outcomes.

Список литературы диссертационного исследования кандидат наук Сарачаков Александр Евгеньевич, 2026 год

210 Bibliography

1. Lin S., Zhou H., Watson M., Govindan R., Cote R.J., Yang C. Impact of stain variation and color normalization for prognostic predictions in pathology // Scientific Reports. — 2025. — Vol. 15. — P. 2369.

2. Watson D.S. Interpretable machine learning for genomics // Human Genetics. — 2022. — Vol. 141. — No. 9. — P. 1499-1513.

3. Kapoor S., Narayanan A. Leakage and the reproducibility crisis in machine-learning-based science // Patterns. — 2023. — Vol. 4. — No. 9.

4. Szymborski J., Emad A. A flaw in using pre-trained pLLMs in protein-protein interaction inference models. — [Electronic resource]. — bioRxiv, 2025. — URL: https://www.biorxiv.org/content/10.1101/2025.04.21.649858v2 (accessed: 18.12.2025).

5. Leli V.M., Rubashevskii A., Sarachakov A., Rogov O., Dylov D.V. Near-Infrared-to-Visible Vein Imaging via Convolutional Neural Networks and Reinforcement Learning // 2020 16th International Conference on Control, Automation, Robotics and Vision (ICARCV) 2020 16th International Conference on Control, Automation, Robotics and Vision (ICARCV). — 2020. — P. 434-441.

6. Leli V.M., Shipitsin V., Rogov O.Y., Sarachakov A., Dylov D.V. Adaptive Denoising and Alignment Agents for Infrared Imaging // IEEE Control Systems Letters. — 2022. — Vol. 6. — P. 1586-1591.

7. Sarachakov A., Varlamova A., Svekolkin V., Polyakova M., Valencia I., Unkenholz C., Pannellini T., Galkin I., Ovcharov P., Tabakov D., Postovalova E., Shin N., Sethi I., Bagaev A., Itkin T., Crane G., Kluk M., Geyer J., Inghirami G., Patel S. Spatial mapping of human hematopoiesis at single-cell resolution reveals aging-associated topographic remodeling // Blood. — 2023. — Vol. 142. — No. 26. — P. 2282-2295.

8. Radtke A.J., Postovalova E., Varlamova A., Bagaev A., Sorokina M., Kudryashova O., Meerson M., Polyakova M., Galkin I., Svekolkin V., Isaev S., Wiebe D., Sharun A., Sarachakov A., Perelman G., Lozinsky Y., Yaniv Z., Lowekamp B.C., Speranza E., Yao L.,

Pittaluga S., Shaffer A.L., Jonigk D., Phelan J.D., Davies-Hill T., Huang D.W., Ovcharov P., Nomie K., Nuzhdina E., Kotlov N., Ataullakhanov R., Fowler N., Kelly M., Muppidi J., Davis J.L., Hernandez J.M., Wilson W.H., Jaffe E.S., Staudt L.M., Roschewski M., Germain R.N. Multi-omic profiling of follicular lymphoma reveals changes in tissue architecture and enhanced stromal remodeling in high-risk patients // Cancer Cell. — 2024. — Vol. 42. — No. 3. — P. 444-463.e10.

9. Hanna G.J., Oakley L.B., Shi R., ONeill A., Shin K.Y., Scarfo N., Sehgal K., Dennis M.J., Quinn N., Jo V.Y., Wong K., Shvyrkova A., Kushnarev V., Shanthappa B.U., Tkachuk A., Kryukov K., Sarachakov A., Svekolkin V., Lennerz J., Waters S., Haddad R.I. Duvelisib with Docetaxel for Patients with Anti-PD-1 Refractory, Recurrent, or Metastatic Head and Neck Squamous Cell Carcinoma // Clinical Cancer Research: An Official Journal of the American Association for Cancer Research. — 2025. — Vol. 31. — No. 4. — P. 619-627.

10. Gubbiotti M.A., McCutcheon I.E., Rao P., Genovese G., Wang L., Tarasov A., Putintsev V., Berlinski A., Stupichev D., Kriukov K., Davitavyan S., Salem B., Sarachakov A., Lebedev D., Hensley M., Bagaev A., Paradiso F., Kushnarev V., Khegai G., Tannir N.M., Msaouel P. A novel case of glial transdifferentiation in renal medullary carcinoma brain metastasis // Acta Neuropathologica Communications. — 2025. — Vol. 13. — No. 1. — P. 12.

11. Bianconi S., Mohseni H. Recent Advances in Infrared Imagers: Toward Thermodynamic and Quantum Limits of Photon Sensitivity // Reports on progress in physics. Physical Society (Great Britain). — 2020. — Vol. 83. — Recent Advances in Infrared Imagers. — No. 4. — P. 044101.

12. Lahiri B.B., Bagavathiappan S., Jayakumar T., Philip J. Medical applications of infrared thermography: A review // Infrared Physics & Technology. — 2012. — Vol. 55. — Medical applications of infrared thermography. — No. 4. — P. 221-235.

13. Setchfield K., Gorman A., Simpson A.H.R.W., Somekh M.G., Wright A.J. Effect of skin color on optical properties and the implications for medical optical technologies: a

review // Journal of Biomedical Optics. — 2024. — Vol. 29. — Effect of skin color on optical properties and the implications for medical optical technologies. — No. 1. — P. 010901.

14. Lane L.A., Xue R., Nie S. Emergence of Two Near-Infrared Windows for In-Vivo and Intraoperative SERS // Current opinion in chemical biology. — 2018. — Vol. 45. — P. 95-103.

15. Ai D., Yang J., Fan J., Zhao Y., Song X., Shen J., Shao L., Wang Y. Augmented reality based real-time subcutaneous vein imaging system // Biomedical Optics Express. — 2016.

— Vol. 7. — No. 7. — P. 2565-2585.

16. Vyas V., Sharma A., Goyal S., Kothari N. Infrared vein visualisation devices for ease of intravenous access in children: hope versus hype // Anesthesiology Intensive Therapy. — 2021. — Vol. 53. — Infrared vein visualisation devices for ease of intravenous access in children. — No. 1. — P. 69-78.

17. Fiedler L.S., Lippert B.M., Adrian L., Meyer T. Perfusion in Pedicled Skin Flaps: Initial Insights from Smartphone-Based Thermal Imaging Protocol // Journal of Personalized Medicine. — 2024. — Vol. 14. — Perfusion in Pedicled Skin Flaps. — No. 7. — P. 730.

18. Hou F., Zhang Y., Zhou Y., Zhang M., Lv B., Wu J. Review on Infrared Imaging Technology // Sustainability. — 2022. — Vol. 14. — No. 18. — P. 11161.

19. Lahiri B.B., Bagavathiappan S., Jayakumar T., Philip J. Medical applications of infrared thermography: A review // Infrared Physics & Technology. — 2012. — Vol. 55.

— Medical applications of infrared thermography. — No. 4. — P. 221-235.

20. Fantini S., Sassaroli A. Near-Infrared Optical Mammography for Breast Cancer Detection with Intrinsic Contrast // Annals of biomedical engineering. — 2012. — Vol. 40. — No. 2. — P. 398-407.

21. Herman C. The role of dynamic infrared imaging in melanoma diagnosis // Expert review of dermatology. — 2013. — Vol. 8. — No. 2. — P. 177-184.

22. Kesztyüs D., Bae H., Wilson C., Schön M.P., Kesztyüs T. Non-invasive infrared

thermography for screening, diagnosis and monitoring of skin cancer. - [Electronic resource]. - JDDG: Journal der Deutschen Dermatologischen Gesellschaft, 2025.-URL: https://onlinelibrary.wiley.com/doi/10.1111/ddg.15598 (accessed: 18.02.2025).

23. Ferri F., Montorfano L., Bordes S.J., Forleiter C., Newman M.I. Near-Infrared Fluorescence Imaging for Sentinel Lymph Node Identification in Melanoma Surgery // Cureus. — Vol. 13. — No. 4. — P. e14550.

24. McKinley E.T., Sui Y., Al-Kofahi Y., Millis B.A., Tyska M.J., Roland J.T., Santamaria-Pang A., Ohland C.L., Jobin C., Franklin J.L., Lau K.S., Gerdes M.J., Coffey R.J. Optimized multiplex immunofluorescence single-cell analysis reveals tuft cell heterogeneity // JCI Insight. — 2017. — Vol. 2. — No. 11.

25. Zhao C., Germain R.N. Multiplex imaging in immuno-oncology // Journal for Immunotherapy of Cancer. — 2023. — Vol. 11. — No. 10. — P. e006923.

26. Parra E.R. Methods to Determine and Analyze the Cellular Spatial Distribution Extracted From Multiplex Immunofluorescence Data to Understand the Tumor Microenvironment // Frontiers in Molecular Biosciences. — 2021. — Vol. 8.

27. Rashid R., Gaglia G., Chen Y.-A., Lin J.-R., Du Z., Maliga Z., Schapiro D., Yapp C., Muhlich J., Sokolov A., Sorger P., Santagata S. Highly multiplexed immunofluorescence images and single-cell data of immune markers in tonsil and lung cancer // Scientific Data. — 2019. — Vol. 6. — No. 1. — P. 323.

28. Schapiro D., Sokolov A., Yapp C., Chen Y.-A., Muhlich J.L., Hess J., Creason A.L., Nirmal A.J., Baker G.J., Nariya M.K., Lin J.-R., Maliga Z., Jacobson C.A., Hodgman M.W., Ruokonen J., Farhi S.L., Abbondanza D., McKinley E.T., Persson D., Betts C., Sivagnanam S., Regev A., Goecks J., Coffey R.J., Coussens L.M., Santagata S., Sorger P.K. MCMICRO: a scalable, modular image-processing pipeline for multiplexed tissue imaging // Nature Methods. — 2022. — Vol. 19. — MCMICRO. — No. 3. — P. 311-315.

29. Bao S., Cui C., Li J., Tang Y., Lee H.H., Deng R., Remedios L.W., Yu X., Yang Q., Chiron S., Patterson N.H., Lau K.S., Liu Q., Roland J.T., Coburn L.A., Wilson K.T.,

Landman B.A., Huo Y. Topological-Preserving Membrane Skeleton Segmentation in Multiplex Immunofluorescence Imaging // Proceedings of SPIE--the International Society for Optical Engineering. — 2023. — Vol. 12471. — P. 124710B.

30. Stringer C., Pachitariu M. Cellpose3: one-click image restoration for improved cellular segmentation // Nature Methods. — 2025. — Cellpose3. — P. 1-8.

31. Han W., Cheung A.M., Yaffe M.J., Martel A.L. Cell segmentation for immunofluorescence multiplexed images using two-stage domain adaptation and weakly labeled data for pre-training // Scientific Reports. — 2022. — Vol. 12. — No. 1. — P. 4399.

32. Maus R.L.G., Leontovich A.A., Moore R.M., Fogarty Z., Guo R., Davidson T.M., Tekin B., Atherton C., Schimke J.M., Dicke B.A., Chen B.J., Markovic S.N. Quantitative spatial evaluation of tumor-immune interactions in the immunotherapy setting of metastatic melanoma lymph nodes // Frontiers in Immunology. — 2022. — Vol. 13. — P. 1024039.

33. Han S., Phasouk K., Zhu J., Fong Y. Optimizing Deep Learning-Based Segmentation of Densely Packed Cells using Cell Surface Markers // Research Square. — 2023. — P. rs.3.rs-3307496.

34. Greenwald N.F., Miller G., Moen E., Kong A., Kagel A., Dougherty T., Fullaway C.C., McIntosh B.J., Leow K.X., Schwartz M.S., Pavelchek C., Cui S., Camplisson I., Bar-Tal O., Singh J., Fong M., Chaudhry G., Abraham Z., Moseley J., Warshawsky S., Soon E., Greenbaum S., Risom T., Hollmann T., Bendall S.C., Keren L., Graf W., Angelo M., Van Valen D. Whole-cell segmentation of tissue images with human-level performance using large-scale data annotation and deep learning // Nature biotechnology. — 2022. — Vol. 40. — No. 4. — P. 555-565.

35. De León Rodríguez S.G., Hernández Herrera P., Aguilar Flores C., Pérez Koldenkova V., Guerrero A., Mantilla A., Fuentes-Pananá E.M., Wood C., Bonifaz L.C. A Machine Learning Workflow of Multiplexed Immunofluorescence Images to Interrogate Activator and Tolerogenic Profiles of Conventional Type 1 Dendritic Cells Infiltrating Melanomas of Disease-Free and Metastatic Patients // Journal of Oncology. — 2022. — Vol. 2022. —

P. 9775736.

36. Stringer C., Pachitariu M. Cellpose3: one-click image restoration for improved cellular segmentation // Nature Methods. — 2025. — Cellpose3. — P. 1-8.

37. Cutler K.J., Stringer C., Lo T.W., Rappez L., Stroustrup N., Brook Peterson S., Wiggins P.A., Mougous J.D. Omnipose: a high-precision morphology-independent solution for bacterial cell segmentation // Nature Methods. — 2022. — Vol. 19. — Omnipose. — No. 11. — P. 1438-1448.

38. Lee M.Y., Bedia J.S., Bhate S.S., Barlow G.L., Phillips D., Fantl W.J., Nolan G.P., Schürch C.M. CellSeg: a robust, pre-trained nucleus segmentation and pixel quantification software for highly multiplexed fluorescence images // BMC Bioinformatics. — 2022. — Vol. 23. — CellSeg. — No. 1. — P. 46.

39. Seal S., Wrobel J., Johnson A.M., Nemenoff R.A., Schenk E.L., Bitler B.G., Jordan K.R., Ghosh D. On clustering for cell-phenotyping in multiplex immunohistochemistry (mIHC) and multiplexed ion beam imaging (MIBI) data // BMC research notes. — 2022. — Vol. 15. — No. 1. — P. 215.

40. Levine J.H., Simonds E.F., Bendall S.C., Davis K.L., Amir E.D., Tadmor M.D., Litvin O., Fienberg H.G., Jager A., Zunder E.R., Finck R., Gedman A.L., Radtke I., Downing J.R., Pe'er D., Nolan G.P. Data-Driven Phenotypic Dissection of AML Reveals Progenitor-like Cells that Correlate with Prognosis // Cell. — 2015. — Vol. 162. — No. 1. — P. 184-197.

41. Van Gassen S., Callebaut B., Van Helden M.J., Lambrecht B.N., Demeester P., Dhaene T., Saeys Y. FlowSOM: Using self-organizing maps for visualization and interpretation of cytometry data // Cytometry Part A. — 2015. — Vol. 87. — FlowSOM. — No. 7. — P. 636-645.

42. Peng H., Wu X., Zhong R., Yu T., Cai X., Liu J., Wen Y., Ao Y., Chen J., Li Y., He M., Li C., Zheng H., Chen Y., Pan Z., He J., Liang W. Profiling Tumor Immune Microenvironment of Non-Small Cell Lung Cancer Using Multiplex Immunofluorescence // Frontiers in Immunology. — 2021. — Vol. 12. — P. 750046.

43. Wilson C.M., Ospina O.E., Townsend M.K., Nguyen J., Moran Segura C., Schildkraut J.M., Tworoger S.S., Peres L.C., Fridley B.L. Challenges and Opportunities in the Statistical Analysis of Multiplex Immunofluorescence Data // Cancers. — 2021. — Vol. 13. — No. 12. — P. 3031.

44. Peng H., Wu X., Zhong R., Yu T., Cai X., Liu J., Wen Y., Ao Y., Chen J., Li Y., He M., Li C., Zheng H., Chen Y., Pan Z., He J., Liang W. Profiling Tumor Immune Microenvironment of Non-Small Cell Lung Cancer Using Multiplex Immunofluorescence // Frontiers in Immunology. — 2021. — Vol. 12. — P. 750046.

45. Gavrielatou N., Liu Y., Vathiotis I., Zugazagoitia J., Aung T.N., Shafi S., Fernandez A., Schalper K., Psyrri A., Rimm D.L. Association of PD-1/PD-L1 Co-location with Immunotherapy Outcomes in Non-Small Cell Lung Cancer // Clinical cancer research : an official journal of the American Association for Cancer Research. — 2022. — Vol. 28. — No. 2. — P. 360-367.

46. Pourmaleki M., Socci N.D., Hollmann T.J., Mellinghoff I.K. Moving Spatially Resolved Multiplexed Protein Profiling toward Clinical Oncology // Cancer Discovery. — 2023. — Vol. 13. — No. 4. — P. 824.

47. Cenerenti M., Saillard M., Romero P., Jandus C. The Era of Cytotoxic CD4 T Cells // Frontiers in Immunology. — 2022. — Vol. 13.

48. Miheecheva N., Postovalova E., Lyu Y., Ramachandran A., Bagaev A., Svekolkin V., Galkin I., Zyrin V., Maximov V., Lozinsky Y., Isaev S., Ovcharov P., Shamsutdinova D., Cheng E.H., Nomie K., Brown J.H., Tsiper M., Ataullakhanov R., Fowler N., Hsieh J.J. Multiregional single-cell proteogenomic analysis of ccRCC reveals cytokine drivers of intratumor spatial heterogeneity // Cell Reports. — 2022. — Vol. 40. — No. 7. — P. 111180.

49. Hoyt C.C. Multiplex Immunofluorescence and Multispectral Imaging: Forming the Basis of a Clinical Test Platform for Immuno-Oncology // Frontiers in Molecular Biosciences. — 2021. — Vol. 8. — Multiplex Immunofluorescence and Multispectral Imaging. — P. 674747.

50. Lipkova J., Chen R.J., Chen B., Lu M.Y., Barbieri M., Shao D., Vaidya A.J., Chen C., Zhuang L., Williamson D.F., Shaban M., Chen T.Y., Mahmood F. Artificial Intelligence for Multimodal Data Integration in Oncology // Cancer cell. — 2022. — Vol. 40. — No. 10.

— P. 1095.

51. Weinstein J.N., Collisson E.A., Mills G.B., Shaw K.R.M., Ozenberger B.A., Ellrott K., Shmulevich I., Sander C., Stuart J.M. The Cancer Genome Atlas Pan-Cancer analysis project // Nature Genetics. — 2013. — Vol. 45. — No. 10. — P. 1113-1120.

52. Menyhart O., Gyorffy B. Multi-omics approaches in cancer research with applications in tumor subtyping, prognosis, and diagnosis // Computational and Structural Biotechnology Journal. — 2021. — Vol. 19. — P. 949.

53. Cai Z., Poulos R.C., Liu J., Zhong Q. Machine learning for multi-omics data integration in cancer // iScience. — 2022. — Vol. 25. — No. 2. — P. 103798.

54. You Y., Fu Y., Li L., Zhang Z., Jia S., Lu S., Ren W., Liu Y., Xu Y., Liu X., Jiang F., Peng G., Sampath Kumar A., Ritchie M.E., Liu X., Tian L. Systematic comparison of sequencing-based spatial transcriptomic methods // Nature Methods. — 2024. — Vol. 21.

— No. 9. — P. 1743-1754.

55. Zhang L., Chen D., Song D., Liu X., Zhang Y., Xu X., Wang X. Clinical and translational values of spatial transcriptomics // Signal Transduction and Targeted Therapy.

— 2022. — Vol. 7. — No. 1. — P. 1-17.

56. Heo Y.J., Hwa C., Lee G.-H., Park J.-M., An J.-Y. Integrative Multi-Omics Approaches in Cancer Research: From Biological Networks to Clinical Subtypes // Molecules and Cells. — 2021. — Vol. 44. — Integrative Multi-Omics Approaches in Cancer Research.

— No. 7. — P. 433-443.

57. Lu C., Shiradkar R., Liu Z. Integrating pathomics with radiomics and genomics for cancer prognosis: A brief review // Chinese Journal of Cancer Research. — 2021. — Vol. 33. — Integrating pathomics with radiomics and genomics for cancer prognosis. — No. 5. — P. 563-573.

58. Dasari S., Chakraborty A., Truong L., Mohan C. A Systematic Review of

Interpathologist Agreement in Histologic Classification of Lupus Nephritis // Kidney International Reports. — 2019. — Vol. 4. — No. 10. — P. 1420-1425.

59. Chen X., Fan H., Girshick R., He K. Improved baselines with momentum contrastive learning. - [Electronic resource]. - arXiv, 2020. - URL: https://arxiv.org/abs/2003.04297 (accessed: 18.10.2025).

60. Oquab M., Darcet T., Moutakanni T., Vo H., Szafraniec M., Khalidov V., Fernandez P., Haziza D., Massa F., El-Nouby A., Assran M., Ballas N., Galuba W., Howes R., Huang P.-Y., Li S.-W., Misra I., Rabbat M., Sharma V., Synnaeve G., Xu H., Jegou H., Mairal J., Labatut P., Joulin A., Bojanowski P. DINOv2: learning robust visual features without supervision.- [Electronic resource].- arXiv, 2024. -URL: https://arxiv.org/abs/2304.07193 (accessed: 18.10.2025).

61. Xu H., Usuyama N., Bagga J., Zhang S., Rao R., Naumann T., Wong C., Gero Z., González J., Gu Y., Xu Y., Wei M., Wang W., Ma S., Wei F., Yang J., Li C., Gao J., Rosemon J., Bower T., Lee S., Weerasinghe R., Wright B.J., Robicsek A., Piening B., Bifulco C., Wang S., Poon H. A whole-slide foundation model for digital pathology from real-world data // Nature. — 2024. — Vol. 630. — No. 8015. — P. 181-188.

62. Schneider L., Laiouar-Pedari S., Kuntz S., Krieghoff-Henning E., Hekler A., Kather J.N., Gaiser T., Fröhling S., Brinker T.J. Integration of deep learning-based image analysis and genomic data in cancer pathology: A systematic review // European Journal of Cancer (Oxford, England: 1990). — 2022. — Vol. 160. — Integration of deep learning-based image analysis and genomic data in cancer pathology. — P. 80-91.

63. Menesatti P., Angelini C., Pallottino F., Antonucci F., Aguzzi J., Costa C. RGB Color Calibration for Quantitative Image Analysis: The "3D Thin-Plate Spline" Warping Approach // Sensors. — 2012. — Vol. 12. — RGB Color Calibration for Quantitative Image Analysis. — No. 6. — P. 7063-7079.

64. Voon W., Hum Y.C., Tee Y.K., Yap W.-S., Nisar H., Mokayed H., Gupta N., Lai K.W. Evaluating the effectiveness of stain normalization techniques in automated grading of invasive ductal carcinoma histopathological images // Scientific Reports. — 2023. —

Vol. 13. — No. 1. — P. 20518.

65. Ding T., Wagner S.J., Song A.H., Chen R.J., Lu M.Y., Zhang A., Vaidya A.J., Jaume G., Shaban M., Kim A., Williamson D.F.K., Chen B., Almagro-Perez C., Doucet P., Sahai S., Chen C., Komura D., Kawabe A., Ishikawa S., Gerber G., Peng T., Le L.P., Mahmood F. Multimodal whole slide foundation model for pathology. - [Electronic resource]. - arXiv, 2024.-URL: https://arxiv.org/abs/2411.19666 (accessed: 18.10.2025).

66. Tang Z., Yang Z., Zhu C., Zeng M., Bansal M. Any-to-any generation via composable diffusion. - [Electronic resource]. - arXiv, 2023. - URL: https://arxiv.org/abs/2305.11846 (accessed: 18.10.2025).

67. He X., Liu X., Zuo F., Shi H., Jing J. Artificial intelligence-based multi-omics analysis fuels cancer precision medicine // Seminars in Cancer Biology. — 2023. — Vol. 88. — P. 187-200.

68. Pawlowski M., Wroblewska A., Sysko-Romanczuk S. Effective Techniques for Multimodal Data Fusion: A Comparative Analysis // Sensors (Basel, Switzerland). — 2023. — Vol. 23. — Effective Techniques for Multimodal Data Fusion. — No. 5. — P. 2381.

69. Mitchel J., Chatlin K., Tong L., Wang M.D. A Translational Pipeline for Overall Survival Prediction of Breast Cancer Patients by Decision-Level Integration of Multi-Omics Data // Proceedings. IEEE International Conference on Bioinformatics and Biomedicine. — 2019. — Vol. 2019. — P. 1573-1580.

70. Vandereyken K., Sifrim A., Thienpont B., Voet T. Methods and applications for single-cell and spatial multi-omics // Nature Reviews. Genetics. — 2023. — P. 1-22.

71. Wang C., Pecot T., Zynger D.L., Machiraju R., Shapiro C.L., Huang K. Identifying survival associated morphological features of triple negative breast cancer using multiple datasets // Journal of the American Medical Informatics Association: JAMIA. — 2013. — Vol. 20. — No. 4. — P. 680-687.

72. Martins F.C., Santiago I. de, Trinh A., Xian J., Guo A., Sayal K., Jimenez-Linan M., Deen S., Driver K., Mack M., Aslop J., Pharoah P.D., Markowetz F., Brenton J.D.

Combined image and genomic analysis of high-grade serous ovarian cancer reveals PTEN loss as a common driver event and prognostic classifier // Genome Biology. — 2014. — Vol. 15. — No. 12. — P. 526.

73. Ash J.T., Darnell G., Munro D., Engelhardt B.E. Joint analysis of expression levels and histological images identifies genes associated with tissue morphology // Nature Communications. — 2021. — Vol. 12. — No. 1. — P. 1609.

74. Subramanian V., Chidester B., Ma J., Do M.N. Correlating cellular features with gene expression using CCA // 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018) 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018). — 2018. — P. 805-808.

75. Unger M., Kather J.N. Deep learning in cancer genomics and histopathology // Genome Medicine. — 2024. — Vol. 16. — No. 1. — P. 44.

76. Mobadersany P., Yousefi S., Amgad M., Gutman D.A., Barnholtz-Sloan J.S., Velazquez Vega J.E., Brat D.J., Cooper L.A.D. Predicting cancer outcomes from histology and genomics using convolutional networks // Proceedings of the National Academy of Sciences of the United States of America. — 2018. — Vol. 115. — No. 13. — P. E2970-E2979.

77. Cheerla A., Gevaert O. Deep learning with multimodal representation for pancancer prognosis prediction // Bioinformatics (Oxford, England). — 2019. — Vol. 35. — No. 14. — P. i446-i454.

78. He B., Bergenstrahle L., Stenbeck L., Abid A., Andersson A., Borg A., Maaskola J., Lundeberg J., Zou J. Integrating spatial gene expression and breast tumour morphology via deep learning // Nature Biomedical Engineering. — 2020. — Vol. 4. — No. 8. — P. 827-834.

79. Vaidya A., Zhang A., Jaume G., Song A.H., Ding T., Wagner S.J., Lu M.Y., Doucet P., Robertson H., AlmagroPerez C., Chen R.J., ElHarouni D., Ayoub G., Bossi C., Ligon K.L., Gerber G., Le L.P., Mahmood F. Molecular-driven foundation model for oncologic pathology. - [Electronic resource]. - arXiv, 2025.-URL: https://arxiv.org/abs/2501.16652

(accessed: 18.10.2025).

80. Chen R.J., Lu M.Y., Wang J., Williamson D.F.K., Rodig S.J., Lindeman N.I., Mahmood

F. Pathomic Fusion: An Integrated Framework for Fusing Histopathology and Genomic Features for Cancer Diagnosis and Prognosis // IEEE transactions on medical imaging. — 2022. — Vol. 41. — Pathomic Fusion. — No. 4. — P. 757-770.

81. Bandyopadhyay S., Duffy M.P., Ahn K.J., Sussman J.H., Pang M., Smith D., Duncan

G., Zhang I., Huang J., Lin Y., Xiong B., Imtiaz T., Chen C.-H., Thadi A., Chen C., Xu J., Reichart M., Martinez Z., Diorio C., Chen C., Pillai V., Snaith O., Oldridge D., Bhattacharyya S., Maillard I., Carroll M., Nelson C., Qin L., Tan K. Mapping the cellular biogeography of human bone marrow niches using single-cell transcriptomics and proteomic imaging // Cell. — 2024. — Vol. 187. — No. 12. — P. 3120-3140.e29.

82. Hagos Y.B., Lecat C.S.Y., Patel D., Mikolajczak A., Castillo S.P., Lyon E.J., Foster K., Tran T.-A., Lee L.S.H., Rodriguez-Justo M., Yong K.L., Yuan Y. Deep Learning Enables Spatial Mapping of the Mosaic Microenvironment of Myeloma Bone Marrow Trephine Biopsies // Cancer Research. — 2024. — Vol. 84. — No. 3. — P. 493-508.

83. Kuswanto W., Nolan G., Lu G. Highly multiplexed spatial profiling with CODEX: bioinformatic analysis and application in human disease // Seminars in Immunopathology. — 2023. — Vol. 45. — Highly multiplexed spatial profiling with CODEX. — No. 1. — P. 145-157.

84. Allam M., Hu T., Cai S., Laxminarayanan K., Hughley R.B., Coskun A.F. Spatially visualized single-cell pathology of highly multiplexed protein profiles in health and disease // Communications Biology. — 2021. — Vol. 4. — No. 1. — P. 1-16.

85. Pang M., Roy T.K., Wu X., Tan K. CelloType: a unified model for segmentation and classification of tissue images // Nature Methods. — 2025. — Vol. 22. — CelloType. — No. 2. — P. 348-357.

86. Keren L., Bosse M., Marquez D., Angoshtari R., Jain S., Varma S., Yang S.-R., Kurian A., Van Valen D., West R., Bendall S.C., Angelo M. A Structured Tumor-Immune Microenvironment in Triple Negative Breast Cancer Revealed by Multiplexed Ion Beam

Imaging // Cell. — 2018. — Vol. 174. — No. 6. — P. 1373-1387.e19.

87. Aoki T., Jiang A., Xu A., Yin Y., Gamboa A., Milne K., Takata K., Miyata-Takata T., Chung S., Rai S., Wu S., Warren M., Strong C., Goodyear T., Morris K., Chong L.C., Hav M., Colombo A.R., Telenius A., Boyle M., Ben-Neriah S., Power M., Gerrie A.S., Weng A.P., Karsan A., Roth A., Farinha P., Scott D.W., Savage K.J., Nelson B.H., Merchant A., Steidl C. Spatially Resolved Tumor Microenvironment Predicts Treatment Outcomes in Relapsed/Refractory Hodgkin Lymphoma // Journal of Clinical Oncology. — 2024. — Vol. 42. — No. 9. — P. 1077-1087.

88. Zhao Y., Xiong S., Ren Q., Wang J., Li M., Yang L., Wu D., Tang K., Pan X., Chen F., Wang W., Jin S., Liu X., Lin G., Yao W., Cai L., Yang Y., Liu J., Wu J., Fu W., Sun K., Li F., Cheng B., Zhan S., Wang H., Yu Z., Liu X., Zhong R., Wang H., He P., Zheng Y., Liang P., Chen L., Hou T., Huang J., He B., Song J., Wu L., Hu C., He J., Yao J., Liang W. Deep learning using histological images for gene mutation prediction in lung cancer: a multicentre retrospective study // The Lancet Oncology. — 2025. — Vol. 26. — Deep learning using histological images for gene mutation prediction in lung cancer. — No. 1. — P. 136-146.

89. Mandair D., Reis-Filho J.S., Ashworth A. Biological insights and novel biomarker discovery through deep learning approaches in breast cancer histopathology // NPJ Breast Cancer. — 2023. — Vol. 9. — P. 21.

90. Chen R.J., Lu M.Y., Williamson D.F.K., Chen T.Y., Lipkova J., Noor Z., Shaban M., Shady M., Williams M., Joo B., Mahmood F. Pan-Cancer Integrative Histology-Genomic Analysis via Multimodal Deep Learning // Cancer cell. — 2022. — Vol. 40. — No. 8. — P. 865-878.e6.

91. Dobersalske C., Rauschenbach L., Hua Y., Berliner C., Steinbach A., Grüneboom A., Kokkaliaris K.D., Heiland D.H., Berger P., Langer S., Tan C.L., Stenzel M., Landolsi S., Weber F., Darkwah Oppong M., Werner R.A., Gull H., Schröder T., Linsenmann T., Buck A.K., Gunzer M., Stuschke M., Keyvani K., Forsting M., Glas M., Kipnis J., Steindler D.A., Reinhardt H.C., Green E.W., Platten M., Tasdogan A., Herrmann K., Rambow F.,

Cima I., Sure U., Scheffler B. Cranioencephalic functional lymphoid units in glioblastoma // Nature Medicine. — 2024. — Vol. 30. — No. 10. — P. 2947-2956.

92. Kang J., Chowdhry A., Pugh S., Park J. Integrating Artificial Intelligence and Machine Learning into Cancer Clinical Trials // Seminars in radiation oncology. — 2023. — Vol. 33. — No. 4. — P. 386-394.

93. Hong J.C., Niedzwiecki D., Palta M., Tenenbaum J.D. Predicting Emergency Visits and Hospital Admissions During Radiation and Chemoradiation: An Internally Validated Pretreatment Machine Learning Algorithm // JCO clinical cancer informatics. — 2018. — Vol. 2. — Predicting Emergency Visits and Hospital Admissions During Radiation and Chemoradiation. — P. 1-11.

94. Hong J.C., Eclov N.C.W., Dalal N.H., Thomas S.M., Stephens S.J., Malicki M., Shields S., Cobb A., Mowery Y.M., Niedzwiecki D., Tenenbaum J.D., Palta M. System for High-Intensity Evaluation During Radiation Therapy (SHIELD-RT): A Prospective Randomized Study of Machine Learning-Directed Clinical Evaluations During Radiation and Chemoradiation // Journal of Clinical Oncology: Official Journal of the American Society of Clinical Oncology. — 2020. — Vol. 38. — System for High-Intensity Evaluation During Radiation Therapy (SHIELD-RT). — No. 31. — P. 3652-3661.

95. Parikh R.B., Manz C., Chivers C., Regli S.H., Braun J., Draugelis M.E., Schuchter L.M., Shulman L.N., Navathe A.S., Patel M.S., O'Connor N.R. Machine Learning Approaches to Predict 6-Month Mortality Among Patients With Cancer // JAMA network open. — 2019. — Vol. 2. — No. 10. — P. e1915997.

96. Kann B.H., Aneja S., Loganadane G.V., Kelly J.R., Smith S.M., Decker R.H., Yu J.B., Park H.S., Yarbrough W.G., Malhotra A., Burtness B.A., Husain Z.A. Pretreatment Identification of Head and Neck Cancer Nodal Metastasis and Extranodal Extension Using Deep Learning Neural Networks // Scientific Reports. — 2018. — Vol. 8. — P. 14036.

97. Ayana G., Lee E., Choe S. Vision Transformers for Breast Cancer Human Epidermal Growth Factor Receptor 2 Expression Staging without Immunohistochemical Staining // The American Journal of Pathology. — 2024. — Vol. 194. — No. 3. — P. 402-414.

98. Gulshan V., Peng L., Coram M., Stumpe M.C., Wu D., Narayanaswamy A., Venugopalan S., Widner K., Madams T., Cuadros J., Kim R., Raman R., Nelson P.C., Mega J.L., Webster D.R. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs // JAMA. — 2016. — Vol. 316. — No. 22. — P. 2402-2410.

99. Beede E., Baylor E., Hersch F., Iurchenko A., Wilcox L., Ruamviboonsuk P., Vardoulakis L.M. A human-centered evaluation of a deep learning system deployed in clinics for the detection of diabetic retinopathy. - [Electronic resource]. - Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems.-2020.-URL: https://dl.acm.org/doi/10.1145/3313831.3376718 (accessed: 15.04.2025).

100. Hanna M.G., Pantanowitz L., Dash R., Harrison J.H., Deebajah M., Pantanowitz J., Rashidi H.H. Future of Artificial Intelligence—Machine Learning Trends in Pathology and Medicine // Modern Pathology. — 2025. — Vol. 38. — No. 4.

101. Matthews G.A., McGenity C., Bansal D., Treanor D. Public evidence on AI products for digital pathology // npj Digital Medicine. — 2024. — Vol. 7. — No. 1. — P. 1-11.

102. Chen J., Lin A., Jiang A., Qi C., Liu Z., Cheng Q., Yuan S., Luo P. Computational frameworks transform antagonism to synergy in optimizing combination therapies // npj Digital Medicine. — 2025. — Vol. 8. — No. 1. — P. 1-22.

103. Jiang P., Huang S., Fu Z., Sun Z., Lakowski T.M., Hu P. Deep graph embedding for prioritizing synergistic anticancer drug combinations // Computational and Structural Biotechnology Journal. — 2020. — Vol. 18. — P. 427-438.

104. Zhang T., Zhang L., Payne P.R.O., Li F. Synergistic Drug Combination Prediction by Integrating Multiomics Data in Deep Learning Models // Methods in Molecular Biology (Clifton, N.J.). — 2021. — Vol. 2194. — P. 223-238.

105. Preuer K., Lewis R.P.I., Hochreiter S., Bender A., Bulusu K.C., Klambauer G. DeepSynergy: predicting anti-cancer drug synergy with Deep Learning // Bioinformatics (Oxford, England). — 2018. — Vol. 34. — DeepSynergy. — No. 9. — P. 1538-1546.

106. Torkamannia A., Omidi Y., Ferdousi R. SYNDEEP: a deep learning approach for the

prediction of cancer drugs synergy // Scientific Reports. — 2023. — Vol. 13. — SYNDEEP. — No. 1. — P. 6184.

107. Rakaee M., Tafavvoghi M., Ricciuti B., Alessi J.V., Cortellini A., Citarella F., Nibid L., Perrone G., Adib E., Fulgenzi C.A.M., Hidalgo Filho C.M., Di Federico A., Jabar F., Hashemi S., Houda I., Richardsen E., Rasmussen Busund L.-T., Donnem T., Bahce I., Pinato D.J., Helland Ä., Sholl L.M., Awad M.M., Kwiatkowski D.J. Deep Learning Model for Predicting Immunotherapy Response in Advanced Non-Small Cell Lung Cancer // JAMA Oncology. — 2025. — Vol. 11. — No. 2. — P. 109-118.

108. Bergstrom E.N., Abbasi A., Diaz-Gay M., Galland L., Ladoire S., Lippman S.M., Alexandrov L.B. Deep Learning Artificial Intelligence Predicts Homologous Recombination Deficiency and Platinum Response From Histologic Slides // Journal of Clinical Oncology. — 2024. — Vol. 42. — No. 30. — P. 3550-3560.

109. Hanna M.G., Pantanowitz L., Dash R., Harrison J.H., Deebajah M., Pantanowitz J., Rashidi H.H. Future of Artificial Intelligence—Machine Learning Trends in Pathology and Medicine // Modern Pathology. — 2025. — Vol. 38. — No. 4.

110. Fu X., Sahai E., Wilkins A. Application of digital pathology-based advanced analytics of tumour microenvironment organisation to predict prognosis and therapeutic response // The Journal of Pathology. — 2023. — Vol. 260. — No. 5. — P. 578-591.

111. 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. — Vol. 14. — No. 1. — P. 100921.

112. Wu C., Andaloussi M.A., Hormuth D.A., Lima E.A.B.F., Lorenzo G., Stowers C.E., Ravula S., Levac B., Dimakis A.G., Tamir J.I., Brock K.K., Chung C., Yankeelov T.E. A critical assessment of artificial intelligence in magnetic resonance imaging of cancer // npj Imaging. — 2025. — Vol. 3. — No. 1. — P. 1-13.

113. Sheller M.J., Edwards B., Reina G.A., Martin J., Pati S., Kotrotsou A., Milchenko M., Xu W., Marcus D., Colen R.R., Bakas S. Federated learning in medicine: facilitating multi-institutional collaborations without sharing patient data // Scientific Reports. —

2020. — Vol. 10. — Federated learning in medicine. — No. 1. — P. 12598.

114. Karargyris A., Umeton R., Sheller M.J., Aristizabal A., George J., Mattson P., et al. Federated benchmarking of medical artificial intelligence with MedPerf // Nature Machine Intelligence. — 2023. — Vol. 5. — No. 7. — P. 799-810.

115. Pati S., Baid U., Edwards B., Sheller M., Wang S.-H., Bakas S., et al. Federated learning enables big data for rare cancer boundary detection // Nature Communications. — 2022. — Vol. 13. — No. 1. — P. 7346.

116. Sweeney S.M., Hamadeh H.K., Abrams N., Adam S.J., Brenner S., Connors D.E., Davis G.J., Fiore L., Gawel S.H., Grossman R.L., Hanlon S.E., Hsu K., Kelloff G.J., Kirsch I.R., Louv B., McGraw D., Meng F., Milgram D., Miller R.S., Morgan E., Mukundan L., O'Brien T., Robbins P., Rubin E.H., Rubinstein W.S., Salmi L., Schaller T., Shi G., Sigman C.C., Srivastava S. Challenges to Using Big Data in Cancer // Cancer Research. — 2023. — Vol. 83. — No. 8. — P. 1175-1182.

117. Chen Y., Schönlieb C.-B., Lio P., Leiner T., Dragotti P.L., Wang G., Rueckert D., Firmin D., Yang G. AI-Based Reconstruction for Fast MRI—A Systematic Review and Meta-Analysis // Proceedings of the IEEE. — 2022. — Vol. 110. — No. 2. — P. 224-245.

118. Daye D., Wiggins W.F., Lungren M.P., Alkasab T., Kottler N., Allen B., Roth C.J., Bizzo B.C., Durniak K., Brink J.A., Larson D.B., Dreyer K.J., Langlotz C.P. Implementation of Clinical Artificial Intelligence in Radiology: Who Decides and How? // Radiology. — 2022. — Implementation of Clinical Artificial Intelligence in Radiology.

119. Kang J., Chowdhry A., Pugh S., Park J. Integrating Artificial Intelligence and Machine Learning into Cancer Clinical Trials // Seminars in radiation oncology. — 2023. — Vol. 33. — No. 4. — P. 386-394.

120. Lotter W., Hassett M.J., Schultz N., Kehl K.L., Van Allen E.M., Cerami E. Artificial Intelligence (AI) in Oncology: Current Landscape, Challenges, and Future Directions // Cancer discovery. — 2024. — Vol. 14. — Artificial Intelligence (AI) in Oncology. — No. 5. — P. 711-726.

121. United States Food and Drug Administration (FDA). Proposed regulatory framework for modifications to artificial intelligence/machine learning (AI/ML)-based software as a medical device (SaMD). - [Electronic resource]. - Silver Spring, MD: FDA, 2019. -URL: https://www.fda.gov/media/122535/download (accessed: 18.10.2025).

122. Krumm N. Organizational and Technical Security Considerations for Laboratory Cloud Computing // The Journal of Applied Laboratory Medicine. — 2023. — Vol. 8. — No. 1. — P. 180-193.

123. Lu M.Y., Chen B., Williamson D.F.K., Chen R.J., Liang I., Ding T., Jaume G., Odintsov I., Le L.P., Gerber G., Parwani A.V., Zhang A., Mahmood F. A visual-language foundation model for computational pathology // Nature Medicine. — 2024. — Vol. 30. — No. 3. — P. 863-874.

124. Bonizzi G., Zattoni L., Fusco N. Biobanking in the digital pathology era // Oncology Research. — 2021. — Vol. 29. — No. 4. — P. 229-233.

125. Brancato V., Esposito G., Coppola L., Cavaliere C., Mirabelli P., Scapicchio C., Borgheresi R., Neri E., Salvatore M., Aiello M. Standardizing digital biobanks: integrating imaging, genomic, and clinical data for precision medicine // Journal of Translational Medicine. — 2024. — Vol. 22. — Standardizing digital biobanks. — No. 1. — P. 136.

126. Frascarelli C., Bonizzi G., Musico C.R., Mane E., Cassi C., Guerini Rocco E., Farina A., Scarpa A., Lawlor R., Reggiani Bonetti L., Caramaschi S., Eccher A., Marletta S., Fusco N. Revolutionizing Cancer Research: The Impact of Artificial Intelligence in Digital Biobanking // Journal of Personalized Medicine. — 2023. — Vol. 13. — Revolutionizing Cancer Research. — No. 9. — P. 1390.

127. Fu X., Sahai E., Wilkins A. Application of digital pathology-based advanced analytics of tumour microenvironment organisation to predict prognosis and therapeutic response // The Journal of Pathology. — 2023. — Vol. 260. — No. 5. — P. 578.

128. Birla M., Rajan, Roy P.G., Gupta I., Malik P.S. Integrating Artificial Intelligence-Driven Wearable Technology in Oncology Decision-Making: A Narrative Review // Oncology. — 2025. — Vol. 103. — Integrating Artificial Intelligence-Driven

Wearable Technology in Oncology Decision-Making. — No. 1. — P. 69-82.

129. Aziz F., Bianchini D., Olawade D.B., Boussios S. The Impact of AI-driven Remote Patient Monitoring on Cancer Care: A Systematic Review // Anticancer Research. — 2025. — Vol. 45. — The Impact of AI-driven Remote Patient Monitoring on Cancer Care.

— No. 2. — P. 407-418.

130. Schulman J., Wolski F., Dhariwal P., Radford A., Klimov O. Proximal policy optimization algorithms. - [Electronic resource]. - arXiv, 2017. -URL: https://arxiv.org/abs/1707.06347 (accessed: 18.10.2025).

131. Baccin C., Al-Sabah J., Velten L., Helbling P.M., Grünschläger F., Hernandez-Malmierca P., Nombela-Arrieta C., Steinmetz L.M., Trumpp A., Haas S. Combined single-cell and spatial transcriptomics reveal the molecular, cellular and spatial bone marrow niche organization // Nature Cell Biology. — 2020. — Vol. 22. — No. 1.

— P. 38-48.

132. Zhang Z., Huang Z., Ong B., Sahu C., Zeng H., Ruan H.-B. Bone marrow adipose tissue-derived stem cell factor mediates metabolic regulation of hematopoiesis // Haematologica. — 2019. — Vol. 104. — No. 9. — P. 1731-1743.

133. Lengefeld J., Cheng C.-W., Maretich P., Blair M., Hagen H., McReynolds M.R., Sullivan E., Majors K., Roberts C., Kang J.H., Steiner J.D., Miettinen T.P., Manalis S.R., Antebi A., Morrison S.J., Lees J.A., Boyer L.A., Yilmaz Ö.H., Amon A. Cell size is a determinant of stem cell potential during aging // Science Advances. — 2021. — Vol. 7.

— No. 46. — P. eabk0271.

134. Heazlewood S.Y., Ahmad T., Cao B., Cao H., Domingues M., Sun X., Heazlewood C.K., Li S., Williams B., Fulton M., White J.F., Nebl T., Nefzger C.M., Polo J.M., Kile B.T., Kraus F., Ryan M.T., Sun Y.B., Choong P.F.M., Ellis S.L., Anko M.-L., Nilsson S.K. High ploidy large cytoplasmic megakaryocytes are hematopoietic stem cells regulators and essential for platelet production // Nature Communications. — 2023. — Vol. 14. — No. 1. — P. 2099.

135. Casulo C., Byrtek M., Dawson K.L., Zhou X., Farber C.M., Flowers C.R., Hainsworth

J.D., Maurer M.J., Cerhan J.R., Link B.K., Zelenetz A.D., Friedberg J.W. Early Relapse of Follicular Lymphoma After Rituximab Plus Cyclophosphamide, Doxorubicin, Vincristine, and Prednisone Defines Patients at High Risk for Death: An Analysis From the National LymphoCare Study // Journal of Clinical Oncology. — 2015. — Vol. 33. — No. 23. — P. 2516-2522.