Chromatin organisation and cancer prognosis: a pan-cancer study

Artificial Intelligence-Powered Quality Assurance: Transforming Diagnostics, Surgery, and Patient Care-Innovations, Limitations, and Future Directions

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.

Clinical implications of DNA ploidy, stroma, and nucleotyping in predicting peritoneal metastasis risk for gastric cancer

BACKGROUND

Gastric cancer peritoneal metastasis lacks effective predictive indices. This article retrospectively explored predictive values of DNA ploidy, stroma, and nucleotyping in gastric cancer peritoneal metastasis.

METHODS

A comprehensive analysis was conducted on specimens obtained from 80 gastric cancer patients who underwent gastric resection at the Department of Gastrointestinal Surgery of Wuhan University Renmin Hospital. Tumor tissues were sectioned and stained. DNA ploidy, stroma, and nucleotyping were quantified using microscopy and digital analysis software. Data analysis was employed by Pearson Chi-square, continuous correction Chi-square, and binary logistic regression.

RESULTS

Using both univariate and multivariate analysis, pathological T stage and nucleotyping exhibited a positive correlation with peritoneal metastasis. DNA ploidy and stroma showed a positive correlation in univariate analysis. Chi-square tests demonstrated a positive correlation of DNA ploidy, stroma, and nucleotyping with peritoneal metastasis. The combined application of these three indicators displayed heightened predictive value for peritoneal metastasis. Non-diploid status, high stroma, and chromosomal heterogeneity emerged as positive factors for peritoneal metastasis in gastric cancer.

CONCLUSIONS

DNA ploidy, stroma, and nucleotyping prove to be predictive factors for peritoneal metastasis, with enhanced predictive efficacy when combined in pairs.

Development of metastasis and survival prediction model of luminal and non-luminal breast cancer with weakly supervised learning based on pathomics

Objective

Breast cancer stands as the most prevalent form of cancer among women globally. This heterogeneous disease exhibits varying clinical behaviors. The stratification of breast cancer patients into risk groups, determined by their metastasis and survival outcomes, is pivotal for tailoring personalized treatments and therapeutic interventions. The pathological sections of radical specimens encompass a diverse range of histological information pertinent to the metastasis and survival of patients. In this study, our objective is to develop a deep learning model utilizing pathological images to predict the metastasis and survival outcomes for breast cancer patients.

Methods

This study utilized pathological sections from 204 radical mastectomy specimens obtained between January 2013 and December 2014 at the Second Affiliated Hospital of the Medical College of Zhejiang University. The 204 pathological slices were scanned and transformed into whole slide imaging (WSI), with manual labeling of all tumor areas. The WSI was then partitioned into smaller tiles measuring 512 × 512 pixels. Three networks, namely Densely Connected Convolutional Network 121 (DenseNet121), Residual Network (ResNet50), and Inception_v3, were assessed. Subsequently, we combined patch-level predictions, probability histograms, and Term Frequency-Inverse Document Frequency (TF-IDF) features to create comprehensive participants representations. These features served as the foundational input for developing a machine learning algorithm for metastasis analysis and a Cox regression model for survival analysis.

Result

Our results show that the Inception_v3 model shows a particularly robust patch recognition ability for estrogen receptor (ER) recognition. Our pathological model shows high accuracy in predicting tumor regions. The train area under the curve (AUC) of the Inception_v3 model based on supervised learning is 0.975, which is higher than the model established by weakly supervised learning. But the AUC of the metastasis prediction in training and testing sets is higher than value based on supervised learning. Furthermore, the C-index of the survival prediction model is 0.710 in the testing sets, which is also better than the value by supervised learning.

Conclusion

Our study demonstrates the significant potential of deep learning models in predicting breast cancer metastasis and prognosis, with the pathomic model showing high accuracy in identifying tumor areas and ER status. The integration of clinical features and pathomics signature into a nomogram further provides a valuable tool for clinicians to make individualized treatment decisions.

Chromosomal instability is associated with prognosis and efficacy of bevacizumab after resection of colorectal cancer liver metastasis

INTRODUCTION

Individualized treatment of colorectal cancer liver metastases (CRLM) remains challenging due to differences in the severity of metastatic disease and tumour biology. Exploring specific prognostic risk subgroups is urgently needed. The current study aimed to investigate the prognostic value of chromosomal instability (CIN) in patients with initially resectable CRLM and the predictive value of CIN for the efficacy of bevacizumab.

METHODS

Ninety-one consecutive patients with initially resectable CRLM who underwent curative liver resection from 2006 to 2018 at Sun Yat-sen University Cancer Center were selected for analysis. CIN was evaluated by automated digital imaging systems. Immunohistochemistry (IHC) was performed to detect interleukin-6 (IL-6), vascular endothelial growth factor A (VEGFA) and CD31 expression in paraffin-embedded specimens. Recurrence-free survival (RFS) and overall survival (OS) were analysed using the Kaplan-Meier method and Cox regression models.

RESULTS

Patients with high chromosomal instability (CIN-H) had a worse 3-year RFS rate (HR, 1.953; 95% CI, 1.001-3.810; p = 0.049) and a worse 3-year OS rate (HR, 2.449; 95% CI, 1.150-5.213; p = 0.016) than those with low chromosomal instability (CIN-L). CIN-H was identified as an independent prognostic factor for RFS (HR, 2.569; 95% CI, 1.078-6.121; p = 0.033) and OS (HR, 3.852; 95% CI, 1.173-12.645; p = 0.026) in the multivariate analysis. The protein levels of IL-6, VEGFA and CD31 were upregulated in patients in the CIN-H group compared to those in the CIN-L group in both primary tumour and liver metastases tissues. Among them, 22 patients with recurrent tumours were treated with first-line bevacizumab treatment and based on the clinical response assessment, disease control rates were adversely associated with chromosomal instability (p = 0.043).

CONCLUSIONS

Our study showed that high chromosomal instability is a negative prognostic factor for patients with initially resectable CRLM after liver resection. CIN may have positive correlations with angiogenesis through expression of IL-6-VEGFA axis and be used as a potential predictor of efficacy of bevacizumab.

OralEpitheliumDB: A Dataset for Oral Epithelial Dysplasia Image Segmentation and Classification

Early diagnosis of potentially malignant disorders, such as oral epithelial dysplasia, is the most reliable way to prevent oral cancer. Computational algorithms have been used as an auxiliary tool to aid specialists in this process. Usually, experiments are performed on private data, making it difficult to reproduce the results. There are several public datasets of histological images, but studies focused on oral dysplasia images use inaccessible datasets. This prevents the improvement of algorithms aimed at this lesion. This study introduces an annotated public dataset of oral epithelial dysplasia tissue images. The dataset includes 456 images acquired from 30 mouse tongues. The images were categorized among the lesion grades, with nuclear structures manually marked by a trained specialist and validated by a pathologist. Also, experiments were carried out in order to illustrate the potential of the proposed dataset in classification and segmentation processes commonly explored in the literature. Convolutional neural network (CNN) models for semantic and instance segmentation were employed on the images, which were pre-processed with stain normalization methods. Then, the segmented and non-segmented images were classified with CNN architectures and machine learning algorithms. The data obtained through these processes is available in the dataset. The segmentation stage showed the F1-score value of 0.83, obtained with the U-Net model using the ResNet-50 as a backbone. At the classification stage, the most expressive result was achieved with the Random Forest method, with an accuracy value of 94.22%. The results show that the segmentation contributed to the classification results, but studies are needed for the improvement of these stages of automated diagnosis. The original, gold standard, normalized, and segmented images are publicly available and may be used for the improvement of clinical applications of CAD methods on oral epithelial dysplasia tissue images.

Combination of DNA ploidy, stroma, and nucleotyping predicting prognosis and tailoring adjuvant chemotherapy duration in stage III colon cancer

Introduction

DNA ploidy (P), stroma fraction (S), and nucleotyping (N) collectively known as PSN, have proven prognostic accuracy in stage II colorectal cancer (CRC). However, few studies have reported on the prognostic value of the PSN panel in stage III colon cancer patients receiving capecitabine and oxaliplatin adjuvant chemotherapy.

Objectives

This study aimed to validate PSN's prognostic impact on stage III colon cancer, identifying candidates for optimized adjuvant chemotherapy duration.

Design

A retrospective analysis was conducted on a cohort of stage III colon cancer patients from April 2008 to June 2020.

Methods

Postoperative pathological samples from stage III colon cancer patients who underwent radical surgery and postoperative adjuvant chemotherapy at Sun Yat-sen University Cancer Center were retrospectively collected. Automated digital imaging assessed PSN, categorizing risk groups. Kaplan-Meier, Cox regression, and time-dependent receiver operating characteristic analysis compared model validity.

Results

Significant differences in 5-year disease-free survival (DFS) and overall survival (OS) were noted among PSN-based low-, moderate-, and high-risk groups (DFS: 92.10% versus 83.62% versus 79.80%, p = 0.029; OS: 96.69% versus 93.99% versus 90.12%, p = 0.016). PSN emerged as an independent prognostic factor for DFS [hazard ratio (HR) = 1.409, 95% confidence interval (CI): 1.002-1.981, p = 0.049] and OS (HR = 1.720, 95% CI: 1.127-2.624, p = 0.012). The PSN model, incorporating perineural invasion and tumor location, displayed superior area under the curve for 5-year (0.692 versus 0.553, p = 0.020) and 10-year (0.694 versus 0.532, p = 0.006) DFS than TNM stage. In the PSN high-risk group, completing eight cycles of adjuvant chemotherapy significantly improved 5-year DFS and OS compared to four to seven cycles (DFS: 89.43% versus 71.52%, p = 0.026; OS: 96.77% versus 85.46%, p = 0.007).

Conclusion

The PSN panel effectively stratifies stage III colon cancer, aiding in optimized adjuvant chemotherapy duration determination.

Exploring prognostic values of DNA ploidy, stroma-tumor fraction and nucleotyping in stage II colon cancer patients

PURPOSE

To assess the prognostic value of three novel biomarkers, DNA ploidy, stroma-tumor fraction, and nucleotyping, seeking for more accurate stratification in stage II colon cancer.

METHODS

A total of 417 patients with complete follow up information were enrolled in this study and divided into three clinical risk groups. IHC was performed to examine MSI status. DNA ploidy, stroma and nucleotyping were estimated using automated digital imaging system. Kaplan-Meier survival curves, Cox proportional hazards regression models, and correlation analyses were carried out to process our data.

RESULTS

In the whole cohort of stage II colon cancer, nucleotyping and DNA ploidy were significant prognostic factors on OS in univariate analyses. The combination of nucleotyping and DNA ploidy signified superior OS and DFS. Difference was not significant between low-stroma and high-stroma patients. In multivariable analyses, nucleotyping and the combination of nucleotyping and DNA ploidy were proven the dominant contributory factors for OS. In the low-risk group, we found the combination of nucleotyping and DNA ploidy as the independent prognostic factor statistically significant in both univariate and multivariable, while in the high-risk group, the nucleotyping.

CONCLUSIONS

Our study has proven nucleotyping and the combination of DNA ploidy and nucleotyping as independent prognostic indicators, thus expanding the application of nucleotyping as a predictor from high risk stage II colon cancer to whole risks.

Morphological diversity of cancer cells predicts prognosis across tumor types

BACKGROUND

Intratumor heterogeneity drives disease progression and treatment resistance, which can lead to poor patient outcomes. Here, we present a computational approach for quantification of cancer cell diversity in routine hematoxylin-eosin-stained histopathology images.

METHODS

We analyzed publicly available digitized whole-slide hematoxylin-eosin images for 2000 patients. Four tumor types were included: lung, head and neck, colon, and rectal cancers, representing major histology subtypes (adenocarcinomas and squamous cell carcinomas). We performed single-cell analysis on hematoxylin-eosin images and trained a deep convolutional autoencoder to automatically learn feature representations of individual cancer nuclei. We then computed features of intranuclear variability and internuclear diversity to quantify tumor heterogeneity. Finally, we used these features to build a machine-learning model to predict patient prognosis.

RESULTS

A total of 68 million cancer cells were segmented and analyzed for nuclear image features. We discovered multiple morphological subtypes of cancer cells (range = 15-20) that co-exist within the same tumor, each with distinct phenotypic characteristics. Moreover, we showed that a higher morphological diversity is associated with chromosome instability and genomic aneuploidy. A machine-learning model based on morphological diversity demonstrated independent prognostic values across tumor types (hazard ratio range = 1.62-3.23, P < .035) in validation cohorts and further improved prognostication when combined with clinical risk factors.

CONCLUSIONS

Our study provides a practical approach for quantifying intratumor heterogeneity based on routine histopathology images. The cancer cell diversity score can be used to refine risk stratification and inform personalized treatment strategies.

Personalizing adjuvant therapy for patients with colorectal cancer

The current standard-of-care adjuvant treatment for patients with colorectal cancer (CRC) comprises a fluoropyrimidine (5-fluorouracil or capecitabine) as a single agent or in combination with oxaliplatin, for either 3 or 6 months. Selection of therapy depends on conventional histopathological staging procedures, which constitute a blunt tool for patient stratification. Given the relatively marginal survival benefits that patients can derive from adjuvant treatment, improving the safety of chemotherapy regimens and identifying patients most likely to benefit from them is an area of unmet need. Patient stratification should enable distinguishing those at low risk of recurrence and a high chance of cure by surgery from those at higher risk of recurrence who would derive greater absolute benefits from chemotherapy. To this end, genetic analyses have led to the discovery of germline determinants of toxicity from fluoropyrimidines, the identification of patients at high risk of life-threatening toxicity, and enabling dose modulation to improve safety. Thus far, results from analyses of resected tissue to identify mutational or transcriptomic signatures with value as prognostic biomarkers have been rather disappointing. In the past few years, the application of artificial intelligence-driven models to digital images of resected tissue has identified potentially useful algorithms that stratify patients into distinct prognostic groups. Similarly, liquid biopsy approaches involving measurements of circulating tumour DNA after surgery are additionally useful tools to identify patients at high and low risk of tumour recurrence. In this Perspective, we provide an overview of the current landscape of adjuvant therapy for patients with CRC and discuss how new technologies will enable better personalization of therapy in this setting.

Artificial intelligence-based risk stratification, accurate diagnosis and treatment prediction in gynecologic oncology

As data-driven science, artificial intelligence (AI) has paved a promising path toward an evolving health system teeming with thrilling opportunities for precision oncology. Notwithstanding the tremendous success of oncological AI in such fields as lung carcinoma, breast tumor and brain malignancy, less attention has been devoted to investigating the influence of AI on gynecologic oncology. Hereby, this review sheds light on the ever-increasing contribution of state-of-the-art AI techniques to the refined risk stratification and whole-course management of patients with gynecologic tumors, in particular, cervical, ovarian and endometrial cancer, centering on information and features extracted from clinical data (electronic health records), cancer imaging including radiological imaging, colposcopic images, cytological and histopathological digital images, and molecular profiling (genomics, transcriptomics, metabolomics and so forth). However, there are still noteworthy challenges beyond performance validation. Thus, this work further describes the limitations and challenges faced in the real-word implementation of AI models, as well as potential solutions to address these issues.

Reactivation of embryonic genetic programs in tissue regeneration and disease

Embryonic genetic programs are reactivated in response to various types of tissue damage, providing cell plasticity for tissue regeneration or disease progression. In acute conditions, these programs remedy the damage and then halt to allow a return to homeostasis. In chronic situations, including inflammatory diseases, fibrosis and cancer, prolonged activation of embryonic programs leads to disease progression and tissue deterioration. Induction of progenitor identity and cell plasticity, for example, epithelial-mesenchymal plasticity, are critical outcomes of reactivated embryonic programs. In this Review, we describe molecular players governing reactivated embryonic genetic programs, their role during disease progression, their similarities and differences and lineage reversion in pathology and discuss associated therapeutics and drug-resistance mechanisms across many organs. We also discuss the diversity of reactivated programs in different disease contexts. A comprehensive overview of commonalities between development and disease will provide better understanding of the biology and therapeutic strategies.

Prognostic Value of Chromatin Structure Typing in Early-Stage Non-Small Cell Lung Cancer

(1) Background: Chromatin structure typing has been used for prognostic risk stratification among cancer survivors. This study aimed to ascertain the prognostic values of ploidy, nucleotyping, and tumor-stroma ratio (TSR) in predicting disease progression for patients with early-stage non-small cell lung cancer (NSCLC), and to explore whether patients with different nucleotyping profiles can benefit from adjuvant chemotherapy. (2) Methods: DNA ploidy, nucleotyping, and TSR were measured by chromatin structure typing analysis (Matrix Analyser, Room4, Kent, UK). Cox proportional hazard regression models were used to assess the relationships of DNA ploidy, nucleotyping, and TSR with a 5-year disease-free survival (DFS). (3) Results: among 154 early-stage NSCLC patients, 102 were non-diploid, 40 had chromatin heterogeneity, and 126 had a low stroma fraction, respectively. Univariable analysis suggested that non-diploidy was associated with a significantly lower 5-year DFS rate. After combining DNA ploidy and nucleotyping for risk stratification and adjusting for potential confounders, the DNA ploidy and nucleotyping (PN) high-risk group and PN medium-risk group had a 4- (95% CI: 1.497-8.754) and 3-fold (95% CI: 1.196-6.380) increase in the risk of disease progression or mortality within 5 years of follow-up, respectively, compared to the PN low-risk group. In PN high-risk patients, adjuvant therapy was associated with a significantly improved 5-year DFS (HR = 0.214, 95% CI: 0.048-0.957, p = 0.027). (4) Conclusions: the non-diploid DNA status and the combination of ploidy and nucleotyping can be useful prognostic indicators to predict long-term outcomes in early-stage NSCLC patients. Additionally, NSCLC patients with non-diploidy and chromatin homogenous status may benefit from adjuvant therapy.

Role of ferroptosis in colorectal cancer

Colorectal cancer (CRC) is the second deadliest cancer and the third-most common malignancy in the world. Surgery, chemotherapy, and targeted therapy have been widely used to treat CRC, but some patients still develop resistance to these treatments. Ferroptosis is a novel non-apoptotic form of cell death. It is an iron-dependent non-apoptotic cell death characterized by the accumulation of lipid reactive oxygen species and has been suggested to play a role in reversing resistance to anticancer drugs. This review summarizes recent advances in the prognostic role of ferroptosis in CRC and the mechanism of action in CRC.

Next-Generation Morphometry for pathomics-data mining in histopathology

Pathology diagnostics relies on the assessment of morphology by trained experts, which remains subjective and qualitative. Here we developed a framework for large-scale histomorphometry (FLASH) performing deep learning-based semantic segmentation and subsequent large-scale extraction of interpretable, quantitative, morphometric features in non-tumour kidney histology. We use two internal and three external, multi-centre cohorts to analyse over 1000 kidney biopsies and nephrectomies. By associating morphometric features with clinical parameters, we confirm previous concepts and reveal unexpected relations. We show that the extracted features are independent predictors of long-term clinical outcomes in IgA-nephropathy. We introduce single-structure morphometric analysis by applying techniques from single-cell transcriptomics, identifying distinct glomerular populations and morphometric phenotypes along a trajectory of disease progression. Our study provides a concept for Next-generation Morphometry (NGM), enabling comprehensive quantitative pathology data mining, i.e., pathomics.

Cancer-associated fibroblast-related gene signatures predict survival and drug response in patients with colorectal cancer

Background: Cancer-associated fibroblasts (CAFs) play an important role in the tumorigenesis, immunosuppression and metastasis of colorectal cancer (CRC), and can predict poor prognosis in patients with CRC. The present study aimed to construct a CAFs-related prognostic signature for CRC. Methods: The clinical information and corresponding RNA data of CRC patients were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. The Estimation of STromal and Immune cells in MAlignant Tumor tissues (ESTIMATES) and xCell methods were applied to evaluate the tumor microenvironment infiltration from bulk gene expression data. Weighted gene co-expression network analysis (WGCNA) was used to construct co-expression modules. The key module was identified by calculating the module-trait correlations. The univariate Cox regression and least absolute shrinkage operator (LASSO) analyses were combined to develop a CAFs-related signature for the prognostic model. Moreover, pRRophetic and Tumor Immune Dysfunction and Exclusion (TIDE) algorithms were utilized to predict chemosensitivity and immunotherapy response. Human Protein Atlas (HPA) databases were employed to evaluate the protein expressions. Results: ESTIMATES and xCell analysis showed that high CAFs infiltration was associated with adverse prognoses. A twenty-gene CAFs-related prognostic signature (CAFPS) was established in the training cohort. Kaplan-Meier survival analyses reveled that CRC patients with higher CAFs risk scores were associated with poor prognosis in each cohort. Univariate and multivariate Cox regression analyses verified that CAFPS was as an independent prognostic factor in predicting overall survival, and a nomogram was built for clinical utility in predicting CRC prognosis. Patients with higher CAFs risk scores tended to not respond to immunotherapy, but were more sensitive to five conventional chemotherapeutic drugs. Conclusion: In summary, the CAFPS could serve as a robust prognostic indicator in CRC patients, which might help to optimize risk stratification and provide a new insight into individual treatments for CRC.

Deep Learning Technology Applied to Medical Image Tissue Classification

Medical image classification is a novel technology that presents a new challenge. It is essential that pathological images are automatically and correctly classified to enable doctors to provide precise treatment. Convolutional neural networks have demonstrated their effectiveness in classifying images in deep learning, which may have dozens or hundreds of layers, to illustrate the relationship between them in terms of their different neural network features. Convolutional layers consisting of small kernels take weights as input and guide them through an activation function as output. The main advantage of using convolutional neural networks (CNNs) instead of traditional neural networks is that they reduce the model parameters for greater accuracy. However, many studies have simply been focused on finding the best CNN model and classification results from a single medical image classification. Therefore, we applied a common deep learning network model in an attempt to identify the best model framework by training and validating different model parameters to classify medical images. After conducting experiments on six publicly available databases of pathological images, including colorectal cancer tissue, chest X-rays, common skin lesions, diabetic retinopathy, pediatric chest X-ray, and breast ultrasound image datasets, we were able to confirm that the recognition accuracy of the Inception V3 method was significantly better than that of other existing deep learning models.

Adversarial attacks and adversarial robustness in computational pathology

Artificial Intelligence (AI) can support diagnostic workflows in oncology by aiding diagnosis and providing biomarkers directly from routine pathology slides. However, AI applications are vulnerable to adversarial attacks. Hence, it is essential to quantify and mitigate this risk before widespread clinical use. Here, we show that convolutional neural networks (CNNs) are highly susceptible to white- and black-box adversarial attacks in clinically relevant weakly-supervised classification tasks. Adversarially robust training and dual batch normalization (DBN) are possible mitigation strategies but require precise knowledge of the type of attack used in the inference. We demonstrate that vision transformers (ViTs) perform equally well compared to CNNs at baseline, but are orders of magnitude more robust to white- and black-box attacks. At a mechanistic level, we show that this is associated with a more robust latent representation of clinically relevant categories in ViTs compared to CNNs. Our results are in line with previous theoretical studies and provide empirical evidence that ViTs are robust learners in computational pathology. This implies that large-scale rollout of AI models in computational pathology should rely on ViTs rather than CNN-based classifiers to provide inherent protection against perturbation of the input data, especially adversarial attacks.

Artificial Intelligence can support diagnostic workflows in oncology, but they are vulnerable to adversarial attacks. Here, the authors show that convolutional neural networks are highly susceptible to white- and black-box adversarial attacks in clinically relevant classification tasks.

Machine learning applications in gynecological cancer: A critical review

Machine Learning (ML) represents a computer science capable of generating predictive models, by exposure to raw, training data, without being rigidly programmed. Over the last few years, ML has gained attention within the field of oncology, with considerable strides in both diagnostic, predictive, and prognostic spectrum of malignancies, but also as a catalyst of cancer research. In this review, we discuss the state of ML applications on gynecologic oncology and systematically address major technical and ethical concerns, with respect to their real-world medical practice translation. Undoubtedly, advances in ML will enable the analysis of large, rather complex, datasets for improved, cost-effective, and efficient clinical decisions.

Artificial intelligence in histopathology: enhancing cancer research and clinical oncology

Artificial intelligence (AI) methods have multiplied our capabilities to extract quantitative information from digital histopathology images. AI is expected to reduce workload for human experts, improve the objectivity and consistency of pathology reports, and have a clinical impact by extracting hidden information from routinely available data. Here, we describe how AI can be used to predict cancer outcome, treatment response, genetic alterations and gene expression from digitized histopathology slides. We summarize the underlying technologies and emerging approaches, noting limitations, including the need for data sharing and standards. Finally, we discuss the broader implications of AI in cancer research and oncology.

Computational pathology in ovarian cancer

Histopathologic evaluations of tissue sections are key to diagnosing and managing ovarian cancer. Pathologists empirically assess and integrate visual information, such as cellular density, nuclear atypia, mitotic figures, architectural growth patterns, and higher-order patterns, to determine the tumor type and grade, which guides oncologists in selecting appropriate treatment options. Latent data embedded in pathology slides can be extracted using computational imaging. Computers can analyze digital slide images to simultaneously quantify thousands of features, some of which are visible with a manual microscope, such as nuclear size and shape, while others, such as entropy, eccentricity, and fractal dimensions, are quantitatively beyond the grasp of the human mind. Applications of artificial intelligence and machine learning tools to interpret digital image data provide new opportunities to explore and quantify the spatial organization of tissues, cells, and subcellular structures. In comparison to genomic, epigenomic, transcriptomic, and proteomic patterns, morphologic and spatial patterns are expected to be more informative as quantitative biomarkers of complex and dynamic tumor biology. As computational pathology is not limited to visual data, nuanced subvisual alterations that occur in the seemingly “normal” pre-cancer microenvironment could facilitate research in early cancer detection and prevention. Currently, efforts to maximize the utility of computational pathology are focused on integrating image data with other -omics platforms that lack spatial information, thereby providing a new way to relate the molecular, spatial, and microenvironmental characteristics of cancer. Despite a dire need for improvements in ovarian cancer prevention, early detection, and treatment, the ovarian cancer field has lagged behind other cancers in the application of computational pathology. The intent of this review is to encourage ovarian cancer research teams to apply existing and/or develop additional tools in computational pathology for ovarian cancer and actively contribute to advancing this important field.

Segmentation of Oral Leukoplakia (OL) and Proliferative Verrucous Leukoplakia (PVL) Using Artificial Intelligence Techniques

PVL (proliferative verrucous leukoplakia) has distinct clinical characteristics. They have a proclivity for multifocality, a high recurrence rate after treatment, and malignant transformation, and they can progress to verrucous or squamous cell carcinoma. AI can aid in the diagnosis and prognosis of cancers and other diseases. Computational algorithms can spot tissue changes that a pathologist might overlook. This method is only used in a few studies to diagnose LB and PVL. To see if their cellular nuclei differed and if this cellular compartment could classify them, researchers used a computational system and a polynomial classifier to compare OLs and PVLs. 161 OL and 3 PVL specimens in the lab were grown, photographed, and used for training and computation. Exam orders revealed patients' sociodemographics and clinical pathologies. The nucleus was segmented using Mask R-CNN, and LB and PVL were classified using a polynomial classifier based on nucleus area, perimeter, eccentricity, orientation, solidity, entropies, and Moran Index (a measure of disorderliness). The majority of OL patients were male smokers; most PVL patients were female, with a third having malignant transformation. The neural network correctly identified cell nuclei 92.95% of the time. Except for solidity, 11 of the 13 nuclear characteristics compared between the PVL and the LB showed significant differences. The 97.6% under the curve of the polynomial classifier was used to classify the two lesions. These results demonstrate that computational methods can aid in diagnosing these two lesions.

A scalable, open-source implementation of a large-scale mechanistic model for single cell proliferation and death signaling

Mechanistic models of how single cells respond to different perturbations can help integrate disparate big data sets or predict response to varied drug combinations. However, the construction and simulation of such models have proved challenging. Here, we developed a python-based model creation and simulation pipeline that converts a few structured text files into an SBML standard and is high-performance- and cloud-computing ready. We applied this pipeline to our large-scale, mechanistic pan-cancer signaling model (named SPARCED) and demonstrate it by adding an IFNγ pathway submodel. We then investigated whether a putative crosstalk mechanism could be consistent with experimental observations from the LINCS MCF10A Data Cube that IFNγ acts as an anti-proliferative factor. The analyses suggested this observation can be explained by IFNγ-induced SOCS1 sequestering activated EGF receptors. This work forms a foundational recipe for increased mechanistic model-based data integration on a single-cell level, an important building block for clinically-predictive mechanistic models.

Alternative polyadenylation associated with prognosis and therapy in colorectal cancer

Colorectal cancer (CRC) is among the most widely spread cancers globally. Aberrant alternative polyadenylation (APA) plays a role in cancer onset and its progression. Consequently, this study focused on highlighting the role of APA events and signals in the prognosis of patients with CRC. The APA events, RNA sequencing (RNA-seq), somatic mutations, copy number variants (CNVs), and clinical information of the CRC cohort were obtained from The Cancer Genome Atlas (TCGA) database and UCSC (University of California-Santa Cruz) Xena database. The whole set was sorted into two sets: a training set and a test set in a ratio of 7:3. 197 prognosis-related APA events were collected by performing univariate Cox regression signature in patients with CRC. Subsequently, a signature for APA events was established by least absolute shrinkage and selection operator (LASSO) and multivariate Cox analysis. The risk scores were measured for individual patients on the basis of the signature and patients were sorted into two groups; the high-risk group and the low-risk group as per their median risk scores. Kaplan–Meier curves, principal component analysis (PCA), and time-dependent receiver operator characteristic (ROC) curves revealed that the signature was able to predict patient prognosis effectively and further validation was provided in the test set and the whole set. The high-risk and low-risk groups displayed various distributions of mutations and CNVs. Tumor mutation burden (TMB) alone and in combination with the signature predicted the prognosis of CRC patients, but the gene frequencies of TMBs and CNVs did not change in the low- and high-risk groups. Moreover, immunotherapy and chemotherapy treatments showed different responses to PD-1 inhibitors and multiple chemotherapeutic agents in the low and high-risk groups based on the tumor immune dysfunction and exclusion (TIDE) and genomics of drugs sensitivity in cancer (GDSC) databases. This study may help in understanding the potential roles of APA in CRC, and the signature for prognosis-related APA events can work as a potential predictor for survival and treatment in patients with CRC.

Can aggressive cancers be identified by the "aggressiveness" of their chromatin?

Phenotypic plasticity is a crucial feature of aggressive cancer, providing the means for cancer progression. Stochastic changes in tumor cell transcriptional programs increase the chances of survival under any condition. I hypothesize that unstable chromatin permits stochastic transitions between transcriptional programs in aggressive cancers and supports non-genetic heterogeneity of tumor cells as a basis for their adaptability. I present a mechanistic model for unstable chromatin which includes destabilized nucleosomes, mobile chromatin fibers and random enhancer-promoter contacts, resulting in stochastic transcription. I suggest potential markers for "unsettled" chromatin in tumors associated with poor prognosis. Although many of the characteristics of unstable chromatin have been described, they were mostly used to explain changes in the transcription of individual genes. I discuss approaches to evaluate the role of unstable chromatin in non-genetic tumor cell heterogeneity and suggest using the degree of chromatin instability and transcriptional noise in tumor cells to predict cancer aggressiveness.

Automatic discrimination of Yamamoto-Kohama classification by machine learning approach for invasive pattern of oral squamous cell carcinoma using digital microscopic images: a retrospective study

OBJECTIVE

The Yamamoto-Kohama criteria are clinically useful for determining the mode of tumor invasion, especially in Japan. However, this evaluation method is based on subjective visual findings and has led to significant differences in determinations between evaluators and facilities. In this retrospective study, we aimed to develop an automatic method of determining the mode of invasion based on the processing of digital medical images.

STUDY DESIGN

Using 101 digitized photographic images of anonymized stained specimen slides, we created a classifier that allowed clinicians to introduce feature values and subjected the cases to machine learning using a random forest approach. We then compared the Yamamoto-Kohama grades (1, 2, 3, 4C, 4D) determined by a human oral and maxillofacial surgeon with those determined using the machine learning approach.

RESULTS

The input of multiple test images into the newly created classifier yielded an overall F-measure value of 87% (grade 1, 93%; grade 2, 67%; grade 3, 89%; grade 4C, 83%; grade 4D, 94%). These results suggest that the output of the classifier was very similar to the judgments of the clinician.

CONCLUSIONS

This system may be valuable for diagnostic support to provide an accurate determination of the mode of invasion.

Automated assessment of DNA ploidy, chromatin organization, and stroma fraction to predict prognosis and adjuvant therapy response in patients with stage II colorectal carcinoma

DNA ploidy, tumor stroma, and chromatin organization have important implications in tumorigenesis and patient outcome. Automated image cytometry tools were developed to quantitatively measure DNA ploidy (P), stroma fraction (S), and chromatin organization or Nucleotyping (N). This study aimed to discover their clinical value in different stages of colorectal cancer (CRC) in a Chinese patient population. A total of 496 CRC patients of stages I, II, and LMCRC (liver metastatic CRC) were enrolled in this study. Stage II CRC patients with diploidy, low-stroma, or chromatin homogenous status predicted significantly higher 5-year OS and DFS. We constructed a PSN-panel enabled the stage II patients to be further stratified into low-, middle-, high-risk groups, the 5-year OS (89.5% vs 67.9% vs 60.9%, P<0.001) and DFS (86.0% vs 62.3% vs 53.6%, P<0.001) were stratified significantly. In addition, when combined the PSN-panel with T stage or MSS status in stage II patients, the PSN-low risk patients showed significant longer 5-year OS and DFS than the PSN-high risk patients in T3 (OS: 86.3% vs 65.3%, P=0.015; DFS: 83.5 vs 59.8%, P=0.013) or MSS (OS: 86.4% vs 63.9%, P=0.005; DFS: 85.5 vs 57.8%, P=0.003) patients. Finally, in the group of stage II patients with at least one high-risk factor (non-diploidy, high-stroma, chromatin heterogenous), patients who received adjuvant therapy showed significantly longer OS (72.1% vs 48.3%, P=0.007) and DFS (64.5% vs 43.9%, P=0.015) than those who did not receive adjuvant therapy. In contrast, P, S, N couldn't predict the prognosis of stage I and LMCRC patients. Overall, our data demonstrate that the PSN panel is an accurate prognostic tool that can guide treatment decisions for Chinese stage II CRC patients.

Pan-cancer investigation of CENPK gene: clinical significance and oncogenic immunology

Many studies have confirmed that the CENPK gene regulates the progression of cancers, but its specific molecular mechanism remains unidentified, as does its significance in the analysis of human cancers. We specify a comprehensive genomic architecture of the CENPK gene associated with the tumor immune microenvironment and its clinical relevance across a broad spectrum of solid tumors. Statistics from The Cancer Genome Atlas (TCGA) and Cancer Cell Line Encyclopedia (CCLE) of over 30 solid tumors were examined. CENPK was expressed differentially in several cancers and is significantly associated in survival outcomes, with higher CENPK signifying a worse prognosis for ACC, KICH, KIRC, KIRP, LGG, LIHC, LUAD, MESO, and SARC. We further examined its clinical relevance with tumor immunogenic features. The expression level of CENPK was not only strongly linked to the tumor infiltration, such as tumor-infiltrating immune cells and immune scores but also linked to microsatellite instability and tumor mutation burden in diverse cancers (P<0.05). I mmune markers such as TNFRSF14 and VSIR were highly expressed on over 20 kinds of human cancer and mismatch repair genes like MLH1, MSH2, MSH6, and PMS2 were positively related with CENPK expression. Moreover, the methyltransferases and functional pathways also seem to have a relationship with the CENPK. CENPK is expected to be a guiding marker gene for clinical prognosis and tumor personalized immunotherapy.

Prognostic value of a novel biomarker combining DNA ploidy and tumor burden score for initially resectable liver metastases from patients with colorectal cancer

Background

Colorectal cancer liver metastases (CRLM) has not been identified as a unified disease entity due to the differences in the severity of metastatic disease and tumor aggressiveness. A screen for specific prognostic risk subgroups is urgently needed. The current study aimed to investigate the prognostic value of DNA ploidy, stroma fraction and nucleotyping of initially resectable liver metastases from patients with CRLM.

Methods

One hundred thirty-nine consecutive patients with initially resectable CRLM who underwent curative liver resection from 2006 to 2018 at Sun Yat-sen University Cancer Center were selected for analysis. DNA ploidy, stroma fraction and nucleotyping of liver metastases were evaluated using automated digital imaging systems. Recurrence-free survival (RFS) and overall survival (OS) were analyzed using the Kaplan-Meier method and Cox regression models.

Results

DNA ploidy was identified as an independent prognostic factor for RFS (HR, 2.082; 95% CI 1.053–4.115; P = 0.035) in the multivariate analysis, while stroma-tumor fraction and nucleotyping were not significant prognostic factors. A significant difference in 3-year RFS was observed among the low-, moderate- and high-risk groups stratified by a novel parameter combined with the tumor burden score (TBS) and DNA ploidy (72.5% vs. 63.2% vs. 37.3%, P = 0.007). The high-risk group who received adjuvant chemotherapy had a significantly better 3-year RFS rate than those without adjuvant chemotherapy (46.7% vs. 24.8%; P = 0.034).

Conclusions

Our study showed that DNA ploidy of liver metastases is an independent prognostic factor for patients with initially resectable CRLM after liver resection. The combination of DNA ploidy and TBS may help to stratify patients into different recurrence risk groups and may guide postoperative treatment among the patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12935-021-02250-x.

An Original Ferroptosis-Related Gene Signature Effectively Predicts the Prognosis and Clinical Status for Colorectal Cancer Patients

Background

Colorectal cancer (CRC) is one of the most common malignant tumors in the world. Ferroptosis is a newly defined form of cell death, distinguished by different morphology, biochemistry, and genetics, and involved in CRC progression and treatment. This study aims to establish a predictive model to elucidate the relationship between ferroptosis and prognosis of CRC patients, to explore the potential value of ferroptosis in therapeutic options.

Methods

The ferroptosis-related genes were obtained from the GeneCards and FerrDb websites. The limma R package was used to screen the differential ferroptosis-related genes (DEGs) in CRC from The Cancer Genome Atlas (TCGA) dataset. The least absolute shrinkage and selection operator (LASSO) and multivariate Cox regressions were to establish the 10-gene prognostic signature. The survival and receiver operating characteristic (ROC) curves were illustrated to evaluate the predictive effect of the signature. Besides, independent prognostic factors, downstream functional enrichment, drug sensitivity, somatic mutation status, and immune feature were analyzed. Moreover, all these conclusions were verified by using multiple datasets in International Cancer Genome Consortium (ICGC) and Gene Expression Omnibus (GEO).

Results

Ten ferroptosis-related gene signature (TFAP2C, SLC39A8, NOS2, HAMP, GDF15, FDFT1, CDKN2A, ALOX12, AKR1C1, ATP6V1G2) was established to predict the prognosis of CRC patients by Lasso cox analysis, demonstrating a good performance on Receiver operating characteristic (ROC) and Kaplan–Meier (K–M) analyses. The CRC patients in the high- or low-risk group showed significantly different fractions of immune cells, such as macrophage cells and CD8+ T cells. Drug sensitivity and somatic mutation status like TP53 were also closely associated with the risk scores.

Conclusions

In this study, we identified a novel ferroptosis-related 10-gene signature, which could effectively predict the prognosis and survival time of CRC patients, and provide meaningful clinical implications for targeted therapy or immunotherapy. Targeting ferroptosis is a good therapeutic option for CRC patients. Further studies are needed to reveal the underlying mechanisms of ferroptosis in CRC.

Exploring prognostic indicators in the pathological images of hepatocellular carcinoma based on deep learning

OBJECTIVE

Tumour pathology contains rich information, including tissue structure and cell morphology, that reflects disease progression and patient survival. However, phenotypic information is subtle and complex, making the discovery of prognostic indicators from pathological images challenging.

DESIGN

An interpretable, weakly supervised deep learning framework incorporating prior knowledge was proposed to analyse hepatocellular carcinoma (HCC) and explore new prognostic phenotypes on pathological whole-slide images (WSIs) from the Zhongshan cohort of 1125 HCC patients (2451 WSIs) and TCGA cohort of 320 HCC patients (320 WSIs). A 'tumour risk score (TRS)' was established to evaluate patient outcomes, and then risk activation mapping (RAM) was applied to visualise the pathological phenotypes of TRS. The multi-omics data of The Cancer Genome Atlas(TCGA) HCC were used to assess the potential pathogenesis underlying TRS.

RESULTS

Survival analysis revealed that TRS was an independent prognosticator in both the Zhongshan cohort (p<0.0001) and TCGA cohort (p=0.0003). The predictive ability of TRS was superior to and independent of clinical staging systems, and TRS could evenly stratify patients into up to five groups with significantly different prognoses. Notably, sinusoidal capillarisation, prominent nucleoli and karyotheca, the nucleus/cytoplasm ratio and infiltrating inflammatory cells were identified as the main underlying features of TRS. The multi-omics data of TCGA HCC hint at the relevance of TRS to tumour immune infiltration and genetic alterations such as the FAT3 and RYR2 mutations.

CONCLUSION

Our deep learning framework is an effective and labour-saving method for decoding pathological images, providing a valuable means for HCC risk stratification and precise patient treatment.

Circulating Tumor Cell Chromosomal Instability and Neuroendocrine Phenotype by Immunomorphology and Poor Outcomes in Men with mCRPC Treated with Abiraterone or Enzalutamide

PURPOSE

While the detection of AR-V7 in circulating tumor cells (CTC) is associated with resistance to abiraterone or enzalutamide in men with metastatic castration-resistant prostate cancer (mCRPC), it only accounts for a minority of this resistance. Neuroendocrine (NE) differentiation or chromosomal instability (CIN) may be additional mechanisms that mediate resistance.

EXPERIMENTAL DESIGN

PROPHECY was a multicenter prospective study of men with high-risk mCRPC starting abiraterone or enzalutamide. A secondary objective was to assess Epic CTC CIN and NE phenotypes before abiraterone or enzalutamide and at progression. The proportional hazards (PH) model was used to investigate the prognostic importance of CIN and NE in predicting progression-free survival and overall survival (OS) adjusting for CTC number (CellSearch), AR-V7, prior therapy, and clinical risk score. The PH model was utilized to validate this association of NE with OS in an external dataset of patients treated similarly at Memorial Sloan Kettering Cancer Center (MSKCC; New York, NY).

RESULTS

We enrolled 118 men with mCRPC starting on abiraterone or enzalutamide; 107 were evaluable on the Epic platform. Of these, 36.4% and 8.4% were CIN positive and NE positive, respectively. CIN and NE were independently associated with worse OS [HR, 2.2; 95% confidence interval (CI), 1.2-4.0 and HR 3.8; 95% CI, 1.2-12.3, respectively] when treated with abiraterone/enzalutamide. The prognostic significance of NE positivity for worse OS was confirmed in the MSKCC dataset (n = 173; HR, 5.7; 95% CI, 2.6-12.7).

CONCLUSIONS

A high CIN and NE CTC phenotype is independently associated with worse survival in men with mCRPC treated with abiraterone/enzalutamide, warranting further prospective controlled predictive studies to inform treatment decisions.

Elucidation of the Clustered Nano-Architecture of Radiation-Induced DNA Damage Sites and Surrounding Chromatin in Cancer Cells: A Single Molecule Localization Microscopy Approach

In cancer therapy, the application of (fractionated) harsh radiation treatment is state of the art for many types of tumors. However, ionizing radiation is a "double-edged sword"-it can kill the tumor but can also promote the selection of radioresistant tumor cell clones or even initiate carcinogenesis in the normal irradiated tissue. Individualized radiotherapy would reduce these risks and boost the treatment, but its development requires a deep understanding of DNA damage and repair processes and the corresponding control mechanisms. DNA double strand breaks (DSBs) and their repair play a critical role in the cellular response to radiation. In previous years, it has become apparent that, beyond genetic and epigenetic determinants, the structural aspects of damaged chromatin (i.e., not only of DSBs themselves but also of the whole damage-surrounding chromatin domains) form another layer of complex DSB regulation. In the present article, we summarize the application of super-resolution single molecule localization microscopy (SMLM) for investigations of these structural aspects with emphasis on the relationship between the nano-architecture of radiation-induced repair foci (IRIFs), represented here by γH2AX foci, and their chromatin environment. Using irradiated HeLa cell cultures as an example, we show repair-dependent rearrangements of damaged chromatin and analyze the architecture of γH2AX repair clusters according to topological similarities. Although HeLa cells are known to have highly aberrant genomes, the topological similarity of γH2AX was high, indicating a functional, presumptively genome type-independent relevance of structural aspects in DSB repair. Remarkably, nano-scaled chromatin rearrangements during repair depended both on the chromatin domain type and the treatment. Based on these results, we demonstrate how the nano-architecture and topology of IRIFs and chromatin can be determined, point to the methodological relevance of SMLM, and discuss the consequences of the observed phenomena for the DSB repair network regulation or, for instance, radiation treatment outcomes.

The multifunctional role of SPANX-A/D protein subfamily in the promotion of pro-tumoural processes in human melanoma

Human sperm protein associated with the nucleus on the X chromosome (SPANX) genes encode a protein family (SPANX-A, -B, -C and -D), whose expression is limited to the testis and spermatozoa in normal tissues and various tumour cells. SPANX-A/D proteins have been detected in metastatic melanoma cells, but their contribution to cancer development and the underlying molecular mechanisms of skin tumourigenesis remain unknown. Combining functional and proteomic approaches, the present work describes the presence of SPANX-A/D in primary and metastatic human melanoma cells and how it promotes pro-tumoural processes such as cell proliferation, motility and migration. We provide insights into the molecular features of skin tumourigenesis, describing for the first time a multifunctional role of the SPANX-A/D protein family in nuclear function, energy metabolism and cell survival, considered key hallmarks of cancer. A better comprehension of the SPANX-A/D protein subfamily and its molecular mechanisms will help to describe new aspects of tumour cell biology and develop new therapeutic targets and tumour-directed pharmacological drugs for skin tumours.

Deep learning in cancer pathology: a new generation of clinical biomarkers

Clinical workflows in oncology rely on predictive and prognostic molecular biomarkers. However, the growing number of these complex biomarkers tends to increase the cost and time for decision-making in routine daily oncology practice; furthermore, biomarkers often require tumour tissue on top of routine diagnostic material. Nevertheless, routinely available tumour tissue contains an abundance of clinically relevant information that is currently not fully exploited. Advances in deep learning (DL), an artificial intelligence (AI) technology, have enabled the extraction of previously hidden information directly from routine histology images of cancer, providing potentially clinically useful information. Here, we outline emerging concepts of how DL can extract biomarkers directly from histology images and summarise studies of basic and advanced image analysis for cancer histology. Basic image analysis tasks include detection, grading and subtyping of tumour tissue in histology images; they are aimed at automating pathology workflows and consequently do not immediately translate into clinical decisions. Exceeding such basic approaches, DL has also been used for advanced image analysis tasks, which have the potential of directly affecting clinical decision-making processes. These advanced approaches include inference of molecular features, prediction of survival and end-to-end prediction of therapy response. Predictions made by such DL systems could simplify and enrich clinical decision-making, but require rigorous external validation in clinical settings.

Detecting Differences of Fluorescent Markers Distribution in Single Cell Microscopy: Textural or Pointillist Feature Space?

We consider the detection of change in spatial distribution of fluorescent markers inside cells imaged by single cell microscopy. Such problems are important in bioimaging since the density of these markers can reflect the healthy or pathological state of cells, the spatial organization of DNA, or cell cycle stage. With the new super-resolved microscopes and associated microfluidic devices, bio-markers can be detected in single cells individually or collectively as a texture depending on the quality of the microscope impulse response. In this work, we propose, via numerical simulations, to address detection of changes in spatial density or in spatial clustering with an individual (pointillist) or collective (textural) approach by comparing their performances according to the size of the impulse response of the microscope. Pointillist approaches show good performances for small impulse response sizes only, while all textural approaches are found to overcome pointillist approaches with small as well as with large impulse response sizes. These results are validated with real fluorescence microscopy images with conventional resolution. This, a priori non-intuitive result in the perspective of the quest of super-resolution, demonstrates that, for difference detection tasks in single cell microscopy, super-resolved microscopes may not be mandatory and that lower cost, sub-resolved, microscopes can be sufficient.

The landscape of long noncoding RNA-involved and tumor-specific fusions across various cancers

The majority of the human genome encodes long noncoding RNA (lncRNA) genes, critical regulators of various cellular processes, which largely outnumber protein-coding genes. However, lncRNA-involved fusions have not been surveyed and characterized yet. Here, we present a systematic study of the lncRNA fusion landscape across cancer types and identify >30 000 high-confidence tumor-specific lncRNA fusions (using 8284 tumor and 6946 normal samples). Fusions positively correlated with DNA damage and cancer stemness and were specifically low in microsatellite instable (MSI)-High or virus-infected tumors. Moreover, fusions distribute differently among cancer molecular subtypes, but with shared enrichment in tumors that are microsatellite stable (MSS), with high somatic copy number alterations (SCNA), and with poor survival. Importantly, we find a potentially new mechanism, mediated by enhancer RNAs (eRNA), which generates secondary fusions that form densely connected fusion networks with many fusion hubs targeted by FDA-approved drugs. Finally, we experimentally validate functions of two tumor-promoting chimeric proteins derived from mRNA-lncRNA fusions, KDM4B-G039927 and EPS15L1-lncOR7C2-1. The EPS15L1 fusion protein may regulate (Gasdermin E) GSDME, critical in pyroptosis and anti-tumor immunity. Our study completes the fusion landscape in cancers, sheds light on fusion mechanisms, and enriches lncRNA functions in tumorigenesis and cancer progression.

Prognostic Value of the Diversity of Nuclear Chromatin Compartments in Gynaecological Carcinomas

Statistical texture analysis of cancer cell nuclei stained for DNA has recently been used to develop a pan-cancer prognostic marker of chromatin heterogeneity. In this study, we instead analysed chromatin organisation by automatically quantifying the diversity of chromatin compartments in cancer cell nuclei. The aim was to investigate the prognostic value of such an assessment in relation to chromatin heterogeneity and as a potential supplement to pathological risk classifications in gynaecological carcinomas. The diversity was quantified by calculating the entropy of both chromatin compartment sizes and optical densities within compartments. We analysed a median of 281 nuclei (interquartile range (IQR), 273 to 289) from 246 ovarian carcinoma patients and a median of 997 nuclei (IQR, 502 to 1452) from 791 endometrial carcinoma patients. The prognostic value of the entropies and chromatin heterogeneity was moderately strongly correlated (r ranged from 0.68 to 0.73), but the novel marker was observed to provide additional prognostic information. In multivariable analysis with clinical and pathological markers, the hazard ratio associated with the novel marker was 2.1 (95% CI, 1.3 to 3.5) in ovarian carcinoma and 2.4 (95% CI, 1.5 to 3.9) in endometrial carcinoma. Integration with pathological risk classifications gave three risk groups with distinctly different prognoses. This suggests that the novel marker of diversity of chromatin compartments might possibly contribute to the selection of high-risk stage I ovarian carcinoma patients for adjuvant chemotherapy and low-risk endometrial carcinoma patients for less extensive surgery.

Convergent network effects along the axis of gene expression during prostate cancer progression

Background

Tumor-specific genomic aberrations are routinely determined by high-throughput genomic measurements. It remains unclear how complex genome alterations affect molecular networks through changing protein levels and consequently biochemical states of tumor tissues.

Results

Here, we investigate the propagation of genomic effects along the axis of gene expression during prostate cancer progression. We quantify genomic, transcriptomic, and proteomic alterations based on 105 prostate samples, consisting of benign prostatic hyperplasia regions and malignant tumors, from 39 prostate cancer patients. Our analysis reveals the convergent effects of distinct copy number alterations impacting on common downstream proteins, which are important for establishing the tumor phenotype. We devise a network-based approach that integrates perturbations across different molecular layers, which identifies a sub-network consisting of nine genes whose joint activity positively correlates with increasingly aggressive tumor phenotypes and is predictive of recurrence-free survival. Further, our data reveal a wide spectrum of intra-patient network effects, ranging from similar to very distinct alterations on different molecular layers.

Conclusions

This study uncovers molecular networks with considerable convergent alterations across tumor sites and patients. It also exposes a diversity of network effects: we could not identify a single sub-network that is perturbed in all high-grade tumor regions.

The Clinical Application of the Deep Learning Technique for Predicting Trigger Origins in Patients With Paroxysmal Atrial Fibrillation With Catheter Ablation

Background:

Non–pulmonary vein (NPV) trigger has been reported as an important predictor of recurrence post–atrial fibrillation ablation. Elimination of NPV triggers can reduce the recurrence of postablation atrial fibrillation. Deep learning was applied to preablation pulmonary vein computed tomography geometric slices to create a prediction model for NPV triggers in patients with paroxysmal atrial fibrillation.

Methods:

We retrospectively analyzed 521 patients with paroxysmal atrial fibrillation who underwent catheter ablation of paroxysmal atrial fibrillation. Among them, pulmonary vein computed tomography geometric slices from 358 patients with nonrecurrent atrial fibrillation (1–3 mm interspace per slice, 20–200 slices for each patient, ranging from the upper border of the left atrium to the bottom of the heart, for a total of 23 683 images of slices) were used in the deep learning process, the ResNet34 of the neural network, to create the prediction model of the NPV trigger. There were 298 (83.2%) patients with only pulmonary vein triggers and 60 (16.8%) patients with NPV triggers±pulmonary vein triggers. The patients were randomly assigned to either training, validation, or test groups, and their data were allocated according to those sets. The image datasets were split into training (n=17 340), validation (n=3491), and testing (n=2852) groups, which had completely independent sets of patients.

Results:

The accuracy of prediction in each pulmonary vein computed tomography image for NPV trigger was up to 82.4±2.0%. The sensitivity and specificity were 64.3±5.4% and 88.4±1.9%, respectively. For each patient, the accuracy of prediction for a NPV trigger was 88.6±2.3%. The sensitivity and specificity were 75.0±5.8% and 95.7±1.8%, respectively. The area under the curve for each image and patient were 0.82±0.01 and 0.88±0.07, respectively.

Conclusions:

The deep learning model using preablation pulmonary vein computed tomography can be applied to predict the trigger origins in patients with paroxysmal atrial fibrillation receiving catheter ablation. The application of this model may identify patients with a high risk of NPV trigger before ablation.

[Artificial intelligence and machine learning in oncologic imaging]

Machine learning (ML) is entering many areas of society, including medicine. This transformation has the potential to drastically change medicine and medical practice. These aspects become particularly clear when considering the different stages of oncologic patient care and the involved interdisciplinary and intermodality interactions. In recent publications, computers-in collaboration with humans or alone-have been outperforming humans regarding tumor identification, tumor classification, estimating prognoses, and evaluation of treatments. In addition, ML algorithms, e.g., artificial neural networks (ANNs), which constitute the drivers behind many of the latest achievements in ML, can deliver this level of performance in a reproducible, fast, and inexpensive manner. In the future, artificial intelligence applications will become an integral part of the medical profession and offer advantages for oncologic diagnostics and treatment.

Morphology-Predicted Large-Scale Transition Number in Circulating Tumor Cells Identifies a Chromosomal Instability Biomarker Associated with Poor Outcome in Castration-Resistant Prostate Cancer

Chromosomal instability (CIN) increases a tumor cell's ability to acquire chromosomal alterations, a mechanism by which tumor cells evolve, adapt, and resist therapeutics. We sought to develop a biomarker of CIN in circulating tumor cells (CTC) that are more likely to reflect the genetic diversity of patient's disease than a single-site biopsy and be assessed rapidly so as to inform treatment management decisions in real time. Large-scale transitions (LST) are genomic alterations defined as chromosomal breakages that generate chromosomal gains or losses of greater than or equal to10 Mb. Here we studied the relationship between the number of LST in an individual CTC determined by direct sequencing and morphologic features of the cells. This relationship was then used to develop a computer vision algorithm that utilizes CTC image features to predict the presence of a high (9 or more) versus low (8 or fewer) LST number in a single cell. As LSTs are a primary functional component of homologous recombination deficient cellular phenotypes, the image-based algorithm was studied prospectively on 10,240 CTCs in 367 blood samples obtained from 294 patients with progressing metastatic castration-resistant prostate cancer taken prior to starting a standard-of-care approved therapy. The resultant computer vision-based biomarker of CIN in CTCs in a pretreatment sample strongly associated with poor overall survival times in patients treated with androgen receptor signaling inhibitors and taxanes. SIGNIFICANCE: A rapidly assessable biomarker of chromosomal instability in CTC is associated with poor outcomes when detected in men with progressing mCRPC.

Prognostic value of nucleotyping, DNA ploidy and stroma in high-risk stage II colon cancer

Background

Heterogeneity with respect to recurrence and survival in high-risk stage II colon cancer patients still exists, and further classification is urgently required. This study aimed to ascertain the prognostic value of DNA ploidy, stroma-tumour fraction and nucleotyping in the prognosis of high-risk stage II colon cancer.

Methods

A total of 188 high-risk stage II colon cancer patients received radical surgery in Peking University Cancer Hospital, from 2009 to 2015. Status of mismatch repair proteins in tumours was analysed using immunohistochemistry. DNA ploidy, stroma-tumour fraction and nucleotyping were estimated by automated digital imaging systems.

Results

Nucleotyping and DNA ploidy were significant prognostic factors, while stroma-tumour fraction were not significantly prognostic in the univariate analysis. In the multivariable model, the dominant contributory factor of disease-free survival was chromatin heterogeneous vs. chromatin homogeneous [HR 3.309 (95% CI: 1.668–6.564), P = 0.001].

Conclusions

Our study indicates that nucleotyping is an independent prognostic factor in high-risk stage II colon cancer. Therefore, it may help subdivide patients into different subgroups and give them different strategies for follow-up and treatment in the future.

Predicting long-term multicategory cause of death in patients with prostate cancer: random forest versus multinomial model

The majority of patients with prostate cancer die of non-cancer causes of death (COD). It is thus important to accurately predict multi-category COD in these patients. Random forest (RF), a popular machine learning model, has been shown useful for predicting binary cancer-specific deaths. However, its accuracy for predicting multi-category COD in cancer patients is unclear. We included patients in Surveillance, Epidemiology, and End Results-18 cancer registry-program with prostate cancer diagnosed in 2004 (followed-up through 2016). They were randomly divided into training and testing sets with equal sizes. We evaluated prediction accuracies of RF and conventional statistical/multinomial models for 6-category COD by data-encoding types using the 2-fold cross-validation approach. Among 49,864 prostate cancer patients, 29,611 (59.4%) were alive at the end of follow-up, and 5,448 (10.9%) died of cardiovascular disease, 4,607 (9.2%) of prostate cancer, 3,681 (7.4%) of non-prostate cancer, 717 (1.4%) of infection, and 5,800 (11.6%) of other causes. We predicted 6-category COD among these patients with a mean accuracy of 59.1% (n=240, 95% CI, 58.7%-59.4%) in RF models with one-hot encoding, and 50.4% (95% CI, 49.7%-51.0%) in multinomial models. Tumor characteristics, prostate-specific antigen level, and diagnosis confirmation-method were important in RF and multinomial models. In RF models, no statistical differences were found between the accuracies of training versus cross-validation phases, and those of categorical versus one-hot encoding. We here report that RF models can outperform multinomial logistic models (absolute accuracy-difference, 8.7%) in predicting long-term 6-category COD among prostate cancer patients, while pathology diagnosis itself and tumor pathology remain important factors.

Trop2 is a driver of metastatic prostate cancer with neuroendocrine phenotype via PARP1

NEPC is a highly aggressive subtype of prostate cancer that is increasing in incidence, likely due to use of new secondary androgen deprivation therapies. Here, we demonstrate that Trop2 is significantly elevated in CRPC and NEPC and represents a driver of metastatic NEPC. Trop2 overexpression increases tumor growth, drives metastasis and neuroendocrine phenotype, and significantly increases PARP1 levels. Inhibition of PARP1 in Trop2-driven NEPC significantly decreases neuroendocrine features, tumor growth, and metastatic colonization in vivo, suggesting that PARP1 inhibitors may represent a promising therapeutic strategy for metastatic prostate cancer expressing high levels of Trop2.

Resistance to androgen deprivation therapy, or castration-resistant prostate cancer (CRPC), is often accompanied by metastasis and is currently the ultimate cause of prostate cancer-associated deaths in men. Recently, secondary hormonal therapies have led to an increase of neuroendocrine prostate cancer (NEPC), a highly aggressive variant of CRPC. Here, we identify that high levels of cell surface receptor Trop2 are predictive of recurrence of localized prostate cancer. Moreover, Trop2 is significantly elevated in CRPC and NEPC, drives prostate cancer growth, and induces neuroendocrine phenotype. Overexpression of Trop2 induces tumor growth and metastasis while loss of Trop2 suppresses these abilities in vivo. Trop2-driven NEPC displays a significant up-regulation of PARP1, and PARP inhibitors significantly delay tumor growth and metastatic colonization and reverse neuroendocrine features in Trop2-driven NEPC. Our findings establish Trop2 as a driver and therapeutic target for metastatic prostate cancer with neuroendocrine phenotype and suggest that high Trop2 levels could identify cancers that are sensitive to Trop2-targeting therapies and PARP1 inhibition.

Machine Learning in Prediction of Second Primary Cancer and Recurrence in Colorectal Cancer

BACKGROUND

Colorectal cancer (CRC) is the third commonly diagnosed cancer worldwide. Recurrence of CRC (Re) and onset of a second primary malignancy (SPM) are important indicators in treating CRC, but it is often difficult to predict the onset of a SPM. Therefore, we used mechanical learning to identify risk factors that affect Re and SPM.

PATIENT AND METHODS

CRC patients with cancer registry database at three medical centers were identified. All patients were classified based on Re or no recurrence (NRe) as well as SPM or no SPM (NSPM). Two classifiers, namely A Library for Support Vector Machines (LIBSVM) and Reduced Error Pruning Tree (REPTree), were applied to analyze the relationship between clinical features and Re and/or SPM category by constructing optimized models.

RESULTS

When Re and SPM were evaluated separately, the accuracy of LIBSVM was 0.878 and that of REPTree was 0.622. When Re and SPM were evaluated in combination, the precision of models for SPM+Re, NSPM+Re, SPM+NRe, and NSPM+NRe was 0.878, 0.662, 0.774, and 0.778, respectively.

CONCLUSIONS

Machine learning can be used to rank factors affecting tumor Re and SPM. In clinical practice, routine checkups are necessary to ensure early detection of new tumors. The success of prediction and early detection may be enhanced in the future by applying "big data" analysis methods such as machine learning.

Disordered chromatin packing regulates phenotypic plasticity

Three-dimensional supranucleosomal chromatin packing plays a profound role in modulating gene expression by regulating transcription reactions through mechanisms such as gene accessibility, binding affinities, and molecular diffusion. Here, we use a computational model that integrates disordered chromatin packing (CP) with local macromolecular crowding (MC) to study how physical factors, including chromatin density, the scaling of chromatin packing, and the size of chromatin packing domains, influence gene expression. We computationally and experimentally identify a major role of these physical factors, specifically chromatin packing scaling, in regulating phenotypic plasticity, determining responsiveness to external stressors by influencing both intercellular transcriptional malleability and heterogeneity. Applying CPMC model predictions to transcriptional data from cancer patients, we identify an inverse relationship between patient survival and phenotypic plasticity of tumor cells.

Chromatin structure changes during various processes from a DNA sequence view

Chromatin mainly consists of protein and DNA, and the sequence information of DNA contributes to controlling the spatial structure of chromatin. Genome-wide contact patterns of chromosome at high precision uncover fine structural properties, conductive to exploring underlying mechanisms on structure establishment and function realization for chromatin. In this short review, we describe changes of chromatin structure during various biological processes from a DNA sequence view, with an increase of the overall domain segregation from birth to senescence and establishment of cell identity related cross-domain contacts. Segregation patterns vary with cell stage and genomic distance. Meanwhile, possible effects of cell cycle, temperature, nuclear lamina and nucleolus on chromatin structure are discussed. At last, important roles of transcription factors and other proteins in proper chromatin organization are also discussed.

Artificial intelligence in digital pathology - new tools for diagnosis and precision oncology

In the past decade, advances in precision oncology have resulted in an increased demand for predictive assays that enable the selection and stratification of patients for treatment. The enormous divergence of signalling and transcriptional networks mediating the crosstalk between cancer, stromal and immune cells complicates the development of functionally relevant biomarkers based on a single gene or protein. However, the result of these complex processes can be uniquely captured in the morphometric features of stained tissue specimens. The possibility of digitizing whole-slide images of tissue has led to the advent of artificial intelligence (AI) and machine learning tools in digital pathology, which enable mining of subvisual morphometric phenotypes and might, ultimately, improve patient management. In this Perspective, we critically evaluate various AI-based computational approaches for digital pathology, focusing on deep neural networks and 'hand-crafted' feature-based methodologies. We aim to provide a broad framework for incorporating AI and machine learning tools into clinical oncology, with an emphasis on biomarker development. We discuss some of the challenges relating to the use of AI, including the need for well-curated validation datasets, regulatory approval and fair reimbursement strategies. Finally, we present potential future opportunities for precision oncology.

ERO1α promotes hypoxic tumor progression and is associated with poor prognosis in pancreatic cancer

Pancreatic cancer is a leading cause of mortality worldwide due to the difficulty of detecting early-stage disease and our poor understanding of the mediators that drive progression of hypoxic solid tumors. We therefore used a heavy isotope 'pulse/trace' proteomic approach to determine how hypoxia (Hx) alters pancreatic tumor expression of proteins that confer treatment resistance, promote metastasis, and suppress host immunity. Using this method, we identified that hypoxia stress stimulates pancreatic cancer cells to rapidly translate proteins that enhance metastasis (NOTCH2, NCS1, CD151, NUSAP1), treatment resistance (ABCB6), immune suppression (NFIL3, WDR4), angiogenesis (ANGPT4, ERO1α, FOS), alter cell metabolic activity (HK2, ENO2), and mediate growth-promoting cytokine responses (CLK3, ANGPTL4). Database mining confirmed that elevated gene expression of these hypoxia-induced mediators is significantly associated with poor patient survival in various stages of pancreatic cancer. Among these proteins, the oxidoreductase enzyme ERO1α was highly sensitive to induction by hypoxia stress across a range of different pancreatic cancer cell lines and was associated with particularly poor prognosis in human patients. Consistent with these data, genetic deletion of ERO1α substantially reduced growth rates and colony formation by pancreatic cancer cells when assessed in a series of functional assays in vitro. Accordingly, when transferred into a mouse xenograft model, ERO1α-deficient tumor cells exhibited severe growth restriction and negligible disease progression in vivo. Together, these data indicate that ERO1α is potential prognostic biomarker and novel drug target for pancreatic cancer therapy.