Conventional histopathology offers an extraordinarily detailed view of tissue, but it does so through a narrow window. A biopsy is typically sliced into thin sections, mounted on glass slides and stained so that cells and tissue architecture can be examined under a microscope. Although these images can reveal crucial diagnostic features, each section represents only a tiny fraction of the original specimen—often less than 1% of the total biopsy volume. A lesion may therefore be missed if it lies between sampled sections or if its most important features are not captured in the selected planes. This problem becomes especially significant when disease is spatially patchy, with abnormal regions distributed unevenly throughout a specimen.

Three-dimensional pathology seeks to overcome that sampling limitation by imaging tissue throughout its depth without destroying it. One of the technologies enabling this shift is open-top light-sheet microscopy, which illuminates tissue with a thin sheet of light and captures fluorescence or other optical signals across broad areas. After suitable preparation, large clinical specimens can be scanned in three dimensions at high resolution, producing a volumetric map in which cells, glands and other structures can be examined across many consecutive depths. The result is more comprehensive than a conventional slide series, but it also creates a formidable data-management problem: a single tissue volume may contain hundreds or thousands of potential viewing levels, each resembling a digital pathology slide.

TRICARE addresses this challenge by assigning a risk score to every two-dimensional level within a 3D tissue volume. The system is based on deep learning, a class of machine-learning methods that uses multiple layers of artificial neural networks to recognize patterns in complex data. In this case, the model does not evaluate each image as an isolated snapshot. Instead, it incorporates information from a selected group of neighbouring depth levels, allowing it to interpret local three-dimensional context. That distinction is technically important because many pathological structures extend across several planes. A suspicious gland, fragmented lesion or evolving tissue boundary may appear ambiguous in one section but become much clearer when adjacent levels are considered together.

The researchers compared this context-aware strategy with models that make predictions from individual 2D levels alone. According to the study, TRICARE performed better than approaches that ignored neighbouring sections. The advantage reflects a basic property of biological tissue: disease is not distributed as a collection of unrelated flat images. Tumours and precancerous changes have continuity, shape and spatial relationships that unfold through depth. By examining a limited neighbourhood around each level, the model can use that continuity as evidence, potentially reducing errors caused by folds, artifacts, staining variation or anatomically misleading views. The system can then rank the full stack of sections, placing the most concerning levels at the front of a pathologist’s review queue.

The team evaluated the framework in two clinically important settings. The first involved prostate cancer biopsies, where disease can be small, irregularly distributed and difficult to characterize from limited sampling. In that use case, TRICARE was used for risk stratification, separating tissue levels according to the likelihood that they contained high-risk pathological features. The second focused on endoscopic biopsies from patients with Barrett’s esophagus, a condition in which the lining of the esophagus changes and can progress to dysplasia or cancer. Detecting these changes is a central goal of surveillance, yet abnormal areas may be sparse and easily overlooked when tissue is assessed through conventional sampling. In both applications, the researchers report that AI-assisted review of 3D pathology showed potential to improve the identification of high-risk disease compared with standard slide-based workflows.

The proposed workflow is deliberately conservative. TRICARE does not issue a final diagnosis or remove the pathologist from the process. Instead, it acts as a filter for an enormous image set, highlighting the levels that deserve priority and allowing clinicians to examine those areas first. A pathologist could then review the flagged sections, inspect surrounding tissue, and make a diagnosis using clinical information and professional judgment. This design may be especially important for the introduction of AI into pathology, a field in which false negatives can have serious consequences and unexplained automated decisions can be difficult to accept. By retaining a human expert at the end of the chain, the technology offers a lower-risk route toward using 3D datasets in routine assessment.

The potential workload reduction is substantial. If a 3D scan contains hundreds of relevant levels, reviewing every one with equal attention could turn a single case into a lengthy, impractical task. A triage system can transform that problem into a ranked investigation, concentrating human effort on the most informative regions while still preserving access to the complete volume. This does not mean that unflagged areas would automatically be considered normal. Rather, the model could help clinicians decide where to begin and which parts of the specimen require especially careful evaluation. In a future clinical laboratory, such prioritization could make comprehensive imaging more compatible with the time pressures of diagnostic medicine.

Important challenges remain before systems such as TRICARE can become routine clinical tools. Deep-learning models must be tested across different hospitals, scanners, staining protocols, tissue-preparation methods and patient populations. Their performance must also be measured not only by technical accuracy but by whether they genuinely improve diagnostic sensitivity, reduce missed lesions and fit safely into existing laboratory workflows. Three-dimensional pathology itself requires specialized imaging, data storage and computational infrastructure, while pathologists will need interfaces that make volumetric information intuitive rather than overwhelming. Even so, the study points toward a new model of pathology in which diagnosis is not restricted to a handful of thin slices. By combining comprehensive tissue imaging with spatially aware AI triage, TRICARE suggests that the future of biopsy analysis may involve seeing more of the specimen while asking humans to spend their time where it matters most.

Subject of Research: Deep-learning triage of three-dimensional pathology datasets for efficient detection and assessment of high-risk tissue regions.

Article Title: Deep-learning triage of three-dimensional pathology datasets for comprehensive and efficient pathologist assessments.

Article References: Gao, G., Yan, R., Song, A.H. et al. “Deep-learning triage of three-dimensional pathology datasets for comprehensive and efficient pathologist assessments.” Nature Biomedical Engineering (2026). https://doi.org/10.1038/s41551-026-01760-1

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41551-026-01760-1

Keywords: 3D pathology, artificial intelligence, deep learning, digital pathology, light-sheet microscopy, cancer detection, prostate cancer, Barrett’s esophagus, dysplasia, pathology triage

The study was conducted against the backdrop of the ring trial, wherein 21 sarcoma centers provided a platform for benchmarking AI against seasoned professionals. As the biomedical community engages more with AI algorithms, many are left pondering: Can an AI outsmart a panel of human experts when faced with complex cancer cases? The idea of AI-driven diagnostics and treatment planning is not entirely new; however, this study represents a methodical investigation into its actual capabilities in real-world applications.

In this experimental setup, the researchers leveraged advanced algorithmic strategies inherent in LLMs, which are designed to interpret vast amounts of medical literature and patient data. By harnessing these sophisticated AI models, they aimed to replicate the decision-making processes typically executed by tumor boards, who often base their diagnoses and treatment recommendations on collective knowledge and experience. The delegation of such responsibilities to AI introduces fascinating possibilities and raises significant ethical questions.

One of the most striking revelations from this comparative study was not just the performance of LLMs in diagnostic accuracy, but how they processed information. Unlike human oncologists, who consider the nuances of patient history and context, AI tends to operate strictly on the data provided. This difference highlights a critical gap between AI capabilities and human faculties. AI’s potential lies significantly in its ability to analyze data patterns rapidly; however, it lacks the empathetic understanding and holistic view that seasoned oncologists bring to the table.

The results of the study indicated that while LLMs achieved commendable accuracy in some diagnostic domains, there were instances where their recommendations diverged from human consensus. Correlations between certain tumor characteristics and treatment efficacy were not as apparent to the AI as they were to human experts, revealing limitations in the way AI interprets innovative and dynamic medical scenarios. These disparities raise essential questions regarding the reliability of AI in oncology and the subsequent impact on patient care.

As healthcare systems worldwide gradually incorporate AI technologies, the findings from this benchmarking study can serve as a guiding light for future advancements. They underscore the necessity for a collaborative framework wherein humans and AI coexist rather than compete. In this envisioned future, the strengths of both can complement each other, leading to optimized treatment protocols and improved patient outcomes. Such a synergy may very well redefine clinical practices and therapeutic approaches in oncology.

Moreover, the researchers noted that transparency in AI decision-making processes will be crucial for gaining the trust of healthcare professionals. Developing an AI system that not only provides answers but also explains its reasoning is paramount. If oncologists can comprehend how an AI arrives at its recommendations, they are more likely to embrace its guidance. This goes beyond mere functionality; it’s about fostering a relationship where doctors feel empowered by AI assistance rather than threatened by it.

One of the areas ripe for further investigation arising from this study is how to enhance LLMs’ learning modalities. As algorithms continue to evolve, integrating experiential learning that includes patient interactions may be critical. Such advancements could enable LLMs to better understand context, subtleties, and patient-specific variables, bridging the divide between human intuition and machine logic. Investing in the convergence of machine learning and practical clinical application could significantly enrich the capabilities of AI in oncology.

The ethical ramifications of implementing AI in national healthcare frameworks are vast and require careful consideration. As AI systems take on more responsibility in clinical environments, issues regarding accountability, decision-making hierarchy, and patient confidentiality arise. This is especially pertinent in oncology, where treatment choices can be life-altering. Ensuring that AI complements rather than replaces human judgment will be essential in developing patient-centered practices.

As the biomedical arena moves towards incorporating AI models into daily practice, substantial work remains to be done in refining these technologies. The continuing development of more nuanced and capable language models could one day lead to remarkable advancements that parallel human expertise. The aim of achieving an inseparable partnership where AI augments human capability rather than competes with it is the ultimate goal.

The advancements observed in this research illuminate the importance of interdisciplinary collaboration, not just within medical teams but also amongst technologists, ethicists, and policymakers. As we venture into uncharted territories, the collective insights brought by varied professions will be indispensable in creating robust guidelines that govern AI usage in healthcare.

In response to the challenge presented by large language models, the oncology field is at a crossroads. The potential of AI is palpable, yet caution and thorough evaluation must accompany this enthusiasm. Only by striking a balance between embracing technological progress and safeguarding patient welfare can we ensure that AI contributes positively to the practice of medicine.

Ultimately, this research presents an intriguing glimpse into the future of healthcare, where artificial intelligence is integrated thoughtfully alongside human expertise. It serves as a reminder that while technology progresses at an astonishing rate, the essence of medicine—understanding, empathy, and nuanced decision-making—remains an irreplaceable component of patient care. The dialogue initiated by these findings will undoubtedly cultivate further exploration and refinement in the interplay between human and machine in oncology.

As the study unfolds, the implications for education, training, and the future workforce in medicine become increasingly clear. Preparing the next generation of oncologists to work alongside AI will be imperative. Education systems must evolve to equip future doctors with not just knowledge but also the skills needed to partner with technology effectively. This collaborative ethos will ensure that patient care remains at the forefront as new tools emerge.

In conclusion, the ambitious research conducted by Li, Kalisa, and Roohani opens up essential discussions on the intersection of AI and medical practice. The journey to harnessing the power of large language models in oncology is just beginning, with infinite potential ahead. However, the commitment to maintaining the compassionate essence of medicine must remain unwavering as we tread further into this transformative age.

Subject of Research: AI in Oncology Decision-Making

Article Title: The imitation game: large language models versus multidisciplinary tumor boards: benchmarking AI against 21 sarcoma centers from the ring trial.

Article References:

Li, CP., Kalisa, A.T., Roohani, S. et al. The imitation game: large language models versus multidisciplinary tumor boards: benchmarking AI against 21 sarcoma centers from the ring trial.
J Cancer Res Clin Oncol 151, 248 (2025). https://doi.org/10.1007/s00432-025-06304-9

Image Credits: AI Generated

DOI: 10.1007/s00432-025-06304-9

Keywords: Artificial Intelligence, Oncology, Large Language Models, Tumor Boards, Sarcoma, Patient Care, Medical Ethics, Collaboration, Machine Learning

The announcement of MARQO, which stands for Multi-Analytical Robust Quantitative Observation, emerges from research recently published in the esteemed journal Nature Biomedical Engineering. The study outlines MARQO’s capabilities in extracting detailed cellular and spatial information from whole-slide images of tumor tissues. This development signals a leap forward in the integration of computational power with traditional pathology, enabling more nuanced examinations of cancerous samples than ever before.

Historically, the process of analyzing stained tissue sections is labor-intensive and often limited in scope. Pathologists typically examine small areas under a microscope, making it challenging to see the broader picture of tumor composition and organization. MARQO addresses this limitation by permitting the analysis of entire slides without the need for manual segmentation into smaller patches. This innovation not only saves time but also improves the accuracy of analyses, which is critical for ensuring precise interpretations and diagnoses.

One of the remarkable features of MARQO is its adaptability to multiple staining techniques. It supports common immunohistochemistry (IHC) and immunofluorescence (IF) staining methods, which are staples in cancer research for identifying and localizing markers in tissues. The ability to unify analyses across different staining protocols enhances reproducibility, making it easier for researchers to draw comparisons across studies and improve the reliability of their findings.

In addition to speed and versatility, MARQO incorporates sophisticated algorithms that automatically detect likely positive cells within the tissue sample. By flagging these cells and recording their precise coordinates and marker intensities, MARQO creates a structured dataset that pathologists can then validate. This combination of automated processing and human expertise fosters a more efficient workflow, allowing medical professionals to focus on interpreting the data and uncovering insights rather than getting bogged down by the tedious aspects of analysis.

Dr. Sacha Gnjatic, the lead researcher and a prominent figure in immunology and immunotherapy at Mount Sinai, emphasized the importance of MARQO in filling a crucial gap in current pathology practices. He stated that the tool was specifically designed to streamline the conversion of complex whole-slide images into actionable, structured datasets swiftly and consistently. By automating the more demanding aspects of slide analysis, MARQO empowers experts to concentrate on what truly matters—their interpretative insights and the advancement of cancer research.

Despite its impressive capabilities, MARQO is still in the research phase and has not yet been validated for clinical diagnostics. However, its compatibility with widely accepted clinical staining methods suggests a promising future where MARQO could enhance routine pathology work. The research team behind MARQO has plans to continue developing the tool, aiming to improve its user interface and introduce advanced analytical capabilities that will facilitate large-scale studies involving immense volumes of digitized tissue slides.

The potential applications of MARQO extend beyond mere analysis. As a platform for biomarker discovery, it could revolutionize how researchers identify and evaluate potential targets for cancer therapies. With accurate data on cellular composition and spatial organization, oncologists and researchers would be better equipped to predict which patients are likely to benefit from specific treatments, thereby supporting the development of personalized medicine approaches.

The implications of MARQO are far-reaching, with the potential to enhance cancer diagnostics significantly. By resolving issues related to speed, accuracy, and user-friendliness in the analysis of whole-slide images, MARQO could change the standard operating procedures within pathology labs and, ultimately, the outcomes for cancer patients. The research team views this tool as a critical step toward more precise and efficient diagnostic processes in the ever-evolving landscape of cancer treatment.

As more focus is placed on technology’s role in healthcare, tools like MARQO come to the forefront, showcasing how artificial intelligence can augment human capabilities in medical settings. The careful integration of such technologies into pathology not only supports current research efforts but also lays the groundwork for future innovations in disease detection and management.

With continued development and validation, MARQO stands to become an indispensable asset in the fight against cancer, bridging the gap between traditional analytical methods and the advancing digital landscape of medicine. As it progresses toward larger studies and clinical applications, the scientific community remains optimistic about the possibilities that MARQO opens up for enhanced cancer tissue analysis.

In conclusion, the introduction of MARQO marks a pivotal moment in how pathologists can examine and interpret cancer tissues. By harnessing the power of AI, this new tool promises to deliver unprecedented speeds and accuracy in slide analysis, encouraging a new era of research that prioritizes rapid and refined insights into cancer biology. The pursuit of precision in cancer diagnostics takes a significant leap forward with innovations like MARQO, ultimately aiming to improve patient outcomes and advance treatment methodologies.

Subject of Research: Cells
Article Title: Multiparametric cellular and spatial organization in cancer tissue lesions with a streamlined pipeline
News Publication Date: 25-Aug-2025
Web References: Nature Biomedical Engineering
References: doi:10.1038/s41551-025-01475-9
Image Credits: Credit: Mount Sinai Health System

Keywords

• Imaging
• Cell pathology