Use of Artificial Intelligence in Radiology: Impact on Pediatric Patients, a White Paper From the ACR Pediatric AI Workgroup

Deep learning MR reconstruction in knees and ankles in children and young adults. Is it ready for clinical use?

OBJECTIVE

To evaluate the diagnostic performance and image quality of accelerated Turbo Spin Echo sequences using deep-learning (DL) reconstructions compared to conventional sequences in knee and ankle MRIs of children and young adults.

MATERIALS AND METHODS

IRB-approved prospective study consisting of 49 MRIs from 48 subjects (10 males, mean age 16.4 years, range 7-29 years), with each MRI consisting of both conventional and DL sequences. Sequences were evaluated blindly to determine predictive values, sensitivity, and specificity of DL sequences using conventional sequences and knee arthroscopy (if available) as references. Physeal patency and appearance were evaluated. Qualitative parameters were compared. Presence of undesired image alterations was assessed.

RESULTS

The prevalence of abnormal findings in the knees and ankles were 11.7% (75/640), and 11.5% (19/165), respectively. Using conventional sequences as reference, sensitivity and specificity of DL sequences in knees were 90.7% and 99.3%, and in ankles were 100.0% and 100.0%. Using arthroscopy as reference, sensitivity and specificity of DL sequences were 80.0% and 95.8%, and of conventional sequences were 80.0% and 97.9%. Agreement of physeal status was 100.0%. DL sequences were qualitatively "same-or-better" compared to conventional (p < 0.032), except for pixelation artifact for the PDFS sequence (p = 0.233). No discrete image alteration was identified in the knee DL sequences. In the ankle, we identified one DL artifact involving a tendon (0.8%, 1/125). DL sequences were faster than conventional sequences by a factor of 2 (p < 0.001).

CONCLUSION

In knee and ankle MRIs, DL sequences provided similar diagnostic performance and "same-or-better" image quality than conventional sequences at half the acquisition time.

Medical Imaging Data Strategies for Catalyzing AI Medical Device Innovation

Continuous and consistent access to quality medical imaging data stimulates innovations in artificial intelligence (AI) technologies for patient care. Breakthrough innovations in data-driven AI technologies are founded on seamless communication between data providers, data managers, data users and regulators or other evaluators to determine the standards for quality data. However, the complexity in imaging data quality and heterogeneous nature of AI-enabled medical devices and their intended uses presents several challenges limiting the clinical translation of novel AI technologies. In this commentary, we discuss these challenges across different characteristics of data, such as data size, data labels, data diversity, data sequestration and reuse, and data drift. We discuss strategies around a data platform that incorporates protocols and checklists for ensuring data quality, tools and interactive guidelines that may help assess data diversity, study design and performance metrics for data usage and monitoring for data analytics. We envision this data platform to catalyze AI-enabled medical device innovation by providing a more efficient development and evaluation environment for bringing safe and effective AI technologies to the clinic.

Optimizing adult-oriented artificial intelligence for pediatric chest radiographs by adjusting operating points

The purpose of this study was to evaluate whether the optimal operating points of adult-oriented artificial intelligence (AI) software differ for pediatric chest radiographs and to assess its diagnostic performance. Chest radiographs from patients under 19 years old, collected between March and November 2021, were divided into test and exploring sets. A commercial adult-oriented AI software was utilized to detect lung lesions, including pneumothorax, consolidation, nodule, and pleural effusion, using a standard operating point of 15%. A pediatric radiologist reviewed the radiographs to establish ground truth for lesion presence. To determine the optimal operating points, receiver operating characteristic (ROC) curve analysis was conducted, varying thresholds to balance sensitivity and specificity by lesion type, age group, and imaging method. The test set (4,727 chest radiographs, mean 7.2 ± 6.1 years) and exploring set (2,630 radiographs, mean 5.9 ± 6.0 years) yielded optimal operating points of 11% for pneumothorax, 14% for consolidation, 15% for nodules, and 6% for pleural effusion. Using a 3% operating point improved pneumothorax sensitivity for children under 2 years, portable radiographs, and anteroposterior projections. Therefore, optimizing operating points of AI based on lesion type, age, and imaging method could improve diagnostic performance for pediatric chest radiographs, building on adult-oriented AI as a foundation.

US FDA Approval of Pediatric Artificial Intelligence and Machine Learning-Enabled Medical Devices

This cross-sectional study analyzes the availability of artificial intelligence and machine learning–enabled devices authorized for children by the US Food and Drug Administration (FDA) and assesses reporting of algorithm validation in the pediatric population.

"How I would like AI used for my imaging": children and young persons' perspectives

OBJECTIVES

Artificial intelligence (AI) tools are becoming more available in modern healthcare, particularly in radiology, although less attention has been paid to applications for children and young people. In the development of these, it is critical their views are heard.

MATERIALS AND METHODS

A national, online survey was publicised to UK schools, universities and charity partners encouraging any child or young adult to participate. The survey was "live" for one year (June 2022 to 2023). Questions about views of AI in general, and in specific circumstances (e.g. bone fractures) were asked.

RESULTS

One hundred and seventy-one eligible responses were received, with a mean age of 19 years (6-23 years) with representation across all 4 UK nations. Most respondents agreed or strongly agreed they wanted to know the accuracy of an AI tool that was being used (122/171, 71.3%), that accuracy was more important than speed (113/171, 66.1%), and that AI should be used with human oversight (110/171, 64.3%). Many respondents (73/171, 42.7%) felt AI would be more accurate at finding problems on bone X-rays than humans, with almost all respondents who had sustained a missed fracture strongly agreeing with that sentiment (12/14, 85.7%).

CONCLUSIONS

Children and young people in our survey had positive views regarding AI, and felt it should be integrated into modern healthcare, but expressed a preference for a "medical professional in the loop" and accuracy of findings over speed. Key themes regarding information on AI performance and governance were raised and should be considered prior to future AI implementation for paediatric healthcare.

CLINICAL RELEVANCE STATEMENT

Artificial intelligence (AI) integration into clinical practice must consider all stakeholders, especially paediatric patients who have largely been ignored. Children and young people favour AI involvement with human oversight, seek assurances for safety, accuracy, and clear accountability in case of failures.

KEY POINTS

Paediatric patient's needs and voices are often overlooked in AI tool design and deployment. Children and young people approved of AI, if paired with human oversight and reliability. Children and young people are stakeholders for developing and deploying AI tools in paediatrics.

Post-deployment performance of a deep learning algorithm for normal and abnormal chest X-ray classification: A study at visa screening centers in the United Arab Emirates

Background

Chest radiographs (CXRs) are widely used to screen for infectious diseases like tuberculosis and COVID-19 among migrants. At such high-volume settings, manual CXR reporting is challenging and integrating artificial intelligence (AI) algorithms into the workflow help to rule out normal findings in minutes, allowing radiologists to focus on abnormal cases.

Methods

In this post-deployment study, all the CXRs acquired during the visa screening process across 33 centers in United Arab Emirates from January 2021 to June 2022 (18 months) were included. The qXR v2.1 chest X-ray interpretation software was used to classify the scans into normal and abnormal, and its agreement against radiologist was evaluated. Additionally, a digital survey was conducted among 20 healthcare professionals with prior AI experience to understand real-world implementation challenges and impact.

Results

The analysis of 1309,443 CXRs from 1309,431 patients (median age: 35 years; IQR [29-42]; 1030,071 males [78.7 %]) in this study revealed a Negative Predictive Value (NPV) of 99.92 % (95 % CI: 99.92, 99.93), Positive Predictive Value (PPV) of 5.06 % (95 % CI: 4.99, 5.13) and overall percent agreement of the AI with radiologists of 72.90 % (95 % CI: 72.82, 72.98). In the survey, majority (88.2 %) of the radiologists agreed to turnaround time reduction after AI integration, while 82 % suggested that the AI improved their diagnostic accuracy.

Discussion

In contrast with the existing studies, this research uses a substantially large data. A high NPV and satisfactory agreement with human readers indicate that AI can reliably identify normal CXRs, making it suitable for routine applications.

Artificial intelligence: a primer for pediatric radiologists

Artificial intelligence (AI) is increasingly recognized for its transformative potential in radiology; yet, its application in pediatric radiology is relatively limited when compared to the whole of radiology. This manuscript introduces pediatric radiologists to essential AI concepts, including topics such as use case, data science, machine learning, deep learning, natural language processing, and generative AI as well as basics of AI training and validating. We outline the unique challenges of applying AI in pediatric imaging, such as data scarcity and distinct clinical characteristics, and discuss the current uses of AI in pediatric radiology, including both image interpretive and non-interpretive tasks. With this overview, we aim to equip pediatric radiologists with the foundational knowledge needed to engage with AI tools and inspire further exploration and innovation in the field.

Liver magnetic resonance imaging, non-alcoholic fatty liver disease and metabolic syndrome risk in pre-pubertal Mexican boys

Rising global pediatric obesity rates, increase non-alcoholic fatty liver disease (NAFLD) and metabolic syndrome (MetS) prevalence, with MetS being a NAFLD risk factor. NAFLD can be asymptomatic, with liver function tests insensitive to mild disease, and liver biopsy, risking complications. Thus, we investigated multiparametric MRI (mpMRI) metrics of liver fat (proton density fat fraction, PDFF) and disease activity (fibro-inflammation; iron-corrected T1, cT1), in a Hispanic pre-pubertal pediatric cohort, with increased risk of NAFLD. Pre-pubertal boys (n = 81) of varying Body-Mass Index (BMI) were recruited in Mexico City. Most children (81%) had normal liver transaminase levels, 38% had high BMI, and 14% had ≥ 3 MetS risk factors. Applying mpMRI thresholds, 12%, 7% and 4% of the cohort had NAFLD, NASH and high-risk NASH respectively. Participants with ≥ 3 MetS risk factors had higher cT1 (834 ms vs. 737 ms, p = 0.004) and PDFF (8.7% vs. 2.2%, p < 0.001) compared to those without risk factors. Those with elevated cT1 tended to have high BMI and high insulin (p = 0.005), HOMA-IR (p = 0.005) and leptin (p < 0.001). The significant association of increased risk of MetS with abnormal mpMRI, particularly cT1, proposes the potential of using mpMRI for routine pediatric NAFLD screening of high-risk (high BMI, high MetS risk score) populations.

Reference Range of Quantitative MRI Metrics Corrected T1 and Liver Fat Content in Children and Young Adults: Pooled Participant Analysis

BACKGROUND

Multiparametric MRI markers of liver health corrected T1 (cT1) and proton density fat fraction (PDFF) have shown utility in the management of various chronic liver diseases. We assessed the normal population reference range of both cT1 and PDFF in healthy child and adult volunteers without any known liver disease.

METHODS

A retrospective multi-centre pooled analysis of 102 child and young adult (9.1 years (6-18)) volunteers from three centres: Children's Memorial Health Institute (N = 21), University Hospital Southampton (N = 28) and Hospital Infantil de Mexico (N = 53). Sex and ethnic differences were investigated for both cT1 and PDFF. Age effects were investigated with comparison to a pooled adult cohort from the UK Biobank (N = 500) and CoverScan (N = 71), covering an age range of 21 to 81 years.

RESULTS

cT1 values were normally distributed with a median of 748 ms (IQR: 725-768 ms; 2.5-97.5 percentiles: 683-820 ms). PDFF values followed a normal distribution with a median of 1.7% (IQR: 1.3-1.9%; 2.5-97.5 percentiles: 1-4.4%). There were no significant age and sex differences in cT1 and PDFF between children and young adults. No differences in cT1 and PDFF were found between ethnicities. Age comparisons showed statistically significant, but clinically negligible, cT1 (748 ms vs. 732 ms) and PDFF (2.4% vs. 1.9%) differences between paediatric and adult groups, respectively.

CONCLUSIONS

Median healthy cT1 and PDFF reference ranges in children and young adults fall within the reported limits for normal of 800 ms and 5%, respectively.

Evaluation of 3D T1-weighted spoiled gradient echo MR image quality using artificial intelligence image reconstruction techniques in the pediatric brain

PURPOSE

To assess image quality and diagnostic confidence of 3D T1-weighted spoiled gradient echo (SPGR) MRI using artificial intelligence (AI) reconstruction.

MATERIALS AND METHODS

This prospective, IRB-approved study enrolled 50 pediatric patients (mean age = 11.8 ± 3.1 years) undergoing clinical brain MRI. In addition to standard of care (SOC) compressed SENSE (CS = 2.5), 3D T1-weighted SPGR images were obtained with higher CS acceleration factors (5 and 8) to evaluate the ability of AI reconstruction to improve image quality and reduce scan time. Images were reviewed independently on dedicated research PACS workstations by two neuroradiologists. Quantitative analysis of signal intensities to calculate apparent grey and white matter signal to noise (aSNR) and grey-white matter apparent contrast to noise ratios (aCNR) was performed.

RESULTS

AI improved overall image quality compared to standard CS reconstruction in 35% (35/100) of evaluations in CS = 2.5 (average scan time = 221 ± 6.9 s), 100% (46/46) of CS = 5 (average scan time = 113.3 ± 4.6 s) and 94% (47/50) of CS = 8 (average scan time = 74.1 ± 0.01 s). Quantitative analysis revealed significantly higher grey matter aSNR, white matter aSNR and grey-white matter aCNR with AI reconstruction compared to standard reconstruction for CS 5 and 8 (all p-values < 0.001), however not for CS 2.5.

CONCLUSIONS

AI reconstruction improved overall image quality and gray-white matter qualitative and quantitative aSNR and aCNR in highly accelerated (CS = 5 and 8) 3D T1W SPGR images in the majority of pediatric patients.

Children Are Not Small Adults: Addressing Limited Generalizability of an Adult Deep Learning CT Organ Segmentation Model to the Pediatric Population

Deep learning (DL) tools developed on adult data sets may not generalize well to pediatric patients, posing potential safety risks. We evaluated the performance of TotalSegmentator, a state-of-the-art adult-trained CT organ segmentation model, on a subset of organs in a pediatric CT dataset and explored optimization strategies to improve pediatric segmentation performance. TotalSegmentator was retrospectively evaluated on abdominal CT scans from an external adult dataset (n = 300) and an external pediatric data set (n = 359). Generalizability was quantified by comparing Dice scores between adult and pediatric external data sets using Mann-Whitney U tests. Two DL optimization approaches were then evaluated: (1) 3D nnU-Net model trained on only pediatric data, and (2) an adult nnU-Net model fine-tuned on the pediatric cases. Our results show TotalSegmentator had significantly lower overall mean Dice scores on pediatric vs. adult CT scans (0.73 vs. 0.81, P < .001) demonstrating limited generalizability to pediatric CT scans. Stratified by organ, there was lower mean pediatric Dice score for four organs (P < .001, all): right and left adrenal glands (right adrenal, 0.41 [0.39-0.43] vs. 0.69 [0.66-0.71]; left adrenal, 0.35 [0.32-0.37] vs. 0.68 [0.65-0.71]); duodenum (0.47 [0.45-0.49] vs. 0.67 [0.64-0.69]); and pancreas (0.73 [0.72-0.74] vs. 0.79 [0.77-0.81]). Performance on pediatric CT scans improved by developing pediatric-specific models and fine-tuning an adult-trained model on pediatric images where both methods significantly improved segmentation accuracy over TotalSegmentator for all organs, especially for smaller anatomical structures (e.g., > 0.2 higher mean Dice for adrenal glands; P < .001).

Capability of multimodal large language models to interpret pediatric radiological images

BACKGROUND

There is a dearth of artificial intelligence (AI) development and research dedicated to pediatric radiology. The newest iterations of large language models (LLMs) like ChatGPT can process image and video input in addition to text. They are thus theoretically capable of providing impressions of input radiological images.

OBJECTIVE

To assess the ability of multimodal LLMs to interpret pediatric radiological images.

MATERIALS AND METHODS

Thirty medically significant cases were collected and submitted to GPT-4 (OpenAI, San Francisco, CA), Gemini 1.5 Pro (Google, Mountain View, CA), and Claude 3 Opus (Anthropic, San Francisco, CA) with a short history for a total of 90 images. AI responses were recorded and independently assessed for accuracy by a resident and attending physician. 95% confidence intervals were determined using the adjusted Wald method.

RESULTS

Overall, the models correctly diagnosed 27.8% (25/90) of images (95% CI=19.5-37.8%), were partially correct for 13.3% (12/90) of images (95% CI=2.7-26.4%), and were incorrect for 58.9% (53/90) of images (95% CI=48.6-68.5%).

CONCLUSION

Multimodal LLMs are not yet capable of interpreting pediatric radiological images.

Assessing the diagnostic accuracy of artificial intelligence in post-endovascular aneurysm repair endoleak detection using dual-energy computed tomography angiography

Purpose

The aim of this study was to evaluate the diagnostic accuracy of an artificial intelligence (AI) tool in detecting endoleaks in patients undergoing endovascular aneurysm repair (EVAR) using dual-energy computed tomography angiography (CTA).

Material and methods

The study involved 95 patients who underwent EVAR and subsequent CTA follow-up. Dualenergy scans were performed, and images were reconstructed as linearly blended (LB) and 40 keV virtual monoenergetic (VMI) images. The AI tool PRAEVAorta®2 was used to assess arterial phase images for endoleaks. Two experienced readers independently evaluated the same images, and their consensus served as the reference standard. Key metrics, including accuracy, precision, recall, F1 score, and area under the receiver operating characteristic (ROC) curve (AUC), were calculated.

Results

The final analysis included 94 patients. The AI tool demonstrated an accuracy of 78.7%, precision of 67.6%, recall of 10 71.9%, F1 score of 69.7%, and an AUC of 0.77 using LB images. However, the tool failed to process 40 keV VMI images correctly, limiting further analysis of these datasets.

Conclusions

The AI tool showed moderate diagnostic accuracy in detecting endoleaks using LB images but failed to achieve the reliability needed for clinical use due to the significant number of misdiagnoses.

MRI-serum-based score accurately identifies patients undergoing liver transplant without rejection avoiding the need for liver biopsy: A multisite European study

Serum liver tests (serum tests) and histological assessment for T-cell-mediated rejection are essential for post-liver transplant monitoring. Liver biopsy carries a risk of complications that are preferably avoided in low-risk patients. Multiparametric magnetic resonance imaging (mpMRI) is a reliable noninvasive diagnostic method that quantifies liver disease activity and has prognostic utility. Our aim was to determine whether using mpMRI in combination with serum tests could noninvasively identify low-risk patients who underwent liver transplants who are eligible to avoid invasive liver biopsies. In a multicenter prospective study (RADIcAL2), including 131 adult and pediatric (children and adolescent) patients with previous liver transplants from the Netherlands, Portugal, and the United Kingdom, concomitant mpMRI and liver biopsies were performed. Biopsies were centrally read by 2 expert pathologists. T-cell-mediated rejection was assessed using the BANFF global assessment. Diagnostic accuracy to discriminate no rejection versus indeterminate or T-cell-mediated liver transplant rejection was performed using the area under the receiver operating characteristic curve. In this study, 52% of patients received a routine (protocol) biopsy, while 48% had a biopsy for suspicion of pathology. Thirty-eight percent of patients had no rejection, while 62% had either indeterminate (21%) or T-cell-mediated rejection (41%). However, there was a high interobserver variability (0 < Cohen's Kappa < 0.85) across all histology scores. The combined score of mpMRI and serum tests had area under the receiver operating characteristic curve 0.7 (negative predictive value 0.8) to identify those without either indeterminate or T-cell-mediated rejection. Combining both imaging and serum biomarkers into a composite biomarker (imaging and serum biomarkers) has the potential to monitor the liver graft to effectively risk stratify patients and identify those most likely to benefit from a noninvasive diagnostic approach, reducing the need for liver biopsy.

Emerging technologies in pediatric radiology: current developments and future prospects

Radiological imaging is a crucial diagnostic tool for the pediatric population. However, it is associated with several unique challenges in this age group compared to adults. These challenges mainly come from the fact that children are not small-sized adults and differ in development, anatomy, physiology, and pathology compared to adults. This paper reviews relevant articles published between January 2015 and October 2023 to analyze challenges associated with imaging technologies currently used in pediatric radiology, emerging technologies, and their role in resolving the challenges and future prospects of pediatric radiology. In recent decades, imaging technologies have advanced rapidly, developing advanced ultrasound, computed tomography, magnetic resonance, nuclear imaging, teleradiology, artificial intelligence, machine learning, three-dimensional printing, radiomics, and radiogenomics, among many others. By prioritizing the unique needs of pediatric patients while developing such technologies, we can significantly alleviate the challenges faced in pediatric radiology.

Evaluation of T2W FLAIR MR image quality using artificial intelligence image reconstruction techniques in the pediatric brain

BACKGROUND

Artificial intelligence (AI) reconstruction techniques have the potential to improve image quality and decrease imaging time. However, these techniques must be assessed for safe and effective use in clinical practice.

OBJECTIVE

To assess image quality and diagnostic confidence of AI reconstruction in the pediatric brain on fluid-attenuated inversion recovery (FLAIR) imaging.

MATERIALS AND METHODS

This prospective, institutional review board (IRB)-approved study enrolled 50 pediatric patients (median age=12 years, Q1=10 years, Q3=14 years) undergoing clinical brain MRI. T2-weighted (T2W) FLAIR images were reconstructed by both standard clinical and AI reconstruction algorithms (strong denoising). Images were independently rated by two neuroradiologists on a dedicated research picture archiving and communication system (PACS) to indicate whether AI increased, decreased, or had no effect on image quality compared to standard reconstruction. Quantitative analysis of signal intensities was also performed to calculate apparent signal to noise (aSNR) and apparent contrast to noise (aCNR) ratios.

RESULTS

AI reconstruction was better than standard in 99% (reader 1, 49/50; reader 2, 50/50) for overall image quality, 99% (reader 1, 49/50; reader 2, 50/50) for subjective SNR, and 98% (reader 1, 49/50; reader 2, 49/50) for diagnostic preference. Quantitative analysis revealed significantly higher gray matter aSNR (30.6±6.5), white matter aSNR (21.4±5.6), and gray-white matter aCNR (7.1±1.6) in AI-reconstructed images compared to standard reconstruction (18±2.7, 14.2±2.8, 4.4±0.8, p<0.001) respectively.

CONCLUSION

We conclude that AI reconstruction improved T2W FLAIR image quality in most patients when compared with standard reconstruction in pediatric patients.

Artificial Intelligence in the Future Landscape of Pediatric Neuroradiology: Opportunities and Challenges

This paper will review how artificial intelligence (AI) will play an increasingly important role in pediatric neuroradiology in the future. A safe, transparent, and human-centric AI is needed to tackle the quadruple aim of improved health outcomes, enhanced patient and family experience, reduced costs, and improved well-being of the healthcare team in pediatric neuroradiology. Equity, diversity and inclusion, data safety, and access to care will need to always be considered. In the next decade, AI algorithms are expected to play an increasingly important role in access to care, workflow management, abnormality detection, classification, response prediction, prognostication, report generation, as well as in the patient and family experience in pediatric neuroradiology. Also, AI algorithms will likely play a role in recognizing and flagging rare diseases and in pattern recognition to identify previously unknown disorders. While AI algorithms will play an important role, humans will not only need to be in the loop, but in the center of pediatric neuroimaging. AI development and deployment will need to be closely watched and monitored by experts in the field. Patient and data safety need to be at the forefront, and the risks of a dependency on technology will need to be contained. The applications and implications of AI in pediatric neuroradiology will differ from adult neuroradiology.