Review••

Artificial intelligence across oncology specialties: current applications and emerging tools

•,,,,,,,,.

...

Abstract

Oncology is becoming increasingly personalised through advancements in precision in diagnostics and therapeutics, with more and more data available on both ends to create individualised plans. The depth and breadth of data are outpacing our natural ability to interpret it. Artificial intelligence (AI) provides a solution to ingest and digest this data deluge to improve detection, prediction and skill development. In this review, we provide multidisciplinary perspectives on oncology applications touched by AI—imaging, pathology, patient triage, radiotherapy, genomics-driven therapy and surgery—and integration with existing tools—natural language processing, digital twins and clinical informatics.

Introduction

In recent years, modern medical machine learning (ML) and artificial intelligence (AI) have become actualised in the clinic, with hospitals and clinics around the globe using AI assistance for a wide range of clinical applications such as diabetic retinopathy screening,1 stroke detection2 and predicting hospital readmissions.3

AI promises to be a gamechanger in oncology as well, yet oncology is no stranger to moonshot promises. Soon after the initiation of the human genome project around the turn of the century, several promising laboratory-developed tests that purported to diagnose or personalise therapy in ovarian and lung cancer were developed and subsequently debunked.4 Despite these well-documented early setbacks,5 continued refinement of multigene assays has resulted in their integration into the standard of care in certain contexts, notably in breast and prostate cancer.6–8

Many of the hard lessons learnt from entrusting decision-making to multigene laboratory tests remain, with both familiar and new questions regarding generalisability, imbalance and missing data and real-world applications.4 Oncology is a multidisciplinary practice with imperfect information flow between subspecialties. Our primary aim in this narrative review is to address this limitation by discussing ongoing and emerging oncology AI efforts from the perspective of practitioner researchers in clinical informatics (CI), computer science, medical oncology, medical physics, pathology, radiation oncology, radiology and surgery. For more details on AI and ML algorithms, we refer readers to prior reviews.9–12 For a more in-depth discussion about data sharing, ethics and validation, we refer readers to prior work.13

Current applications of AI in oncology

Radiological imaging

Cancer imaging, especially for screening and early detection, presents the archetype for integrating deep learning and AI algorithms into clinical practice. In theory, these algorithms should be able to identify patterns within medical images imperceptible to the human eye and be able to help identify areas that may represent malignant findings. Examples of ongoing efforts for early cancer detection on medical imaging using AI include mammography screening, CT lung cancer screening and early detection on prostate MRI.

Early breast cancer detection

The most studied area for AI and cancer imaging is the use of AI algorithms for earlier breast cancer detection on both two-dimensional and three-dimensional mammography. The DREAM Digital Mammography Challenge, the large-crowdsourced effort in deep learning algorithm development for medical imaging interpretation, launched the development of multiple competing commercial AI algorithms.14 Thus far, early cohort studies suggest that AI improves overall accuracy when used as an adjunct tool by radiologists interpreting mammograms.15 A few studies suggest that AI as a standalone tool may be equivalent to human interpretative accuracy for breast cancer screening.16 Of note, there is a paucity of data demonstrating that there is improved accuracy and outcomes from commercially available, Food and Drug Administration (FDA)-approved AI algorithms for mammography in large, diverse patient populations.17 Initial interim results from a prospective trial in Sweden comparing AI-supported screening versus standard double reading suggests that AI-supported screening leads to similar cancer detection rates.18 Currently, many commercially available algorithms remain unreimbursed given the lack of prospective, clinical effectiveness data.

Tumour characterisation

AI algorithms are being developed to aid in the automated characterisation of intratumoral heterogeneity, potentially allowing for more precise monitoring of disease progression and treatment efficacy.19 Traditional tumour segmentation correlates quantitative imaging features with biological data, including genetic data and molecular signatures. AI can remove the manually laborious work of supervised tumour segmentation, providing the benefits of unsupervised biologic characterisation of tumours. With deep learning and radiomic evaluation of tumour morphology and heterogeneity, more precise monitoring of treatment response of solid tumours can be determined through serial imaging (eg, CT scan, MRI, positron emission tomography (PET)/CT scan) that surpasses current oversimplified criteria such as the Response Evaluation Criteria in Solid Tumors that relies solely on size changes.20 Such efforts in AI-driven advanced imaging treatment response evaluation hold promise for more precise and personalised treatment decision-making for multiple cancer types, including cancer of the lung,21 22 liver,23 pancreas24 25 and head and neck,26–28 as well as for metastatic cancer.29

Pathology

The field of oncologic pathology has been transformed with the introduction of whole slide image (WSI) scanners. By digitising entire glass slides, this major technological advancement has enabled the application of AI and other advanced computational methods to whole tissue specimens in a manner not possible via traditional light microscopy.30

Feature-based extraction

The emerging field of pathomics is a leading example of WSI applications. Pathomics is a high-throughput approach to digital tissue phenotyping, where computational methods are employed to transcribe otherwise unstructured imaging data (ie, pixel-level WSI information) into structured imaging features and actionable knowledge.31 Analogous to radiomics—where quantitative features are calculated from radiology images—pathomics is based on the extraction and analysis of quantitative features derived from digitised tissue samples.31

The pathomic process typically starts with deep learning algorithms that detect and segment different tissue compartments (eg, tumour, stroma, etc) and/or cell types (eg, lymphocytes, cancer cells, etc), from which quantitative features are extracted. These features collectively describe unique patterns, which can serve as digital fingerprints of the tumour microenvironment and may capture imaging characteristics that are either difficult or impossible to characterise by humans. For example, morphological features can describe the size, shape and orientation of individual nuclei; topological features can provide measures of the spatial tissue architecture and interaction between different cell types; and texture features can quantify spatially encoded pixel intensity patterns of chromatin. Illustrating examples in oncology range from pathomic-based prognostic signatures for gastric cancer,32 bladder cancer,33 renal cell carcinoma,34 oropharyngeal cancer35 and breast cancer36 to predictors of microsatellite instability in colorectal cancer37 and mutational status of BRAF-mutated melanomas.38

Deep learning representations

The advantage of hand-crafted feature engineering is that features are easily interpretable. However, they are limited by human intuition. In contrast, deep learning provides an alternative form of pathomic feature extraction, where features are learnt from imaging filters in an unsupervised fashion. While deep pathomic features are difficult to interpret, they can go beyond human intuition. Although deep learning is often thought of as a complete classifier,39–41 it is primarily a kind of image representation.42 As stated above, deep learning is a particularly powerful tool for pixel-level object detection and segmentation (which is often the first step of a pathomics problem). Illustrating examples in oncology include segmentation of epithelial tissue in prostate cancer,43 cellular detection and classification in colon cancer44 and bone marrow,45 counting tumour infiltrating lymphocytes46 and immune cell composition47 and segmentation of nuclei in cervical tissue for squamous epithelium cervical intraepithelial neoplasia grading.48

Patient triage

The dual tasks of prognostication and triage can drive goals of care discussion or referrals to palliative care and hospice, yet remain challenging for oncologists, who tend to overestimate patients’ survival.49 50 Several groups have performed prospective trials based on data-driven triage frameworks which we discuss here (table 1).51

Table 1

Examples of prospective AI-driven clinical trials in oncology

Predicting unplanned hospital visits

Hong et al at Duke University aimed to triage patients undergoing radiation therapy by predicting acute care visits—emergency department or hospital admissions—to improve outcomes and decrease costs. After developing a gradient tree boosting (GTB) model using tabular electronic health record (EHR) data,52 they performed a randomised controlled trial for patients at high risk of acute care visits to show that two times per week visits during radiation (compared with standard one time per week clinic visits) decrease such visits from 22% to 12%.53

Predicting survival

Parikh et al at the University of Pennsylvania aimed to predict which patients with metastatic cancer were at high risk of 180-day mortality to trigger timely serious illness conversations (SIC). After fixing the event rate (the proportion of patients flagged as high risk) to 2% to avoid alert fatigue, they developed a model with approximately 50% positive predictive value (precision).54 After running a silent trial (ie, without alerting clinicians) of their GTB model confirming their prior results (table 2),55 the authors followed up with an interventional randomised stepped-wedge cluster trial.56 The clusters received either a behavioural nudge (weekly emails including up to six high-risk patients) or usual care. In the subset of high-risk encounters, nudges increased SIC from 3.6% to 15.2%, more than tripling the increase seen in all encounters. Of note, the model sensitivity was approximately 25%, but event rate can be tuned to allow more high-risk predictions at the cost of higher risk of alert fatigue due to more false positives.57

Table 2

Confusion matrix summarising performance of the prospective silent trial of the Penn gradient boosted tree model to predict patients at high risk of death within 180 days16

Gensheimer et al at Stanford used clinical notes in addition to EHR, inpatient billing and registry data to train a Cox proportional hazards model to predict overall survival for metastatic solid tumours.58 Text was modelled through a bag of words approach by tallying the top 100 000 one to two words phrases. In a subsequent study, life expectancy estimation was compared between this text-driven model, oncologists and a baseline performance status-based model, with the natural language processing (NLP)-driven model shown to be superior.59 In a quality improvement interventional trial, the rate of advance care planning in clinics that received weekly emails of high-risk patients (predicted survival<2 years) increased to 35% compared with 3% in a control cohort of clinics.60 See the ‘Large language models’ section for further developments in NLP.

Radiation oncology

Organ at risk segmentation

In radiation oncology, AI-assisted autosegmentation of tumour and normal organ contours has made major advancements from the early era of intensity analysis and shape modelling to the modern techniques using deep learning.61 An early randomised trial by Walker et al at MD Anderson demonstrated that resident physicians using autosegmentation assistance had time savings compared with manual segmentation.62 Autosegmentation is also critical to facilitate online adaptive radiotherapy where real-time treatment modification occurs to account for daily changes in setup, anatomy or density.63 Proprietary online adaptive radiotherapy platforms incorporate AI-driven autosegmentation, deformable registration and treatment planning64 with ongoing trials in several disease sites, including a randomised trial in lung cancer (table 3).65

Table 3

Examples of prospective AI-driven clinical trials for radiotherapy planning

Treatment de-escalation

In head and neck cancer, radiation with or without concurrent systemic therapy is a standard of care for curative intent treatment, though not without significant toxicity due to sensitive adjacent organs. Standard radiation fields may include neck regions that are not grossly involved due to concern they may be harbouring subclinical disease. Chen et al at UT Southwestern sought to use computer vision to pinpoint which nodes were involved and only target these to decrease the toxicity of elective nodal radiation. They developed and prospectively validated a hybrid radiomics and convolutional neural network (CNN) model to predict neck node involvement in INRT-AIR de-escalation trial that performed involved node radiation (INRT) with promising early results showing no solitary elective nodal failure.66 Building on INRT-AIR, Sher et al are using INRT with or without near marginless daily adaptation in the DARTBOARD trial (table 3).67

The presence of extranodal extension (ENE) in neck nodes portends more aggressive cancer, though it is very challenging to detect on imaging.68 Kann et al developed and validated a deep CNN model to predict pathologic ENE on CT imaging69 and used ECOG-ACRIN E3311 data in a quality improvement study to suggest better ability to predict pathologic ENE compared with head and neck radiologists.70

Genomics-driven precision oncology

Several categories of precision oncology using clinicogenomic data are emerging, ranging from improving prognostication to biomarker selection to drug development. Although still mainly in the research stage, there are several areas where AI may improve outcomes through refining treatment selection.

Prognostic and predictive tools through clinicogenomic data modelling

Complex clinicogenomic data can be synthesised to improve assessments of natural history and predict therapy benefits. For example, a predictive model that integrates molecular data with traditional clinicopathologic features can provide more nuanced stratification of leukaemic transformation than the Revised International Prognostic Scoring System.71 Clinicogenomic signatures can also predict benefit from adjuvant chemotherapy in patients with gastric cancer,72 and individual liver metastases with HER2-amplified metastatic colorectal cancer treated with dual HER2-targeted therapy with potential for detecting non-responding lesions.73

One emerging trend is to generate complex signatures predicting therapy utility by merging multimodal data.74 Multimodal models can predict response to immune checkpoint inhibitors (ICI) more precisely in patients with lung cancer, through combining mutational burdens, radiomic features and PD-L1 immunohistochemistry assay.75 In patients with castration-resistant prostate cancer, genomics and transcriptomics data have been combined using ML to predict a response to androgen receptor antagonists.76

Biomarker selection with non-destructive virtual assays

AI methods can derive surrogate biomarkers for targeting the molecular alterations of clinical significance in tumour tissues.77 This is helpful when universal testing is not possible or too costly, or in rare variants where a specific biomarker may not be available. Using non-destructive ‘virtual assays’ also minimises serial confirmatory assays, thus preserving biopsy tissues and resulting in better turnaround time. Several studies have demonstrated the potential of using digital pathology and radiomics in such applications. As an example, screening of mismatch repair status on H&E-stained pathology slides and other imaging methods without immunohistochemistry has shown promising results in practice.78–80 This has significant practical implications in identifying ICI response in a small number of patients with a microsatellite unstable phenotype without universal testing. Applying radiogenomic assessment on CT images can identify non-small cell lung cancer (NSCLC) harbouring EGFR mutations based on texture features from pretreatment CT and PET/CT images, suggesting the potential to screen for precise actionable molecular alterations without needing invasive biopsy.81 Radiogenomic analysis has demonstrated a link between CT imaging and radiation-induced changes in cell-free DNA obtained via liquid biopsy of locally advanced lung cancers.82 83

Structured insights for drug development through systematic repurposing

AI can assist in drug development by performing high-dimensional database analysis to screen for candidate drugs and regimens for trial development. In the secondary analysis of the SHIVA01 trial cohort, AI-assisted prioritisation of targeted agents was achieved through tumour molecular profiling.84 Deep learning on a pharmacogenomics database such as Genomic of Drug Sensitivity in Cancer can predict drug response and patients' survival by examining gene expression pathways85 to associate disease control with molecular profiles. In a retrospective analysis, the NetBio framework has shown better prediction of treatment response to ICI in select cancers compared with traditional biomarkers.86

Surgery

In surgery, exploring AI’s ability to guide surgical decision-making has been of great interest. Surgical decision-making can lean heavily on visual cues and images in addition to a surgeon’s understanding of the patient’s clinical picture. For surgical applications, AI could take a virtual form or provide direct physical support (including smart operating rooms, nanorobots or patient-assistance systems). Regardless of the form, three fields have emerged as the main application of ML in surgery: surgical skills assessment, supporting intraoperative decision-making and surgical outcomes prediction.

Assessment of surgical skills

AI has been applied to laparoscopic videos and trained to assist in anatomy identification, segmenting live surgical images as high-risk areas versus low risk areas to minimise adverse events from misinterpretation of anatomy.87 AI algorithms can recognise the operative steps of a laparoscopic sleeve gastrectomy obtained from surgical videos with 85.6% accuracy.88 When additional data was captured involving instrument handling, surgeon eye tracking and motion perturbations, AI has also been used to predict surgeon skill level, procedure length and potentially patient outcomes.88 In fact, experienced surgeons can be differentiated from beginners within the first 10 s, a task with 90% accuracy.89 This has very practical relevance, as currently, assessment of surgical quality largely relies on surgical case volumes, without granular objective data from intraoperative data.

Intraoperative decision-making

The application of AI in assisting intraoperative decision-making has been of great interest. In surgical oncology, the assessment of margins is critical to achieving optimal oncological outcomes. At the same time, the aim of modern cancer surgery is to spare healthy tissue as much as possible to minimise side effects and improve long-term quality of life. With the application of ML and various enhanced imaging systems including Raman spectroscopy, there is the potential for real-time intraoperative support while in the operating room.90 AI has also been used to determine optimal margins for cancer resection, such as in hepatic metastasectomy.91 These techniques are ripe for further evaluation as part of clinical trials.

Outcomes prediction

In predicting outcomes in surgical oncology, radiomics has been applied to predict response after neoadjuvant therapy and to select patients for surgery, for example, according to quantitative features associated with microinvasion,92 or to distinguish high-risk versus low-risk precancerous lesions.93 ML has also been used to predict major complications after cancer surgery94 and to identify patients who are safe to be discharged 2 days following major gastrointestinal cancer surgery.95

These represent the main themes of AI application within surgery. With ongoing technological advances, surgery will move towards a smart operating room and more precise perioperative decision-making.

Emerging tools

Large language models (LLM)

NLP has been applied to detect real-world cancer outcomes such as metastatic progression from radiology reports, pathology reports and clinical notes.96–101 As discussed above, Gensheimer et al have incorporated NLP into clinical triage.58 NLP systems for clinical trial matching are in test deployment at large healthcare networks.102

Since the late 2010s, advanced NLP models called LLM have leveraged transformer-based architectures with massive datasets (on the order of terabytes of text) to make significant breakthroughs in language understanding. While not specifically trained on biomedical or clinical text, LLMs such as Generative Pre-trained Transformer 4 (GPT-4) are capable of encoding clinical knowledge, as evidenced by their remarkable performance on medical licensing examinations and challenge problems,103–105 as well as to highly specialised topics such as medical physics.106 Researchers are now experimenting with fine-tuning pretrained models to focus specifically on medical applications, as seen in Med-PaLM 2 for question/answering, RadOnc-GPT for oncology-specific tasks and LLMs for clinical trial matching.102 107 108

In some cases, models are even trained from scratch using solely medical datasets to optimise their performance in the healthcare context. A prime example of domain-specific NLP utilisation is the Clinical Bidirectional Encoder Representations from Transformers (ClinicalBERT) model, which was trained on a large corpus of deidentified intensive care unit notes and fine-tuned to predict short term readmission.109

The technological leaps provided by LLM have led to high-profile healthcare collaborations. Recent collaborations include between Epic Systems and Microsoft to integrate GPT-4 by drafting of patient communication responses and providing data visualisation recommendations and between Mayo Clinic and Google Cloud, which builds on a partnership formed in 2019.

Research in this area has grown rapidly with several research groups training their own BERT-type and GPT-type models on their own datasets for various applications, and we recommend keeping a close eye on this space.

3.1.2. Challenges of NLP in healthcare

Despite showing strong performance on benchmark tasks, significant risks still remain for the use of LLM in healthcare and oncology. One prominent concern is the presence of bias within these models, which may inadvertently perpetuate or exacerbate existing healthcare disparities. Additionally, LLM are known to experience ‘hallucinations,’ generating plausible sounding yet incorrect or unrelated information, which could potentially lead to detrimental clinical decisions. Furthermore, these models may still exhibit inaccurate yet plausible reasoning, thereby making it difficult to catch errors and omissions. It is crucial to address these challenges and ensure these technologies are used responsibly, with human oversight remaining integral to decision-making processes.

Digital twins (DTs)

A DT is a virtual replica of a physical system which is not only created to mirror the real-world system but is also capable of analysis and prediction. DTs continuously monitor patients in real time, integrating data from wearable devices, sensors and electronic health records and thus is complemented by other technologies, including transfer learning, the Internet of Things, edge computing and cloud computing.110 DTs are being explored in oncology as a promising approach to enhance cancer care and may be used in various aspects of oncology including drug discovery and personalised treatment planning.111

Drug discovery

DTs have demonstrated the potential to streamline pharmaceutical processes and generate realistic input–output predictions for biochemical reactions. Through in silico techniques, several drugs have been identified and successfully brought to market for various diseases, including anticancer agent raltitrexed.112 In silico trials are currently being investigated, initially focusing on synthetic control arms and eventually expanding to predict clinical intervention. Both the US FDA and the European Medicines Agency have taken steps to support the integration of in silico approaches into control arms. For instance, a synthetic control arm consisting of 68 patients was used to extend the coverage of targeted therapy for NSCLC specifically alectinib, across 20 European countries.113 Synthetic controls have also played a role in expanding the indication of palbociclib, a kinase inhibitor, to include men with HR-positive HER2-negative advanced or metastatic breast cancer, as well as facilitating accelerated approval for blinatumomab which treats acute lymphoblastic leukaemia.114 In a phase I trial, existing quantitative systems pharmacology model of the anti-CD20/CD3 T-cell engaging bispecific antibody, mosunetuzumab, were used to incorporate different dosing regimens and patient heterogeneity within the trial.115

Treatment planning and prognosis monitoring

DTs may be created to personalise treatment planning as they enable the simulation and optimisation of treatments by integrating patient-specific genomics, imaging and clinical information. The utilisation of NLP for large-scale labelling of CT reports presents an opportunity to advance the development of DTs in oncology. In a recent study, NLP was used to perform consecutive multireport prediction of metastases, enabling highly detailed representations that effectively model a cancer patient’s disease progression over time.116 These approaches facilitate the generation of a comprehensive database consisting of patterns of disease spread, facilitating early detection and prediction of an individual patient’s progression. In another study, DTs of patients were generated, and clinical trials were simulated to anticipate the optimal salvage therapy following progressive disease while on pembrolizumab.117 For spine metastases, DT was used to simulate vertebroplasty and its impact on mechanical stability of the vertebra.118 Finally, so called ‘virtual imaging trials’ aim to simulate the entire radiological imaging process using realistic digital phantoms, simulated image acquisition and reconstruction and AI-driven readers/computational observer models to improve the precision and accuracy of imaging systems and downstream biomarkers on DTs.119 Further research, validation and clinical trials are needed to fully establish the effectiveness and integration of DTs into routine clinical practice in oncology.

Clinical informatics

There are several resources that can help non-AI expert clinicians in the interpretation and ethical application of AI tools in the clinic120 121 and the applications within specific disease sites, both during and after clinical training.

Education

In the USA, the clinical subspecialty of CI was recognised by the American Board of Medical Specialties (ABMS) in 2011 and the first physicians were board certified in CI in 2013.122 CI subspecialty fellowships are open to physicians from any ABMS specialty and allow fellows to spend 2 years dedicated to studying and practicing CI. NIH National Library of Medicine informatics fellowships can also provide physicians with opportunities to gain experience in programming and application of AI tools in clinical practice. There are other pathways123 and less formal educational resources, including master’s degrees or certificate programmes, American Medical Informatics Association 10×10 programs and massive open online courses on ML such as those on Coursera by Andrew Ng.

Research infrastructure and community

Sharing of data and conference resources in oncology are increasing, even outside of data access statements in publications. Federated learning can facilitate AI from much larger datasets while protecting data privacy by decentralising raw data, which has the potential to speed up validation of models. Work continues to try to standardise oncology data elements—mCODE, a collaboration between ASCO, CancerLinQ and MITRE—and interoperability—FHIR by HL7. The National Cancer Institute (NCI) is a major organiser of cancer datasets like the Cancer Research Data Commons, which includes The Cancer Genome Atlas, and the NCI Data Catalog and NCI Cancer Imaging Archive. Academic groups are building free software packages and platforms such as MultiAssayExperiment and CURATE.AI.124 AI-driven data fusion techniques that intelligently combine data from these different source domains (eg, clinical, imaging, omics, etc) can integrate knowledge to provide insight that is greater than the sum of the parts.125 More than a dozen technology companies are building platforms and software as a service (SaaS) tools to try to facilitate precision oncology and data analysis, including ConcertAI, Onc.Ai, Azra AI, ArteraAI and PreciseDx. It is imperative that oncologists become comfortable with critiquing, interpreting and applying these tools in clinical practice as well as research.

Conclusion

Much like oncologists 10–15 years ago would have been hard pressed to predict the paradigm shifts provided by advances in targeted therapy, immunotherapy, stereotactic radiation therapy, minimally invasive surgery, digital pathology and theranostics, it seems we are at or nearing an inflection point for AI in medicine given the amount of investment by hospitals, companies and governments. Table 4 summarises AI applications across fields of oncology described in this review. We see common themes for AI used in disease detection, outcome prediction and education across specialties, and over the last few years, prospective and randomised evidence is accumulating in the domains of patient triage and radiological imaging. The primary limitation of AI in oncology has been a lack of validation. In the past several years, we are seeing more prospective trials and randomised trials, though these still remain largely single institutional. As higher levels of evidence lead to improved outcomes, we expect further coverage for AI tools by payors. Further advancements in these fields supported by the rise of NLP, DTs and CI will pave the way for the actualisation of AI in oncology in the next 5–10 years.

Table 4

Summary of illustrative AI applications across domains of oncology described in this review

Miliard M. Google, Verily Using AI to Screen for Diabetic Retinopathy in India. Healthcare IT News 2019;
Available: here [Accessed 14 May 2023]
Google Scholar
Morey JR, Zhang X, Yaeger KA, et al. Real-world experience with artificial intelligence-based triage in transferred large vessel occlusion stroke patients. Cerebrovasc Dis 2021; 50:450–5.
doi:10.1159/000515320•Google Scholar
Hwang AB, Schuepfer G, Pietrini M, et al. External validation of EPIC’s risk of unplanned readmission model, the LACE+ index and Sqlape as predictors of unplanned hospital readmissions: a monocentric, retrospective, diagnostic cohort study in Switzerland. PLoS ONE 2021; 16.
doi:10.1371/journal.pone.0258338•Google Scholar
Kang J, Rancati T, Lee S, et al. Machine learning and radiogenomics: lessons learned and future directions. Front Oncol 2018; 8.
doi:10.3389/fonc.2018.00228•Google Scholar
Micheel C, Nass SJ, Omenn GS, et al. Committee on the review of Omics-based tests for predicting patient outcomes in clinical trials, Evolution of Translational Omics: Lessons Learned and the Path Forward. National Academies Press 2012;
doi:10.17226/13297•Google Scholar
Lehman CD, Wellman RD, Buist DSM, et al. Diagnostic accuracy of digital screening mammography with and without computer-aided detection. JAMA Intern Med 2015; 175:1828–37.
doi:10.1001/jamainternmed.2015.5231•Google Scholar
Sparano JA, Gray RJ, Makower DF, et al. Adjuvant chemotherapy guided by a 21-gene expression assay in breast cancer. N Engl J Med 2018; 379:111–21.
doi:10.1056/NEJMoa1804710•Google Scholar
Zhao SG, Chang SL, Spratt DE, et al. Development and validation of a 24-gene predictor of response to postoperative radiotherapy in prostate cancer: a matched, retrospective analysis. Lancet Oncol 2016; 17:1612–20.
doi:10.1016/S1470-2045(16)30491-0•Google Scholar
Kang J, Schwartz R, Flickinger J, et al. Machine learning approaches for predicting radiation therapy outcomes: a clinician’s perspective. Int J Radiat Oncol Biol Phys 2015; 93:1127–35.
doi:10.1016/j.ijrobp.2015.07.2286•Google Scholar
Kann BH, Hosny A, Aerts H, et al. Artificial intelligence for clinical oncology. Cancer Cell 2021; 39:916–27.
doi:10.1016/j.ccell.2021.04.002•Google Scholar
Bhinder B, Gilvary C, Madhukar NS, et al. Artificial intelligence in cancer research and precision medicine. Cancer Discov 2021; 11:900–15.
doi:10.1158/2159-8290.CD-21-0090•Google Scholar
Shimizu H, Nakayama KI. Artificial intelligence in oncology. Cancer Sci 2020; 111:1452–60.
Available: here
doi:10.1111/cas.14377•Google Scholar
Kang J, Thompson RF, Aneja S, et al. National Cancer Institute workshop on artificial intelligence in radiation oncology: training the next generation. Practical Radiation Oncology 2021; 11:74–83.
doi:10.1016/j.prro.2020.06.001•Google Scholar
Schaffter T, Buist DSM, Lee CI, et al. Evaluation of combined artificial intelligence and radiologist assessment to interpret screening mammograms. JAMA Netw Open 2020; 3.
doi:10.1001/jamanetworkopen.2020.0265•Google Scholar
Houssami N, Kirkpatrick-Jones G, Noguchi N, et al. Artificial intelligence (AI) for the early detection of breast cancer: a scoping review to assess AI’s potential in breast screening practice. Expert Rev Med Devices 2019; 16:351–62.
doi:10.1080/17434440.2019.1610387•Google Scholar
Anderson AW, Marinovich ML, Houssami N, et al. Independent external validation of artificial intelligence algorithms for automated interpretation of screening mammography: a systematic review. J Am Coll Radiol 2022; 19:259–73.
doi:10.1016/j.jacr.2021.11.008•Google Scholar
Lee CI, Houssami N, Elmore JG, et al. Pathways to breast cancer screening artificial intelligence algorithm validation. The Breast 2020; 52:146–9.
doi:10.1016/j.breast.2019.09.005•Google Scholar
Lång K, Josefsson V, Larsson A-M, et al. Artificial intelligence-supported screen reading versus standard double reading in the mammography screening with artificial intelligence trial (MASAI): a clinical safety analysis of a randomised, controlled, non-inferiority, single-blinded, screening accuracy study. Lancet Oncol 2023; 24:936–44.
doi:10.1016/S1470-2045(23)00298-X•Google Scholar
Paudyal R, Shah AD, Akin O, et al. Artificial intelligence in CT and MR imaging for oncological applications. Cancers (Basel) 2023; 15.
doi:10.3390/cancers15092573•Google Scholar
Jaffe CC. Measures of response: RECIST, WHO, and new alternatives. JCO 2006; 24:3245–51.
doi:10.1200/JCO.2006.06.5599•Google Scholar
Owens C, Hindocha S, Lee R, et al. The lung cancers: staging and response, CT, 18F-FDG PET/CT, MRI, DWI: review and new perspectives. Br J Radiol 2023; 96.
doi:10.1259/bjr.20220339•Google Scholar
Lafata KJ, Hong JC, Geng R, et al. Association of pre-treatment radiomic features with lung cancer recurrence following stereotactic body radiation therapy. Phys Med Biol 2019; 64.
doi:10.1088/1361-6560/aaf5a5•Google Scholar
Fahmy D, Alksas A, Elnakib A, et al. The role of radiomics and AI technologies in the segmentation, detection, and management of hepatocellular carcinoma. Cancers (Basel) 2022; 14.
doi:10.3390/cancers14246123•Google Scholar
Rigiroli F, Hoye J, Lerebours R, et al. CT Radiomic features of superior mesenteric artery involvement in pancreatic ductal adenocarcinoma: a pilot study. Radiology 2021; 301:610–22.
doi:10.1148/radiol.2021210699•Google Scholar
Rigiroli F, Hoye J, Lerebours R, et al. Exploratory analysis of mesenteric-portal axis CT radiomic features for survival prediction of patients with pancreatic ductal adenocarcinoma. Eur Radiol 2023; 33:5779–91.
doi:10.1007/s00330-023-09532-0•Google Scholar
Lafata KJ, Chang Y, Wang C, et al. Intrinsic radiomic expression patterns after 20 GY demonstrate early metabolic response of oropharyngeal cancers. Medical Physics 2021; 48:3767–77.
Available: here
doi:10.1002/mp.14926•Google Scholar
Wang C, Liu C, Chang Y, et al. Dose-distribution-driven PET image-based outcome prediction (DDD-PIOP): a deep learning study for oropharyngeal cancer IMRT application. Front Oncol 2020; 10:1592.
doi:10.3389/fonc.2020.01592•Google Scholar
Ji H, Lafata K, Mowery Y, et al. Post-radiotherapy PET image outcome prediction by deep learning under biological model guidance: a feasibility study of oropharyngeal cancer application. Front Oncol 2022; 12:895544.
doi:10.3389/fonc.2022.895544•Google Scholar
Rompianesi G, Pegoraro F, Ceresa CD, et al. Artificial intelligence in the diagnosis and management of colorectal cancer liver metastases. WJG 2022; 28:108–22.
doi:10.3748/wjg.v28.i1.108•Google Scholar
Bera K, Schalper KA, Rimm DL, et al. Artificial intelligence in digital pathology - new tools for diagnosis and precision oncology. Nat Rev Clin Oncol 2019; 16:703–15.
doi:10.1038/s41571-019-0252-y•Google Scholar
Lafata KJ, Wang Y, Konkel B, et al. Radiomics: a primer on high-throughput image phenotyping. Abdom Radiol (NY) 2022; 47:2986–3002.
doi:10.1007/s00261-021-03254-x•Google Scholar
Chen D, Fu M, Chi L, et al. Prognostic and predictive value of a pathomics signature in gastric cancer. Nat Commun 2022; 13.
doi:10.1038/s41467-022-34703-w•Google Scholar
Chen S, Jiang L, Zheng X, et al. Clinical use of machine learning-based pathomics signature for diagnosis and survival prediction of bladder cancer. Cancer Sci 2021; 112:2905–14.
doi:10.1111/cas.14927•Google Scholar
Chen S, Jiang L, Gao F, et al. Machine learning-based pathomics signature could act as a novel Prognostic marker for patients with clear cell renal cell carcinoma. Br J Cancer 2022; 126:771–7.
doi:10.1038/s41416-021-01640-2•Google Scholar
Ali S, Lewis J, Madabhushi A, et al. Spatially aware cell cluster(spacc1) graphs: predicting outcome in oropharyngeal Pl6+ tumors. Med Image Comput Comput-Assist Interv MICCAI Int Conf Med Image Comput Comput-Assist Interv 2013; 16:412–9.
doi:10.1007/978-3-642-38709-8•Google Scholar
Heindl A, Sestak I, Naidoo K, et al. Relevance of spatial heterogeneity of immune infiltration for predicting risk of recurrence after endocrine therapy of ER+ breast cancer. J Natl Cancer Inst 2018; 110.
doi:10.1093/jnci/djx137•Google Scholar
Cao R, Yang F, Ma S-C, et al. Development and interpretation of a pathomics-based model for the prediction of microsatellite instability in colorectal cancer. Theranostics 2020; 10:11080–91.
doi:10.7150/thno.49864•Google Scholar
Kim RH, Nomikou S, Coudray N, et al. Deep learning and pathomics analyses reveal cell nuclei as important features for Mutation prediction of BRAF-mutated melanomas. J Invest Dermatol 2022; 142:1650–8.
doi:10.1016/j.jid.2021.09.034•Google Scholar
Campanella G, Hanna MG, Geneslaw L, et al. Clinical-grade computational pathology using weakly supervised deep learning on whole slide images. Nat Med 2019; 25:1301–9.
doi:10.1038/s41591-019-0508-1•Google Scholar
Song Z, Zou S, Zhou W, et al. Clinically applicable histopathological diagnosis system for gastric cancer detection using deep learning. Nat Commun 2020; 11.
doi:10.1038/s41467-020-18147-8•Google Scholar
Wetstein SC, de Jong VMT, Stathonikos N, et al. Deep learning-based breast cancer grading and survival analysis on whole-slide Histopathology images. Sci Rep 2022; 12.
doi:10.1038/s41598-022-19112-9•Google Scholar
Barisoni L, Lafata KJ, Hewitt SM, et al. Digital Pathology and computational image analysis in Nephropathology. Nat Rev Nephrol 2020; 16:669–85.
doi:10.1038/s41581-020-0321-6•Google Scholar
Bulten W, Bándi P, Hoven J, et al. Epithelium segmentation using deep learning in H&Amp;E-stained prostate specimens with immunohistochemistry as reference standard. Sci Rep 2019; 9.
doi:10.1038/s41598-018-37257-4•Google Scholar
Sirinukunwattana K, Ahmed Raza SE. Locality sensitive deep learning for detection and classification of nuclei in routine colon cancer histology images. IEEE Trans Med Imaging 2016; 35:1196–206.
doi:10.1109/TMI.2016.2525803•Google Scholar
Wang C-W, Huang S-C, Lee Y-C, et al. Deep learning for bone marrow cell detection and classification on whole-slide images. Med Image Anal 2022; 75:102270.
doi:10.1016/j.media.2021.102270•Google Scholar
Linder N, Taylor JC, Colling R, et al. Deep learning for detecting tumour-infiltrating lymphocytes in Testicular germ cell tumours. J Clin Pathol 2019; 72:157–64.
doi:10.1136/jclinpath-2018-205328•Google Scholar
Aprupe L, Litjens G, Brinker TJ, et al. Robust and accurate quantification of biomarkers of immune cells in lung cancer micro-environment using deep convolutional neural networks. PeerJ 2019; 7.
doi:10.7717/peerj.6335•Google Scholar
Sornapudi S, Stanley RJ, Stoecker WV, et al. Deep learning nuclei detection in digitized histology images by superpixels. J Pathol Inform 2018; 9.
doi:10.4103/jpi.jpi_74_17•Google Scholar
Taniyama TK, Hashimoto K, Katsumata N, et al. Can oncologists predict survival for patients with progressive disease after standard Chemotherapies. Curr Oncol 2014; 21:84–90.
doi:10.3747/co.21.1743•Google Scholar
Kim YJ, Yoon SJ, Suh S-Y, et al. Performance of clinician prediction of survival in oncology outpatients with advanced cancer. PLOS ONE 2022; 17.
doi:10.1371/journal.pone.0267467•Google Scholar
Kang J, Chowdhry AK, Pugh SL, et al. Integrating artificial intelligence and machine learning into cancer clinical trials. Semin Radiat Oncol 2023; 33:386–94.
doi:10.1016/j.semradonc.2023.06.004•Google Scholar
Hong JC, Niedzwiecki D, Palta M, et al. Predicting emergency visits and hospital admissions during radiation and chemoradiation: an internally validated pretreatment machine learning algorithm. JCO Clin Cancer Inform 2018; 2:1–11.
doi:10.1200/CCI.18.00037•Google Scholar
Hong JC, Eclov NCW, Dalal NH, et al. System for high-intensity evaluation during radiation therapy (SHIELD-RT): a prospective randomized study of machine learning–directed clinical evaluations during radiation and chemoradiation. J Clin Oncol 2020; 38:3652–61.
doi:10.1200/JCO.20.01688•Google Scholar
Parikh RB, Manz C, Chivers C, et al. Machine learning approaches to predict 6-month mortality among patients with cancer. JAMA Netw Open 2019; 2.
doi:10.1001/jamanetworkopen.2019.15997•Google Scholar
Manz CR, Chen J, Liu M, et al. Validation of a machine learning algorithm to predict 180-day mortality for outpatients with cancer. JAMA Oncol 2020; 6:1723.
doi:10.1001/jamaoncol.2020.4331•Google Scholar
Manz CR, Parikh RB, Small DS, et al. Effect of integrating machine learning mortality estimates with behavioral nudges to clinicians on serious illness conversations among patients with cancer: a stepped-wedge cluster randomized clinical trial. JAMA Oncol 2020; 6.
doi:10.1001/jamaoncol.2020.4759•Google Scholar
Kang J, Morin O, Hong JC, et al. Closing the gap between machine learning and clinical cancer care—first steps into a larger world. JAMA Oncol 2020; 6:1731–2.
doi:10.1001/jamaoncol.2020.4314•Google Scholar
Gensheimer MF, Henry AS, Wood DJ, et al. Automated survival prediction in metastatic cancer patients using high-dimensional electronic medical record data. J Natl Cancer Inst 2019; 111:568–74.
doi:10.1093/jnci/djy178•Google Scholar
Gensheimer MF, Aggarwal S, Benson KRK, et al. Automated model versus treating physician for predicting survival time of patients with metastatic cancer. J Am Med Inform Assoc 2021; 28:1108–16.
doi:10.1093/jamia/ocaa290•Google Scholar
Gensheimer MF, Gupta D, Patel MI, et al. Use of machine learning and lay care coaches to increase advance care planning conversations for patients with metastatic cancer. JCO Oncol Pract 2023; 19:e176–84.
doi:10.1200/OP.22.00128•Google Scholar
Harrison K, Pullen H, Welsh C, et al. Machine learning for auto-segmentation in radiotherapy planning. Clinical Oncology 2022; 34:74–88.
doi:10.1016/j.clon.2021.12.003•Google Scholar
Walker GV, Awan M, Tao R, et al. Prospective randomized double-blind study of atlas-based organ-at-risk autosegmentation-assisted radiation planning in head and neck cancer. Radiotherapy and Oncology 2014; 112:321–5.
doi:10.1016/j.radonc.2014.08.028•Google Scholar
Heukelom J, Fuller CD. Head and neck cancer adaptive radiation therapy (ART): conceptual considerations for the informed clinician. Semin Radiat Oncol 2019; 29:258–73.
doi:10.1016/j.semradonc.2019.02.008•Google Scholar
Archambault Y, Boylan C, Bullock D, et al. Making on-line adaptive radiotherapy possible using artificial intelligence and machine learning for efficient daily re-planning. Med Phys Int 2020; 8:77–86.
Google Scholar
Varian, a Siemens Healthineers Company. Daily adaptive vs non-adaptive external beam radiation therapy with concurrent chemotherapy for locally advanced non-small cell lung cancer: a prospective randomized trial of an individualized approach for toxicity reduction (ARTIA-lung). ClinicaltrialsGov 2022;
Available: here
Google Scholar
Sher DJ, Avkshtol V, Moon D, et al. Recurrence and quality-of-life following involved node radiotherapy for head and neck squamous cell carcinoma: initial results from the phase II INRT-air trial. Int J Radia Oncol*Biolo*Phy 2021; 111.
doi:10.1016/j.ijrobp.2021.07.1155•Google Scholar
Sher D. DARTBOARD: a prospective randomized phase II study of daily adaptive radiotherapy to better organ-at-risk doses in head and neck cancer. ClinicaltrialsGov 2022;
Available: here
Google Scholar
Sahin O, Multidisciplinary OPC ENE Group. Multi-specialty expert physician identification of Extranodal extension in computed tomography scans of oropharyngeal cancer patients: prospective blinded human inter-observer performance evaluation. Oncology 2023;
doi:10.1101/2023.02.25.23286432•Google Scholar
Kann BH, Hicks DF, Payabvash S, et al. Multi-institutional validation of deep learning for pretreatment identification of extranodal extension in head and neck squamous cell carcinoma. J Clin Oncol 2020; 38:1304–11.
doi:10.1200/JCO.19.02031•Google Scholar
Kann BH, Likitlersuang J, Ye Z, et al. Screening for extranodal extension with deep learning: evaluation in ECOG-ACRIN E3311, a randomized de-escalation trial for HPV-associated oropharyngeal carcinoma. Int J Radia Oncol*Biol*Phy 2022; 114:S26–7.
doi:10.1016/j.ijrobp.2022.07.379•Google Scholar
Nazha A, Komrokji R, Meggendorfer M, et al. Personalized prediction model to risk stratify patients with myelodysplastic syndromes. JCO 2021; 39:3737–46.
doi:10.1200/JCO.20.02810•Google Scholar
Sundar R, Barr Kumarakulasinghe N, Huak Chan Y, et al. Machine-learning model derived gene signature predictive of paclitaxel survival benefit in gastric cancer: results from the randomised phase III SAMIT trial. Gut 2022; 71:676–85.
doi:10.1136/gutjnl-2021-324060•Google Scholar
Giannini V, Rosati S, Defeudis A, et al. Radiomics predicts response of individual Her2-amplified colorectal cancer liver metastases in patients treated with Her2-targeted therapy. Int J Cancer 2020; 147:3215–23.
doi:10.1002/ijc.33271•Google Scholar
Lipkova J, Chen RJ, Chen B, et al. Artificial intelligence for multimodal data integration in oncology. Cancer Cell 2022; 40:1095–110.
doi:10.1016/j.ccell.2022.09.012•Google Scholar
Vanguri RS, Luo J, Aukerman AT, et al. Multimodal integration of radiology, pathology and genomics for prediction of response to PD-(L)1 blockade in patients with non-small cell lung cancer. Nat Cancer 2022; 3:1151–64.
doi:10.1038/s43018-022-00416-8•Google Scholar
de Jong AC, Danyi A, van Riet J, et al. Predicting response to enzalutamide and abiraterone in metastatic prostate cancer using whole-omics machine learning. Nat Commun 2023; 14.
doi:10.1038/s41467-023-37647-x•Google Scholar
Cifci D, Foersch S, Kather JN, et al. Artificial intelligence to identify genetic alterations in conventional Histopathology. J Pathol 2022; 257:430–44.
doi:10.1002/path.5898•Google Scholar
Echle A, Ghaffari Laleh N, Quirke P, et al. Artificial intelligence for detection of microsatellite instability in colorectal cancer-a multicentric analysis of a pre-screening tool for clinical application. ESMO Open 2022; 7.
doi:10.1016/j.esmoop.2022.100400•Google Scholar
Gerwert K, Schörner S, Großerueschkamp F, et al. Fast and label-free automated detection of microsatellite status in early colon cancer using artificial intelligence integrated infrared imaging. European Journal of Cancer 2023; 182:122–31.
doi:10.1016/j.ejca.2022.12.026•Google Scholar
Schrammen PL, Ghaffari Laleh N, Echle A, et al. Weakly supervised annotation-free cancer detection and prediction of genotype in routine Histopathology. J Pathol 2022; 256:50–60.
doi:10.1002/path.5800•Google Scholar
Nair JKR, Saeed UA, McDougall CC, et al. Radiogenomic models using machine learning techniques to predict EGFR mutations in non-small cell lung cancer. Can Assoc Radiol J 2021; 72:109–19.
doi:10.1177/0846537119899526•Google Scholar
Lafata KJ, Corradetti MN, Gao J, et al. Radiogenomic analysis of locally advanced lung cancer based on CT imaging and Intratreatment changes in cell-free DNA. Radiol Imaging Cancer 2021; 3.
doi:10.1148/rycan.2021200157•Google Scholar
Corradetti MN, Torok JA, Hatch AJ, et al. Dynamic changes in circulating tumor DNA during chemoradiation for locally advanced lung cancer. Adv Radiat Oncol 2019; 4:748–52.
doi:10.1016/j.adro.2019.05.004•Google Scholar
Petak I, Kamal M, Dirner A, et al. A computational method for prioritizing targeted therapies in precision oncology: performance analysis in the Shiva01 trial. NPJ Precis Oncol 2021; 5.
doi:10.1038/s41698-021-00191-2•Google Scholar
Sakellaropoulos T, Vougas K, Narang S, et al. A deep learning framework for predicting response to therapy in cancer. Cell Rep 2019; 29:3367–73.
doi:10.1016/j.celrep.2019.11.017•Google Scholar
Kong J, Ha D, Lee J, et al. Network-based machine learning approach to predict immunotherapy response in cancer patients. Nat Commun 2022; 13.
doi:10.1038/s41467-022-31535-6•Google Scholar
Madani A, Namazi B, Altieri MS, et al. Artificial intelligence for intraoperative guidance: using semantic segmentation to identify surgical anatomy during laparoscopic cholecystectomy. Ann Surg 2022; 276:363–9.
doi:10.1097/SLA.0000000000004594•Google Scholar
Hashimoto DA, Rosman G, Witkowski ER, et al. Computer vision analysis of intraoperative video: automated recognition of operative steps in laparoscopic sleeve gastrectomy. Ann Surg 2019; 270:414–21.
doi:10.1097/SLA.0000000000003460•Google Scholar
French A, Lendvay TS, Sweet RM, et al. Predicting surgical skill from the first N seconds of a task: value over task time using the Isogony principle. Int J Comput Assist Radiol Surg 2017; 12:1161–70.
doi:10.1007/s11548-017-1606-5•Google Scholar
Haifler M, Pence I, Sun Y, et al. Discrimination of malignant and normal kidney tissue with short wave infrared dispersive Raman spectroscopy. J Biophotonics 2018; 11.
doi:10.1002/jbio.201700188•Google Scholar
Bertsimas D, Margonis GA, Sujichantararat S, et al. Using artificial intelligence to find the optimal margin width in hepatectomy for colorectal cancer liver metastases. JAMA Surg 2022; 157.
doi:10.1001/jamasurg.2022.1819•Google Scholar
Zheng J, Chakraborty J, Chapman WC, et al. Preoperative prediction of microvascular invasion in hepatocellular carcinoma using quantitative image analysis. J Am Coll Surg 2017; 225:778–788e1.
doi:10.1016/j.jamcollsurg.2017.09.003•Google Scholar
Attiyeh MA, Chakraborty J, Gazit L, et al. Preoperative risk prediction for intraductal papillary mucinous neoplasms by quantitative CT image analysis. HPB 2019; 21:212–8.
doi:10.1016/j.hpb.2018.07.016•Google Scholar
Deng H, Eftekhari Z, Carlin C, et al. Development and validation of an explainable machine learning model for major complications after cytoreductive surgery. JAMA Netw Open 2022; 5.
doi:10.1001/jamanetworkopen.2022.12930•Google Scholar
van de Sande D, van Genderen ME, Verhoef C, et al. Optimizing discharge after major surgery using an artificial intelligence-based decision support tool (DESIRE): an external validation study. Surgery 2022; 172:663–9.
doi:10.1016/j.surg.2022.03.031•Google Scholar
Kehl KL, Elmarakeby H, Nishino M, et al. Assessment of deep natural language processing in ascertaining oncologic outcomes from radiology reports. JAMA Oncol 2019; 5:1421–9.
doi:10.1001/jamaoncol.2019.1800•Google Scholar
Kehl KL, Xu W, Lepisto E, et al. Natural language processing to ascertain cancer outcomes from medical oncologist notes. JCO Clin Cancer Inform 2020; 4:680–90.
doi:10.1200/CCI.20.00020•Google Scholar
Kehl KL, Xu W, Gusev A, et al. Artificial intelligence-aided clinical annotation of a large multi-cancer genomic dataset. Nat Commun 2021; 12.
doi:10.1038/s41467-021-27358-6•Google Scholar
Banerjee I, Bozkurt S, Caswell-Jin JL, et al. Natural language processing approaches to detect the timeline of metastatic recurrence of breast cancer. JCO Clin Cancer Inform 2019; 3:1–12.
doi:10.1200/CCI.19.00034•Google Scholar
Sanyal J, Tariq A, Kurian AW, et al. Weakly supervised temporal model for prediction of breast cancer distant recurrence. Sci Rep 2021; 11.
doi:10.1038/s41598-021-89033-6•Google Scholar
Karimi YH, Blayney DW, Kurian AW, et al. Development and use of natural language processing for identification of distant cancer recurrence and sites of distant recurrence using unstructured electronic health record data. JCO Clin Cancer Inform 2021; 5:469–78.
doi:10.1200/CCI.20.00165•Google Scholar
Wong C, Zhang S, Gu Y, et al. Scaling clinical trial matching using large language models: a case study in oncology. 2023;
Available: here
Google Scholar
Bhayana R, Bleakney RR, Krishna S, et al. GPT-4 in radiology: improvements in advanced reasoning. Radiology 2023; 307.
doi:10.1148/radiol.230987•Google Scholar
Nori H, King N, McKinney SM, et al. Capabilities of GPT-4 on medical challenge problems. 2023;
Available: here
Google Scholar
Singhal K, Azizi S, Tu T, et al. Large language models encode clinical knowledge. Nature 2023; 620:172–80.
doi:10.1038/s41586-023-06291-2•Google Scholar
Holmes J, Liu Z, Zhang L, et al. Evaluating large language models on a highly-specialized topic, radiation oncology physics. Front Oncol 2023; 13.
doi:10.3389/fonc.2023.1219326•Google Scholar
Singhal K, Tu T, Gottweis J, et al. Towards expert-level medical question answering with large language models. arXiv.org
Available: here [Accessed 6 Jun 2023]
Google Scholar
Liu Z, Wang P, Li Y, et al. Radonc-GPT: a large language model for radiation oncology. 2023;
Available: here
Google Scholar
Huang K, Altosaar J, Ranganath R, et al. Clinicalbert: modeling clinical notes and predicting hospital readmission. ArXiv190405342 Cs 2020;
Available: here [Accessed 28 May 2021]
Google Scholar
Moztarzadeh O, Jamshidi MB, Sargolzaei S, et al. Metaverse and healthcare: machine learning-enabled digital twins of cancer. Bioengineering (Basel) 2023; 10.
doi:10.3390/bioengineering10040455•Google Scholar
Sahal R, Alsamhi SH, Brown KN, et al. Personal digital twin: a close look into the present and a step towards the future of personalised healthcare industry. Sensors (Basel) 2022; 22:15.
doi:10.3390/s22155918•Google Scholar
Shaker B, Ahmad S, Lee J, et al. In silico methods and tools for drug discovery. Comput Biol Med 2021; 137:104851.
doi:10.1016/j.compbiomed.2021.104851•Google Scholar
Davies J, Martinec M, Delmar P, et al. Comparative effectiveness from a single-arm trial and real-world data: alectinib versus Ceritinib. J Comp Eff Res 2018; 7:855–65.
doi:10.2217/cer-2018-0032•Google Scholar
Wedam S, Fashoyin-Aje L, Bloomquist E, et al. FDA approval summary: palbociclib for male patients with metastatic breast cancer. Clin Cancer Res 2020; 26:1208–12.
doi:10.1158/1078-0432.CCR-19-2580•Google Scholar
Susilo ME, Li C-C, Gadkar K, et al. Systems-based digital twins to help characterize clinical dose-response and propose predictive biomarkers in a phase I study of Bispecific antibody, Mosunetuzumab, in NHL. Clin Transl Sci 2023; 16:1134–48.
doi:10.1111/cts.13501•Google Scholar
Batch KE, Yue J, Darcovich A, et al. Developing a cancer digital twin: supervised metastases detection from consecutive structured radiology reports. Front Artif Intell 2022; 5.
doi:10.3389/frai.2022.826402•Google Scholar
Qi T, Cao Y. Virtual clinical trials: a tool for predicting patients who may benefit from treatment beyond progression with pembrolizumab in non-small cell lung cancer. CPT Pharmacom & Syst Pharma 2023; 12:236–49.
Available: here
doi:10.1002/psp4.12896•Google Scholar
Ahmadian H, Mageswaran P, Walter BA, et al. A digital twin for simulating the vertebroplasty procedure and its impact on mechanical stability of vertebra in cancer patients. Numer Methods Biomed Eng 2022; 38.
Available: here
doi:10.1002/cnm.3600•Google Scholar
Abadi E, Segars WP, Tsui BMW, et al. Virtual clinical trials in medical imaging: a review. J Med Imag 2020; 7:1.
doi:10.1117/1.JMI.7.4.042805•Google Scholar
Shreve JT, Khanani SA, Haddad TC, et al. Artificial intelligence in oncology: current capabilities, future opportunities, and ethical considerations. Am Soc Clin Oncol Educ Book 2022; 42:1–10.
doi:10.1200/EDBK_350652•Google Scholar
Lahmi L, Mamzer M-F, Burgun A, et al. Ethical aspects of artificial intelligence in radiation oncology. Semin Radiat Oncol 2022; 32:442–8.
doi:10.1016/j.semradonc.2022.06.013•Google Scholar
Detmer DE, Lumpkin JR, Williamson JJ, et al. Defining the medical subspecialty of clinical informatics. J Am Med Infor Assoc 2009; 16:167–8.
doi:10.1197/jamia.M3094•Google Scholar
Kim E, Bitterman DS, Kann BH, et al. Hidden in plain sight: clinical Informaticians are the oncology subspecialists you did not know you needed. Clinical Oncology 2022; 34:135–40.
doi:10.1016/j.clon.2021.11.018•Google Scholar
Blasiak A, Khong J, Kee T, et al. CURATE.AI: optimizing personalized medicine with artificial intelligence. SLAS Technol 2020; 25:95–105.
doi:10.1177/2472630319890316•Google Scholar
Wang Y, Li X, Konanur M, et al. Towards optimal deep fusion of imaging and clinical data via a model-based description of fusion quality. Med Phys 2023; 50:3526–37.
Available: here
doi:10.1002/mp.16181•Google Scholar

Overview

Abstract
Introduction
Current applications of AI in oncology
Emerging tools
Conclusion
References

Footnotes
Publication history
Rapid Responses

Article metrics

Altmetric

Dimensions