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 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.

Self-supervised enhanced thyroid nodule detection in ultrasound examination video sequences with multi-perspective evaluation.

Objective.Ultrasound is the most commonly used examination for the detection and identification of thyroid nodules. Since manual detection is time-consuming and subjective, attempts to introduce machine learning into this process are ongoing. However, the performance of these methods is limited by the low signal-to-noise ratio and tissue contrast of ultrasound images. To address these challenges, we extend thyroid nodule detection from image-based to video-based using the temporal context information in ultrasound videos.Approach.We propose a video-based deep learning model with adjacent frame perception (AFP) for accurate and real-time thyroid nodule detection. Compared to image-based methods, AFP can aggregate semantically similar contextual features in the video. Furthermore, considering the cost of medical image annotation for video-based models, a patch scale self-supervised model (PASS) is proposed. PASS is trained on unlabeled datasets to improve the performance of the AFP model without additional labelling costs.Main results.The PASS model is trained by 92 videos containing 23 773 frames, of which 60 annotated videos containing 16 694 frames were used to train and evaluate the AFP model. The evaluation is performed from the video, frame, nodule, and localization perspectives. In the evaluation of the localization perspective, we used the average precision metric with the intersection-over-union threshold set to 50% (AP@50), which is the area under the smoothed Precision-Recall curve. Our proposed AFP improved AP@50 from 0.256 to 0.390, while the PASS-enhanced AFP further improved the AP@50 to 0.425. AFP and PASS also improve the performance in the valuations of other perspectives based on the localization results.Significance.Our video-based model can mitigate the effects of low signal-to-noise ratio and tissue contrast in ultrasound images and enable the accurate detection of thyroid nodules in real-time. The evaluation from multiple perspectives of the ablation experiments demonstrates the effectiveness of our proposed AFP and PASS models.

CT features associated with underlying malignancy in patients with diagnosed mesenteric panniculitis.

PURPOSE: The purpose of this study was to identify abdominal computed tomography (CT) features associated with underlying malignancy in patients with mesenteric panniculitis (MP). MATERIALS AND METHODS: This single-institution retrospective longitudinal cohort study included patients with MP and a minimum 1-year abdominopelvic CT follow-up or 2-year clinical follow-up after initial abdominopelvic CT examination. Two radiologists, blinded to patients' medical records, conjointly reviewed CT-based features of MP. Electronic medical records were reviewed for newly diagnosed malignancies with the following specific details: type (lymphoproliferative disease or solid malignancy), location (possible mesenteric drainage or distant), stage, time to diagnosis. An expert panel of three radiologists and one hemato-oncologist, who were blinded to the initial CT-based MP features, assessed the probability of association between MP and malignancy based on the malignancy characteristics. RESULTS: From 2006 to 2016, 444 patients with MP were included. There were 272 men and 172 women, with a median age of 64 years (age range: 25-89); the median overall follow-up was 36 months (IQR: 22, 60; range: 12-170). A total of 34 (8%) patients had a diagnosis of a new malignancy; 5 (1%) were considered possibly related to the MP, all being low-grade B-cell non-Hodgkin lymphomas. CT features associated with the presence of an underlying malignancy were the presence of an MP soft-tissue nodule with a short axis >10 mm (P < 0.0001) or lymphadenopathy in another abdominopelvic region (P < 0.0001). Associating these two features resulted in high diagnostic performance (sensitivity 100%; [95% CI: 57-100]; specificity 99% [95% CI: 98-100]). All related malignancies were identified. CONCLUSION: Further workup to rule out an underlying malignancy is only necessary in the presence of an MP soft-tissue nodule >10 mm or associated abdominopelvic lymphadenopathy.

Incidence of infectious complications following ultrasound-guided percutaneous musculoskeletal interventions with the use of an uncovered transducer footprint.

OBJECTIVES: To determine the incidence of infectious complications following ultrasound-guided musculoskeletal interventions performed with a disinfected uncovered ultrasound transducer footprint. METHODS: Electronic medical records of all patients who underwent an ultrasound-guided musculoskeletal procedure (including injection, calcific lavage, or ganglion cyst aspiration) performed by any of the 14 interventional musculoskeletal radiologists at our institution between January 2013 and December 2018 were retrospectively reviewed to identify procedure site infections. Biopsies and joint aspirations were excluded. The procedures were performed using a disinfected uncovered transducer footprint. First, an automated chart review identified cases with (1) positive answers to the nurse's post-procedure call, (2) an International Classification of Diseases (ICD) diagnostic code related to a musculoskeletal infection, or (3) an antibiotic prescription within 30 days post-procedure. Then, these cases were manually reviewed for evidence of procedure site infection. RESULTS: In total, 6511 procedures were included. The automated chart review identified 3 procedures (2 patients) in which post-procedural fever was reported during the nurse's post-procedure call, 33 procedures (28 patients) with an ICD code for a musculoskeletal infection, and 220 procedures (216 patients) with an antibiotic prescription within 30 post-procedural days. The manual chart review of these patients revealed no cases of confirmed infection and 1 case (0.015%) of possible site infection. CONCLUSIONS: The incidence of infectious complications after an ultrasound-guided musculoskeletal procedure performed with an uncovered transducer footprint is extremely low. This information allows radiologists to counsel their patients more precisely when obtaining informed consent. KEY POINTS: • Infectious complications after ultrasound-guided musculoskeletal procedures performed with a disinfected uncovered transducer footprint are extremely rare.

Impact of sarcopenia on outcomes of patients treated with trimodal therapy for muscle invasive bladder cancer.

OBJECTIVE: We sought to investigate the incidence of sarcopenia and its impact on main oncological outcomes in patients with muscle invasive bladder cancer (MIBC) treated with trimodal therapy (TMT). PATIENTS AND METHODS: This was a retrospective analysis of 141 MIBC patients treated with TMT in the period 2002 to 2018. Sarcopenia was identified through pretreatment computed tomography scans and defined as a skeletal muscle index of <55 cm2/m2 for men and <39 cm2/m2 for women. Body mass index (BMI)-adjusted definition of sarcopenia was used to evaluate for sarcopenic obesity. Uni- and multivariable analyses were performed to assess the impact of sarcopenia on initial complete response and overall survival (OS) to TMT. RESULTS: Median age at diagnosis was 73 years [range: 65-81] and median follow up was 32 months (Inter Quartile Range: 18-66). Median OS was 67 months (95% CI: 53-83). The incidence of sarcopenia and BMI-adjusted sarcopenia was 56.7% and 40.4%, respectively. On multivariable analysis, Eastern Cooperative Oncology Group performance status (HR = 2.37, 95% CI: 2.1-5.67, P = 0.001) and complete response to treatment (HR = 0.26, 95% CI: 0.14-0.049, P = 0.001] were independently associated with improved OS. Sarcopenia and BMI-adjusted sarcopenia were not independently associated with either complete response to TMT or OS. Similarly, in a subpopulation of 74 patients considered fit for radical cystectomy, we found that neither sarcopenia (P = 0.49) nor BMI-adjusted sarcopenia (P = 0.22) had an impact on OS. CONCLUSION: Sarcopenia and BMI-adjusted sarcopenia are prevalent in patients with MIBC undergoing TMT. TMT is a suitable treatment modality for patients with MIBC irrespective of their sarcopenia status.

Weakly-Supervised teacher-Student network for liver tumor segmentation from non-enhanced images.

Accurate liver tumor segmentation without contrast agents (non-enhanced images) avoids the contrast-agent-associated time-consuming and high risk, which offers radiologists quick and safe assistance to diagnose and treat the liver tumor. However, without contrast agents enhancing, the tumor in liver images presents low contrast and even invisible to naked eyes. Thus the liver tumor segmentation from non-enhanced images is quite challenging. We propose a Weakly-Supervised Teacher-Student network (WSTS) to address the liver tumor segmentation in non-enhanced images by leveraging additional box-level-labeled data (labeled with a tumor bounding-box). WSTS deploys a weakly-supervised teacher-student framework (TCH-ST), namely, a Teacher Module learns to detect and segment the tumor in enhanced images during training, which facilitates a Student Module to detect and segment the tumor in non-enhanced images independently during testing. To detect the tumor accurately, the WSTS proposes a Dual-strategy DRL (DDRL), which develops two tumor detection strategies by creatively introducing a relative-entropy bias in the DRL. To accurately predict a tumor mask for the box-level-labeled enhanced image and thus improve tumor segmentation in non-enhanced images, the WSTS proposes an Uncertainty-Sifting Self-Ensembling (USSE). The USSE exploits the weakly-labeled data with self-ensembling and evaluates the prediction reliability with a newly-designed Multi-scale Uncertainty-estimation. WSTS is validated with a 2D MRI dataset, where the experiment achieves 83.11% of Dice and 85.12% of Recall in 50 patient testing data after training by 200 patient data (half amount data is box-level-labeled). Such a great result illustrates the competence of WSTS to segment the liver tumor from non-enhanced images. Thus, WSTS has excellent potential to assist radiologists by liver tumor segmentation without contrast-agents.

Synthesis of gadolinium-enhanced liver tumors on nonenhanced liver MR images using pixel-level graph reinforcement learning.

If successful, synthesis of gadolinium (Gd)-enhanced liver tumors on nonenhanced liver MR images will be critical for liver tumor diagnosis and treatment. This synthesis will offer a safe, efficient, and low-cost clinical alternative to eliminate the use of contrast agents in the current clinical workflow and significantly benefit global healthcare systems. In this study, we propose a novel pixel-level graph reinforcement learning method (Pix-GRL). This method directly takes regular nonenhanced liver images as input and outputs AI-enhanced liver tumor images, thereby making them comparable to traditional Gd-enhanced liver tumor images. In Pix-GRL, each pixel has a pixel-level agent, and the agent explores the pixels features and outputs a pixel-level action to iteratively change the pixel value, ultimately generating AI-enhanced liver tumor images. Most importantly, Pix-GRL creatively embeds a graph convolution to represent all the pixel-level agents. A graph convolution is deployed to the agent for feature exploration to improve the effectiveness through the aggregation of long-range contextual features, as well as outputting the action to enhance the efficiency through shared parameter training between agents. Moreover, in our Pix-GRL method, a novel reward is used to measure pixel-level action to significantly improve the performance by considering the improvement in each action in each pixel with its own future state, as well as those of neighboring pixels. Pix-GRL significantly upgrades the existing medical DRL methods from a single agent to multiple pixel-level agents, becoming the first DRL method for medical image synthesis. Comprehensive experiments on three types of liver tumor datasets (benign, cancerous, and healthy controls) with 325 patients (24,375 images) show that our novel Pix-GRL method outperforms existing medical image synthesis learning methods. It achieved an SSIM of 0.85 ± 0.06 and a Pearson correlation coefficient of 0.92 in terms of the tumor size. These results prove that the potential exists to develop a successful clinical alternative to Gd-enhanced liver MR imaging.

ACR Appropriateness Criteria® Recurrent Lower Urinary Tract Infections in Females.

Urinary tract infections (UTIs) in women are common, with an overall lifetime risk over >50%. UTIs are considered recurrent when they follow complete clinical resolution of a previous UTI and are usually defined as at least three episodes of infection within the preceding 12 months. An uncomplicated UTI is classified as a UTI without structural or functional abnormalities of the urinary tract and without relevant comorbidities. Complicated UTIs are those occurring in patients with underlying structural or medical problems. In women with recurrent uncomplicated UTIs, cystoscopy and imaging are not routinely used. In women suspected of having a recurrent complicated UTI, cystoscopy and imaging should be considered. CT urography or MR urography are usually appropriate for the evaluation of recurrent complicated lower urinary tract infections or for women who are nonresponders to conventional therapy, develop frequent reinfections or relapses, or have known underlying risk factors. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

Tripartite-GAN: Synthesizing liver contrast-enhanced MRI to improve tumor detection.

Contrast-enhanced magnetic resonance imaging (CEMRI) is crucial for the diagnosis of patients with liver tumors, especially for the detection of benign tumors and malignant tumors. However, it suffers from high-risk, time-consuming, and expensive in current clinical diagnosis due to the use of the gadolinium-based contrast agent (CA) injection. If the CEMRI can be synthesized without CA injection, there is no doubt that it will greatly optimize the diagnosis. In this study, we propose a Tripartite Generative Adversarial Network (Tripartite-GAN) as a non-invasive, time-saving, and inexpensive clinical tool by synthesizing CEMRI to detect tumors without CA injection. Specifically, our innovative Tripartite-GAN combines three associated-networks (an attention-aware generator, a convolutional neural network-based discriminator, and a region-based convolutional neural network-based detector) for the first time, which achieves CEMRI synthesis and tumor detection promoting each other in an end-to-end framework. The generator facilitates detector for accurate tumor detection via synthesizing tumor-specific CEMRI. The detector promotes the generator for accurate CEMRI synthesis via the back-propagation. In order to synthesize CEMRI of equivalent clinical value to real CEMRI, the attention-aware generator expands the receptive field via hybrid convolution, and enhances feature representation and context learning of multi-class liver MRI via dual attention mechanism, and improves the performance of convergence of loss via residual learning. Moreover, the attention maps obtained from the generator newly added into the detector improve the performance of tumor detection. The discriminator promotes the generator to synthesize high-quality CEMRI via the adversarial learning strategy. This framework is tested on a large corpus of axial T1 FS Pre-Contrast MRI and axial T1 FS Delay MRI of 265 subjects. Experimental results and quantitative evaluation demonstrate that the Tripartite-GAN achieves high-quality CEMRI synthesis that peak signal-to-noise rate of 28.8 and accurate tumor detection that accuracy of 89.4%, which reveals that Tripartite-GAN can aid in the clinical diagnosis of liver tumors.

Pattern Recognition in Musculoskeletal Imaging Using Artificial Intelligence.

Artificial intelligence (AI) has the potential to affect every step of the radiology workflow, but the AI application that has received the most press in recent years is image interpretation, with numerous articles describing how AI can help detect and characterize abnormalities as well as monitor disease response. Many AI-based image interpretation tasks for musculoskeletal (MSK) pathologies have been studied, including the diagnosis of bone tumors, detection of osseous metastases, assessment of bone age, identification of fractures, and detection and grading of osteoarthritis. This article explores the applications of AI for image interpretation of MSK pathologies.

ACR Appropriateness Criteria® Lower Urinary Tract Symptoms-Suspicion of Benign Prostatic Hyperplasia.

Lower urinary tract symptoms due to benign prostatic enlargement have a high prevalence in men over 50 years of age. Diagnosis is made with a combination of focused history and physician examination and validated symptom questionnaires. Urodynamic studies can help to differentiate storage from voiding abnormalities. Pelvic ultrasound may be indicated to assess bladder volume and wall thickness. Other imaging modalities, including prostate MRI, are usually not indicated in the initial workup and evaluation of uncomplicated lower urinary tract symptoms from an enlarged prostate. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

ACR Appropriateness Criteria® Penetrating Trauma-Lower Abdomen and Pelvis.

Lower urinary tract injury is most commonly the result of blunt trauma but can also result from penetrating or iatrogenic trauma. Clinical findings in patients with a mechanism of penetrating trauma to the lower urinary tract include lacerations or puncture wounds of the pelvis, perineum, buttocks, or genitalia, as well as gross hematuria or inability to void. CT cystography or fluoroscopy retrograde cystography are usually the most appropriate initial imaging procedures in patients with a mechanism of penetrating trauma to the lower urinary tract. CT of the pelvis with intravenous contrast, pelvic radiography, fluoroscopic retrograde urethrography, and CT of the pelvis without intravenous contrast may be appropriate in some cases. Arteriography, radiographic intravenous urography, CT of the pelvis without and with intravenous contrast, ultrasound, MRI, and nuclear scintigraphy are usually not appropriate. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

ACR Appropriateness Criteria® Post-Treatment Surveillance of Bladder Cancer.

Urothelial cancer is the second most common cancer, and cause of cancer death, related to the genitourinary tract. The goals of surveillance imaging after the treatment of urothelial cancer of the urinary bladder are to detect new or previously undetected urothelial tumors, to identify metastatic disease, and to evaluate for complications of therapy. For surveillance, patients can be stratified into one of three groups: (1) nonmuscle invasive bladder cancer with no symptoms or additional risk factors; (2) nonmuscle invasive bladder cancer with symptoms or additional risk factors; and (3) muscle invasive bladder cancer. This article is a review of the current literature for urothelial cancer and resulting recommendations for surveillance imaging. The American College of Radiology Appropriateness Criteria are evidence-based guidelines for specific clinical conditions that are reviewed annually by a multidisciplinary expert panel. The guideline development and revision include an extensive analysis of current medical literature from peer reviewed journals and the application of well-established methodologies (RAND/UCLA Appropriateness Method and Grading of Recommendations Assessment, Development, and Evaluation or GRADE) to rate the appropriateness of imaging and treatment procedures for specific clinical scenarios. In those instances where evidence is lacking or equivocal, expert opinion may supplement the available evidence to recommend imaging or treatment.

Image-based biomarkers for solid tumor quantification.

The last few decades have witnessed tremendous technological developments in image-based biomarkers for tumor quantification and characterization. Initially limited to manual one- and two-dimensional size measurements, image biomarkers have evolved to harness developments not only in image acquisition technology but also in image processing and analysis algorithms. At the same time, clinical validation remains a major challenge for the vast majority of these novel techniques, and there is still a major gap between the latest technological developments and image biomarkers used in everyday clinical practice. Currently, the imaging biomarker field is attracting increasing attention not only because of the tremendous interest in cutting-edge therapeutic developments and personalized medicine but also because of the recent progress in the application of artificial intelligence (AI) algorithms to large-scale datasets. Thus, the goal of the present article is to review the current state of the art for image biomarkers and their use for characterization and predictive quantification of solid tumors. Beginning with an overview of validated imaging biomarkers in current clinical practice, we proceed to a review of AI-based methods for tumor characterization, such as radiomics-based approaches and deep learning.Key Points• Recent years have seen tremendous technological developments in image-based biomarkers for tumor quantification and characterization.• Image-based biomarkers can be used on an ongoing basis, in a non-invasive (or mildly invasive) way, to monitor the development and progression of the disease or its response to therapy.• We review the current state of the art for image biomarkers, as well as the recent developments in artificial intelligence (AI) algorithms for image processing and analysis.

Features from Computerized Texture Analysis of Breast Cancers at Pretreatment MR Imaging Are Associated with Response to Neoadjuvant Chemotherapy.

Purpose To evaluate whether features from texture analysis of breast cancers were associated with pathologic complete response (pCR) after neoadjuvant chemotherapy and to explore the association between texture features and tumor subtypes at pretreatment magnetic resonance (MR) imaging. Materials and Methods Institutional review board approval was obtained. This retrospective study included 85 patients with 85 breast cancers who underwent breast MR imaging before neoadjuvant chemotherapy between April 10, 2008, and March 12, 2015. Two-dimensional texture analysis was performed by using software at T2-weighted MR imaging and contrast material-enhanced T1-weighted MR imaging. Quantitative parameters were compared between patients with pCR and those with non-pCR and between patients with triple-negative breast cancer and those with non-triple-negative cancer. Multiple logistic regression analysis was used to determine independent parameters. Results Eighteen tumors (22%) were triple-negative breast cancers. pCR was achieved in 30 of the 85 tumors (35%). At univariate analysis, mean pixel intensity with spatial scaling factor (SSF) of 2 and 4 on T2-weighted images and kurtosis on contrast-enhanced T1-weighted images showed a significant difference between triple-negative breast cancer and non-triple-negative breast cancer (P = .009, .003, and .001, respectively). Kurtosis (SSF, 2) on T2-weighted images showed a significant difference between pCR and non-pCR (P = .015). At multiple logistic regression, kurtosis on T2-weighted images was independently associated with pCR in non-triple-negative breast cancer (P = .033). A multivariate model incorporating T2-weighted and contrast-enhanced T1-weighted kurtosis showed good performance for the identification of triple-negative breast cancer (area under the receiver operating characteristic curve, 0.834). Conclusion At pretreatment MR imaging, kurtosis appears to be associated with pCR to neoadjuvant chemotherapy in non-triple-negative breast cancer and may be a promising biomarker for the identification of triple-negative breast cancer. © RSNA, 2017.