Developing, Purchasing, Implementing and Monitoring AI Tools in Radiology: Practical Considerations. A Multi-Society Statement From the ACR, CAR, ESR, RANZCR & RSNA.

Artificial Intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever­growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones. This multi­society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools.

Interstitial lung abnormalities in a large clinical lung cancer screening cohort: association with mortality and ILD diagnosis.

BACKGROUND: Interstitial lung abnormalities (ILA) are CT findings suggestive of interstitial lung disease in individuals without a prior diagnosis or suspicion of ILD. Previous studies have demonstrated that ILA are associated with clinically significant outcomes including mortality. The aim of this study was to determine the prevalence of ILA in a large CT lung cancer screening program and the association with clinically significant outcomes including mortality, hospitalizations, cancer and ILD diagnosis. METHODS: This was a retrospective study of individuals enrolled in a CT lung cancer screening program from 2012 to 2014. Baseline and longitudinal CT scans were scored for ILA per Fleischner Society guidelines. The primary analyses examined the association between baseline ILA and mortality, all-cause hospitalization, and incidence of lung cancer. Kaplan-Meier plots were generated to visualize the associations between ILA and lung cancer and all-cause mortality. Cox regression proportional hazards models were used to test for this association in both univariate and multivariable models. RESULTS: 1699 subjects met inclusion criteria. 41 (2.4%) had ILA and 101 (5.9%) had indeterminate ILA on baseline CTs. ILD was diagnosed in 10 (24.4%) of 41 with ILA on baseline CT with a mean time from baseline CT to diagnosis of 4.47 ± 2.72 years. On multivariable modeling, the presence of ILA remained a significant predictor of death, HR 3.87 (2.07, 7.21; p < 0.001) when adjusted for age, sex, BMI, pack years and active smoking, but not of lung cancer and all-cause hospital admission. Approximately 50% with baseline ILA had progression on the longitudinal scan. CONCLUSIONS: ILA identified on baseline lung cancer screening exams are associated with all-cause mortality. In addition, a significant proportion of patients with ILA are subsequently diagnosed with ILD and have CT progression on longitudinal scans. TRIAL REGISTRATION NUMBER: ClinicalTrials.gov; No.: NCT04503044.

Stochastic programming for outpatient scheduling with flexible inpatient exam accommodation.

This study is concerned with the determination of an optimal appointment schedule in an outpatient-inpatient hospital system where the inpatient exams can be cancelled based on certain rules while the outpatient exams cannot be cancelled. Stochastic programming models were formulated and solved to tackle the stochasticity in the procedure durations and patient arrival patterns. The first model, a two-stage stochastic programming model, is formulated to optimize the slot size. The second model further optimizes the inpatient block (IPB) placement and slot size simultaneously. A computational method is developed to solve the second optimization problem. A case study is conducted using the data from Magnetic Resonance Imaging (MRI) centers of Lahey Hospital and Medical Center (LHMC). The current schedule and the schedules obtained from the optimization models are evaluated and compared using simulation based on FlexSim Healthcare. Results indicate that the overall weighted cost can be reduced by 11.6% by optimizing the slot size and can be further reduced by an additional 12.6% by optimizing slot size and IPB placement simultaneously. Three commonly used sequencing rules (IPBEG, OPBEG, and a variant of ALTER rule) were also evaluated. The results showed that when optimization tools are not available, ALTER variant which evenly distributes the IPBs across the day has the best performance. Sensitivity analysis of weights for patient waiting time, machine idle time and exam cancellations further supports the superiority of ALTER variant sequencing rules compared to the other sequencing methods. A Pareto frontier was also developed and presented between patient waiting time and machine idle time to enable medical centers with different priorities to obtain solutions that accurately reflect their respective optimal tradeoffs. An extended optimization model was also developed to incorporate the emergency patient arrivals. The optimal schedules from the extended model show only minor differences compared to those from the original model, thus proving the robustness of the scheduling solutions obtained from our optimal models against the impacts of emergency patient arrivals.

HIGHLIGHTS: Timestamped operational data was analyzed to identify sources of uncertainty and delays. Stochastic programming models were developed to optimize slot size and inpatient block placement. A case study showed that the optimized schedules can reduce overall costs by 23%. Distributing inpatient and outpatient slots evenly throughout the day provides the best performance. A Pareto frontier was developed to allow practitioners to choose their own best tradeoffs between multiple objectives.

Identifying Areas for Operational Improvement and Growth in IR Workflow Using Workflow Modeling, Simulation, and Optimization Techniques.

Identifying areas for workflow improvement and growth is essential for an interventional radiology (IR) department to stay competitive. Deployment of traditional methods such as Lean and Six Sigma helped in reducing the waste in workflows at a strategic level. However, achieving efficient workflow needs both strategic and tactical approaches. Uncertainties about patient arrivals, staff availability, and variability in procedure durations pose hindrances to efficient workflow and lead to delayed patient care and staff overtime. We present an alternative approach to address both tactical and strategic needs using discrete event simulation (DES) and simulation based optimization methods. A comprehensive digital model of the patient workflow in a hospital-based IR department was modeled based on expert interviews with the incumbent personnel and analysis of 192 days' worth of electronic medical record (EMR) data. Patient arrival patterns and process times were derived from 4393 individual patient appointments. Exactly 196 unique procedures were modeled, each with its own process time distribution and rule-based procedure-room mapping. Dynamic staff schedules for interventional radiologists, technologists, and nurses were incorporated in the model. Stochastic model simulation runs revealed the resource "computed tomography (CT) suite" as the major workflow bottleneck during the morning hours. This insight compelled the radiology department leadership to re-assign time blocks on a diagnostic CT scanner to the IR group. Moreover, this approach helped identify opportunities for additional appointments at times of lower diagnostic scanner utilization. Demand for interventional service from Outpatients during late hours of the day required the facility to extend hours of operations. Simulation-based optimization methods were used to model a new staff schedule, stretching the existing pool of resources to support the additional 2.5 h of daily operation. In conclusion, this study illustrates that the combination of workflow modeling, stochastic simulations, and optimization techniques is a viable and effective approach for identifying workflow inefficiencies and discovering and validating improvement options through what-if scenario testing.

Quantitative Pectoralis Muscle Area is Associated with the Development of Lung Cancer in a Large Lung Cancer Screening Cohort.

BACKGROUND: Studies have demonstrated an inverse relationship between body mass index (BMI) and the risk of developing lung cancer. We conducted a retrospective cohort study evaluating baseline quantitative computed tomography (CT) measurements of body composition, specifically muscle and fat area in a large CT lung screening cohort (CTLS). We hypothesized that quantitative measurements of baseline body composition may aid in risk stratification for lung cancer. METHODS: Patients who underwent baseline CTLS between January 1st, 2012 and September 30th, 2014 and who had an in-network primary care physician were included. All patients met NCCN Guidelines eligibility criteria for CTLS. Quantitative measurements of pectoralis muscle area (PMA) and subcutaneous fat area (SFA) were performed on a single axial slice of the CT above the aortic arch with the Chest Imaging Platform Workstation software. Cox multivariable proportional hazards model for cancer was adjusted for variables with a univariate p < 0.2. Data were dichotomized by sex and then combined to account for baseline differences between sexes. RESULTS: One thousand six hundred and ninety six patients were included in this study. A total of 79 (4.7%) patients developed lung cancer. There was an association between the 25th percentile of PMA and the development of lung cancer [HR 1.71 (1.07, 2.75), p < 0.025] after adjusting for age, BMI, qualitative emphysema, qualitative coronary artery calcification, and baseline Lung-RADS® score. CONCLUSIONS: Quantitative assessment of PMA on baseline CTLS was associated with the development of lung cancer. Quantitative PMA has the potential to be incorporated as a variable in future lung cancer risk models.

How We Do It: Operationalizing Just Culture in a Radiology Department.

OBJECTIVE. The purpose of this article is to describe how establishing routine practice sessions facilitates adoption by modality operations managers of the just culture model of error management in a radiology department. CONCLUSION. Implementation of ongoing just culture training among radiology operations managers can help them approach uniformity, equity, and transparency in managing errors. Managers see the just culture method as an effective tool that helps improve the safety of patient care.

Automated Tracking of Follow-Up Imaging Recommendations.

OBJECTIVE. Radiology reports often contain follow-up imaging recommendations. Failure to comply with these recommendations in a timely manner can lead to poor patient outcomes, complications, and legal liability. As such, the primary objective of this research was to determine adherence rates to follow-up recommendations. MATERIALS AND METHODS. Radiology-related examination data, including report text, for examinations performed between June 1, 2015, and July 31, 2017, were extracted from the radiology departments at the University of Washington (UW) and Lahey Hospital and Medical Center (LHMC). The UW dataset contained 923,885 examinations, and the LHMC dataset contained 763,059 examinations. A 1-year period was used for detection of imaging recommendations and up to 14-months for the follow-up examination to be performed. RESULTS. On the basis of an algorithm with 97.9% detection accuracy, the follow-up imaging recommendation rate was 11.4% at UW and 20.9% at LHMC. Excluding mammography examinations, the overall follow-up imaging adherence rate was 51.9% at UW (range, 44.4% for nuclear medicine to 63.0% for MRI) and 52.0% at LHMC (range, 30.1% for fluoroscopy to 63.2% for ultrasound) using a matcher algorithm with 76.5% accuracy. CONCLUSION. This study suggests that follow-up imaging adherence rates vary by modality and between sites. Adherence rates can be influenced by various legitimate factors. Having the capability to identify patients who can benefit from patient engagement initiatives is important to improve overall adherence rates. Monitoring of follow-up adherence rates over time and critical evaluation of variation in recommendation patterns across the practice can inform measures to standardize and help mitigate risk.

Probabilistic Modeling of Exam Durations in Radiology Procedures.

In this paper, we model the statistical properties of imaging exam durations using parametric probability distributions such as the Gaussian, Gamma, Weibull, lognormal, and log-logistic. We establish that in a majority of radiology procedures, the underlying distribution of exam durations is best modeled by a log-logistic distribution, while the Gaussian has the poorest fit among the candidates. Further, through illustrative examples, we show how business insights and workflow analytics can be significantly impacted by making the correct (log-logistic) versus incorrect (Gaussian) model choices.

NCCN Guidelines as a Model of Extended Criteria for Lung Cancer Screening.

Background: This review assessed the performance of patients in NCCN high-risk group 2 in a clinical CT lung screening (CTLS) program. Methods: We retrospectively reviewed screening results for all patients from our institution undergoing clinical CTLS from January 2012 through December 2016, with follow-up through June 2017. To qualify for screening, patients had to meet the NCCN Guidelines high-risk criteria for CTLS, have a physician order for screening, be asymptomatic, be lung cancer-free for 5 years, and have no known metastatic disease. We compared demographics and screening performance of NCCN high-risk groups 1 and 2 across >4 rounds of screening. Screening metrics assessed included rates of positive and suspicious examinations, significant incidental and infectious/inflammatory findings, false negatives, and cancer detection. We also compared cancer stage and histology detected in each NCCN high-risk group. Results: A total of 2,927 individuals underwent baseline screening, of which 698 (24%) were in NCCN group 2. On average, group 2 patients were younger (60.6 vs 63.1 years), smoked less (38.8 vs 50.8 pack-years), had quit longer (18.1 vs 6.3 years), and were more often former smokers (61.4% vs 44.2%). Positive and suspicious examination rates, false negatives, and rates of infectious/inflammatory findings were equivalent in groups 1 and 2 across all rounds of screening. An increased rate of cancer detection was observed in group 2 during the second annual (T2) screening round (2.7% vs 0.5%; P=.005), with no difference in the other screening rounds: baseline (T0; 2% vs 2.3%; P=.61), first annual (T1; 1.2% vs 1.7%; P=.41), and third annual and beyond (≥T3; 1.2% vs 1.1%; P=1.00). Conclusions: CTLS appears to be equally effective in both NCCN high-risk groups.

Inpatient Complexity in Radiology-a Practical Application of the Case Mix Index Metric.

With ongoing healthcare payment reforms in the USA, radiology is moving from its current state of a revenue generating department to a new reality of a cost-center. Under bundled payment methods, radiology does not get reimbursed for each and every inpatient procedure, but rather, the hospital gets reimbursed for the entire hospital stay under an applicable diagnosis-related group code. The hospital case mix index (CMI) metric, as defined by the Centers for Medicare and Medicaid Services, has a significant impact on how much hospitals get reimbursed for an inpatient stay. Oftentimes, patients with the highest disease acuity are treated in tertiary care radiology departments. Therefore, the average hospital CMI based on the entire inpatient population may not be adequate to determine department-level resource utilization, such as the number of technologists and nurses, as case length and staffing intensity gets quite high for sicker patients. In this study, we determine CMI for the overall radiology department in a tertiary care setting based on inpatients undergoing radiology procedures. Between April and September 2015, CMI for radiology was 1.93. With an average of 2.81, interventional neuroradiology had the highest CMI out of the ten radiology sections. CMI was consistently higher across seven of the radiology sections than the average hospital CMI of 1.81. Our results suggest that inpatients undergoing radiology procedures were on average more complex in this hospital setting during the time period considered. This finding is relevant for accurate calculation of labor analytics and other predictive resource utilization tools.

Imaging techniques for the diagnosis of hepatocellular carcinoma and the evaluation of response to treatment.

Imaging plays a critical role in the diagnosis of hepatocellular carcinoma (HCC). In the USA, non-invasive imaging based diagnosis of HCC has largely replaced biopsy because of the high specificity and positive predictive value of imaging features for HCC. Because of the important role of imaging and the need to promote standardization of the management of HCC, several imaging-based algorithms for the diagnosis of HCC in at-risk patients have been developed.Imaging also plays a vital role in the assessment of HCC response to locoregional therapies (LRT) such as ablative and endovascular therapies. Standard imaging response criteria of solid tumors that rely solely on change in tumor size for determination of therapeutic success are not applicable to HCC undergoing LRT. Therefore, several systems have been developed over the years to objectively evaluate HCC response to LRT.In this review, we will describe major and ancillary imaging features of HCC, how these features are incorporated into the various imaging based algorithms, discuss the differences between algorithms, and address the emerging role of new imaging techniques and contrast agents in the diagnosis of HCC. We will also discuss the importance of assessment of HCC response to LRT, describe patterns of imaging response to the various therapies including newer volumetric and functional response measures, and examine and compare proposed response criteria of HCC to LRT.

Projected Growth in FDA-Approved Artificial Intelligence Products Given Venture Capital Funding.

PURPOSE: Medical imaging accounts for 85% of digital health's venture capital funding. As funding grows, it is expected that artificial intelligence (AI) products will increase commensurately. The study's objective is to project the number of new AI products given the statistical association between historical funding and FDA-approved AI products. METHODS: The study used data from the ACR Data Science Institute and for the number of FDA-approved AI products (2008-2022) and data from Rock Health for AI funding (2013-2022). Employing a 6-year lag between funding and product approved, we used linear regression to estimate the association between new products approved in a certain year, based on the lagged funding (ie, product-year funding). Using this statistical relationship, we forecasted the number of new FDA-approved products. RESULTS: The results show that there are 11.33 (95% confidence interval: 7.03-15.64) new AI products for every $1 billion in funding assuming a 6-year lag between funding and product approval. In 2022 there were 69 new FDA-approved products associated with $4.8 billion in funding. In 2035, product-year funding is projected to reach $30.8 billion, resulting in 350 new products that year. CONCLUSIONS: FDA-approved AI products are expected to grow from 69 in 2022 to 350 in 2035 given the expected funding growth in the coming years. AI is likely to change the practice of diagnostic radiology as new products are developed and integrated into practice. As more AI products are integrated, it may incentivize increased investment for future AI products.

Proceedings From the 2022 ACR-RSNA Workshop on Safety, Effectiveness, Reliability, and Transparency in AI.

Despite the surge in artificial intelligence (AI) development for health care applications, particularly for medical imaging applications, there has been limited adoption of such AI tools into clinical practice. During a 1-day workshop in November 2022, co-organized by the ACR and the RSNA, participants outlined experiences and problems with implementing AI in clinical practice, defined the needs of various stakeholders in the AI ecosystem, and elicited potential solutions and strategies related to the safety, effectiveness, reliability, and transparency of AI algorithms. Participants included radiologists from academic and community radiology practices, informatics leaders responsible for AI implementation, regulatory agency employees, and specialty society representatives. The major themes that emerged fell into two categories: (1) AI product development and (2) implementation of AI-based applications in clinical practice. In particular, participants highlighted key aspects of AI product development to include clear clinical task definitions; well-curated data from diverse geographic, economic, and health care settings; standards and mechanisms to monitor model reliability; and transparency regarding model performance, both in controlled and real-world settings. For implementation, participants emphasized the need for strong institutional governance; systematic evaluation, selection, and validation methods conducted by local teams; seamless integration into the clinical workflow; performance monitoring and support by local teams; performance monitoring by external entities; and alignment of incentives through credentialing and reimbursement. Participants predicted that clinical implementation of AI in radiology will continue to be limited until the safety, effectiveness, reliability, and transparency of such tools are more fully addressed.

Real-World Performance of Large Vessel Occlusion Artificial Intelligence-Based Computer-Aided Triage and Notification Algorithms-What the Stroke Team Needs to Know.

PURPOSE: To evaluate the real-world performance of two FDA-approved artificial intelligence (AI)-based computer-aided triage and notification (CADt) detection devices and compare them with the manufacturer-reported performance testing in the instructions for use. MATERIALS AND METHODS: Clinical performance of two FDA-cleared CADt large-vessel occlusion (LVO) devices was retrospectively evaluated at two separate stroke centers. Consecutive "code stroke" CT angiography examinations were included and assessed for patient demographics, scanner manufacturer, presence or absence of CADt result, CADt result, and LVO in the internal carotid artery (ICA), horizontal middle cerebral artery (MCA) segment (M1), Sylvian MCA segments after the bifurcation (M2), precommunicating part of cerebral artery, postcommunicating part of the cerebral artery, vertebral artery, basilar artery vessel segments. The original radiology report served as the reference standard, and a study radiologist extracted the above data elements from the imaging examination and radiology report. RESULTS: At hospital A, the CADt algorithm manufacturer reports assessment of intracranial ICA and MCA with sensitivity of 97% and specificity of 95.6%. Real-world performance of 704 cases included 79 in which no CADt result was available. Sensitivity and specificity in ICA and M1 segments were 85.3% and 91.9%. Sensitivity decreased to 68.5% when M2 segments were included and to 59.9% when all proximal vessel segments were included. At hospital B the CADt algorithm manufacturer reports sensitivity of 87.8% and specificity of 89.6%, without specifying the vessel segments. Real-world performance of 642 cases included 20 cases in which no CADt result was available. Sensitivity and specificity in ICA and M1 segments were 90.7% and 97.9%. Sensitivity decreased to 76.4% when M2 segments were included and to 59.4% when all proximal vessel segments are included. DISCUSSION: Real-world testing of two CADt LVO detection algorithms identified gaps in the detection and communication of potentially treatable LVOs when considering vessels beyond the intracranial ICA and M1 segments and in cases with absent and uninterpretable data.

Developing, Purchasing, Implementing and Monitoring AI Tools in Radiology: Practical Considerations. A Multi-Society Statement from the ACR, CAR, ESR, RANZCR and RSNA.

Artificial Intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever-growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones. This multi-society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools. This article is simultaneously published in Insights into Imaging (DOI 10.1186/s13244-023-01541-3), Journal of Medical Imaging and Radiation Oncology (DOI 10.1111/1754-9485.13612), Canadian Association of Radiologists Journal (DOI 10.1177/08465371231222229), Journal of the American College of Radiology (DOI 10.1016/j.jacr.2023.12.005), and Radiology: Artificial Intelligence (DOI 10.1148/ryai.230513). Keywords: Artificial Intelligence, Radiology, Automation, Machine Learning Published under a CC BY 4.0 license. ©The Author(s) 2024. Editor's Note: The RSNA Board of Directors has endorsed this article. It has not undergone review or editing by this journal.

Developing, purchasing, implementing and monitoring AI tools in radiology: Practical considerations. A multi-society statement from the ACR, CAR, ESR, RANZCR & RSNA.

Artificial Intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever-growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones. This multi-society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools.

Developing, Purchasing, Implementing and Monitoring AI Tools in Radiology: Practical Considerations. A Multi-Society Statement From the ACR, CAR, ESR, RANZCR & RSNA.

Artificial intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever-growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones. This multi-society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools. KEY POINTS.

Developing, purchasing, implementing and monitoring AI tools in radiology: practical considerations. A multi-society statement from the ACR, CAR, ESR, RANZCR & RSNA.

Artificial Intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever-growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones.This multi-society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools.Key points • The incorporation of artificial intelligence (AI) in radiological practice demands increased monitoring of its utility and safety.• Cooperation between developers, clinicians, and regulators will allow all involved to address ethical issues and monitor AI performance.• AI can fulfil its promise to advance patient well-being if all steps from development to integration in healthcare are rigorously evaluated.

Keeping Patient Data Secure in the Age of Radiology Artificial Intelligence: Cybersecurity Considerations and Future Directions.

Artificial intelligence (AI)-based solutions are increasingly being incorporated into radiology workflows. Implementation of AI comes along with cybersecurity risks and challenges that practices should be aware of and mitigate for a successful and secure deployment. In this article, these cybersecurity issues are examined through the lens of the "CIA" triad framework-confidentiality, integrity, and availability. We discuss the implications of implementation configurations and development approaches on data security and confidentiality and the potential impact that the insertion of AI can have on the truthfulness of data, access to data, and the cybersecurity attack surface. Finally, we provide a checklist to address important security considerations before deployment of an AI application, and discuss future advances in AI addressing some of these security concerns.