Interpretable deep learning model to predict the molecular classification of endometrial cancer from haematoxylin and eosin-stained whole-slide images: a combined analysis of the PORTEC randomised trials and clinical cohorts

Prediction of recurrence risk in endometrial cancer with multimodal deep learning

Predicting distant recurrence of endometrial cancer (EC) is crucial for personalized adjuvant treatment. The current gold standard of combined pathological and molecular profiling is costly, hampering implementation. Here we developed HECTOR (histopathology-based endometrial cancer tailored outcome risk), a multimodal deep learning prognostic model using hematoxylin and eosin-stained, whole-slide images and tumor stage as input, on 2,072 patients from eight EC cohorts including the PORTEC-1/-2/-3 randomized trials. HECTOR demonstrated C-indices in internal (n = 353) and two external (n = 160 and n = 151) test sets of 0.789, 0.828 and 0.815, respectively, outperforming the current gold standard, and identified patients with markedly different outcomes (10-year distant recurrence-free probabilities of 97.0%, 77.7% and 58.1% for HECTOR low-, intermediate- and high-risk groups, respectively, by Kaplan-Meier analysis). HECTOR also predicted adjuvant chemotherapy benefit better than current methods. Morphological and genomic feature extraction identified correlates of HECTOR risk groups, some with therapeutic potential. HECTOR improves on the current gold standard and may help delivery of personalized treatment in EC.

Prediction of Mismatch Repair Status in Endometrial Cancer from Histological Slide Images Using Various Deep Learning-Based Algorithms

The application of deep learning algorithms to predict the molecular profiles of various cancers from digital images of hematoxylin and eosin (H&E)-stained slides has been reported in recent years, mainly for gastric and colon cancers. In this study, we investigated the potential use of H&E-stained endometrial cancer slide images to predict the associated mismatch repair (MMR) status. H&E-stained slide images were collected from 127 cases of the primary lesion of endometrial cancer. After digitization using a Nanozoomer virtual slide scanner (Hamamatsu Photonics), we segmented the scanned images into 5397 tiles of 512 × 512 pixels. The MMR proteins (PMS2, MSH6) were immunohistochemically stained, classified into MMR proficient/deficient, and annotated for each case and tile. We trained several neural networks, including convolutional and attention-based networks, using tiles annotated with the MMR status. Among the tested networks, ResNet50 exhibited the highest area under the receiver operating characteristic curve (AUROC) of 0.91 for predicting the MMR status. The constructed prediction algorithm may be applicable to other molecular profiles and useful for pre-screening before implementing other, more costly genetic profiling tests.

LncRNAs in necroptosis: Deciphering their role in cancer pathogenesis and therapy

Necroptosis, a controlled type of cell death that is different from apoptosis, has become a key figure in the aetiology of cancer and offers a possible target for treatment. A growing number of biological activities, including necroptosis, have been linked to long noncoding RNAs (lncRNAs), a varied family of RNA molecules with limited capacity to code for proteins. The complex interactions between LncRNAs and important molecular effectors of necroptosis, including mixed lineage kinase domain-like pseudokinase (MLKL) and receptor-interacting protein kinase 3 (RIPK3), will be investigated. We will explore the many methods that LncRNAs use to affect necroptosis, including protein-protein interactions, transcriptional control, and post-transcriptional modification. Additionally, the deregulation of certain LncRNAs in different forms of cancer will be discussed, highlighting their dual function in influencing necroptotic processes as tumour suppressors and oncogenes. The goal of this study is to thoroughly examine the complex role that LncRNAs play in controlling necroptotic pathways and how that regulation affects the onset and spread of cancer. In the necroptosis for cancer treatment, this review will also provide insight into the possible therapeutic uses of targeting LncRNAs. Techniques utilising LncRNA-based medicines show promise in controlling necroptotic pathways to prevent cancer from spreading and improve the effectiveness of treatment.

Predicting Mismatch Repair Deficiency Status in Endometrial Cancer through Multi-Resolution Ensemble Learning in Digital Pathology

For molecular classification of endometrial carcinoma, testing for mismatch repair (MMR) status is becoming a routine process. Mismatch repair deficiency (MMR-D) is caused by loss of expression in one or more of the 4 major MMR proteins: MLH1, MSH2, MSH6, PHS2. Over 30% of patients with endometrial cancer have MMR-D. Determining the MMR status holds significance as individuals with MMR-D are potential candidates for immunotherapy. Pathological whole slide image (WSI) of endometrial cancer with immunohistochemistry results of MMR proteins were gathered. Color normalization was applied to the tiles using a CycleGAN-based network. The WSI was divided into tiles at three different magnifications (2.5 × , 5 × , and 10 ×). Three distinct networks of the same architecture were employed to include features from all three magnification levels and were stacked for ensemble learning. Three architectures, InceptionResNetV2, EfficientNetB2, and EfficientNetB3 were employed and subjected to comparison. The per-tile results were gathered to classify MMR status in the WSI, and prediction accuracy was evaluated using the following performance metrics: AUC, accuracy, sensitivity, and specificity. The EfficientNetB2 was able to make predictions with an AUC of 0.821, highest among the three architectures, and an overall AUC range of 0.767 - 0.821 was reported across the three architectures. In summary, our study successfully predicted MMR classification from pathological WSIs in endometrial cancer through a multi-resolution ensemble learning approach, which holds the potential to facilitate swift decisions on tailored treatment, such as immunotherapy, in clinical settings.

Precision treatment in advanced hepatocellular carcinoma

The past decade has witnessed significant advances in the systemic treatment of advanced hepatocellular carcinoma (HCC). Nevertheless, the newly developed treatment strategies have not achieved universal success and HCC patients frequently exhibit therapeutic resistance to these therapies. Precision treatment represents a paradigm shift in cancer treatment in recent years. This approach utilizes the unique molecular characteristics of individual patient to personalize treatment modalities, aiming to maximize therapeutic efficacy while minimizing side effects. Although precision treatment has shown significant success in multiple cancer types, its application in HCC remains in its infancy. In this review, we discuss key aspects of precision treatment in HCC, including therapeutic biomarkers, molecular classifications, and the heterogeneity of the tumor microenvironment. We also propose future directions, ranging from revolutionizing current treatment methodologies to personalizing therapy through functional assays, which will accelerate the next phase of advancements in this area.

Morphology-based molecular classification of spinal cord ependymomas using deep neural networks

Based on DNA-methylation, ependymomas growing in the spinal cord comprise two major molecular types termed spinal (SP-EPN) and myxopapillary ependymomas (MPE(-A/B)), which differ with respect to their clinical features and prognosis. Due to the existing discrepancy between histomorphogical diagnoses and classification using methylation data, we asked whether deep neural networks can predict the DNA methylation class of spinal cord ependymomas from hematoxylin and eosin stained whole-slide images. Using explainable AI, we further aimed to prospectively improve the consistency of histology-based diagnoses with DNA methylation profiling by identifying and quantifying distinct morphological patterns of these molecular ependymoma types. We assembled a case series of 139 molecularly characterized spinal cord ependymomas (nMPE  = 84, nSP-EPN  = 55). Self-supervised and weakly-supervised neural networks were used for classification. We employed attention analysis and supervised machine-learning methods for the discovery and quantification of morphological features and their correlation to the diagnoses of experienced neuropathologists. Our best performing model predicted the DNA methylation class with 98% test accuracy and used self-supervised learning to outperform pretrained encoder-networks (86% test accuracy). In contrast, the diagnoses of neuropathologists matched the DNA methylation class in only 83% of cases. Domain-adaptation techniques improved model generalization to an external validation cohort by up to 22%. Statistically significant morphological features were identified per molecular type and quantitatively correlated to human diagnoses. The approach was extended to recently defined subtypes of myxopapillary ependymomas (MPE-(A/B), 80% test accuracy). In summary, we demonstrated the accurate prediction of the DNA methylation class of spinal cord ependymomas (SP-EPN, MPE(-A/B)) using hematoxylin and eosin stained whole-slide images. Our approach may prospectively serve as a supplementary resource for integrated diagnostics and may even help to establish a standardized, high-quality level of histology-based diagnostics across institutions-in particular in low-income countries, where expensive DNA-methylation analyses may not be readily available.

Prediction of anticancer drug resistance using a 3D microfluidic bladder cancer model combined with convolutional neural network-based image analysis

Bladder cancer is the most common urological malignancy worldwide, and its high recurrence rate leads to poor survival outcomes. The effect of anticancer drug treatment varies significantly depending on individual patients and the extent of drug resistance. In this study, we developed a validation system based on an organ-on-a-chip integrated with artificial intelligence technologies to predict resistance to anticancer drugs in bladder cancer. As a proof-of-concept, we utilized the gemcitabine-resistant bladder cancer cell line T24 with four distinct levels of drug resistance (parental, early, intermediate, and late). These cells were co-cultured with endothelial cells in a 3D microfluidic chip. A dataset comprising 2,674 cell images from the chips was analyzed using a convolutional neural network (CNN) to distinguish the extent of drug resistance among the four cell groups. The CNN achieved 95.2% accuracy upon employing data augmentation and a step decay learning rate with an initial value of 0.001. The average diagnostic sensitivity and specificity were 90.5% and 96.8%, respectively, and all area under the curve (AUC) values were over 0.988. Our proposed method demonstrated excellent performance in accurately identifying the extent of drug resistance, which can assist in the prediction of drug responses and in determining the appropriate treatment for bladder cancer patients.

AI-enabled organoids: Construction, analysis, and application

Organoids, miniature and simplified in vitro model systems that mimic the structure and function of organs, have attracted considerable interest due to their promising applications in disease modeling, drug screening, personalized medicine, and tissue engineering. Despite the substantial success in cultivating physiologically relevant organoids, challenges remain concerning the complexities of their assembly and the difficulties associated with data analysis. The advent of AI-Enabled Organoids, which interfaces with artificial intelligence (AI), holds the potential to revolutionize the field by offering novel insights and methodologies that can expedite the development and clinical application of organoids. This review succinctly delineates the fundamental concepts and mechanisms underlying AI-Enabled Organoids, summarizing the prospective applications on rapid screening of construction strategies, cost-effective extraction of multiscale image features, streamlined analysis of multi-omics data, and precise preclinical evaluation and application. We also explore the challenges and limitations of interfacing organoids with AI, and discuss the future direction of the field. Taken together, the AI-Enabled Organoids hold significant promise for advancing our understanding of organ development and disease progression, ultimately laying the groundwork for clinical application.

One label is all you need: Interpretable AI-enhanced histopathology for oncology

Artificial Intelligence (AI)-enhanced histopathology presents unprecedented opportunities to benefit oncology through interpretable methods that require only one overall label per hematoxylin and eosin (H&E) slide with no tissue-level annotations. We present a structured review of these methods organized by their degree of verifiability and by commonly recurring application areas in oncological characterization. First, we discuss morphological markers (tumor presence/absence, metastases, subtypes, grades) in which AI-identified regions of interest (ROIs) within whole slide images (WSIs) verifiably overlap with pathologist-identified ROIs. Second, we discuss molecular markers (gene expression, molecular subtyping) that are not verified via H&E but rather based on overlap with positive regions on adjacent tissue. Third, we discuss genetic markers (mutations, mutational burden, microsatellite instability, chromosomal instability) that current technologies cannot verify if AI methods spatially resolve specific genetic alterations. Fourth, we discuss the direct prediction of survival to which AI-identified histopathological features quantitatively correlate but are nonetheless not mechanistically verifiable. Finally, we discuss in detail several opportunities and challenges for these one-label-per-slide methods within oncology. Opportunities include reducing the cost of research and clinical care, reducing the workload of clinicians, personalized medicine, and unlocking the full potential of histopathology through new imaging-based biomarkers. Current challenges include explainability and interpretability, validation via adjacent tissue sections, reproducibility, data availability, computational needs, data requirements, domain adaptability, external validation, dataset imbalances, and finally commercialization and clinical potential. Ultimately, the relative ease and minimum upfront cost with which relevant data can be collected in addition to the plethora of available AI methods for outcome-driven analysis will surmount these current limitations and achieve the innumerable opportunities associated with AI-driven histopathology for the benefit of oncology.

Translation of tissue-based artificial intelligence into clinical practice: from discovery to adoption

Digital pathology (DP), or the digitization of pathology images, has transformed oncology research and cancer diagnostics. The application of artificial intelligence (AI) and other forms of machine learning (ML) to these images allows for better interpretation of morphology, improved quantitation of biomarkers, introduction of novel concepts to discovery and diagnostics (such as spatial distribution of cellular elements), and the promise of a new paradigm of cancer biomarkers. The application of AI to tissue analysis can take several conceptual approaches, within the domains of language modelling and image analysis, such as Deep Learning Convolutional Neural Networks, Multiple Instance Learning approaches, or the modelling of risk scores and their application to ML. The use of different approaches solves different problems within pathology workflows, including assistive applications for the detection and grading of tumours, quantification of biomarkers, and the delivery of established and new image-based biomarkers for treatment prediction and prognostic purposes. All these AI formats, applied to digital tissue images, are also beginning to transform our approach to clinical trials. In parallel, the novelty of DP/AI devices and the related computational science pipeline introduces new requirements for manufacturers to build into their design, development, regulatory and post-market processes, which may need to be taken into account when using AI applied to tissues in cancer discovery. Finally, DP/AI represents challenge to the way we accredit new diagnostic tools with clinical applicability, the understanding of which will allow cancer patients to have access to a new generation of complex biomarkers.

Predicting acute pancreatitis severity with enhanced computed tomography scans using convolutional neural networks

This study aimed to evaluate acute pancreatitis (AP) severity using convolutional neural network (CNN) models with enhanced computed tomography (CT) scans. Three-dimensional DenseNet CNN models were developed and trained using the enhanced CT scans labeled with two severity assessment methods: the computed tomography severity index (CTSI) and Atlanta classification. Each labeling method was used independently for model training and validation. Model performance was evaluated using confusion matrices, areas under the receiver operating characteristic curve (AUC-ROC), accuracy, precision, recall, F1 score, and respective macro-average metrics. A total of 1,798 enhanced CT scans met the inclusion criteria were included in this study. The dataset was randomly divided into a training dataset (n = 1618) and a test dataset (n = 180) with a ratio of 9:1. The DenseNet model demonstrated promising predictions for both CTSI and Atlanta classification-labeled CT scans, with accuracy greater than 0.7 and AUC-ROC greater than 0.8. Specifically, when trained with CT scans labeled using CTSI, the DenseNet model achieved good performance, with a macro-average F1 score of 0.835 and a macro-average AUC-ROC of 0.980. The findings of this study affirm the feasibility of employing CNN models to predict the severity of AP using enhanced CT scans.

A consolidated working classification of gastric cancer for histopathologists (Review)

Gastric cancer (GC) remains a disease with poor prognosis despite increasing availability of more effective targeted treatment. This may be in part due to the difficulty in selecting patients for appropriate treatment. Conventional taxonomic classifications of GC are ill-suited to make full use of recent advances in personalised therapy. In the past decade a number of molecular classifications have been proposed to address this; however, to date, there has been little implementation in the diagnostic routine. The lack of harmonisation between these classifications, the complexity and unavailability of some of the tests required plus the demands on time and resources, all contribute to poor uptake in the diagnostic routine. In the present study, these classifications were reviewed and an inclusive working classification that includes their main points, focuses on prognosis and treatment options and can be delivered using four on-slide tests (in situ hybridization for Epstein-Barr encoding region and immunohistochemistry for mismatch repair, E-cadherin and p53) is proposed. These tests can be performed on paraffin-embedded tissue and could be available in the majority of histopathology laboratories. The proposed classification also includes reflex testing for specific biomarkers relevant to treatment selection.

Integrative models of histopathological images and multi-omics data predict prognosis in endometrial carcinoma

Objective

This study aimed to predict the molecular features of endometrial carcinoma (EC) and the overall survival (OS) of EC patients using histopathological imaging.

Methods

The patients from The Cancer Genome Atlas (TCGA) were separated into the training set (n = 215) and test set (n = 214) in proportion of 1:1. By analyzing quantitative histological image features and setting up random forest model verified by cross-validation, we constructed prognostic models for OS. The model performance is evaluated with the time-dependent receiver operating characteristics (AUC) over the test set.

Results

Prognostic models based on histopathological imaging features (HIF) predicted OS in the test set (5-year AUC = 0.803). The performance of combining histopathology and omics transcends that of genomics, transcriptomics, or proteomics alone. Additionally, multi-dimensional omics data, including HIF, genomics, transcriptomics, and proteomics, attained the largest AUCs of 0.866, 0.869, and 0.856 at years 1, 3, and 5, respectively, showcasing the highest discrepancy in survival (HR = 18.347, 95% CI [11.09-25.65], p < 0.001).

Conclusions

The results of this experiment indicated that the complementary features of HIF could improve the prognostic performance of EC patients. Moreover, the integration of HIF and multi-dimensional omics data might ameliorate survival prediction and risk stratification in clinical practice.

The slow-paced digital evolution of pathology: lights and shadows from a multifaceted board

Objective

The digital revolution in pathology represents an invaluable resource fto optimise costs, reduce the risk of error and improve patient care, even though it is still adopted in a minority of laboratories. Barriers include concerns about initial costs, lack of confidence in using whole slide images for primary diagnosis, and lack of guidance on transition. To address these challenges and develop a programme to facilitate the introduction of digital pathology (DP) in Italian pathology departments, a panel discussion was set up to identify the key points to be considered.

Methods

On 21 July 2022, an initial conference call was held on Zoom to identify the main issues to be discussed during the face-to-face meeting. The final summit was divided into four different sessions: (I) the definition of DP, (II) practical applications of DP, (III) the use of AI in DP, (IV) DP and education.

Results

Essential requirements for the implementation of DP are a fully tracked and automated workflow, selection of the appropriate scanner based on the specific needs of each department, and a strong commitment combined with coordinated teamwork (pathologists, technicians, biologists, IT service and industries). This could reduce human error, leading to the application of AI tools for diagnosis, prognosis and prediction. Open challenges are the lack of specific regulations for virtual slide storage and the optimal storage solution for large volumes of slides.

Conclusion

Teamwork is key to DP transition, including close collaboration with industry. This will ease the transition and help bridge the gap that currently exists between many labs and full digitisation. The ultimate goal is to improve patient care.

Weakly supervised deep learning to predict recurrence in low-grade endometrial cancer from multiplexed immunofluorescence images

Predicting recurrence in low-grade, early-stage endometrial cancer (EC) is both challenging and clinically relevant. We present a weakly-supervised deep learning framework, NaroNet, that can learn, without manual expert annotation, the complex tumor-immune interrelations at three levels: local phenotypes, cellular neighborhoods, and tissue areas. It uses multiplexed immunofluorescence for the simultaneous visualization and quantification of CD68 + macrophages, CD8 + T cells, FOXP3 + regulatory T cells, PD-L1/PD-1 protein expression, and tumor cells. We used 489 tumor cores from 250 patients to train a multilevel deep-learning model to predict tumor recurrence. Using a tenfold cross-validation strategy, our model achieved an area under the curve of 0.90 with a 95% confidence interval of 0.83-0.95. Our model predictions resulted in concordance for 96,8% of cases (κ = 0.88). This method could accurately assess the risk of recurrence in EC, outperforming current prognostic factors, including molecular subtyping.

MALDI HiPLEX-IHC: multiomic and multimodal imaging of targeted intact proteins in tissues

Matrix-assisted laser desorption/ionization mass spectrometry imaging (MALDI-MSI) is one of the most widely used methods for imaging the spatial distribution of unlabeled small molecules such as metabolites, lipids and drugs in tissues. Recent progress has enabled many improvements including the ability to achieve single cell spatial resolution, 3D-tissue image reconstruction, and the precise identification of different isomeric and isobaric molecules. However, MALDI-MSI of high molecular weight intact proteins in biospecimens has thus far been difficult to achieve. Conventional methods normally require in situ proteolysis and peptide mass fingerprinting, have low spatial resolution, and typically detect only the most highly abundant proteins in an untargeted manner. In addition, MSI-based multiomic and multimodal workflows are needed which can image both small molecules and intact proteins from the same tissue. Such a capability can provide a more comprehensive understanding of the vast complexity of biological systems at the organ, tissue, and cellular levels of both normal and pathological function. A recently introduced top-down spatial imaging approach known as MALDI HiPLEX-IHC (MALDI-IHC for short) provides a basis for achieving this high-information content imaging of tissues and even individual cells. Based on novel photocleavable mass-tags conjugated to antibody probes, high-plex, multimodal and multiomic MALDI-based workflows have been developed to image both small molecules and intact proteins on the same tissue sample. Dual-labeled antibody probes enable multimodal mass spectrometry and fluorescent imaging of targeted intact proteins. A similar approach using the same photocleavable mass-tags can be applied to lectin and other probes. We detail here several examples of MALDI-IHC workflows designed to enable high-plex, multiomic and multimodal imaging of tissues at a spatial resolution as low as 5 µm. This approach is compared to other existing high-plex methods such as imaging mass cytometry, MIBI-TOF, GeoMx and CODEX. Finally, future applications of MALDI-IHC are discussed.