Improving the proportion of patients diagnosed with early-stage cancer is a key priority of the World Health Organisation. In many tumour groups, screening programmes have led to improvements in survival, but patient selection and risk stratification are key challenges. In addition, there are concerns about limited diagnostic workforces, particularly in light of the COVID-19 pandemic, placing a strain on pathology and radiology services. In this review, we discuss how artificial intelligence algorithms could assist clinicians in (1) screening asymptomatic patients at risk of cancer, (2) investigating and triaging symptomatic patients, and (3) more effectively diagnosing cancer recurrence. We provide an overview of the main artificial intelligence approaches, including historical models such as logistic regression, as well as deep learning and neural networks, and highlight their early diagnosis applications. Many data types are suitable for computational analysis, including electronic healthcare records, diagnostic images, pathology slides and peripheral blood, and we provide examples of how these data can be utilised to diagnose cancer. We also discuss the potential clinical implications for artificial intelligence algorithms, including an overview of models currently used in clinical practice. Finally, we discuss the potential limitations and pitfalls, including ethical concerns, resource demands, data security and reporting standards.

Lung cancer is the most commonly diagnosed cancer and the leading cause of cancer deaths in both sexes combined. Recent years have seen major advances in the diagnostic and treatment options for patients with non-small-cell lung cancer (NSCLC), including the routine use of 2-deoxy-2[¹⁸F]-fluoro-D-glucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) in staging and response evaluation, minimally invasive endoscopic biopsy, targeted radiotherapy, minimally invasive surgery, and molecular and immunotherapies. In this review, the central roles of CT and ¹⁸F-FDG PET/CT in staging and response in both NSCLC and malignant pleural mesothelioma (MPM) are critically assessed. The Tumour Node Metastases (TNM-8) staging systems for NSCLC and MPM are presented with critical appraisal of the strengths and pitfalls of imaging. Overviews of the Response Evaluation Criteria in Solid Tumours (RECIST 1.1) for NSCLC and the modified RECIST criteria for MPM are provided, together with discussion of the benefits and limitations of these anatomical-based tools. Metabolic response assessment (not evaluated by RECIST 1.1) will be explored. We introduce the Positron Emission Tomography Response Criteria in Solid Tumours (PERCIST 1.0) to include its advantages and challenges. The limitations of both anatomical and metabolic assessment criteria when applied to NSCLC treated with immunotherapy and the important concept of pseudoprogression are addressed with reference to immune RECIST (iRECIST). Separate consideration is given to the diagnosis and follow up of solitary pulmonary nodules with reference to the British Thoracic Society guidelines and Fleischner guidelines and use of the Brock (CT-based) and Herder (addition of ¹⁸F-FDG PET/CT) models for assessing malignant potential. We discuss how these models inform decisions by the multidisciplinary team, including referral of suspicious nodules for non-surgical management in patients unsuitable for surgery. We briefly outline current lung screening systems being used in the UK, Europe and North America. Emerging roles for MRI in lung cancer imaging are reviewed. The use of whole-body MRI in diagnosing and staging NSCLC is discussed with reference to the recent multicentre Streamline L trial. The potential use of diffusion-weighted MRI to distinguish tumour from radiotherapy-induced lung toxicity is discussed. We briefly summarise the new PET-CT radiotracers being developed to evaluate specific aspects of cancer biology, other than glucose uptake. Finally, we describe how CT, MRI and ¹⁸F-FDG PET/CT are moving from primarily diagnostic tools for lung cancer towards having utility in prognostication and personalised medicine with the agency of artificial intelligence.

Surveillance is universally recommended for non-small cell lung cancer (NSCLC) patients treated with curative-intent radiotherapy. High-quality evidence to inform optimal surveillance strategies is lacking. Machine learning demonstrates promise in accurate outcome prediction for a variety of health conditions. The purpose of this study was to utilise readily available patient, tumour, and treatment data to develop, validate and externally test machine learning models for predicting recurrence, recurrence-free survival (RFS) and overall survival (OS) at 2 years from treatment. A retrospective, multicentre study of patients receiving curative-intent radiotherapy for NSCLC was undertaken. A total of 657 patients from 5 hospitals were eligible for inclusion. Data pre-processing derived 34 features for predictive modelling. Combinations of 8 feature reduction methods and 10 machine learning classification algorithms were compared, producing risk-stratification models for predicting recurrence, RFS and OS. Models were compared with 10-fold cross validation and an external test set and benchmarked against TNM-stage and performance status. Youden Index was derived from validation set ROC curves to distinguish high and low risk groups and Kaplan-Meier analyses performed. Median follow-up time was 852 days. Parameters were well matched across training-validation and external test sets: Mean age was 73 and 71 respectively, and recurrence, RFS and OS rates at 2 years were 43% vs 34%, 54% vs 47% and 54% vs 47% respectively. The respective validation and test set AUCs were as follows: 1) RFS: 0·682 (0·575–0·788) and 0·681 (0·597–0·766), 2) Recurrence: 0·687 (0·582–0·793) and 0·722 (0·635–0·81), and 3) OS: 0·759 (0·663–0·855) and 0·717 (0·634–0·8). Our models were superior to TNM stage and performance status in predicting recurrence and OS. This robust and ready to use machine learning method, validated and externally tested, sets the stage for future clinical trials entailing quantitative personalised risk-stratification and surveillance following curative-intent radiotherapy for NSCLC. A full list of funding bodies that contributed to this study can be found in the Acknowledgements section.

Artificial Intelligence (AI) continues to pervade several aspects of healthcare with pace and scale. The need for an ethical framework in AI to address this has long been recognized, but to date most efforts have delivered only high-level principles and value statements. Herein, we explain the need for an ethical framework in healthcare AI, the different moral theories that may serve as its basis, the rationale for why we believe this should be built around virtue ethics, and explore this in the context of five key ethical concerns for the introduction of AI in healthcare. Some existing work has suggested that AI may replace clinicians. We argue to the contrary, that the clinician will not be replaced, nor their role attenuated. Rather, they will be integral to the responsible design, deployment, and regulation of AI in healthcare, acting as the moral exemplar for the virtuous machine. We collate relevant points from the literature and formulate our own to present a coherent argument for the central role of clinicians in ethical AI and propose ideas to help advance efforts to employ ML-based solutions within healthcare. Finally, we highlight the responsibility of not only clinicians, but also data scientists, tech companies, ethicists, and regulators to act virtuously in realising the vision of ethical and accountable AI in healthcare.

Handcrafted and deep learning (DL) radiomics are popular techniques used to develop computed tomography (CT) imaging-based artificial intelligence models for COVID-19 research. However, contrast heterogeneity from real-world datasets may impair model performance. Contrast-homogenous datasets present a potential solution. We developed a 3D patch-based cycle-consistent generative adversarial network (cycle-GAN) to synthesize non-contrast images from contrast CTs, as a data homogenization tool. We used a multi-centre dataset of 2078 scans from 1,650 patients with COVID-19. Few studies have previously evaluated GAN-generated images with handcrafted radiomics, DL and human assessment tasks. We evaluated the performance of our cycle-GAN with these three approaches. In a modified Turing-test, human experts identified synthetic vs acquired images, with a false positive rate of 67% and Fleiss’ Kappa 0.06, attesting to the photorealism of the synthetic images. However, on testing performance of machine learning classifiers with radiomic features, performance decreased with use of synthetic images. Marked percentage difference was noted in feature values between pre- and post-GAN non-contrast images. With DL classification, deterioration in performance was observed with synthetic images. Our results show that whilst GANs can produce images sufficient to pass human assessment, caution is advised before GAN-synthesized images are used in medical imaging applications.

The rich chemical information from tissue metabolomics provides a powerful means to elaborate tissue physiology or tumor characteristics at cellular and tumor microenvironment levels. However, the process of obtaining such information requires invasive biopsies, is costly, and can delay clinical patient management. Conversely, computed tomography (CT) is a clinical standard of care but does not intuitively harbor histological or prognostic information. Furthermore, the ability to embed metabolome information into CT to subsequently use the learned representation for classification or prognosis has yet to be described. This study develops a deep learning-based framework -- tissue-metabolomic-radiomic-CT (TMR-CT) by combining 48 paired CT images and tumor/normal tissue metabolite intensities to generate ten image embeddings to infer metabolite-derived representation from CT alone. In clinical NSCLC settings, we ascertain whether TMR-CT results in an enhanced feature generation model solving histology classification/prognosis tasks in an unseen international CT dataset of 742 patients. TMR-CT non-invasively determines histological classes - adenocarcinoma/squamous cell carcinoma with an F1-score = 0.78 and further asserts patients’ prognosis with a c-index = 0.72, surpassing the performance of radiomics models and deep learning on single modality CT feature extraction. Additionally, our work shows the potential to generate informative biology-inspired CT-led features to explore connections between hard-to-obtain tissue metabolic profiles and routine lesion-derived image data.