RESUMO
RATIONALE AND OBJECTIVES: Large Language Models can capture the context of radiological reports, offering high accuracy in detecting unexpected findings. We aim to fine-tune a Robustly Optimized BERT Pretraining Approach (RoBERTa) model for the automatic detection of unexpected findings in radiology reports to assist radiologists in this relevant task. Second, we compared the performance of RoBERTa with classical convolutional neural network (CNN) and with GPT4 for this goal. MATERIALS AND METHODS: For this study, a dataset consisting of 44,631 radiological reports for training and 5293 for the initial test set was used. A smaller subset comprising 100 reports was utilized for the comparative test set. The complete dataset was obtained from our institution's Radiology Information System, including reports from various dates, examinations, genders, ages, etc. For the study's methodology, we evaluated two Large Language Models, specifically performing fine-tuning on RoBERTa and developing a prompt for ChatGPT. Furthermore, extending previous studies, we included a CNN in our comparison. RESULTS: The results indicate an accuracy of 86.15% in the initial test set using the RoBERTa model. Regarding the comparative test set, RoBERTa achieves an accuracy of 79%, ChatGPT 64%, and the CNN 49%. Notably, RoBERTa outperforms the other systems by 30% and 15%, respectively. CONCLUSION: Fine-tuned RoBERTa model can accurately detect unexpected findings in radiology reports outperforming the capability of CNN and ChatGPT for this task.
RESUMO
Nowadays, the use of advanced MRI sequences such as Diffusion-Weighted Imaging (DWI) or Perfusion-Weighted Imaging (PWI) in the field of musculoskeletal (MSK) radiology remains limited compared to other anatomical regions and subspecialties. Several reasons underpin this, primarily technical challenges, and a longstanding reliance on conventional and morphological evaluations of soft tissue and bone lesions. Experienced radiologists often assert that these advanced sequences don't offer added diagnostic value, claiming that a morphological approach suffices. However, in our opinion, the role of these advanced MRI sequences extends beyond merely naming an MSK lesion. In this commentary, we elucidate how these sequences can aid radiologists in various scenarios, from determining patient prognosis and tracking treatment progress to enhancing clinical-radiological correlations or guiding less experienced radiologists in evaluating soft tissues or bone tumors.
RESUMO
BACKGROUND AND OBJECTIVES: In the last decade, there has been a growing interest in applying artificial intelligence (AI) systems to breast cancer assessment, including breast density evaluation. However, few models have been developed to integrate textual mammographic reports and mammographic images. Our aims are (1) to generate a natural language processing (NLP)-based AI system, (2) to evaluate an external image-based software, and (3) to develop a multimodal system, using the late fusion approach, by integrating image and text inferences for the automatic classification of breast density according to the American College of Radiology (ACR) guidelines in mammograms and radiological reports. METHODS: We first compared different NLP models, three based on n-gram term frequency - inverse document frequency and two transformer-based architectures, using 1533 unstructured mammogram reports as a training set and 303 reports as a test set. Subsequently, we evaluated an external image-based software using 303 mammogram images. Finally, we assessed our multimodal system taking into account both text and mammogram images. RESULTS: Our best NLP model achieved 88 % accuracy, while the external software and the multimodal system achieved 75 % and 80 % accuracy, respectively, in classifying ACR breast densities. CONCLUSION: Although our multimodal system outperforms the image-based tool, it currently does not improve the results offered by the NLP model for ACR breast density classification. Nevertheless, the promising results observed here open the possibility to more comprehensive studies regarding the utilization of multimodal tools in the assessment of breast density.
RESUMO
This paper presents the implementation of two automated text classification systems for prostate cancer findings based on the PI-RADS criteria. Specifically, a traditional machine learning model using XGBoost and a language model-based approach using RoBERTa were employed. The study focused on Spanish-language radiological MRI prostate reports, which has not been explored before. The results demonstrate that the RoBERTa model outperforms the XGBoost model, although both achieve promising results. Furthermore, the best-performing system was integrated into the radiological company's information systems as an API, operating in a real-world environment.
RESUMO
Despite not being the first imaging modality for thyroid gland assessment, Magnetic Resonance Imaging (MRI), thanks to its optimal tissue contrast and spatial resolution, has provided some advancements in detecting and characterizing thyroid abnormalities. Recent research has been focused on improving MRI sequences and employing advanced techniques for a more comprehensive understanding of thyroid pathology. Although not yet standard practice, advanced MRI sequences have shown high accuracy in preliminary studies, correlating well with histopathological results. They particularly show promise in determining malignancy risk in thyroid lesions, which may reduce the need for invasive procedures like biopsies. In this line, functional MRI sequences like Diffusion Weighted Imaging (DWI), Dynamic Contrast-Enhanced MRI (DCE-MRI), and Arterial Spin Labeling (ASL) have demonstrated their potential usefulness in evaluating both diffuse thyroid conditions and focal lesions. Multicompartmental DWI models, such as Intravoxel Incoherent Motion (IVIM) and Diffusion Kurtosis Imaging (DKI), and novel methods like Amide Proton Transfer (APT) imaging or artificial intelligence (AI)-based analyses are being explored for their potential valuable insights into thyroid diseases. This manuscript reviews the critical physical principles and technical requirements for optimal functional MRI sequences of the thyroid and assesses the clinical utility of each technique. It also considers future prospects in the context of advanced MR thyroid imaging and analyzes the current role of advanced MRI sequences in routine practice.
Assuntos
Imageamento por Ressonância Magnética , Humanos , Imageamento por Ressonância Magnética/métodos , Neoplasias da Glândula Tireoide/diagnóstico por imagem , Doenças da Glândula Tireoide/diagnóstico por imagem , Meios de ContrasteRESUMO
Susceptibility-weighted imaging (SWI) has become a standard component of most brain MRI protocols. While traditionally used for detecting and characterising brain hemorrhages typically associated with stroke or trauma, SWI has also shown promising results in glioma assessment. Numerous studies have highlighted SWI's role in differentiating gliomas from other brain lesions, such as primary central nervous system lymphomas or metastases. Additionally, SWI aids radiologists in non-invasively grading gliomas and predicting their phenotypic profiles. Various researchers have suggested incorporating SWI as an adjunct sequence for predicting treatment response and for post-treatment monitoring. A significant focus of these studies is on the detection of intratumoural susceptibility signals (ITSSs) in gliomas, which are indicative of microhemorrhages and vessels within the tumour. The quantity, distribution, and characteristics of these ITSSs can provide radiologists with more precise information for evaluating and characterising gliomas. Furthermore, the potential benefits and added value of performing SWI after the administration of gadolinium-based contrast agents (GBCAs) have been explored. This review offers a comprehensive, educational, and practical overview of the potential applications and future directions of SWI in the context of glioma assessment. CLINICAL RELEVANCE STATEMENT: SWI has proven effective in evaluating gliomas, especially through assessing intratumoural susceptibility signal changes, and is becoming a promising, easily integrated tool in MRI protocols for both pre- and post-treatment assessments. KEY POINTS: ⢠Susceptibility-weighted imaging is the most sensitive sequence for detecting blood and calcium inside brain lesions. ⢠This sequence, acquired with and without gadolinium, helps with glioma diagnosis, characterisation, and grading through the detection of intratumoural susceptibility signals. ⢠There are ongoing challenges that must be faced to clarify the role of susceptibility-weighted imaging for glioma assessment.
RESUMO
OBJECTIVES: This study addresses the critical need for accurate summarization in radiology by comparing various Large Language Model (LLM)-based approaches for automatic summary generation. With the increasing volume of patient information, accurately and concisely conveying radiological findings becomes crucial for effective clinical decision-making. Minor inaccuracies in summaries can lead to significant consequences, highlighting the need for reliable automated summarization tools. METHODS: We employed two language models - Text-to-Text Transfer Transformer (T5) and Bidirectional and Auto-Regressive Transformers (BART) - in both fine-tuned and zero-shot learning scenarios and compared them with a Recurrent Neural Network (RNN). Additionally, we conducted a comparative analysis of 100 MRI report summaries, using expert human judgment and criteria such as coherence, relevance, fluency, and consistency, to evaluate the models against the original radiologist summaries. To facilitate this, we compiled a dataset of 15,508 retrospective knee Magnetic Resonance Imaging (MRI) reports from our Radiology Information System (RIS), focusing on the findings section to predict the radiologist's summary. RESULTS: The fine-tuned models outperform the neural network and show superior performance in the zero-shot variant. Specifically, the T5 model achieved a Rouge-L score of 0.638. Based on the radiologist readers' study, the summaries produced by this model were found to be very similar to those produced by a radiologist, with about 70% similarity in fluency and consistency between the T5-generated summaries and the original ones. CONCLUSIONS: Technological advances, especially in NLP and LLM, hold great promise for improving and streamlining the summarization of radiological findings, thus providing valuable assistance to radiologists in their work.
Assuntos
Estudos de Viabilidade , Imageamento por Ressonância Magnética , Processamento de Linguagem Natural , Redes Neurais de Computação , Humanos , Sistemas de Informação em Radiologia , Joelho/diagnóstico por imagem , Estudos RetrospectivosRESUMO
The integration of AI in radiology raises significant legal questions about responsibility for errors. Radiologists fear AI may introduce new legal challenges, despite its potential to enhance diagnostic accuracy. AI tools, even those approved by regulatory bodies like the FDA or CE, are not perfect, posing a risk of failure. The key issue is how AI is implemented: as a stand-alone diagnostic tool or as an aid to radiologists. The latter approach could reduce undesired side effects. However, it's unclear who should be held liable for AI failures, with potential candidates ranging from engineers and radiologists involved in AI development to companies and department heads who integrate these tools into clinical practice. The EU's AI Act, recognizing AI's risks, categorizes applications by risk level, with many radiology-related AI tools considered high risk. Legal precedents in autonomous vehicles offer some guidance on assigning responsibility. Yet, the existing legal challenges in radiology, such as diagnostic errors, persist. AI's potential to improve diagnostics raises questions about the legal implications of not using available AI tools. For instance, an AI tool improving the detection of pediatric fractures could reduce legal risks. This situation parallels innovations like car turn signals, where ignoring available safety enhancements could lead to legal problems. The debate underscores the need for further research and regulation to clarify AI's role in radiology, balancing innovation with legal and ethical considerations.
Assuntos
Inteligência Artificial , Responsabilidade Legal , Radiologia , Humanos , Radiologia/legislação & jurisprudência , Radiologia/ética , Inteligência Artificial/legislação & jurisprudência , Erros de Diagnóstico/legislação & jurisprudência , Erros de Diagnóstico/prevenção & controle , Radiologistas/legislação & jurisprudênciaRESUMO
Noninvasive differential diagnosis of brain tumors is currently based on the assessment of magnetic resonance imaging (MRI) coupled with dynamic susceptibility contrast (DSC). However, a definitive diagnosis often requires neurosurgical interventions that compromise patients' quality of life. We apply deep learning on DSC images from histology-confirmed patients with glioblastoma, metastasis, or lymphoma. The convolutional neural network trained on â¼50,000 voxels from 40 patients provides intratumor probability maps that yield clinical-grade diagnosis. Performance is tested in 400 additional cases and an external validation cohort of 128 patients. The tool reaches a three-way accuracy of 0.78, superior to the conventional MRI metrics cerebral blood volume (0.55) and percentage of signal recovery (0.59), showing high value as a support diagnostic tool. Our open-access software, Diagnosis In Susceptibility Contrast Enhancing Regions for Neuro-oncology (DISCERN), demonstrates its potential in aiding medical decisions for brain tumor diagnosis using standard-of-care MRI.
Assuntos
Neoplasias Encefálicas , Aprendizado Profundo , Humanos , Qualidade de Vida , Neoplasias Encefálicas/diagnóstico por imagem , Imageamento por Ressonância Magnética/métodos , PerfusãoRESUMO
Artificial Intelligence (AI) applied to radiology is so vast that it provides applications ranging from becoming a complete replacement for radiologists (a potential threat) to an efficient paperwork-saving time assistant (an evident strength). Nowadays, there are AI applications developed to facilitate the diagnostic process of radiologists without directly influencing (or replacing) the proper diagnostic decision step. These tools may help to reduce administrative workload, in different scenarios ranging from assisting in scheduling, study prioritization, or report communication, to helping with patient follow-up, including recommending additional exams. These are just a few of the highly time-consuming tasks that radiologists have to deal with every day in their routine workflow. These tasks hinder the time that radiologists should spend evaluating images and caring for patients, which will have a direct and negative impact on the quality of reports and patient attention, increasing the delay and waiting list of studies pending to be performed and reported. These types of AI applications should help to partially face this worldwide shortage of radiologists.
Assuntos
Inteligência Artificial , Radiologia , Humanos , Radiologia/métodos , Radiologistas , Fluxo de Trabalho , Carga de TrabalhoRESUMO
PURPOSE: The conclusion section of a radiology report is crucial for summarizing the primary radiological findings in natural language and essential for communicating results to clinicians. However, creating these summaries is time-consuming, repetitive, and prone to variability and errors among different radiologists. To address these issues, we evaluated a fine-tuned Text-To-Text Transfer Transformer (T5) model for abstractive summarization to automatically generate conclusions for neuroradiology MRI reports in a low-resource language. METHODS: We retrospectively applied our method to a dataset of 232,425 neuroradiology MRI reports in Spanish. We compared various pre-trained T5 models, including multilingual T5 and those newly adapted for Spanish. For precise evaluation, we employed BLEU, METEOR, ROUGE-L, CIDEr, and cosine similarity metrics alongside expert radiologist assessments. RESULTS: The findings are promising, with the models specifically fine-tuned for neuroradiology MRI achieving scores of 0.46, 0.28, 0.52, 2.45, and 0.87 in the BLEU-1, METEOR, ROUGE-L, CIDEr, and cosine similarity metrics, respectively. In the radiological experts' evaluation, they found that in 75% of the cases evaluated, the conclusions generated by the system were as good as or even better than the manually generated conclusions. CONCLUSION: The methods demonstrate the potential and effectiveness of customizing state-of-the-art pre-trained models for neuroradiology, yielding automatic MRI report conclusions that nearly match expert quality. Furthermore, these results underscore the importance of designing and pre-training a dedicated language model for radiology report summarization.
Assuntos
Processamento de Linguagem Natural , Radiologia , Humanos , Estudos Retrospectivos , Idioma , Imageamento por Ressonância MagnéticaRESUMO
Patients presenting with visual disturbances often require a neuroimaging approach. The spectrum of visual disturbances includes three main categories: vision impairment, ocular motility dysfunction, and abnormal pupillary response. Decreased vision is usually due to an eye abnormality. However, it can also be related to other disorders affecting the visual pathway, from the retina to the occipital lobe. Ocular motility dysfunction may follow disorders of the cranial nerves responsible for eye movements (ie, oculomotor, trochlear, and abducens nerves); may be due to any abnormality that directly affects the extraocular muscles, such as tumor or inflammation; or may result from any orbital disease that can alter the anatomy or function of these muscles, leading to diplopia and strabismus. Given that pupillary response depends on the normal function of the sympathetic and parasympathetic pathways, an abnormality affecting these neuronal systems manifests, respectively, as pupillary miosis or mydriasis, with other related symptoms. In some cases, neuroimaging studies must complement the clinical ophthalmologic examination to better assess the anatomic and pathologic conditions that could explain the symptoms. US has a major role in the assessment of diseases of the eye and anterior orbit. CT is usually the first-line imaging modality because of its attainability, especially in trauma settings. MRI offers further information for inflammatory and tumoral cases. ©RSNA, 2024 Test Your Knowledge questions for this article are available in the supplemental material.
Assuntos
Músculos Oculomotores , Transtornos da Visão , Humanos , Transtornos da Visão/diagnóstico por imagem , Músculos Oculomotores/inervação , Músculos Oculomotores/patologia , Órbita , Imageamento por Ressonância MagnéticaRESUMO
OBJECTIVES: The differential between high-grade glioma (HGG) and metastasis remains challenging in common radiological practice. We compare different natural language processing (NLP)-based deep learning models to assist radiologists based on data contained in radiology reports. METHODS: This retrospective study included 185 MRI reports between 2010 and 2022 from two different institutions. A total of 117 reports were used for the training and 21 were reserved for the validation set, while the rest were used as a test set. A comparison of the performance of different deep learning models for HGG and metastasis classification has been carried out. Specifically, Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), a hybrid version of BiLSTM and CNN, and a radiology-specific Bidirectional Encoder Representations from Transformers (RadBERT) model were used. RESULTS: For the classification of MRI reports, the CNN network provided the best results among all tested, showing a macro-avg precision of 87.32%, a sensitivity of 87.45%, and an F1 score of 87.23%. In addition, our NLP algorithm detected keywords such as tumor, temporal, and lobe to positively classify a radiological report as HGG or metastasis group. CONCLUSIONS: A deep learning model based on CNN enables radiologists to discriminate between HGG and metastasis based on MRI reports with high-precision values. This approach should be considered an additional tool in diagnosing these central nervous system lesions. CLINICAL RELEVANCE STATEMENT: The use of our NLP model enables radiologists to differentiate between patients with high-grade glioma and metastasis based on their MRI reports and can be used as an additional tool to the conventional image-based approach for this challenging task. KEY POINTS: ⢠Differential between high-grade glioma and metastasis is still challenging in common radiological practice. ⢠Natural language processing (NLP)-based deep learning models can assist radiologists based on data contained in radiology reports. ⢠We have developed and tested a natural language processing model for discriminating between high-grade glioma and metastasis based on MRI reports that show high precision for this task.
Assuntos
Aprendizado Profundo , Glioma , Humanos , Processamento de Linguagem Natural , Estudos Retrospectivos , Glioma/diagnóstico por imagem , Redes Neurais de ComputaçãoRESUMO
MRI evaluation of the diabetic foot is still a challenge not only from an interpretative but also from a technical point of view. The incorporation of advanced sequences such as diffusion-weighted imaging (DWI) and dynamic contrast-enhanced (DCE) MRI into standard protocols for diabetic foot assessment could aid radiologists in differentiating between neuropathic osteoarthropathy (Charcot's foot) and osteomyelitis. This distinction is crucial as both conditions can coexist in diabetic patients, and they require markedly different clinical management and have distinct prognoses. Over the past decade, several studies have explored the effectiveness of DWI and dynamic contrast-enhanced MRI (DCE-MRI) in distinguishing between septic and reactive bone marrow, as well as soft tissue involvement in diabetic patients, yielding promising results. DWI, without the need for exogenous contrast, can provide insights into the cellularity of bone marrow and soft tissues. DCE-MRI allows for a more precise evaluation of soft tissue and bone marrow perfusion compared to conventional post-gadolinium imaging. The data obtained from these sequences will complement the traditional MRI approach in assessing the diabetic foot. The objective of this review is to familiarize readers with the fundamental concepts of DWI and DCE-MRI, including technical adjustments and practical tips for image interpretation in diabetic foot cases.
RESUMO
OBJECTIVES: Adult solitary intra-axial cerebellar tumors are uncommon. Their presurgical differentiation based on neuroimaging is crucial, since management differs substantially. Comprehensive full assessment of MR dynamic-susceptibility-contrast perfusion-weighted imaging (DSC-PWI) may reveal key differences between entities. This study aims to provide new insights on perfusion patterns of these tumors and to explore the potential of DSC-PWI in their presurgical discrimination. METHODS: Adult patients with a solitary cerebellar tumor on presurgical MR and confirmed histological diagnosis of metastasis, medulloblastoma, hemangioblastoma, or pilocytic astrocytoma were retrospectively retrieved (2008-2023). Volumetric segmentation of tumors and normal-appearing white matter (for normalization) was semi-automatically performed on CE-T1WI and coregistered with DSC-PWI. Mean normalized values per patient tumor-mask of relative cerebral blood volume (rCBV), percentage of signal recovery (PSR), peak height (PH), and normalized time-intensity curves (nTIC) were extracted. Statistical comparisons were done. Then, the dataset was split into training (75%) and test (25%) cohorts and a classifier was created considering nTIC, rCBV, PSR, and PH in the model. RESULTS: Sixty-eight patients (31 metastases, 13 medulloblastomas, 13 hemangioblastomas, and 11 pilocytic astrocytomas) were included. Relevant differences between tumor types' nTICs were demonstrated. Hemangioblastoma showed the highest rCBV and PH, pilocytic astrocytoma the highest PSR. All parameters showed significant differences on the Kruskal-Wallis tests (p < 0.001). The classifier yielded an accuracy of 98% (47/48) in the training and 85% (17/20) in the test sets. CONCLUSIONS: Intra-axial cerebellar tumors in adults have singular and significantly different DSC-PWI signatures. The combination of perfusion metrics through data-analysis rendered excellent accuracies in discriminating these entities. CLINICAL RELEVANCE STATEMENT: In this study, the authors constructed a classifier for the non-invasive imaging presurgical diagnosis of adult intra-axial cerebellar tumors. The resultant tool can be a support for decision-making in the clinical practice and enables optimal personalized patient management. KEY POINTS: ⢠Adult intra-axial cerebellar tumors exhibit specific, singular, and statistically significant different MR dynamic-susceptibility-contrast perfusion-weighted imaging (DSC-PWI) signatures. ⢠Data-analysis, applied to MR DSC-PWI, could provide added value in the presurgical diagnosis of solitary cerebellar metastasis, medulloblastoma, hemangioblastoma, and pilocytic astrocytoma. ⢠A classifier based on DSC-PWI metrics yields excellent accuracy rates and could be used as a support tool for radiologic diagnosis with clinician-friendly displays.