Generative Artificial Intelligence Biases, Limitations and Risks in Nuclear Medicine: An Argument for Appropriate Use Framework and Recommendations.

Generative artificial intelligence (AI) algorithms for both text-to-text and text-to-image applications have seen rapid and widespread adoption in the general and medical communities. While limitations of generative AI have been widely reported, there remain valuable applications in patient and professional communities. Here, the limitations and biases of both text-to-text and text-to-image generative AI are explored using purported applications in medical imaging as case examples. A direct comparison of the capabilities of four common text-to-image generative AI algorithms is reported and recommendations for the most appropriate use, DALL-E 3, justified. The risks use and biases are outlined, and appropriate use guidelines framed for use of generative AI in nuclear medicine. Generative AI text-to-text and text-to-image generation includes inherent biases, particularly gender and ethnicity, that could misrepresent nuclear medicine. The assimilation of generative AI tools into medical education, image interpretation, patient education, health promotion and marketing in nuclear medicine risks propagating errors and amplification of biases. Mitigation strategies should reside inside appropriate use criteria and minimum standards for quality and professionalism for the application of generative AI in nuclear medicine.

Social Determinants of Health Framework to Identify and Reduce Barriers to Imaging in Marginalized Communities.

Social determinants of health (SDOH) are conditions influencing individuals' health based on their environment of birth, living, working, and aging. Addressing SDOH is crucial for promoting health equity and reducing health outcome disparities. For conditions such as stroke and cancer screening where imaging is central to diagnosis and management, access to high-quality medical imaging is necessary. This article applies a previously described structural framework characterizing the impact of SDOH on patients who require imaging for their clinical indications. SDOH factors can be broadly categorized into five sectors: economic stability, education access and quality, neighborhood and built environment, social and community context, and health care access and quality. As patients navigate the health care system, they experience barriers at each step, which are significantly influenced by SDOH factors. Marginalized communities are prone to disparities due to the inability to complete the required diagnostic or screening imaging work-up. This article highlights SDOH that disproportionately affect marginalized communities, using stroke and cancer as examples of disease processes where imaging is needed for care. Potential strategies to mitigate these disparities include dedicating resources for clinical care coordinators, transportation, language assistance, and financial hardship subsidies. Last, various national and international health initiatives are tackling SDOH and fostering health equity.

Impact of the COVID-19 Pandemic and Vaccine Availability on Utilization of Breast Imaging in a Multistate Radiology Practice.

Method: Data were obtained from medical health records across 77 Radiology Partners practices in the US. The data provided us with the total monthly mammography, breast ultrasound, and breast MRI procedures from January 2019 to September 2022. An interrupted time-series (ITS) analysis was conducted to evaluate the effect of the COVID-19 pandemic and the COVID-19 vaccination. We chose March 2020 and December 2020 as critical time points in the pandemic and analyzed trends before and after these dates. Results: The starting level (at baseline in January 2019) of the total breast imaging procedure volume was estimated at 114,901.5, and this volume appeared to significantly increase every month prior to March 2020 by 4,864.0 (p < 0.0001, CI = [3,077.1, 6,650.9]). In March 2020, there appeared to be a significant decrease in volume by 104,446.3 (p=0.003, CI = [-172,063.1, -36,829.5]), followed by a significant increase in the monthly trend of service volume (relative to the pre-COVID trend) of 20,660.7 per month (p=0.001, CI = [8,828.5, 32,493.0]). In December 2020, there appeared to be a significant decrease in service volume by 69,791.2 (p=0.012, CI = [-123,602.6, -15,979.7]). Compared to the period from March to November 2020, there was a decrease in the monthly trend of service volumes per month by 24,213.9 (p < 0.0001, CI = [-36,027.6, -12,400.2]). After March 2020, the total service volume increased at the rate of 25,524.7 per month (p < 0.0001, CI = [13,828.2, 37,221.2]). In contrast, the service volumes after December 2020 appeared to grow steadily and slowly at a rate of 1,310.8 per month (p=0.118, CI = [-348.8, 2970.3]). Conclusion: Our study revealed that there has been a recovery and a further increase in breast imaging service volumes compared to prepandemic levels. The increase can be best explained by vaccination rollout, reopening of elective/nonemergency healthcare services, insurance coverage expansion, the decline in the US uninsured rate due to government interventions and policies, and the recovery of jobs with employer-provided medical insurance post-pandemic.

Radiation Dosimetry, Artificial Intelligence and Digital Twins: Old Dog, New Tricks.

Developments in artificial intelligence, particularly convolutional neural networks and deep learning, have the potential for problem solving that has previously confounded human intelligence. Accurate prediction of radiation dosimetry pre-treatment with scope to adjust dosing for optimal target and non-target tissue doses is consistent with striving for improved the outcomes of precision medicine. The combination of artificial intelligence and production of digital twins could provide an avenue for an individualised therapy doses and enhanced outcomes in theranostics. While there are barriers to overcome, the maturity of individual technologies (i.e. radiation dosimetry, artificial intelligence, theranostics and digital twins) places these approaches within reach.

The transformational potential of molecular radiomics.

Conventional radiomics in nuclear medicine involve hand-crafted and computer-assisted regions of interest. Recent developments in artificial intelligence (AI) have seen the emergence of AI-augmented segmentation and extraction of lower order traditional radiomic features. Deep learning (DL) affords the opportunity to extract abstract radiomic features directly from input tensors (images) without the need for segmentation. These fourth-order, high dimensional radiomics produce deep radiomics and are well suited to the data density associated with the molecular nature of hybrid imaging. Molecular radiomics and deep molecular radiomics provide insights beyond images and quantitation typical of semantic reporting. While the application of molecular radiomics using hand-crafted and computer-generated features is integrated into decision-making in nuclear medicine, the acceptance of deep molecular radiomics is less universal. This manuscript aims to provide an understanding of the language and principles associated with radiomics and deep radiomics in nuclear medicine.

Intelligent imaging: Applications of machine learning and deep learning in radiology.

Artificial intelligence (AI) in radiology is transforming medical image analysis. While applications in triaging for priority reporting and radiomic feature analysis have been widely reported, perhaps the most important applications lie in noise reduction, image optimization following dose reduction strategies, image reconstruction direct from projection data and generation of pseudo-CT for attenuation correction. There are common beneficial applications, and potential risks, between human radiology and veterinary radiology. Artificial intelligence may see recrafting of some responsibilities but offers AI augmentation of human driven systems. The redundancy afforded by human augmentation of AI and AI autonomy are not on the horizon, but rather are already here.

The deep radiomic analytics pipeline.

Radiomics refers to the process of extracting useful imaging features from radiological data. Conventional radiomics like standard uptake value, intensity histograms, or phase images involve hand-crafted (manual) or automated regions of interest (computer generated), however, artificial intelligence (AI) segmentation (AI-augmented radiomics) has recently emerged. Radiomic feature extraction extends image insights beyond simply data quantitation and provides additional insights to aid semantic reporting. Deeper layers of a convolutional neural network produce more abstract radiomic features that are referred to as deep radiomics. The application of radiomics in veterinary radiology is already firmly entrenched using hand-crafted and automated computer-generated radiomic features in X-ray, nuclear medicine, CT, ultrasound, and MRI. There is an opportunity for veterinary radiology to capitalize on advances in AI, machine learning, and deep learning to enrich imaging interpretation using deep radiomic feature extraction. This manuscript aims to provide a general understanding of radiomics and deep radiomics, and to arm readers with the vernacular to progress discussion and development of deep radiomics in veterinary imaging.

An Unusual Case of Breast Implant-Associated Anaplastic Large Cell Lymphoma.

Introduction: Breast implant-associated anaplastic large cell lymphoma (BIA-ALCL) is a rare disease entity associated with textured breast implants. Though the clinical course is typically indolent, BIA-ALCL can occasionally invade through the capsule into the breast parenchyma with spread to the regional lymph nodes and beyond including chest wall invasive disease. Case: We present the case of a 51-year-old female with a history of bilateral silicone breast implants placed approximately twenty years ago who presented with two months of progressively enlarging right breast mass. Ultrasound-guided biopsy of right breast mass and right axillary lymph node showed CD 30-positive ALK-negative anaplastic large cell lymphoma, and staging work up showed extension of the tumor to chest wall and ribs consistent with advanced disease. She received CHP-BV (cyclophosphamide, doxorubicin, prednisone, and brentuximab vedotin) for six cycles with complete metabolic response. This was followed by extensive surgical extirpation and reconstruction, radiation for residual disease and consolidation with autologous stem cell transplant. She is currently on maintenance brentuximab vedotin with no evidence of active disease post autologous stem cell transplant. Conclusion: Treatment guidelines for advanced chest wall invasive BIA-ALCL are not well defined. Lack of predictive factors warrants mutation analysis and genetic sequencing to identify those at highest risk of progression to chest wall invasive disease. This rare case highlights the need for definitive consensus on the optimal management of chest wall invasive BIA-ALCL.

Social Asymmetry, Artificial Intelligence and the Medical Imaging Landscape.

Social and health care equity and justice should be prioritized by the mantra of medicine, first do no harm. Despite highly motivated national and global health strategies, there remains significant health care inequity. Intrinsic and extrinsic factors, including a number of biases, are key drivers of ongoing health inequity including equity of access and opportunity for nuclear medicine and radiology services. There is a substantial gap in the global practice of nuclear medicine in particular, but also radiology, between developed health economies and those considered developing or undeveloped. At a local level, even in developed health economies, there can be a significant disparity between health services, including medical imaging, between communities based on socioeconomic, cultural or geographic differences. Artificial intelligence (AI) has the potential to either widen the health inequity divide or substantially reduce it. Distributed generally, AI technology could be used to overcome geographic boundaries to health care, thus bringing general and specialist care into underserved communities. However, should AI technology be limited to localities already enjoying ample healthcare access and direct access to health infrastructure, like radiology and nuclear medicine, it could then accentuate the gap. There are a number of challenges across the AI pipeline that need careful attention to ensure beneficence over maleficence. Fully realized, AI augmented health care could be crafted as an integral part of the broader strategy convergence on local, national and global health equity. The applications of AI in nuclear medicine and radiology could emerge as a powerful tool in social and health equity.

DOTATATE Uptake in an Axillary Lymph Node After COVID-19 Vaccination.

ABSTRACT: A 58-year-old man underwent DOTATATE PET/CT scan for follow-up of pulmonary neuroendocrine tumor after resection and adjuvant chemotherapy. On screening paperwork, the patient indicated having received the Johnson & Johnson/Janssen COVID-19 vaccine (Janssen Biotech, Inc) 1 day previously, administered in the right deltoid muscle. Reactive changes in regional lymph nodes is a known response for all 3 currently Food and Drug Administration-approved COVID-19 vaccines. Recent published data have demonstrated FDG PET-avid axillary lymphadenopathy subsequent to COVID-19 vaccination, and included here is a report of DOTATATE PET-avid axillary lymph node after injection of the Johnson & Johnson COVID-19 vaccine.

Delayed FDG PET Provides Superior Glioblastoma Conspicuity Compared to Conventional Image Timing.

Background: Glioblastomas are malignant, often incurable brain tumors. Reliable discrimination between recurrent disease and treatment changes is a significant challenge. Prior work has suggested glioblastoma FDG PET conspicuity is improved at delayed time points vs. conventional imaging times. This study aimed to determine the ideal FDG imaging time point in a population of untreated glioblastomas in preparation for future trials involving the non-invasive assessment of true progression vs. pseudoprogression in glioblastoma. Methods: Sixteen pre-treatment adults with suspected glioblastoma received FDG PET at 1, 5, and 8 h post-FDG injection within the 3 days prior to surgery. Maximum standard uptake values were measured at each timepoint for the central enhancing component of the lesion and the contralateral normal-appearing brain. Results: Sixteen patients (nine male) had pathology confirmed IDH-wildtype, glioblastoma. Our results revealed statistically significant improvements in the maximum standardized uptake values and subjective conspicuity of glioblastomas at later time points compared to the conventional (1 h time point). The tumor to background ratio at 1, 5, and 8 h was 1.4 ± 0.4, 1.8 ± 0.5, and 2.1 ± 0.6, respectively. This was statistically significant for the 5 h time point over the 1 h time point (p > 0.001), the 8 h time point over the 1 h time point (p = 0.026), and the 8 h time point over the 5 h time point (p = 0.036). Conclusions: Our findings demonstrate that delayed imaging time point provides superior conspicuity of glioblastoma compared to conventional imaging. Further research based on these results may translate into improvements in the determination of true progression from pseudoprogression.

Pitfalls in Interpretation of PET/CT in the Chest.

Whole body positron emission tomography (PET)/computed tomography (CT) imaging with [18F]-fluoro-2-deoxy-D-glucose (FDG) is widely used in oncologic imaging. In the chest, common PET/CT applications include the evaluation of solitary pulmonary nodules, cancer staging, assessment of response to therapy, and detection of residual or recurrent disease. Knowledge of the technical artifacts and potential pitfalls that radiologists may encounter in the interpretation of PET/CT in the thorax is important to avoid misinterpretation and optimize patient management. This article will review pitfalls in the interpretation of PET/CT in the chest related to technical factors, physiologic uptake, false positive findings, false negative findings, and iatrogenic conditions.

Radioiodine Imaging and Treatment in Thyroid Disorders.

Thyroid hormones T3 and T4 are crucial for development and differentiation of various cells in the body. They are also essential for regulating metabolism in nearly all tissues. Iodine is an integral element in the synthesis of thyroid hormone and is actively transported into the thyroid by a Na+/I- symporter. The thyroid can take up radioactive iodine just like it would take iodine and hence can be used to evaluate and treat several thyroid diseases. Radioactive iodine is one of the first radioisotopes to be used in medicine.

Intelligent Imaging in Nuclear Medicine: the Principles of Artificial Intelligence, Machine Learning and Deep Learning.

The emergence of artificial intelligence (AI) in nuclear medicine has occurred over the last 50 years but more recent developments in machine learning (ML) and deep learning (DL) have driven new capabilities of AI in nuclear medicine. In nuclear medicine, the artificial neural network (ANN) is the backbone of ML and DL. The inputs may be radiomic features that have been extracted from the image files or, if using a convolutional neural network (CNN), may be the images themselves. AI in nuclear medicine re-engineers and re-imagines clinical and research capabilities. An understanding of the principles of AI, ML and DL contextualised to nuclear medicine allows richer engagement in clinical and research applications, and capacity for problem solving where required. Simple applications of ML include quality assurance, risk assessment, business analytics and rudimentary classifications. More complex applications of DL for detection, localisation, classification, segmentation, quantitation and radiomic feature extraction using CNNs can be applied to general nuclear medicine, SPECT, PET, CT and MRI. There are also applications of ANNs and ML that allow small datasets (and larger ones) to be analysed in parallel to conventional statistical analysis. AI has assimilated into the clinical and research practice of nuclear medicine with little disruption. The emergence of ML and DL applications, however, has produced a seismic significant shift in the clinical and research landscape that demands at least rudimentary understanding of the principles of AI, ANNs and CNNs among nuclear medicine professionals.

ACR-ACNM-ASTRO-SNMMI Practice Parameter for the Performance of Therapy With Radium-223.

AIM/OBJECTIVES/BACKGROUND: The goal of therapy with unsealed radiopharmaceutical sources is to provide either cure or significant prolongation of disease-specific survival, and effective reduction and/or prevention of adverse disease-related symptoms or untoward events while minimizing treatment-associated side effects and complications. Radium-223 dichloride (radium-223) is an alpha particle-emitting isotope used for targeted bone therapy. This practice parameter is intended to guide appropriately trained and licensed physicians performing therapy with radium-223. Such therapy requires close cooperation and communication between the physicians who are responsible for the clinical management of the patient and those who administer radiopharmaceutical therapy and manage the attendant side effects. Adherence to this parameter should help to maximize the efficacious use of radium-223, maintain safe conditions, and ensure compliance with applicable regulations. METHODS: This practice parameter was developed according to the process described on the American College of Radiology (ACR) website ("The Process for Developing ACR Practice Parameters and Technical Standards," www.acr.org/ClinicalResources/Practice-Parameters-and-Technical-Standards) by the Committee on Practice Parameters of the ACR Commission on Radiation Oncology in collaboration with the American College of Nuclear Medicine (ACNM), the American Society for Radiation Oncology (ASTRO), and the Society of Nuclear Medicine and Molecular Imaging (SNMMI). All these societies contributed to the development of the practice parameter and approved the final document. RESULTS: This practice parameter addresses the many factors which contribute to appropriate, safe, and effective clinical use of radium-223. Topics addressed include qualifications and responsibilities of personnel, specifications of patient examination and treatment; documentation, radiation safety, quality control/improvement, infection control, and patient education. CONCLUSIONS: This practice parameter is intended as a tool to guide clinical use of radium-223 with the goal of facilitating safe and effective medical care based on current knowledge, available resources and patient needs. The sole purpose of this document is to assist practitioners in achieving this objective.

Low-dose versus High-dose Carfilzomib with Dexamethasone (S1304) in Patients with Relapsed-Refractory Multiple Myeloma.

PURPOSE: Treatment of multiple myeloma has evolved tremendously and optimal utilization of available therapies will ensure maximal patient benefits. PATIENTS AND METHODS: We report the Southwest Oncology Group randomized phase II trial (S1304) comparing twice weekly low-dose (27 mg/m2; arm 1) to high-dose carfilzomib (56 mg/m2; arm 2), both with dexamethasone, administered for 12 cycles (11 months) for relapsed and/or refractory multiple myeloma with up to six prior lines of therapy (NCT01903811). The primary endpoint was progression-free survival (PFS), and patients on arm 1 could cross-over to arm 2 after progression on treatment. RESULTS: Among 143 enrolled patients, of whom 121 were eligible and analyzable, the overall response rate was 42.8%, with no significant difference between the arms (P = 0.113). Also, neither the median PFS [5 months and 8 months, respectively; HR, 1.061; 80% Wald confidence interval (CI), 0.821-1.370; P = 0.384] nor the median overall survival were significantly different (26 and 22 months, respectively; HR, 1.149, 80% Wald CI, 0.841-.571; P = 0.284). Sixteen patients crossed over to arm 2 with a median PFS benefit of 3 months. Certain adverse events (AE) were more frequent in arm 2, including fatigue, thrombocytopenia, and peripheral neuropathy, but there was no significant difference in cardiopulmonary AEs. CONCLUSIONS: This randomized trial did not support a benefit of fixed duration, twice weekly 56 mg/m2 dosing of carfilzomib over the 27 mg/m2 dose for the treatment of relapsed and/or refractory multiple myeloma. However, treatment to progression in earlier patient populations with high-dose carfilzomib using different schedules should still be considered as part of the standard of care.

Machine Learning and Deep Learning in Medical Imaging: Intelligent Imaging.

Artificial intelligence (AI) in medical imaging is a potentially disruptive technology. An understanding of the principles and application of radiomics, artificial neural networks, machine learning, and deep learning is an essential foundation to weave design solutions that accommodate ethical and regulatory requirements, and to craft AI-based algorithms that enhance outcomes, quality, and efficiency. Moreover, a more holistic perspective of applications, opportunities, and challenges from a programmatic perspective contributes to ethical and sustainable implementation of AI solutions.