Genomics can inform both tissue-of-origin (TOO) and precision treatments for patients with cancer of unknown primary (CUP). Here, we use whole genome and transcriptome sequencing (WGTS) for 72 patients and show diagnostic superiority of WGTS over panel testing (386-523 genes) in 71 paired cases. WGTS detects all reportable DNA features found by panel as well as additional mutations of diagnostic or therapeutic relevance in 76% of cases. Curated WGTS features and a CUP prediction algorithm (CUPPA) trained on WGTS data of known cancer types informs TOO in 71% of cases otherwise undiagnosed by clinicopathology review. WGTS informs treatments for 79% of patients, compared to 59% by panel testing. Finally, WGS of cell-free DNA (cfDNA) from patients with a high cfDNA tumour fraction (>7%), enables high-likelihood CUPPA predictions in 41% of cases. WGTS is therefore superior to panel testing, broadens treatment options, and is feasible using routine pathology samples and cfDNA.

Patients with metastatic disease and no identifiable primary tumor, thus diagnosed with cancer of unknown primary (CUP), typically have a poor prognosis. Tumor DNA sequencing has recently shown promise in identifying the molecular tissue of origin. This study evaluated the value of whole genome sequencing (WGS) in the CUP care pathway, by comparing patient outcomes with a historical control cohort. Also, the value of whole transcriptome sequencing (WTS) was explored.

A prospective intervention cohort was established of provisional CUP patients (≥18 years of age) who had WGS carried out on metastatic tissue between August 2021 and August 2023. A control cohort was established of CUP patients (≥18 years of age) diagnosed between December 2016 and April 2021 without the availability of WGS. The CUP predicting algorithm was applied to WGS data, and data on definitive diagnosis, molecular actionable alterations [ESMO Scale for Clinical Actionability of molecular Targets (ESCAT) tier 1-3], therapy, diagnostic timelines, and overall survival (OS) were captured.

A total of 159 provisional CUP patients (n = 54 intervention cohort, n = 105 control cohort) were included. WGS and WTS were successfully carried out in 46 (85%) and 27 patients (50%). A primary tumor diagnosis was established in 76% of the intervention cohort compared with 16% of the control cohort (P < 0.001). WGS contributed to a primary tumor diagnosis in 34 patients (63%) and identified an actionable alteration in 34 patients (63%). WTS contributed to a primary tumor diagnosis in three patients (6%). Following WGS, treatment recommendations could be made for 38 patients (70%), of whom 22 started the recommended therapies (58%). The median OS was 11 and 9 months in the intervention and control cohorts, respectively (P = 0.345).

Incorporation of WGS into the CUP care pathway is of significant value for diagnosing a primary tumor of origin and contributed to the identification of actionable alterations in the majority of patients.

In the dynamic landscape of targeted therapeutics, drug discovery has pivoted towards understanding underlying disease mechanisms, placing a strong emphasis on molecular perturbations and target identification. This paradigm shift, crucial for drug discovery, is underpinned by big data, a transformative force in the current era. Omics data, characterized by its heterogeneity and enormity, has ushered biological and biomedical research into the big data domain. Acknowledging the significance of integrating diverse omics data strata, known as multi-omics studies, researchers delve into the intricate interrelationships among various omics layers. This review navigates the expansive omics landscape, showcasing tailored assays for each molecular layer through genomes to metabolomes. The sheer volume of data generated necessitates sophisticated informatics techniques, with machine-learning (ML) algorithms emerging as robust tools. These datasets not only refine disease classification but also enhance diagnostics and foster the development of targeted therapeutic strategies. Through the integration of high-throughput data, the review focuses on targeting and modeling multiple disease-regulated networks, validating interactions with multiple targets, and enhancing therapeutic potential using network pharmacology approaches. Ultimately, this exploration aims to illuminate the transformative impact of multi-omics in the big data era, shaping the future of biological research.

Early diagnosis and accurate prognosis play a pivotal role in the clinical management of cancer and in preventing cancer-related mortalities. The burgeoning population of Asia in general and South Asian countries like India in particular pose significant challenges to the healthcare system. Regrettably, the demand for healthcare services in India far exceeds the available resources, resulting in overcrowded hospitals, prolonged wait times, and inadequate facilities. The scarcity of trained manpower in rural settings, lack of awareness and low penetrance of screening programs further compounded the problem. Artificial Intelligence (AI), driven by advancements in machine learning, deep learning, and natural language processing, can profoundly transform the underlying shortcomings in the healthcare industry, more for populous nations like India. With about 1.4 million cancer cases reported annually and 0.9 million deaths, India has a significant cancer burden that surpassed several nations. Further, India's diverse and large ethnic population is a data goldmine for healthcare research. Under these circumstances, AI-assisted technology, coupled with digital health solutions, could support effective oncology care and reduce the economic burden of GDP loss in terms of years of potential productive life lost (YPPLL) due to India's stupendous cancer burden. This review explores different aspects of cancer management, such as prevention, diagnosis, precision treatment, prognosis, and drug discovery, where AI has demonstrated promising clinical results. By harnessing the capabilities of AI in oncology research, healthcare professionals can enhance their ability to diagnose cancers at earlier stages, leading to more effective treatments and improved patient outcomes. With continued research and development, AI and digital health can play a transformative role in mitigating the challenges posed by the growing population and advancing the fight against cancer in India. Moreover, AI-driven technologies can assist in tailoring personalized treatment plans, optimizing therapeutic strategies, and supporting oncologists in making well-informed decisions. However, it is essential to ensure responsible implementation and address potential ethical and privacy concerns associated with using AI in healthcare.

Molecular subtypes, such as defined by The Cancer Genome Atlas (TCGA), delineate a cancer's underlying biology, bringing hope to inform a patient's prognosis and treatment plan. However, most approaches used in the discovery of subtypes are not suitable for assigning subtype labels to new cancer specimens from other studies or clinical trials. Here, we address this barrier by applying five different machine learning approaches to multi-omic data from 8,791 TCGA tumor samples comprising 106 subtypes from 26 different cancer cohorts to build models based upon small numbers of features that can classify new samples into previously defined TCGA molecular subtypes-a step toward molecular subtype application in the clinic. We validate select classifiers using external datasets. Predictive performance and classifier-selected features yield insight into the different machine-learning approaches and genomic data platforms. For each cancer and data type we provide containerized versions of the top-performing models as a public resource.

High-performance prediction of prostate cancer metastasis based on single nucleotide variations remains a challenge. Therefore, we developed a novel biologically informed deep learning framework, named M-NET, for the prediction of prostate cancer metastasis. Within the framework, we transformed single nucleotide variations into patient feature images that are optimal for fitting convolutional neural networks. Moreover, we identified significant pathways associated with the metastatic status. The experimental results showed that M-NET significantly outperformed other comparison methods based on single nucleotide variations, achieving improvements in accuracy, precision, recall, F1-score, area under the receiver operating characteristics curve, and area under the precision-recall curve by 6.3%, 8.4%, 5.1%, 0.070, 0.041, and 0.026, respectively. Furthermore, M-NET identified some important pathways associated with the metastatic status, such as signaling by the hedgehog pathway. In summary, compared with other comparative methods, M-NET exhibited a better performance in the prediction of prostate cancer metastasis.

Next-generation sequencing (NGS) has revolutionized personalized oncology care by providing exceptional insights into the complex genomic landscape. NGS offers comprehensive cancer profiling, which enables clinicians and researchers to better understand the molecular basis of cancer and to tailor treatment strategies accordingly. Targeted therapies based on genomic alterations identified through NGS have shown promise in improving patient outcomes across various cancer types, circumventing resistance mechanisms and enhancing treatment efficacy. Moreover, NGS facilitates the identification of predictive biomarkers and prognostic indicators, aiding in patient stratification and personalized treatment approaches. By uncovering driver mutations and actionable alterations, NGS empowers clinicians to make informed decisions regarding treatment selection and patient management. However, the full potential of NGS in personalized oncology can only be realized through bioinformatics analyses. Bioinformatics plays a crucial role in processing raw sequencing data, identifying clinically relevant variants, and interpreting complex genomic landscapes. This comprehensive review investigates the diverse NGS techniques, including whole-genome sequencing (WGS), whole-exome sequencing (WES), and single-cell RNA sequencing (sc-RNA-Seq), elucidating their roles in understanding the complex genomic/transcriptomic landscape of cancer. Furthermore, the review explores the integration of NGS data with bioinformatics tools to facilitate personalized oncology approaches, from understanding tumor heterogeneity to identifying driver mutations and predicting therapeutic responses. Challenges and future directions in NGS-based cancer research are also discussed, underscoring the transformative impact of these technologies on cancer diagnosis, management, and treatment strategies.

Autism is a developmental disability. Research demonstrated that children with autism benefit from early diagnosis and early intervention. Genetic factors are considered major contributors to the development of autism. Machine learning (ML), including deep learning (DL), has been evaluated in phenotype prediction, but this method has been limited in its application to autism. We developed a DL model, the Separate Translated Autism Research Neural Network (STAR-NN) model to predict autism status. The model was trained and tested using whole exome sequencing data from 43,203 individuals (16,809 individuals with autism and 26,394 non-autistic controls). Polygenic scores from common variants and the aggregated count of rare variants on genes were used as input. In STAR-NN, protein truncating variants, possibly damaging missense variants and mild effect missense variants on the same gene were separated at the input level and merged to one gene node. In this way, rare variants with different level of pathogenic effects were treated separately. We further validated the performance of STAR-NN using an independent dataset, including 13,827 individuals with autism and 14,052 non-autistic controls. STAR-NN achieved a modest ROC-AUC of 0.7319 on the testing dataset and 0.7302 on the independent dataset. STAR-NN outperformed other traditional ML models. Gene Ontology analysis on the selected gene features showed an enrichment for potentially informative pathways including calcium ion transport.

Hemangioblastoma (HB) is a highly vascularized tumor most commonly occurring in the posterior cranial fossa, requiring accurate preoperative diagnosis to avoid accidental intraoperative hemorrhage and even death.

To accurately distinguish HBs from other cerebellar-and-brainstem tumors using a convolutional neural network model based on a contrast-enhanced brain MRI dataset.

Retrospective.

Four hundred five patients (182 = HBs; 223 = other cerebellar-and brainstem tumors): 305 cases for model training, and 100 for evaluation.

3 T/contrast-enhanced T1-weighted imaging (T1WI + C).

A CNN-based 2D classification network was trained by using sliced data along the z-axis. To improve the performance of the network, we introduced demographic information, various data-augmentation methods and an auxiliary task to segment tumor region. Then, this method was compared with the evaluations performed by experienced and intermediate-level neuroradiologists, and the heatmap of deep feature, which indicates the contribution of each pixel to model prediction, was visualized by Grad-CAM for analyzing the misclassified cases.

The Pearson chi-square test and an independent t-test were used to test for distribution difference in age and sex. And the independent t-test was exploited to evaluate the performance between experts and our proposed method. P value <0.05 was considered significant.

The trained network showed a higher accuracy for identifying HBs (accuracy = 0.902 ± 0.031, F1 = 0.891 ± 0.035, AUC = 0.926 ± 0.040) than experienced (accuracy = 0.887 ± 0.013, F1 = 0.868 ± 0.011, AUC = 0.881 ± 0.008) and intermediate-level (accuracy = 0.827 ± 0.037, F1 = 0.768 ± 0.068, AUC = 0.810 ± 0.047) neuroradiologists. The recall values were 0.910 ± 0.050, 0.659 ± 0.084, and 0.828 ± 0.019 for the trained network, intermediate and experienced neuroradiologists, respectively. Additional ablation experiments verified the utility of the introduced demographic information, data augmentation, and the auxiliary-segmentation task.

Our proposed method can successfully distinguish HBs from other cerebellar-and-brainstem tumors and showed diagnostic efficiency comparable to that of experienced neuroradiologists.

3 TECHNICAL EFFICACY: Stage 2.

Metastasis driven by cancer cell migration is the leading cause of cancer-related deaths. It involves significant changes in the organization of the cytoskeleton, which includes the actin microfilaments and the vimentin intermediate filaments. Understanding how these filament change cells from normal to invasive offers insights that can be used to improve cancer diagnosis and therapy. We have developed a computational, transparent, large-scale and imaging-based pipeline, that can distinguish between normal human cells and their isogenically matched, oncogenically transformed, invasive and metastasizing counterparts, based on the spatial organization of actin and vimentin filaments in the cell cytoplasm. Due to the intricacy of these subcellular structures, manual annotation is not trivial to automate. We used established deep learning methods and our new multi-attention channel architecture. To ensure a high level of interpretability of the network, which is crucial for the application area, we developed an interpretable global explainable approach correlating the weighted geometric mean of the total cell images and their local GradCam scores. The methods offer detailed, objective and measurable understanding of how different components of the cytoskeleton contribute to metastasis, insights that can be used for future development of novel diagnostic tools, such as a nanometer level, vimentin filament-based biomarker for digital pathology, and for new treatments that significantly can increase patient survival.

Cancer of unknown primary (CUP) is a rare type of metastatic cancer in which the origin of the tumor is unknown. Since the treatment strategy for patients with metastatic tumors depends on knowing the primary site, accurate identification of the origin site is important. Here, we developed an image-based deep-learning model that utilizes a vision transformer algorithm for predicting the origin of CUP. Using DNA methylation dataset of 8,233 primary tumors from The Cancer Genome Atlas (TCGA), we categorized 29 cancer types into 18 organ classes and extracted 2,312 differentially methylated CpG sites (DMCs) from non-squamous cancer group and 420 DMCs from squamous cell cancer group. Using these DMCs, we created organ-specific DNA methylation images and used them for model training and testing. Model performance was evaluated using 394 metastatic cancer samples from TCGA (TCGA-meta) and 995 samples (693 primary and 302 metastatic cancers) obtained from 20 independent external studies. We identified that the DNA methylation image reveals a distinct pattern based on the origin of cancer. Our model achieved an overall accuracy of 96.95 % in the TCGA-meta dataset. In the external validation datasets, our classifier achieved overall accuracies of 96.39 % and 94.37 % in primary and metastatic tumors, respectively. Especially, the overall accuracies for both primary and metastatic samples of non-squamous cell cancer were exceptionally high, with 96.79 % and 96.85 %, respectively.

In recent decades, machine-learning (ML) technologies have advanced the management of high-dimensional and complex cancer data by developing reliable and user-friendly automated diagnostic tools for clinical applications. Immunohistochemistry (IHC) is an essential staining method that enables the identification of cellular origins by analyzing the expression of specific antigens within tissue samples. The aim of this study was to identify a model that could predict histopathological diagnoses based on specific immunohistochemical markers.

The XGBoost learning model was applied, where the input variable (target variable) was the histopathological diagnosis and the predictors (independent variables influencing the target variable) were the immunohistochemical markers.

Our study demonstrated a precision rate of 85.97% within the dataset, indicating a high level of performance and suggesting that the model is generally reliable in producing accurate predictions.

This study demonstrated the feasibility and clinical efficacy of utilizing the probabilistic decision tree algorithm to differentiate tumor diagnoses according to immunohistochemistry profiles.

With significant advancements of next generation sequencing technologies, large amounts of multi-omics data, including genomics, epigenomics, transcriptomics, proteomics, and metabolomics, have been accumulated, offering an unprecedented opportunity to explore the heterogeneity and complexity of cancer across various molecular levels and scales. One of the promising aspects of multi-omics lies in its capacity to offer a holistic view of the biological networks and pathways underpinning cancer, facilitating a deeper understanding of its development, progression, and response to treatment. However, the exponential growth of data generated by multi-omics studies present significant analytical challenges. Processing, analyzing, integrating, and interpreting these multi-omics datasets to extract meaningful insights is an ambitious task that stands at the forefront of current cancer research. The application of artificial intelligence (AI) has emerged as a powerful solution to these challenges, demonstrating exceptional capabilities in deciphering complex patterns and extracting valuable information from large-scale, intricate omics datasets. This review delves into the synergy of AI and multi-omics, highlighting its revolutionary impact on oncology. We dissect how this confluence is reshaping the landscape of cancer research and clinical practice, particularly in the realms of early detection, diagnosis, prognosis, treatment and pathology. Additionally, we elaborate the latest AI methods for multi-omics integration to provide a comprehensive insight of the complex biological mechanisms and inherent heterogeneity of cancer. Finally, we discuss the current challenges of data harmonization, algorithm interpretability, and ethical considerations. Addressing these challenges necessitates a multidisciplinary collaboration, paving the promising way for more precise, personalized, and effective treatments for cancer patients.

Deep learning holds immense potential for histopathology, automating tasks that are simple for expert pathologists and revealing novel biology for tasks that were previously considered difficult or impossible to solve by eye alone. However, the extent to which the visual strategies learned by deep learning models in histopathological analysis are trustworthy or not has yet to be systematically analysed. Here, we systematically evaluate deep neural networks (DNNs) trained for histopathological analysis in order to understand if their learned strategies are trustworthy or deceptive.

We trained a variety of DNNs on a novel data set of 221 whole-slide images (WSIs) from lung adenocarcinoma patients, and evaluated their effectiveness at (1) molecular profiling of KRAS versus EGFR mutations, (2) determining the primary tissue of a tumour and (3) tumour detection. While DNNs achieved above-chance performance on molecular profiling, they did so by exploiting correlations between histological subtypes and mutations, and failed to generalise to a challenging test set obtained through laser capture microdissection (LCM). In contrast, DNNs learned robust and trustworthy strategies for determining the primary tissue of a tumour as well as detecting and localising tumours in tissue.

Our work demonstrates that DNNs hold immense promise for aiding pathologists in analysing tissue. However, they are also capable of achieving seemingly strong performance by learning deceptive strategies that leverage spurious correlations, and are ultimately unsuitable for research or clinical work. The framework we propose for model evaluation and interpretation is an important step towards developing reliable automated systems for histopathological analysis.

Classifying tumor types using machine learning approaches is not always trivial, particularly for challenging cases such as cancers of unknown primary. In this issue of Cancer Discovery, Darmofal and colleagues describe a new tool that uses information from a clinical sequencing panel to diagnose tumor type, and show that the model is particularly robust. See related article by Darmofal et al., p. 1064 (1).

Tumor type guides clinical treatment decisions in cancer, but histology-based diagnosis remains challenging. Genomic alterations are highly diagnostic of tumor type, and tumor-type classifiers trained on genomic features have been explored, but the most accurate methods are not clinically feasible, relying on features derived from whole-genome sequencing (WGS), or predicting across limited cancer types. We use genomic features from a data set of 39,787 solid tumors sequenced using a clinically targeted cancer gene panel to develop Genome-Derived-Diagnosis Ensemble (GDD-ENS): a hyperparameter ensemble for classifying tumor type using deep neural networks. GDD-ENS achieves 93% accuracy for high-confidence predictions across 38 cancer types, rivaling the performance of WGS-based methods. GDD-ENS can also guide diagnoses of rare type and cancers of unknown primary and incorporate patient-specific clinical information for improved predictions. Overall, integrating GDD-ENS into prospective clinical sequencing workflows could provide clinically relevant tumor-type predictions to guide treatment decisions in real time.

We describe a highly accurate tumor-type prediction model, designed specifically for clinical implementation. Our model relies only on widely used cancer gene panel sequencing data, predicts across 38 distinct cancer types, and supports integration of patient-specific nongenomic information for enhanced decision support in challenging diagnostic situations. See related commentary by Garg, p. 906. This article is featured in Selected Articles from This Issue, p. 897.

In the last decades, the development of high-throughput molecular assays has revolutionised cancer diagnostics, paving the way for the concept of personalised cancer medicine. This progress has been driven by the introduction of such technologies through biomarker-driven oncology trials. In this review, strengths and limitations of various state-of-the-art sequencing technologies, including gene panel sequencing (DNA and RNA), whole-exome/whole-genome sequencing and whole-transcriptome sequencing, are explored, focusing on their ability to identify clinically relevant biomarkers with diagnostic, prognostic and/or predictive impact. This includes the need to assess complex biomarkers, for example microsatellite instability, tumour mutation burden and homologous recombination deficiency, to identify patients suitable for specific therapies, including immunotherapy. Furthermore, the crucial role of biomarker analysis and multidisciplinary molecular tumour boards in selecting patients for trial inclusion is discussed in relation to various trial concepts, including drug repurposing. Recognising that today's exploratory techniques will evolve into tomorrow's routine diagnostics and clinical study inclusion assays, the importance of emerging technologies for multimodal diagnostics, such as proteomics and in vivo drug sensitivity testing, is also discussed. In addition, key regulatory aspects and the importance of patient engagement in all phases of a clinical trial are described. Finally, we propose a set of recommendations for consideration when planning a new precision cancer medicine trial.

Machine learning (ML) has been successfully implemented for classification tasks (e.g., cancer diagnosis). ML performance for more challenging predictions is largely unexplored. This study's objective was to compare machine learning vs. expert-informed predictions for surgical outcome in patients undergoing major liver surgery.

Single tertiary center data on preoperative parameters and postoperative complications for elective hepatic surgery patients were included (2008-2021). Expert-informed prediction models were established on 14 parameters identified by two expert liver surgeons to impact on postoperative outcome. ML models used all available preoperative patient variables (n = 62). Model performance was compared for predicting 3-month postoperative overall morbidity. Temporal validation and additional analysis in major liver resection patients were conducted.

889 patients included. Expert-informed models showed low average bias (2-5 CCI points) with high over/underprediction. ML models performed similarly: average prediction 5-10 points higher than observed CCI values with high variability (95% CI -30 to 50). No performance improvement for major liver surgery patients.

No clinical relevance in the application of ML for predicting postoperative overall morbidity was found. Despite being a novel hype, ML has the potential for application in clinical practice. However, at this stage it does not replace established approaches of prediction modelling.

Multiple primary cancers once thought to be rare have become increasingly common as the lifespan of cancer survivors has increased with availability of better and more effective cancer treatment. However, their exact incidence is not known and data on their epidemiological characteristics are not available.

The aim of this study is to study the epidemiologic characteristics of multiple primary cancers in the eastern region of India.

The study was conducted in the Department of Surgical Oncology, Medical College, Kolkata, from 2017 to 2020 over a period of 3 years. All patients with a diagnosis of second primary as per International Agency for Research on Cancer (IARC) definition or those developing a second primary within the study period were included for analysis. Data were recorded in form of preformed questionnaires. All the cases were followed up for at least 12 months.

Fifty cases of multiple primary tumors were identified, out of which 21 were synchronous while rest 29 were metachronous type. The male-female ratio was 1:1.2. The median age at presentation for index malignancy was 50 years. The most common malignancy in the synchronous group was a combination of variety of GI cancers (six cases). In the metachronous category, a combination of reproductive cancers (breast, ovary, cervix, and endometrium) along with Gastrointestinal cancer (GI) cancers (colon, rectum) was most frequently found (eight cases). Definite risk factors for multiple primary tumors were identifiable in 10 cases: arsenic exposure in 5 cases, hereditary in 4 cases, and immunosuppression in 1, while in 8 cases, risk factors were only speculative (radiation 5 cases, chemotherapy 3). At the time of the last follow-up, 36 subjects were alive and 3 dead while the status of 11 subjects was unknown.

This is the first comprehensive study on multiple primary cancers and the largest so far in India. Our study overcomes the shortcoming of previous case series from our subcontinent. The merits of our study include the use of the most accepted IARC definition, updated staging guidelines with long follow-up, and reliable survival data. Additionally, we could identify risk factors in 50% of our subjects. And our study shows various new combinations of cancers not reported before. Clustering of cases in the young adolescent group (25-49) years is also a new finding. We also highlight the existing ambiguity in the way this entity is defined. Demerits include the loss of follow-up data in a significant number of patients.

Studying the molecular heterogeneity of cancer is essential for achieving personalized therapy. At the same time, understanding the biological processes that drive cancer development can lead to the identification of valuable therapeutic targets. Therefore, achieving accurate and interpretable clinical predictions requires paramount attention to thoroughly characterizing patients at both the molecular and biological pathway levels.

Here, we present GraphPath, a biological knowledge-driven graph neural network with multi-head self-attention mechanism that implements the pathway-pathway interaction network. We train GraphPath to classify the cancer status of patients with prostate cancer based on their multi-omics profiling. Experiment results show that our method outperforms P-NET and other baseline methods. Besides, two external cohorts are used to validate that the model can be generalized to unseen samples with adequate predictive performance. We reduce the dimensionality of latent pathway embeddings and visualize corresponding classes to further demonstrate the optimal performance of the model. Additionally, since GraphPath's predictions are interpretable, we identify target cancer-associated pathways that significantly contribute to the model's predictions. Such a robust and interpretable model has the potential to greatly enhance our understanding of cancer's biological mechanisms and accelerate the development of targeted therapies.

https://github.com/amazingma/GraphPath.

Histopathology and genomic profiling are cornerstones of precision oncology and are routinely obtained for patients with cancer. Traditionally, histopathology slides are manually reviewed by highly trained pathologists. Genomic data, on the other hand, is evaluated by engineered computational pipelines. In both applications, the advent of modern artificial intelligence methods, specifically machine learning (ML) and deep learning (DL), have opened up a fundamentally new way of extracting actionable insights from raw data, which could augment and potentially replace some aspects of traditional evaluation workflows. In this review, we summarize current and emerging applications of DL in histopathology and genomics, including basic diagnostic as well as advanced prognostic tasks. Based on a growing body of evidence, we suggest that DL could be the groundwork for a new kind of workflow in oncology and cancer research. However, we also point out that DL models can have biases and other flaws that users in healthcare and research need to know about, and we propose ways to address them.

The interrelation and complementary nature of multi-omics data can provide valuable insights into the intricate molecular mechanisms underlying diseases. However, challenges such as limited sample size, high data dimensionality and differences in omics modalities pose significant obstacles to fully harnessing the potential of these data. The prior knowledge such as gene regulatory network and pathway information harbors useful gene-gene interaction and gene functional module information. To effectively integrate multi-omics data and make full use of the prior knowledge, here, we propose a Multilevel-graph neural network (GNN): a hierarchically designed deep learning algorithm that sequentially leverages multi-omics data, gene regulatory networks and pathway information to extract features and enhance accuracy in predicting survival risk. Our method achieved better accuracy compared with existing methods. Furthermore, key factors nonlinearly associated with the tumor pathogenesis are prioritized by employing two interpretation algorithms (i.e. GNN-Explainer and IGscore) for neural networks, at gene and pathway level, respectively. The top genes and pathways exhibit strong associations with disease in survival analyses, many of which such as SEC61G and CYP27B1 are previously reported in the literature.

Inferring the cancer-type specificities of ultra-rare, genome-wide somatic mutations is an open problem. Traditional statistical methods cannot handle such data due to their ultra-high dimensionality and extreme data sparsity. To harness information in rare mutations, we have recently proposed a formal multilevel multilogistic "hidden genome" model. Through its hierarchical layers, the model condenses information in ultra-rare mutations through meta-features embodying mutation contexts to characterize cancer types. Consistent, scalable point estimation of the model can incorporate 10s of millions of variants across thousands of tumors and permit impressive prediction and attribution. However, principled statistical inference is infeasible due to the volume, correlation, and noninterpretability of mutation contexts. In this paper, we propose a novel framework that leverages topic models from computational linguistics to effectuate dimension reduction of mutation contexts producing interpretable, decorrelated meta-feature topics. We propose an efficient MCMC algorithm for implementation that permits rigorous full Bayesian inference at a scale that is orders of magnitude beyond the capability of existing out-of-the-box inferential high-dimensional multi-class regression methods and software. Applying our model to the Pan Cancer Analysis of Whole Genomes dataset reveals interesting biological insights including somatic mutational topics associated with UV exposure in skin cancer, aging in colorectal cancer, and strong influence of epigenome organization in liver cancer. Under cross-validation, our model demonstrates highly competitive predictive performance against blackbox methods of random forest and deep learning.

Histopathologic diagnosis and classification of cancer plays a critical role in guiding treatment. Advances in next-generation sequencing have ushered in new complementary molecular frameworks. However, existing approaches do not independently assess both site-of-origin (e.g. prostate) and lineage (e.g. adenocarcinoma) and have minimal validation in metastatic disease, where classification is more difficult. Utilizing gradient-boosted machine learning, we developed ATLAS, a pair of separate AI Tumor Lineage and Site-of-origin models from RNA expression data on 8249 tumor samples. We assessed performance independently in 10,376 total tumor samples, including 1490 metastatic samples, achieving an accuracy of 91.4% for cancer site-of-origin and 97.1% for cancer lineage. High confidence predictions (encompassing the majority of cases) were accurate 98-99% of the time in both localized and remarkably even in metastatic samples. We also identified emergent properties of our lineage scores for tumor types on which the model was never trained (zero-shot learning). Adenocarcinoma/sarcoma lineage scores differentiated epithelioid from biphasic/sarcomatoid mesothelioma. Also, predicted lineage de-differentiation identified neuroendocrine/small cell tumors and was associated with poor outcomes across tumor types. Our platform-independent single-sample approach can be easily translated to existing RNA-seq platforms. ATLAS can complement and guide traditional histopathologic assessment in challenging situations and tumors of unknown primary.

The use of data-driven high-throughput analytical techniques, which has given rise to computational oncology, is undisputed. The widespread use of machine learning (ML) and mathematical modeling (MM)-based techniques is widely acknowledged. These two approaches have fueled the advancement in cancer research and eventually led to the uptake of telemedicine in cancer care. For diagnostic, prognostic, and treatment purposes concerning different types of cancer research, vast databases of varied information with manifold dimensions are required, and indeed, all this information can only be managed by an automated system developed utilizing ML and MM. In addition, MM is being used to probe the relationship between the pharmacokinetics and pharmacodynamics (PK/PD interactions) of anti-cancer substances to improve cancer treatment, and also to refine the quality of existing treatment models by being incorporated at all steps of research and development related to cancer and in routine patient care. This review will serve as a consolidation of the advancement and benefits of ML and MM techniques with a special focus on the area of cancer prognosis and anticancer therapy, leading to the identification of challenges (data quantity, ethical consideration, and data privacy) which are yet to be fully addressed in current studies.

Cancer of unknown primary (CUP) presents a complex diagnostic challenge, characterized by metastatic tumors of unknown tissue origin and a dismal prognosis. This review delves into the emerging significance of artificial intelligence (AI) and machine learning (ML) in transforming the landscape of CUP diagnosis, classification, and treatment. ML approaches, trained on extensive molecular profiling data, have shown promise in accurately predicting tissue of origin. Genomic profiling, encompassing driver mutations and copy number variations, plays a pivotal role in CUP diagnosis by providing insights into tumor type-specific oncogenic alterations. Mutational signatures (MS), reflecting somatic mutation patterns, offer further insights into CUP diagnosis. Known MS with established etiology, such as ultraviolet (UV) light-induced DNA damage and tobacco exposure, have been identified in cases of dedifferentiated/transdifferentiated melanoma and carcinoma. Deep learning models that integrate gene expression data and DNA methylation patterns offer insights into tissue lineage and tumor classification. In digital pathology, machine learning algorithms analyze whole-slide images to aid in CUP classification. Finally, precision oncology, guided by molecular profiling, offers targeted therapies independent of primary tissue identification. Clinical trials assigning CUP patients to molecularly guided therapies, including targetable alterations and tumor mutation burden as an immunotherapy biomarker, have resulted in improved overall survival in a subset of patients. In conclusion, AI- and ML-driven approaches are revolutionizing CUP management by enhancing diagnostic accuracy. Precision oncology utilizing enhanced molecular profiling facilitates the identification of targeted therapies that transcend the need to identify the tissue of origin, ultimately improving patient outcomes.

Mutations in human cells exhibit increased burden in heterochromatic, late DNA replication time (RT) chromosomal domains, with variation in mutation rates between tissues mirroring variation in heterochromatin and RT. We observed that regional mutation risk further varies between individual tumors in a manner independent of cell type, identifying three signatures of domain-scale mutagenesis in >4,000 tumor genomes. The major signature reflects remodeling of heterochromatin and of the RT program domains seen across tumors, tissues and cultured cells, and is robustly linked with higher expression of cell proliferation genes. Regional mutagenesis is associated with loss of activity of the tumor-suppressor genes RB1 and TP53, consistent with their roles in cell cycle control, with distinct mutational patterns generated by the two genes. Loss of regional heterogeneity in mutagenesis is associated with deficiencies in various DNA repair pathways. These mutation risk redistribution processes modify the mutation supply towards important genes, diverting the course of somatic evolution.

Large-scale genomic data are well suited to analysis by deep learning algorithms. However, for many genomic datasets, labels are at the level of the sample rather than for individual genomic measures. Machine learning models leveraging these datasets generate predictions by using statically encoded measures that are then aggregated at the sample level. Here we show that a single weakly supervised end-to-end multiple-instance-learning model with multi-headed attention can be trained to encode and aggregate the local sequence context or genomic position of somatic mutations, hence allowing for the modelling of the importance of individual measures for sample-level classification and thus providing enhanced explainability. The model solves synthetic tasks that conventional models fail at, and achieves best-in-class performance for the classification of tumour type and for predicting microsatellite status. By improving the performance of tasks that require aggregate information from genomic datasets, multiple-instance deep learning may generate biological insight.

Hierarchical classification offers a more specific categorization of data and breaks down large classification problems into subproblems, providing improved prediction accuracy and predictive power for undefined categories, while also mitigating the impact of poor-quality data. Despite these advantages, its application in predicting primary cancer is rare. To leverage the similarity of cancers and the specificity of methylation patterns among them, we developed the Cancer Hierarchy Classification Tool (CHCT) using the idea of hierarchical classification, with methylation data from 30 cancer types and 8239 methylome samples downloaded from publicly available databases (The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO)). We used unsupervised clustering to divide the classification subproblems and screened differentially methylated sites using Analysis of variance (ANOVA) test, Tukey-kramer test, and Boruta algorithms to construct models for each classifier module. After validation, CHCT accurately classified 1568 out of 1660 cases in the test set, with an average accuracy of 94.46%. We further curated an independent validation cohort of 677 cancer samples from GEO and assigned a diagnosis using CHCT, which showed high diagnostic potential with generally high accuracies (an average accuracy of 91.40%). Moreover, CHCT demonstrates predictive capability for additional cancer types beyond its original classifier scope as demonstrated in the medulloblastoma and pituitary tumor datasets. In summary, CHCT can hierarchically classify primary cancer by methylation profile, by splitting a large-scale classification of 30 cancer types into ten smaller classification problems. These results indicate that cancer hierarchical classification has the potential to be an accurate and robust cancer classification method.

Machine learning (ML) represents a powerful tool to capture relationships between molecular alterations and cancer types and to extract biological information. Here, we developed a plain ML model aimed at distinguishing cancer types based on genetic lesions, providing an additional tool to improve cancer diagnosis, particularly for tumors of unknown origin.

TCGA data from 9,927 samples spanning 32 different cancer types were downloaded from cBioportal. A vector space model type data transformation technique was designed to build consistently homogeneous new datasets containing, as predictive features, calls for somatic point mutations and copy number variations at chromosome arm-level, thus allowing the use of the XGBoost classifier models. Considering the imbalance in the dataset, due to large difference in the number of cases for each tumor, two preprocessing strategies were considered: i) setting a percentage cut-off threshold to remove less represented cancer types, ii) dividing cancer types into different groups based on biological criteria and training a specific XGBoost model for each of them. The performance of all trained models was mainly assessed by the out-of-sample balanced accuracy (BACC) and the AUC scores.

The XGBoost classifier achieved the best performance (BACC 77%; AUC 97%) on a dataset containing the 10 most represented tumor types. Moreover, dividing the 18 most represented cancers into three different groups (endocrine-related carcinomas, other carcinomas and other cancers),such analysis models achieved 78%, 71% and 86% BACC, respectively, with AUC scores greater than 96%. In addition, the model capable of linking each group to a specific cancer type reached 81% BACC and 94% AUC. Overall, the diagnostic potential of our model was comparable/higher with respect to others already described in literature and based on similar molecular data and ML approaches.

A boosted ML approach able to accurately discriminate different cancer types was developed. The methodology builds datasets simpler and more interpretable than the original data, while keeping enough information to accurately train standard ML models without resorting to sophisticated Deep Learning architectures. In combination with histopathological examinations, this approach could improve cancer diagnosis by using specific DNA alterations, processed by a replicable and easy-to-use automated technology. The study encourages new investigations which could further increase the classifier's performance, for example by considering more features and dividing tumors into their main molecular subtypes.

Cancer is the second major cause of disease-related death worldwide, and its accurate early diagnosis and therapeutic intervention are fundamental for saving the patient's life. Cancer, as a complex and heterogeneous disorder, results from the disruption and alteration of a wide variety of biological entities, including genes, proteins, mRNAs, miRNAs, and metabolites, that eventually emerge as clinical symptoms. Traditionally, diagnosis is based on clinical examination, blood tests for biomarkers, the histopathology of a biopsy, and imaging (MRI, CT, PET, and US). Additionally, omics biotechnologies help to further characterize the genome, metabolome, microbiome traits of the patient that could have an impact on the prognosis and patient's response to the therapy. The integration of all these data relies on gathering of several experts and may require considerable time, and, unfortunately, it is not without the risk of error in the interpretation and therefore in the decision. Systems biology algorithms exploit Artificial Intelligence (AI) combined with omics technologies to perform a rapid and accurate analysis and integration of patient's big data, and support the physician in making diagnosis and tailoring the most appropriate therapeutic intervention. However, AI is not free from possible diagnostic and prognostic errors in the interpretation of images or biochemical-clinical data. Here, we first describe the methods used by systems biology for combining AI with omics and then discuss the potential, challenges, limitations, and critical issues in using AI in cancer research.

Identifying the primary site of metastatic cancer is critical to guiding the subsequent treatment. Approximately 3-9% of metastatic patients are diagnosed with cancer of unknown primary sites (CUP) even after a comprehensive diagnostic workup. However, a widely accepted molecular test is still not available. Here, we report a method that applies formalin-fixed, paraffin-embedded tissues to construct reduced representation bisulfite sequencing libraries (FFPE-RRBS). We then generate and systematically evaluate 28 molecular classifiers, built on four DNA methylation scoring methods and seven machine learning approaches, using the RRBS library dataset of 498 fresh-frozen tumor tissues from primary cancer patients. Among these classifiers, the beta value-based linear support vector (BELIVE) performs the best, achieving overall accuracies of 81-93% for identifying the primary sites in 215 metastatic patients using top-k predictions (k = 1, 2, 3). Coincidentally, BELIVE also successfully predicts the tissue of origin in 81-93% of CUP patients (n = 68).

Tumor type guides clinical treatment decisions in cancer, but histology-based diagnosis remains challenging. Genomic alterations are highly diagnostic of tumor type, and tumor type classifiers trained on genomic features have been explored, but the most accurate methods are not clinically feasible, relying on features derived from whole genome sequencing (WGS), or predicting across limited cancer types. We use genomic features from a dataset of 39,787 solid tumors sequenced using a clinical targeted cancer gene panel to develop Genome-Derived-Diagnosis Ensemble (GDD-ENS): a hyperparameter ensemble for classifying tumor type using deep neural networks. GDD-ENS achieves 93% accuracy for high-confidence predictions across 38 cancer types, rivalling performance of WGS-based methods. GDD-ENS can also guide diagnoses on rare type and cancers of unknown primary, and incorporate patient-specific clinical information for improved predictions. Overall, integrating GDD-ENS into prospective clinical sequencing workflows has enabled clinically-relevant tumor type predictions to guide treatment decisions in real time.

Single nucleotide mutation rates have critical implications for human evolution and genetic diseases. Importantly, the rates vary substantially across the genome and the principles underlying such variations remain poorly understood. A recent model explained much of this variation by considering higher-order nucleotide interactions in the 7-mer sequence context around mutated nucleotides. This model's success implicates a connection between DNA shape and mutation rates. DNA shape, i.e. structural properties like helical twist and tilt, is known to capture interactions between nucleotides within a local context. Thus, we hypothesized that changes in DNA shape features at and around mutated positions can explain mutation rate variations in the human genome. Indeed, DNA shape-based models of mutation rates showed similar or improved performance over current nucleotide sequence-based models. These models accurately characterized mutation hotspots in the human genome and revealed the shape features whose interactions underlie mutation rate variations. DNA shape also impacts mutation rates within putative functional regions like transcription factor binding sites where we find a strong association between DNA shape and position-specific mutation rates. This work demonstrates the structural underpinnings of nucleotide mutations in the human genome and lays the groundwork for future models of genetic variations to incorporate DNA shape.

DNA methylation analysis has been applied to determine the primary site of cancer; however, robust and accurate prediction of cancer types with a minimum number of sites is still a significant scientific challenge. To build an accurate and robust cancer type prediction tool with a minimum number of DNA methylation sites, we internally benchmarked different DNA methylation site selection and ranking procedures, as well as different classification models. We used The Cancer Genome Atlas dataset (26 cancer types with 8296 samples) to train and test models and used an independent dataset (17 cancer types with 2738 samples) for model validation. A deep neural network model using a combined feature selection procedure (named MethyDeep) can predict 26 cancer types using 30 methylation sites with superior performance compared with the known methods for both primary and metastatic cancers in independent validation datasets. In conclusion, MethyDeep is an accurate and robust cancer type predictor with the minimum number of DNA methylation sites; it could help the cost-effective clarification of cancer of unknown primary patients and the liquid biopsy-based early screening of cancers.

Introduction: Existing large-scale preclinical cancer drug response databases provide us with a great opportunity to identify and predict potentially effective drugs to combat cancers. Deep learning models built on these databases have been developed and applied to tackle the cancer drug-response prediction task. Their prediction has been demonstrated to significantly outperform traditional machine learning methods. However, due to the "black box" characteristic, biologically faithful explanations are hardly derived from these deep learning models. Interpretable deep learning models that rely on visible neural networks (VNNs) have been proposed to provide biological justification for the predicted outcomes. However, their performance does not meet the expectation to be applied in clinical practice. Methods: In this paper, we develop an XMR model, an eXplainable Multimodal neural network for drug Response prediction. XMR is a new compact multimodal neural network consisting of two sub-networks: a visible neural network for learning genomic features and a graph neural network (GNN) for learning drugs' structural features. Both sub-networks are integrated into a multimodal fusion layer to model the drug response for the given gene mutations and the drug's molecular structures. Furthermore, a pruning approach is applied to provide better interpretations of the XMR model. We use five pathway hierarchies (cell cycle, DNA repair, diseases, signal transduction, and metabolism), which are obtained from the Reactome Pathway Database, as the architecture of VNN for our XMR model to predict drug responses of triple negative breast cancer. Results: We find that our model outperforms other state-of-the-art interpretable deep learning models in terms of predictive performance. In addition, our model can provide biological insights into explaining drug responses for triple-negative breast cancer. Discussion: Overall, combining both VNN and GNN in a multimodal fusion layer, XMR captures key genomic and molecular features and offers reasonable interpretability in biology, thereby better predicting drug responses in cancer patients. Our model would also benefit personalized cancer therapy in the future.

Cancer of unknown primary (CUP) is a type of cancer that cannot be traced back to its primary site and accounts for 3-5% of all cancers. Established targeted therapies are lacking for CUP, leading to generally poor outcomes. We developed OncoNPC, a machine-learning classifier trained on targeted next-generation sequencing (NGS) data from 36,445 tumors across 22 cancer types from three institutions. Oncology NGS-based primary cancer-type classifier (OncoNPC) achieved a weighted F1 score of 0.942 for high confidence predictions ([Formula: see text]) on held-out tumor samples, which made up 65.2% of all the held-out samples. When applied to 971 CUP tumors collected at the Dana-Farber Cancer Institute, OncoNPC predicted primary cancer types with high confidence in 41.2% of the tumors. OncoNPC also identified CUP subgroups with significantly higher polygenic germline risk for the predicted cancer types and with significantly different survival outcomes. Notably, patients with CUP who received first palliative intent treatments concordant with their OncoNPC-predicted cancers had significantly better outcomes (hazard ratio (HR) = 0.348; 95% confidence interval (CI) = 0.210-0.570; P = [Formula: see text]). Furthermore, OncoNPC enabled a 2.2-fold increase in patients with CUP who could have received genomically guided therapies. OncoNPC thus provides evidence of distinct CUP subgroups and offers the potential for clinical decision support for managing patients with CUP.

The molecular mechanisms of the liver metastasis of colorectal cancer (CRLM) remain poorly understood. Here, we applied machine learning and bioinformatics trajectory inference to analyze a gene expression dataset of CRLM. We studied the co-regulation patterns at the gene level, the potential paths of tumor development, their functional context, and their prognostic relevance. Our analysis confirmed the subtyping of five liver metastasis subtypes (LMS). We provide gene-marker signatures for each LMS, and a comprehensive functional characterization that considers both the hallmarks of cancer and the tumor microenvironment. The ordering of CRLMs along a pseudotime-tree revealed a continuous shift in expression programs, suggesting a developmental relationship between the subtypes. Notably, trajectory inference and personalized analysis discovered a range of epigenetic states that shape and guide metastasis progression. By constructing prognostic maps that divided the expression landscape into regions associated with favorable and unfavorable prognoses, we derived a prognostic expression score. This was associated with critical processes such as epithelial-mesenchymal transition, treatment resistance, and immune evasion. These factors were associated with responses to neoadjuvant treatment and the formation of an immuno-suppressive, mesenchymal state. Our machine learning-based molecular profiling provides an in-depth characterization of CRLM heterogeneity with possible implications for treatment and personalized diagnostics.