According to current guideline, patients with resected specimens showing high-risk features are recommended additional surgery after local excision (LE) of T1 colorectal cancer, despite the low incidence of recurrence. However, surgical resection in patients with low rectal cancer (RC) is challenging and may compromise anal function, leading to a low quality of life. To reduce unnecessary surgical resection in these patients, we used artificial intelligence (AI) to develop and validate a prediction model for the risk of recurrence after LE.

We constructed an artificial neural network (ANN) to predict recurrence using pathological images from endoscopically or transanal surgically resected T1 RC specimens. Data were retrospectively obtained from two hospitals between 2001 and 2015. The model was constructed using 496 images obtained from the Guangdong Provincial People's Hospital (GDPH), and then validated using independent external datasets (150 images from Sun Yat-sen Memorial Hospital [SYSMH]) to verify its generalizability.

The ANN model yielded good discrimination, achieving areas under the receiver operating characteristic curves (AUC) of 0.979 in the training cohort (GDPH). The AUC for the validation cohort (SYSMH) was 0.978. More importantly, the AI-based prediction model avoided more than 34.9 % of unnecessary additional surgeries compared with the current US guideline in all enrolled patients.

We propose a novel ANN model for the risk of recurrence prediction in patients with T1 RC to provide physicians and patients guidance for decisions after LE. Furthermore, this may lead to a reduction in unnecessary invasive surgeries in patients with T1 RC.

In computational pathology, extracting and representing spatial features from gigapixel whole slide images (WSIs) are fundamental tasks, but due to their large size, WSIs are typically segmented into smaller tiles. A critical aspect of analyzing WSIs is how information across tiles is aggregated to predict outcomes such as patient prognosis. We introduce a model that combines a message-passing graph neural network (GNN) with a state space model (Mamba) to capture both local and global spatial relationships among the tiles in WSIs. The model's effectiveness was demonstrated in predicting progression-free survival among patients with early-stage lung adenocarcinomas (LUAD). We compared the model with other state-of-the-art methods for tile-level information aggregation in WSIs, including statistics-based, multiple instance learning (MIL)-based, GNN-based, and GNN-transformer-based aggregation. Our model achieved the highest c-index (0.70) and has the largest number of parameters among comparison models yet maintained a short inference time. Additional experiments showed the impact of different types of node features and different tile sampling strategies on model performance. Code: https://github.com/rina-ding/gat-mamba .

Artificial intelligence (AI) holds the potential to fundamentally transform how public health authorities use risk communication, community engagement and infodemic management (RCCE-IM) to prepare for, manage and mitigate public health emergencies. As research on this crucial transformation remains limited, we conducted a modified Delphi study on the impact of AI on RCCE-IM.

In two successive surveys, 54 experts-scholars with expertise in public health, digital health, health communication, risk communication and AI, as well as RCCE-IM professionals-from 27 countries assessed opportunities, challenges and risks of AI, anticipated future scenarios, and identified principles and actions to facilitate the responsible use of AI. The first Delphi round followed an open, exploratory approach, while the second sought to prioritise and rank key findings from the initial phase. Qualitative thematic analysis and statistical methods were applied to evaluate responses.

According to the expert panel, AI could be highly beneficial, particularly for risk communication (eg, tailoring messages) and infodemic management (eg, social listening), while its utility for fostering community engagement was viewed more critically. Challenges and risks affect all three components of RCCE-IM equally, with algorithmic bias and privacy breaches being of particular concern. Panellists anticipated both optimistic (eg, democratisation of information) and pessimistic (eg, erosion of public trust) future scenarios. They identified seven principles for the responsible use of AI for public health practices, with equity and transparency being the most important. Prioritised actions ranged from regulatory measures, resource allocation and feedback loops to capacity building, public trust initiatives and educational training.

To responsibly navigate the multifaceted opportunities, challenges and risks of AI for RCCE-IM in public health emergencies, clear guiding principles, ongoing critical evaluation and training as well as societal collaboration across countries are needed.

The role of artificial intelligence (AI) in cancer drug discovery and development has garnered significant attention due to its potential to transform the traditionally time-consuming and expensive processes involved in bringing new therapies to market. AI technologies, such as machine learning (ML) and deep learning (DL), enable the efficient analysis of vast datasets, facilitate faster identification of drug targets, optimization of compounds, and prediction of clinical outcomes. This review explores the multifaceted applications of AI across various stages of cancer drug development, from early-stage discovery to clinical trial design, development. In early-stage discovery, AI-driven methods support target identification, virtual screening (VS), and molecular docking, offering precise predictions that streamline the identification of promising compounds. Additionally, AI is instrumental in de novo drug design and lead optimization, where algorithms can generate novel molecular structures and optimize their properties to enhance drug efficacy and safety profiles. Preclinical development benefits from AI's predictive modeling capabilities, particularly in assessing a drug's toxicity through in silico simulations. AI also plays a pivotal role in biomarker discovery, enabling the identification of specific molecular signatures that can inform patient stratification and personalized treatment approaches. In clinical development, AI optimizes trial design by leveraging real-world data (RWD), improving patient selection, and reducing the time required to bring new drugs to market. Despite its transformative potential, challenges remain, including issues related to data quality, model interpretability, and regulatory hurdles. Addressing these limitations is critical for fully realizing AI's potential in cancer drug discovery and development. As AI continues to evolve, its integration with other technologies, such as genomics and clustered regularly interspaced short palindromic repeats (CRISPR), holds promise for advancing personalized cancer therapies. This review provides a comprehensive overview of AI's impact on the cancer drug discovery and development and highlights future directions for this rapidly evolving field.

To develop an integrative radiopathomic model based on deep learning to predict overall survival (OS) in locally advanced nasopharyngeal carcinoma (LANPC) patients.

A cohort of 343 LANPC patients with pretreatment MRI and whole slide image (WSI) were randomly divided into training (n = 202), validation (n = 91), and external test (n = 50) sets. For WSIs, a self-attention mechanism was employed to assess the significance of different patches for the prognostic task, aggregating them into a WSI-level representation. For MRI, a multilayer perceptron was used to encode the extracted radiomic features, resulting in an MRI-level representation. These were combined in a multimodal fusion model to produce prognostic predictions. Model performances were evaluated using the concordance index (C-index), and Kaplan-Meier curves were employed for risk stratification. To enhance model interpretability, attention-based and Integrated Gradients techniques were applied to explain how WSIs and MRI features contribute to prognosis predictions.

The radiopathomics model achieved high predictive accuracy in predicting the OS, with a C-index of 0.755 (95 % CI: 0.673-0.838) and 0.744 (95 % CI: 0.623-0.808) in the training and validation sets, respectively, outperforming single-modality models (radiomic signature: 0.636, 95 % CI: 0.584-0.688; deep pathomic signature: 0.736, 95 % CI: 0.684-0.810). In the external test, similar findings were observed for the predictive performance of the radiopathomics, radiomic signature, and deep pathomic signature, with their C-indices being 0.735, 0.626, and 0.660 respectively. The radiopathomics model effectively stratified patients into high- and low-risk groups (P < 0.001). Additionally, attention heatmaps revealed that high-attention regions corresponded with tumor areas in both risk groups.

n: The radiopathomics model holds promise for predicting clinical outcomes in LANPC patients, offering a potential tool for improving clinical decision-making.

Combining molecular classification with clinicopathologic methods improves risk assessment and chooses therapies for endometrial cancer (EC). Detecting mismatch repair (MMR) deficiencies in EC is crucial for screening Lynch syndrome and identifying immunotherapy candidates. An affordable and accessible tool is urgently needed to determine MMR status in EC patients. We introduce MMRNet, a deep convolutional neural network designed to predict MMR-deficient EC from whole-slide images stained with hematoxylin and eosin. MMRNet demonstrates strong performance, achieving an average area under the receiver operating characteristic curve (AUROC) of 0.897, with a sensitivity of 0.628 and a specificity of 0.949 in internal cross-validation. External validation using three additional datasets results in AUROCs of 0.790, 0.807, and 0.863. Employing a human-machine fusion approach notably improves diagnostic accuracy. MMRNet presents an effective method for identifying EC cases for confirmatory MMR testing and may assist in selecting candidates for immunotherapy.

Gold standard for immunohistochemical analyses is the manual assessment by two specialist pathologists. This process is time-consuming, highly dependent on the respective evaluator and often difficult to reproduce. The use of image analysis software, such as ImageJ, QuPath, or CellProfiler, which employ machine learning and/or deep learning mechanisms to perform biomarker analyses, offers a potential solution to these problems. The objective of our study is to evaluate whether digital assessment using the open-source software QuPath is comparable to manual evaluation and to examine the inter-evaluator variability between the two manual evaluators and two software-based evaluations. Six tissue microarrays (TMAs) were constructed for a cohort of 309 patients with primary oral squamous cell carcinoma (OSCC). The tumor tissue and corresponding non-lesional squamous epithelial mucosa specimen were immunohistochemically stained for the biomarkers Ki67, as a nuclear marker; the epidermal growth factor receptor (EGF-R), as a membranous marker; and the major histocompatibility complex class I (MHC-I) heavy chain (HC) expressed on the membrane and in the cytoplasm. The staining pattern was analyzed by two experienced, independent manual evaluators and by QuPath. The percentage of positive cells, for Ki67, and the histoscore (H-score) based on the percentage of positive cells and their staining intensity, for EGF-R and MHC-I, were determined as final values. The results yielded high to excellent spearman correlation coefficients for all three biomarkers (p<0.001) in lesional and non-lesional tissues. The Bland-Altman plots demonstrated a high degree of agreement between manual and software-based analysis, as well as inter-evaluator variability demonstrating a high comparability of the evaluation methods. However, a prerequisite for a proper software-based analysis is an accurate, time-consuming annotation of the single specimen, which requires users with a comprehensive understanding of histology and extensive training in QuPath. Once these requirements are met, the software-based analysis offers advantages for large-scale biomarker studies due to objective and reproducible comparability of the stainings leading to a greater accuracy as well as the reuse of established conditions across similar analyses without requiring further operator input.

Measuring the chromatin state of a tumor provides a powerful map of its epigenetic commitments; however, as these are generally bulk measurements, it has not yet been possible to connect changes in chromatin accessibility to the pathological signatures of complex tumors. In parallel, recent advances in computational pathology have enabled the identification of spatial features and immune cells within oral cavity tumors and their microenvironment.

Here, we present pathogenomic fingerprinting (PaGeFin), a novel method that integrates morphological tumor features with chromatin states using ATAC-seq. This framework links spatial morphologic and epigenetic features, offering insights into tumor progression and immune evasion within and across tumors. Morphologic features describing spatial relationships between tumor and lymphocyte cells that are prognostic of oral cavity squamous cell carcinoma (OSCC) were identified through AI-driven pathology analysis. These pathomic features were spatially colocalized within the epigenome of 4 distinct sections of 4 OSCC tumors.

These key features pinpointed chromatin regions responsible for critical immune cell function through peak locations and enrichment analysis, highlighting loci of CD27+ memory B cells, helper CD4+ T cells, and cytotoxic CD8 naïve T cells that likely drive morphologic changes in the distribution of lymphocytes in the tumor microenvironment and promote aggressive tumor behavior. Gene Ontology analysis revealed that the CTLA4, CD79A, CD3D, and CCR7 genes were embedded in these regions.

This computational approach is the first to assess the correlation between pathomic and epigenetic features in the context of cancer.

Spatial transcriptomics has transformed our understanding of tissue architecture by preserving the spatial context of gene expression patterns. Simultaneously, advances in imaging AI have enabled extraction of morphological features describing the tissue. This review introduces a framework for categorizing methods that combine spatial transcriptomics with tissue morphology, focusing on either translating or integrating morphological features into spatial transcriptomics. Translation involves using morphology to predict gene expression, creating super-resolution maps or inferring genetic information from H&E-stained samples. Integration enriches spatial transcriptomics by identifying morphological features that complement gene expression. We also explore learning strategies and future directions for this emerging field.

Tumor-stroma ratio (TSR), which is the area ratio between two components within tumor beds, namely tumor cells and tumor stroma, has been suggested as a promising prognostic feature in breast cancers. However, due to imperfect datasets, and the similarity between tumor stroma and non-tumor stroma, previous algorithms struggle to delineate tumor beds, especially those of histomorphologies with a fibrotic focus. To overcome these limitations, we propose a novel ray-aided quadruple affiliation network (RQA-Net) for calculating TSRs in breast cancers. RQA-Net uses quadruple branches to segment tumor cells and tumor beds simultaneously, where a crisscross task subtraction module (CTS-Module) is designed to locate tumor stroma, grounded on its affiliation relationships with tumor beds. Moreover, we propose an affiliation loss (Aff-Loss) to force identified tumor beds to incorporate tumor cells to enhance their affiliation relationships. Furthermore, we propose a ray-based hypothesis testing (RH-Testing) to obtain line segments from ray equations in tumor beds that can decorate identified tumor beds by overlapping. In summary, RQA-Net precisely predicts tumor cells and tumor beds, and thus supports the calculation of TSRs. We also create a cancerous dataset (CrD-Set) containing 100 slides with an average resolution of $50,000\times 50,000$ pixels from real breast cancer cases, which is the first dataset with pixel-wise tumor bed annotations. Experimental results on existing datasets and CrD-Set demonstrate that compared with previous methods, RQA-Net better calculates breast cancer TSRs by precisely identifying tumor cells and tumor beds. The created CrD-Set and codes in this work will be available online at https://github.com/Kunpingyang1992/Breast-Cancer-TSR-Calculation.

Diagnostic certainty is the cornerstone of modern medicine and critical for maximal treatment benefit. When evaluating biomarker expression by immunohistochemistry (IHC), however, pathologists are hindered by complex scoring methodologies, unique positivity cut-offs and subjective staining interpretation. Artificial intelligence (AI) can potentially eliminate diagnostic uncertainty, especially when AI "trustworthiness" is proven by expert pathologists in the context of real-world clinical practice. Building on an IHC foundation model, we employed pathologists-in-the-loop finetuning to produce a programmed cell death ligand 1 (PD-L1) CPS AI Model. We devised a multi-head augmented reality microscope (ARM) system overlayed with the PD-L1 CPS AI Model to assess interobserver variability and gauge the pathologists' trust in AI model outputs. Using difficult to interpret regions on gastroesophageal biopsies, we show that AI-assistance improved case agreement between any 2 pathologists by 14% (agreement on 77% vs 91%) and among 11 pathologists by 26% (agreement on 43% vs 69%). At a clinical cutoff of PD-L1 CPS ≥ 5, the number of cases diagnosed as positive by all 11 pathologists increased by 31%. Our findings underscore the benefits of fully engaging pathologists as active participants in the development and deployment of IHC AI models and frame the roadmap for trustworthy AI as a bridge to increased adoption in routine pathology practice.

Whole slide pathological images have shown significant potential for patient prognostication. Graph representation learning provides a robust framework for in-depth analysis of whole-slide images to construct predictive models. In this study, we introduce SurvGraph, an innovative graph-based deep learning network designed for gastric cancer survival prediction using whole slide pathological images. SurvGraph employs a hybrid graph construction approach that integrates multiple feature types, including color, texture, and deep learning features extracted from the pathological images to build node representations. SurvGraph utilizes a multi-head attention graph network, which performs survival prediction based on the graph structure. We evaluate the SurvGraph model on a large dataset of 708 gastric cancer patients from three independent cohorts for overall survival prediction. To assess the impact of various feature sets, we examine their performance when used individually and in combination. With five-fold cross-validation, our results demonstrate that the SurvGraph model achieves an average concordance index (C-index) of 0.706 with a standard deviation (SD) of 0.019. The proposed SurvGraph model has also attained a C-index of 0.708 (SD = 0.040) in the external testing set. In addition to baseline comparisons, we conducted a comprehensive benchmarking study comparing SurvGraph against established graph neural network architectures and multiple instance learning-based deep learning frameworks. The results indicate that the SurvGraph model outperforms the compared prediction models, suggesting its potential as a valuable tool for enhancing gastric cancer prognosis estimation.

Immune checkpoint inhibitors (ICIs) have revolutionized cancer treatment by significantly improving the efficacy of treatments and tolerability for patients with non-small cell lung cancer (NSCLC). However, even after meticulous selection based on molecular criteria, only 20-30% of the patients respond to ICIs. This highlights the urgent clinical need to develop more precise biomarkers to better identify individuals who will benefit from these expensive therapies.

Data from NSCLC patients treated with immunotherapy were collected from two institutions. From the histological images of tumors, pathomics features were extracted. We employed six machine learning models and seven feature selection methods to predict expression of the programmed death-ligand 1 (PD-L1), a current biomarker used to select patients for immunotherapy, and progression-free survival (PFS). The association between pathomics features and biological pathways was explored to validate pathomics-based signatures. We performed gene set enrichment analysis to identify the pathways enriched with the predictive signatures.

Handcrafted histological features were extracted from the whole slide images (WSI). The Support Vector Machines model with the SurfStar feature selection method, offered the best results, with an area under the curve (AUC) of around 0.66 for both the training and validation sets to predict PD-L1. For the prediction of PFS, the most effective model was linear discriminant analysis using the Multi Surf feature selection method with an AUC of 0.71 for the training set and 0.62 for the validation set. We found immune pathways to be upregulated in the high PD-L1 and high PFS groups, confirming the utility of image analysis for predicting clinical endpoints in patients treated with immunotherapy.

Our models, based on the analysis of histological images, can serve as predictive biomarkers for PD-L1 and PFS. This approach, focused on histological images, enables the distinction of patients based on treatment response, thus providing clinicians with a valuable tool for patient management. With further validation on external cohorts, these models could enhance clinical decision-making through analysis of routine medical images.

Epidermal growth factor receptor (EGFR) mutations play a pivotal role in guiding targeted therapy for lung cancer, making their accurate detection essential for personalized treatment. Recently, artificial intelligence (AI) has emerged as a promising tool for identifying EGFR mutation status from digital pathology images. This systematic review and meta-analysis evaluate the diagnostic accuracy of AI models in predicting EGFR mutations from whole slide images (WSIs) in lung cancer patients.

A comprehensive search was conducted across four databases (EMBASE, PubMed, Web of Science, and Scopus) for studies published up to June 20th, 2024. Studies employing AI algorithms, including machine learning and deep learning techniques, to predict EGFR mutations from digital pathology images were included. The risk of bias and applicability concerns were assessed using the QUADAS-AI tool. Diagnostic accuracy metrics such as sensitivity, specificity, and the Area Under the Curve (AUC) were extracted. Random-effects models were applied to synthesize the AI model performance. This study is registered with PROSPERO (CRD42024570496).

Out of 1,828 identified studies, 16 met the inclusion criteria, with 4 eligible for meta-analysis. The pooled results demonstrated that AI algorithms achieved an AUC of 0.756 (95% CI: 0.669-0.824), a sensitivity of 66.3% (95% CI: X-Y), and a specificity of 68.1% (95% CI: X-Y). Subgroup analyses revealed that AI models using in-house datasets, surgical resection samples, the ResNet architecture, and tumor-focused regions exhibited improved predictive performance.

AI models exhibit potential for non-invasive prediction of EGFR mutations in lung cancer patients using WSIs, although current accuracy and precision warrant further refinement. Future research should aim to enhance AI algorithms, validate findings on larger datasets, and integrate these tools into clinical workflows to optimize lung cancer management.

Integrating multimodal data, such as pathology images and genomics, is crucial for understanding cancer heterogeneity, personalized treatment complexity, and enhancing survival prediction. However, most current prognostic methods are limited to a single domain of histopathology or genomics, inevitably reducing their potential for accurate patient outcome prediction. Despite advancements in the concurrent analysis of pathology and genomic data, existing approaches inadequately address the intricate intermodal relationships.

This paper introduces the CPathomic method for multimodal data-based survival prediction. By leveraging whole slide pathology images to guide local pathological features, the method effectively mitigates significant intermodal differences through a cross-modal representational contrastive learning module. Furthermore, it facilitates interactive learning between different modalities through cross-modal and gated attention modules.

The extensive experiments on five public TCGA datasets demonstrate that CPathomic framework effectively bridges modality gaps, consistently outperforming alternative multimodal survival prediction methods.

The model we propose, CPathomic, unveils the potential of contrastive learning and cross-modal attention in the representation and fusion of multimodal data, enhancing the performance of patient survival prediction.

Large language models like ChatGPT have demonstrated potential in medical image interpretation, but their efficacy in liver histopathological analysis remains largely unexplored. This study aims to assess ChatGPT-4-vision's diagnostic accuracy, compared to liver pathologists' performance, in evaluating liver fibrosis (stage) in metabolic dysfunction-associated steatohepatitis.

Digitized Sirius Red-stained images for 59 metabolic dysfunction-associated steatohepatitis tissue biopsy specimens were evaluated by ChatGPT-4 and 4 pathologists using the NASH-CRN staging system. Fields of view at increasing magnification levels, extracted by a senior pathologist or randomly selected, were shown to ChatGPT-4, asking for fibrosis staging. The diagnostic accuracy of ChatGPT-4 was compared with pathologists' evaluations and correlated to the collagen proportionate area for additional insights. All cases were further analyzed by an in-context learning approach, where the model learns from exemplary images provided during prompting.

ChatGPT-4's diagnostic accuracy was 81% when using images selected by a pathologist, while it decreased to 54% with randomly cropped fields of view. By employing an in-context learning approach, the accuracy increased to 88% and 77% for selected and random fields of view, respectively. This method enabled the model to fully and correctly identify the tissue structures characteristic of F4 stages, previously misclassified. The study also highlighted a moderate to strong correlation between ChatGPT-4's fibrosis staging and collagen proportionate area.

ChatGPT-4 showed remarkable results with a diagnostic accuracy overlapping those of expert liver pathologists. The in-context learning analysis, applied here for the first time to assess fibrosis deposition in metabolic dysfunction-associated steatohepatitis samples, was crucial in accurately identifying the key features of F4 cases, critical for early therapeutic decision-making. These findings suggest the potential for integrating large language models as supportive tools in diagnostic pathology.

Bladder cancer, when diagnosed at an advanced stage, often necessitates inevitable invasive intervention. Consequently, non-invasive biosensor-based cancer detection and AI-based precision screening are being actively employed. However, the misclassification of cancer patients as normal-referred to as false negatives-remains a significant concern, as it could lead to fatal outcomes in lifespan. Moreover, while ensemble techniques such as soft voting and other methods can improve model accuracy and reduce misclassification, their effectiveness is limited and not applicable to all diagnostic tasks. Here, we developed a double stage cancer screening system that utilizes a sensitive urinary electrical biosensor implemented with an AI model and XAI interpretation tools. This system is designed for screening bladder cancer, well-known for its notable recurrence and high tendency to advance from non-invasive muscle tumors to muscle-invasive tumors. Four urinary biomarkers (CK8, CK18, PD-1, PD-L1) were measured by a field-effect transistor biosensor, and along with gender and age information, patients underwent initial screening by the CatBoost classification model. Patients initially classified as normal were reclassified using local explanations from neural networks offering a different perspective than CatBoost. After the second-stage screening, all of the false negatives from the initial screening could be correctly reclassified as cancer patients. Furthermore, global explanation guides the improvement of the AI model to be trained on an appropriate set of biomarker features to achieve high accuracy.

This study aims to develop a radiopathomics model based on preoperative ultrasound and fine-needle aspiration cytology (FNAC) images to enable accurate, non-invasive preoperative risk stratification for patients with papillary thyroid carcinoma (PTC). The model seeks to enhance clinical decision-making by optimizing preoperative treatment strategies.

A retrospective analysis was conducted on data from PTC patients who underwent thyroidectomy between October 2022 and May 2024 across six centers. Based on lymph node dissection outcomes, patients were categorized into high-risk and low-risk groups. Initially, a clinical predictive model was established based on the maximum diameter of the thyroid nodules. Radiomics features were extracted from preoperative two-dimensional ultrasound images, and pathomics features were extracted from 400x magnification H&E-stained tumor cell images from FNAC. The most predictive radiomics and pathomics features were identified through univariate analysis, Pearson correlation analysis and LASSO algorithm. The most valuable radiopathomics features were then selected by combining these predictive features. Finally, machine learning with the XGBoost algorithm was employed to construct radiomics, pathomics, and radiopathomics models. The performance of the models was evaluated using the area under the curve (AUC), decision curve analysis, accuracy, specificity, sensitivity, positive predictive value, and negative predictive value.

A total of 688 PTC patients were included, with 344 classified as intermediate/high-risk and 344 as low-risk. The multimodal radiopathomics model demonstrated excellent predictive performance, with AUCs of 0.886 (95% CI: 0.829-0.924) and 0.828 (95% CI: 0.751-0.879) in two external validation cohorts, significantly outperforming the clinical model (AUCs of 0.662 and 0.601), radiomics model (AUCs of 0.702 and 0.697), and pathomics model (AUCs of 0.741 and 0.712).

The radiopathomics model exhibits significant advantages in accurately predicting preoperative risk stratification in PTC patients. Its application is expected to reduce unnecessary lymph node dissection surgeries, optimize treatment strategies, and improve therapeutic outcomes.

The 2030 French Innovation Healthcare national plan defined "Digital Healthcare" as one of its priority. The deployment of digital pathology and artificial intelligence fits perfectly into this framework. Therefore we evaluated the pathologist's interest in those fields by conducting an online survey among members of the French Society of Pathology, mainly composed of senior pathologists and trainees (n=2301). We collected 123 answers originating nationwide and representative of all status of practice. A heavy majority of participants (83%) declared themselves "interested or very interested" in digital transition and AI. Twenty-six percent of participants took part in academic research projects and 20% in industrial research projects. With digital pathology becoming increasingly common nationwide, there is a growing interest of pathologists in the development and validation of AI algorithms. In order to support this dynamic, a commission for research on artificial intelligence has been created under the authority of the French Society of Pathology.

The increasing use of Deep Learning (DL) in healthcare has highlighted the critical need for improved transparency and interpretability. While Explainable Artificial Intelligence (XAI) methods provide insights into model predictions, reliability cannot be guaranteed by simply relying on explanations.

This position paper proposes the integration of Uncertainty Quantification (UQ) with XAI methods to improve model reliability and trustworthiness in healthcare applications.

We examine state-of-the-art XAI and UQ techniques, discuss implementation challenges, and suggest solutions to combine UQ with XAI methods. We propose a framework for estimating both aleatoric and epistemic uncertainty in the XAI context, providing illustrative examples of their potential application.

Our analysis indicates that integrating UQ with XAI could significantly enhance the reliability of DL models in practice. This approach has the potential to reduce interpretation biases and over-reliance, leading to more cautious and conscious use of AI in healthcare.

Ttyrosine kinase inhibitors (TKIs) represent the standard first-line treatment for patients with epidermal growth factor receptor (EGFR)-mutant lung adenocarcinoma. However, not all patients with EGFR mutations respond to TKIs. This study aims to develop a deep learning radiological-pathological-clinical (DLRPC) model that integrates computed tomography (CT) images, hematoxylin and eosin (H&E)-stained aspiration biopsy samples, and clinical data to predict the response in EGFR-mutant lung adenocarcinoma patients undergoing TKIs treatment.

We retrospectively analyzed data from 214 lung adenocarcinoma patients who received TKIs treatment from two medical centers between September 2013 and June 2023. The DLRPC model leverages paired CT, pathological images and clinical data, incorporating a clinical-based attention mask to further explore the cross-modality associations. To evaluate its diagnostic performance, we compared the DLRPC model against single-modality models and a decision level fusion model based on Dempster-Shafer theory. Model performances metrics, including area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), were used for evaluation. The Delong test assessed statistically significantly differences in AUC among models.

The DLRPC model demonstrated strong performance, achieving an AUC value of 0.8424. It outperformed the single-modality models (AUC = 0.6894, 0.7753, 0.8052 for CT model, pathology model and clinical model, respectively. P < 0.05). Additionally, the DLRPC model surpassed the decision level fusion model (AUC = 0.8132, P < 0.05).

The DLRPC model effectively predicts the response of EGFR-mutant lung adenocarcinoma patients to TKIs, providing a promising tool for personalized treatment decisions in lung cancer management.

Artificial intelligence (AI) is rapidly transforming the fields of pathology and oncology, offering novel opportunities for advancing diagnosis, prognosis, and treatment of cancer.

Through a systematic review-based approach, the representatives from the European Society for Medical Oncology (ESMO) Precision Oncology Working Group (POWG) and international experts identified studies in pathology and oncology that applied AI-based algorithms for tumour diagnosis, molecular biomarker detection, and cancer prognosis assessment. These findings were synthesised to provide a comprehensive overview of current AI applications and future directions in cancer pathology.

The integration of AI tools in digital pathology is markedly improving the accuracy and efficiency of image analysis, allowing for automated tumour detection and classification, identification of prognostic molecular biomarkers, and prediction of treatment response and patient outcomes. Several barriers for the adoption of AI in clinical workflows, such as data availability, explainability, and regulatory considerations, still persist. There are currently no prognostic or predictive AI-based biomarkers supported by level IA or IB evidence. The ongoing advancements in AI algorithms, particularly foundation models, generalist models and transformer-based deep learning, offer immense promise for the future of cancer research and care. AI is also facilitating the integration of multi-omics data, leading to more precise patient stratification and personalised treatment strategies.

The application of AI in pathology is poised to not only enhance the accuracy and efficiency of cancer diagnosis and prognosis but also facilitate the development of personalised treatment strategies. Although barriers to implementation remain, ongoing research and development in this field coupled with addressing ethical and regulatory considerations will likely lead to a future where AI plays an integral role in cancer management and precision medicine. The continued evolution and adoption of AI in pathology and oncology are anticipated to reshape the landscape of cancer care, heralding a new era of precision medicine and improved patient outcomes.

Pathology is the cornerstone of modern cancer care. With the advancement of precision oncology, the demand for histopathologic diagnosis and stratification of patients is increasing as personalized cancer therapy relies on accurate biomarker assessment. Recently, rapid development of whole slide imaging technology has enabled digitalization of traditional histologic slides at high resolution, holding promise to improve both the precision and efficiency of histopathologic evaluation. In particular, deep learning approaches, such as Convolutional Neural Network, Graph Convolutional Network, and Transformer, have shown great promise in enhancing the sensitivity and accuracy of whole slide image (WSI) analysis in cancer pathology because of their ability to handle high-dimensional and complex image data. The integration of deep learning models with WSIs enables us to explore and mine morphologic features beyond the visual perception of pathologists, which can help automate clinical diagnosis, assess histopathologic grade, predict clinical outcomes, and even discover novel morphologic biomarkers. In this review, we present a comprehensive framework for incorporating deep learning with WSIs, highlighting how deep learning-driven WSI analysis advances clinical tasks in cancer care. Furthermore, we critically discuss the opportunities and challenges of translating deep learning-based digital pathology into clinical practice, which should be considered to support personalized treatment of cancer patients.

Artificial intelligence (AI) has the potential to enhance clinical practice, particularly in the early and accurate diagnosis of diseases like breast cancer. However, for AI models to be effective in medical settings, they must not only be accurate but also interpretable and reliable. This study aims to analyze how variations in different model parameters affect the performance of a weakly supervised deep learning model used for breast cancer detection.

In this work, we apply Analysis of Variance (ANOVA) to investigate how changes in different parameters impact the performance of the deep learning model. The model is built using attention mechanisms, which both perform classification and identify the most relevant regions in medical images, improving the interpretability of the model. ANOVA is used to determine the significance of each parameter in influencing the model's outcome, offering insights into the specific factors that drive its decision-making.

Our analysis reveals that certain parameters significantly affect the model's performance, with some configurations showing higher sensitivity and specificity than others. By using ANOVA, we identify the key factors that influence the model's ability to classify images correctly. This approach allows for a deeper understanding of how the model works and highlights areas where improvements can be made to enhance its reliability in clinical practice.

The study demonstrates that applying ANOVA to deep learning models in medical applications provides valuable insights into the parameters that influence performance. This analysis helps make AI models more interpretable and trustworthy, which is crucial for their adoption in real-world medical environments like breast cancer detection. Understanding these factors enables the development of more transparent and efficient AI tools for clinical use.

The integration of advanced computational methods into precision medicine represents a transformative advancement in healthcare, enabling highly personalized treatment strategies based on individual genetic, environmental, and lifestyle factors. These methodologies have significantly enhanced disease diagnostics, genomic analysis, and drug discovery. However, rapid expansion in this field has resulted in fragmented understandings of its evolution and persistent knowledge gaps. This study employs a scientometric approach to systematically map the research landscape, identify key contributors, and highlight emerging trends in precision medicine. Methods: A scientometric analysis was conducted using data retrieved from the Scopus database, covering publications from 2019 to 2024. Tools such as VOSviewer and R-bibliometrix package (version 4.3.0) were used to perform co-authorship analysis, co-citation mapping, and keyword evolution tracking. The study examined annual publication growth, citation impact, research productivity by country and institution, and thematic clustering to identify core research areas. Results: The analysis identified 4574 relevant publications, collectively amassing 70,474 citations. A rapid growth trajectory was observed, with a 34.3% increase in publications in 2024 alone. The United States, China, and Germany emerged as the top contributors, with Harvard Medical School, the Mayo Clinic, and Sichuan University leading in institutional productivity. Co-citation and keyword analysis revealed three primary research themes: diagnostics and medical imaging, genomic and multi-omics data integration, and personalized treatment strategies. Recent trends indicate a shift toward enhanced clinical decision support systems and precision drug discovery. Conclusions: Advanced computational methods are revolutionizing precision medicine, spurring increased global research collaboration and rapidly evolving methodologies. This study provides a comprehensive knowledge framework, highlighting key developments and future directions. The insights derived can inform policy decisions, funding allocations, and interdisciplinary collaborations, driving further advancements in healthcare solutions.

The application of digital pathology tools has expanded in recent years, but non-neoplastic human brain tissue presents unique challenges due to its complexity. This study evaluated HALO and QuPath tau quantification performance in the hippocampus and mid-frontal gyrus across various tauopathies. Percent positivity emerged as the most reliable measure, showing strong correlations with Braak stages and CERAD scores, outperforming object and optical densities. QuPath demonstrated superior correlations with Braak stages, while HALO excelled in aligning with CERAD scoring. However, HALO's optical density was less consistent. Paired t-tests revealed significant differences in object and optical densities between platforms, though percent positivity was consistent across both. QuPath's threshold-based object density showed similar agreement with manual counts compared to HALO's AI-dependent approach (all ρ > 0.70). Reanalysis of QuPath further improved its agreement with manual measurements and correlations with Braak and CERAD scores (all ρ > 0.70). HALO offers a user-friendly interface and excels in certain metrics but is hindered by frequent software malfunctions and more limited flexibility. In contrast, QuPath's customizable workflows and superior performance in Braak staging make it more suitable for advanced and larger-scale analyses. Overall, our study highlights the strengths and limitations of these platforms, helping guide their application in neuropathology.

Lymphoma is a malignant tumor of the immune system and its incidence is increasing year after year, causing a major threat to people's health. Conventional diagnosis of lymphoma basically depends on histological images consuming long-time and tedious manipulations (e.g., 7-15 days) and large-field view (e.g., > 1000 × 1000 μm2). Artificial intelligence has recently revolutionized cancer diagnosis by training pathological image databases via deep learning. Current approaches, however, remain dependent on analyzing wide-field pathological images to detect distinct nuclear, cytologic, and histomorphologic traits for diagnostic categorization, limiting their applicability to minimally invasive lesion.

Herein, we develop a molecular imaging strategy for minimally invasive lymphoma diagnosis. By spreading lymphoma tissue sections tightly on a surface-enhanced Raman scattering (SERS) chip, label-free images of DNA double strand breaks (DSBs) in 30 × 30 μm2 tissue sections could be achieved in ~ 15 min. To establish a proof of concept, the Raman image datasets collected from clinical samples of normal lymphatic tissues and non-Hodgkin's lymphoma (NHL) tissues were well organized and trained in a deep convolutional neural network model, finally achieving a recognition rate of ~ 91.7 ± 2.1%.

The molecular imaging strategy for minimally invasive lymphoma diagnosis that can achieve a recognition rate of ~ 91.7 ± 2.1%. We anticipate that these results will catalyze the development of a series of histological SERS-AI technologies for diagnosing various diseases, including other types of cancer. In this work, we present a reliable tool to facilitate clinicians in the diagnosis of lymphoma.

Artificial intelligence is rapidly transforming quality assurance in healthcare, driving advancements in diagnostics, surgery, and patient care. This review presents a comprehensive analysis of artificial intelligence integration-particularly convolutional and recurrent neural networks-across key clinical domains, significantly enhancing diagnostic accuracy, surgical performance, and pathology evaluation. Artificial intelligence-based approaches have demonstrated clear superiority over conventional methods: convolutional neural networks achieved 91.56% accuracy in scanner fault detection, surpassing manual inspections; endoscopic lesion detection sensitivity rose from 2.3% to 6.1% with artificial intelligence assistance; and gastric cancer invasion depth classification reached 89.16% accuracy, outperforming human endoscopists by 17.25%. In pathology, artificial intelligence achieved 93.2% accuracy in identifying out-of-focus regions and an F1 score of 0.94 in lymphocyte quantification, promoting faster and more reliable diagnostics. Similarly, artificial intelligence improved surgical workflow recognition with over 81% accuracy and exceeded 95% accuracy in skill assessment classification. Beyond traditional diagnostics and surgical support, AI-powered wearable sensors, drug delivery systems, and biointegrated devices are advancing personalized treatment by optimizing physiological monitoring, automating care protocols, and enhancing therapeutic precision. Despite these achievements, challenges remain in areas such as data standardization, ethical governance, and model generalizability. Overall, the findings underscore artificial intelligence's potential to outperform traditional techniques across multiple parameters, emphasizing the need for continued development, rigorous clinical validation, and interdisciplinary collaboration to fully realize its role in precision medicine and patient safety.

Advances in the various applications of artificial intelligence will have important implications for medical training and practice. The advances in ChatGPT-4 alongside the introduction of the medical licensing assessment (MLA) provide an opportunity to compare GPT-4's medical competence against the expected level of a United Kingdom junior doctor and discuss its potential in clinical practice. Using 191 freely available questions in MLA style, we assessed GPT-4's accuracy with and without offering multiple-choice options. We compared single and multi-step questions, which targeted different points in the clinical process, from diagnosis to management. A chi-squared test was used to assess statistical significance. GPT-4 scored 86.3% and 89.6% in papers one-and-two respectively. Without the multiple-choice options, GPT's performance was 61.5% and 74.7% in papers one-and-two respectively. There was no significant difference between single and multistep questions, but GPT-4 answered 'management' questions significantly worse than 'diagnosis' questions with no multiple-choice options (p = 0.015). GPT-4's accuracy across categories and question structures suggest that LLMs are competently able to process clinical scenarios but remain incapable of understanding these clinical scenarios. Large-Language-Models incorporated into practice alongside a trained practitioner may balance risk and benefit as the necessary robust testing on evolving tools is conducted.

Programmed death-ligand 1 (PD-L1) is an important biomarker increasingly used as a predictive marker in breast cancer immunotherapy. Immunohistochemical quantification remains the standard method for assessment. However, it presents challenges related to time, cost, and reliability. Hematoxylin and eosin (H&E) staining is a routine method in cancer pathology, known for its accessibility and consistently reliability. Deep learning has shown the potential in predicting biomarkers in cancer histopathology. This study employs a weakly supervised multiple instance learning (MIL) approach to predict PD-L1 expression from H&E-stained images using deep learning techniques. In the internal test set, the TransMIL method achieved an area under the curve (AUC) of 0.833, and in an independent external test set, it achieved an AUC of 0.799. Additionally, since RNA sequencing results indicate a threshold that allows for the separation of H&E pathology images, we further validated our approach using the public TCGA-TNBC dataset, achieving an AUC of 0.721. These findings demonstrates that the Transformer-based TransMIL model can effectively capture highly heterogeneous features within the MIL framework, exhibiting strong cross-center generalization capabilities. Our study highlights that appropriate deep learning techniques can enable effective PD-L1 prediction even with limited data, and across diverse regions and centers. This not only underscores the significant potential of deep learning in pathological artificial intelligence (AI) but also provides valuable insights for the rational and efficient allocation of medical resources.

The management of colorectal carcinoma (CRC) relies on pathological interpretation. Digital pathology approaches allow for development of new potent artificial intelligence-based prognostic parameters. The study aimed to develop an artificial intelligence-based image analysis platform allowing fully automatized, quantitative, and explainable tumor microenvironment analysis and extraction of prognostic information from hematoxylin and eosin-stained whole-slide images of CRC patients. Three well--characterized, multi-institutional patient cohorts were included (patient n = 1438, whole-slide image n > 2400). The developed image analysis platform implements quality control and established algorithms to segment tissue and detect cell types. It enabled systematic analysis of immune infiltrate, assessing its prognostic relevance, intratumoral heterogeneity, and biological concepts across multiple survival end points. Analyzing single-cell types and their combinations reveals independent, prognostic parameters, highlighting significant intratumoral heterogeneity, especially in the biopsy setting, which must be accounted for. A key morphologic concept related to tumor control by the immune system is described, resulting in a capable, independent prognostic parameter (tumor "out of control"). Our findings have direct clinical implications and can be used as a foundation for updating the existing CRC grading systems.

The relative abundance of cancer-associated fibroblast (CAF) subtypes influences a tumor's response to treatment, especially immunotherapy. However, the gene expression signatures associated with these CAF subtypes have yet to realize their potential as clinical biomarkers. Here, we describe an interpretable machine learning approach, additive multiple instance learning (aMIL), to predict bulk gene expression signatures from hematoxylin and eosin-stained whole-slide images, focusing on an immunosuppressive LRRC15+ CAF-enriched TGFβ-CAF signature. aMIL models accurately predicted TGFβ-CAF across various cancer types. Tissue regions contributing most highly to slide-level predictions of TGFβ-CAF were evaluated by machine learning models characterizing spatial distributions of diverse cell and tissue types, stromal subtypes, and nuclear morphology. In breast cancer, regions contributing most to TGFβ-CAF-high predictions ("excitatory") were localized to cancer stroma with high fibroblast density and mature collagen fibers. Regions contributing most to TGFβ-CAF-low predictions ("inhibitory") were localized to cancer epithelium and densely inflamed stroma. Fibroblast and lymphocyte nuclear morphology also differed between excitatory and inhibitory regions. Thus, aMIL enables a data-driven link between histologic features and transcription, offering biological interpretability beyond typical black-box models.

The past decade has seen the introduction of artificial intelligence (AI)-based approaches aimed at optimizing several workflows across many medical specialties. In clinical oncology, the most promising applications include those involving image analysis, such as digital pathology. In this Perspective, we provide a comprehensive examination of the developments in AI in digital pathology between 2019 and 2024. We evaluate the current landscape from the lens of technological innovations, regulatory trends, deployment and implementation, reimbursement and commercial implications. We assess the technological advances that have driven improvements in AI, enabling more robust and scalable solutions for digital pathology. We also examine regulatory developments, in particular those affecting in-house devices and laboratory-developed tests, which are shaping the landscape of AI-based tools in digital pathology. Finally, we discuss the role of reimbursement frameworks and commercial investment in the clinical adoption of AI-based technologies. In this Perspective, we highlight both the progress and challenges in AI-driven digital pathology over the past 5 years, outlining the path forward for its adoption into routine practice in clinical oncology.

The response of the kidney after induction treatment is one of the determinants of prognosis in lupus nephritis, but effective predictive tools are lacking. Here, we sought to apply deep learning approaches on kidney biopsies for treatment response prediction in lupus nephritis. Patients who received cyclophosphamide or mycophenolate mofetil as induction treatment were included, and the primary outcome was 12-month treatment response, complete response defined as 24-h urinary protein under 0.5 g with normal estimated glomerular filtration rate or within 10% of normal range. The model development cohort included 245 patients (880 digital slides), and the external test cohort had 71 patients (258 digital slides). Deep learning models were trained independently on hematoxylin and eosin-, periodic acid-Schiff-, periodic Schiff-methenamine silver- and Masson's trichrome-stained slides at multiple magnifications and integrated to predict the primary outcome of complete response to therapy at 12 months. Single-stain models showed area under the curves of 0.813, 0.841, 0.823, and 0.862, respectively. Further, integration of the four models into a multi-stain model achieved area under the curves of 0.901 and 0.840 on internal validation and external testing, respectively, which outperformed conventional clinicopathologic parameters including estimated glomerular filtration rate, chronicity index and reduction in proteinuria at three months. Decisive features uncovered by visualization for model prediction included tertiary lymphoid structures, glomerulosclerosis, interstitial fibrosis and tubular atrophy. Our study demonstrated the feasibility of utilizing deep learning on kidney pathology to predict treatment response for lupus patients. Further validation is required before the model could be implemented for risk stratification and to aid in making therapeutic decisions in clinical practice.

The incidence of anal squamous cell carcinoma (SCC) has increased, and treatment has shifted from surgery to chemoradiotherapy (CRT), with salvage abdominoperineal resection being reserved for persistent/recurrent cases. This study evaluates the utility of different tumor regression scoring systems in predicting survival in anal SCC patients, using pathologists' observations and digital pathology. Data of cases managed surgically from 2005 to 2019 were collected. Residual tumor was assessed by multiple methods (gross tumor size, largest focus of tumor on hematoxylin and eosin (H&E) slide, average of residual tumor in all submitted H&E slides, Japanese Esophageal Society, Chirieac, Schneider, Hermann, and College of American Pathologists scoring system). Three expert pathologists individually estimated ("eyeballed") the residual tumor percentage based on residual tumor/tumor bed (single representative H&E slide). The QuPath software was used to measure tumor volume on the same slide. The American Joint Committee on Cancer eighth staging and outcome data were retrieved from electronic medical records. The study involved 48 participants, predominantly female (56%), with a median age of 57 years. Most were Caucasian. Human papillomavirus-positive was present in 77% of those assessed (17/22). Initial treatment included CRT, followed by abdominoperineal resection (79%) or pelvic exenteration (21%). Complications (13%), persistent disease (33%), and recurrence (54%) led to surgical interventions. Fifty-one percent had moderately differentiated SCC, whereas 42% were poorly differentiated. Lymphovascular invasion (44%), perineural invasion (38%), and lymph node metastasis (13%) were present. Distant metastasis was rare (2%). Median overall survival was 3.2 years. Positive margins (hazard ratio, 4.12; 95% CI, 1.83-9.28) and larger tumor size (hazard ratio, 1.02; 95% CI, 1.01-1.03) were associated with an increased hazard of death. Most residual tumor measurement methods were not significantly associated with overall survival. Interobserver agreement (based on "eyeballing") was moderate (kappa, 0.4). Computational pathology-based residual tumor percentage was the only method significantly associated with outcome, with each 10% increase in the residual tumor percentage corresponding to a 1.23-fold higher hazard death (95% CI, 1.03; 1.46; P = .024). This study highlights computational pathology's important role in predicting outcomes in anal SCC treated with CRT and surgery. Specifically, the computational assessment of the residual tumor percentage proves to be a strong predictor of overall survival, outperforming other established tumor regression scoring systems methods.

Breast cancer (BC) remains a significant threat to women's health worldwide. The oncology field had an exponential growth in the abundance of medical images, clinical information, and genomic data. With its continuous advancement and refinement, artificial intelligence (AI) has demonstrated exceptional capabilities in processing intricate multidimensional BC-related data. AI has proven advantageous in various facets of BC management, encompassing efficient screening and diagnosis, precise prognosis assessment, and personalized treatment planning. However, the implementation of AI into precision medicine and clinical practice presents ongoing challenges that necessitate enhanced regulation, transparency, fairness, and integration of multiple clinical pathways. In this review, we provide a comprehensive overview of the current research related to AI in BC, highlighting its extensive applications throughout the whole BC cycle management and its potential for innovative impact. Furthermore, this article emphasizes the significance of constructing patient-oriented AI algorithms. Additionally, we explore the opportunities and potential research directions within this burgeoning field.

Pathology departments generate large volumes of unstructured data as free-text diagnostic reports. Converting these reports into structured formats for analytics or artificial intelligence projects requires substantial manual effort by specialized personnel. While recent studies show promise in using advanced language models for structuring pathology data, they primarily rely on proprietary models, raising cost and privacy concerns. Additionally, important aspects such as prompt engineering and model quantization for deployment on consumer-grade hardware remain unaddressed.

We created a dataset of 579 annotated pathology reports in German and English versions. Six language models (proprietary: GPT-4; open-source: Llama2 13B, Llama2 70B, Llama3 8B, Llama3 70B, and Qwen2.5 7B) were evaluated for their ability to extract eleven key parameters from these reports. Additionally, we investigated model performance across different prompt engineering strategies and model quantization techniques to assess practical deployment scenarios.

Here we show that open-source language models extract structured data from pathology reports with high precision, matching the accuracy of proprietary GPT-4 model. The precision varies significantly across different models and configurations. These variations depend on specific prompt engineering strategies and quantization methods used during model deployment.

Open-source language models demonstrate comparable performance to proprietary solutions in structuring pathology report data. This finding has significant implications for healthcare institutions seeking cost-effective, privacy-preserving data structuring solutions. The variations in model performance across different configurations provide valuable insights for practical deployment in pathology departments. Our publicly available bilingual dataset serves as both a benchmark and a resource for future research.