RESUMO
OBJECTIVES: To investigate the use of the score-based diffusion model to accelerate breast MRI reconstruction. MATERIALS AND METHODS: We trained a score-based model on 9549 MRI examinations of the female breast and employed it to reconstruct undersampled MRI images with undersampling factors of 2, 5, and 20. Images were evaluated by two experienced radiologists who rated the images based on their overall quality and diagnostic value on an independent test set of 100 additional MRI examinations. RESULTS: The score-based model produces MRI images of high quality and diagnostic value. Both T1- and T2-weighted MRI images could be reconstructed to a high degree of accuracy. Two radiologists rated the images as almost indistinguishable from the original images (rating 4 or 5 on a scale of 5) in 100% (radiologist 1) and 99% (radiologist 2) of cases when the acceleration factor was 2. This fraction dropped to 88% and 70% for an acceleration factor of 5 and to 5% and 21% with an extreme acceleration factor of 20. CONCLUSION: Score-based models can reconstruct MRI images at high fidelity, even at comparatively high acceleration factors, but further work on a larger scale of images is needed to ensure that diagnostic quality holds. CLINICAL RELEVANCE STATEMENT: The number of MRI examinations of the breast is expected to rise with MRI screening recommended for women with dense breasts. Accelerated image acquisition methods can help in making this examination more accessible. KEY POINTS: Accelerating breast MRI reconstruction remains a significant challenge in clinical settings. Score-based diffusion models can achieve near-perfect reconstruction for moderate undersampling factors. Faster breast MRI scans with maintained image quality could revolutionize clinic workflows and patient experience.
RESUMO
Background The level of background parenchymal enhancement (BPE) at breast MRI provides predictive and prognostic information and can have diagnostic implications. However, there is a lack of standardization regarding BPE assessment. Purpose To investigate how well results of quantitative BPE assessment methods correlate among themselves and with assessments made by radiologists experienced in breast MRI. Materials and Methods In this pseudoprospective analysis of 5773 breast MRI examinations from 3207 patients (mean age, 60 years ± 10 [SD]), the level of BPE was prospectively categorized according to the Breast Imaging Reporting and Data System by radiologists experienced in breast MRI. For automated extraction of BPE, fibroglandular tissue (FGT) was segmented in an automated pipeline. Four different published methods for automated quantitative BPE extractions were used: two methods (A and B) based on enhancement intensity and two methods (C and D) based on the volume of enhanced FGT. The results from all methods were correlated, and agreement was investigated in comparison with the respective radiologist-based categorization. For surrogate validation of BPE assessment, how accurately the methods distinguished premenopausal women with (n = 50) versus without (n = 896) antihormonal treatment was determined. Results Intensity-based methods (A and B) exhibited a correlation with radiologist-based categorization of 0.56 ± 0.01 and 0.55 ± 0.01, respectively, and volume-based methods (C and D) had a correlation of 0.52 ± 0.01 and 0.50 ± 0.01 (P < .001). There were notable correlation differences (P < .001) between the BPE determined with the four methods. Among the four quantitation methods, method D offered the highest accuracy for distinguishing women with versus without antihormonal therapy (P = .01). Conclusion Results of different methods for quantitative BPE assessment agree only moderately among themselves or with visual categories reported by experienced radiologists; intensity-based methods correlate more closely with radiologists' ratings than volume-based methods. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Mann in this issue.
Assuntos
Neoplasias da Mama , Mama , Imageamento por Ressonância Magnética , Humanos , Feminino , Pessoa de Meia-Idade , Imageamento por Ressonância Magnética/métodos , Neoplasias da Mama/diagnóstico por imagem , Mama/diagnóstico por imagem , Interpretação de Imagem Assistida por Computador/métodos , Adulto , Estudos Prospectivos , Aumento da Imagem/métodos , Idoso , Reprodutibilidade dos Testes , Estudos RetrospectivosRESUMO
OBJECTIVES: Clinical experience regarding the use of dedicated photon-counting breast CT (PC-BCT) for diagnosis of breast microcalcifications is scarce. This study systematically compares the detection and classification of breast microcalcifications using a dedicated breast photon-counting CT, especially designed for examining the breast, in comparison with digital breast tomosynthesis (DBT). MATERIALS AND METHODS: This is a prospective intraindividual study on women with DBT screening-detected BI-RADS-4/-5 microcalcifications who underwent PC-BCT before biopsy. PC-BCT images were reconstructed with a noninterpolated spatial resolution of 0.15 × 0.15 × 0.15 mm (reconstruction mode 1 [RM-1]) and with 0.3 × 0.3 × 0.3 mm (reconstruction mode 2 [RM-2]), plus thin-slab maximum intensity projection (MIP) reconstructions. Two radiologists independently rated the detection of microcalcifications in direct comparison with DBT on a 5-point scale. The distribution and morphology of microcalcifications were then rated according to BI-RADS. The size of the smallest discernible microcalcification particle was measured. For PC-BCT, the average glandular dose was determined by Monte Carlo simulations; for DBT, the information provided by the DBT system was used. RESULTS: Between September 2022 and July 2023, 22 participants (mean age, 61; range, 42-85 years) with microcalcifications (16 malignant; 6 benign) were included. In 2/22 with microcalcifications in the posterior region, microcalcifications were not detectable on PC-BCT, likely because they were not included in the PC-BCT volume. In the remaining 20 participants, microcalcifications were detectable. With high between-reader agreement (κ > 0.8), conspicuity of microcalcifications was rated similar for DBT and MIPs of RM-1 (mean, 4.83 ± 0.38 vs 4.86 ± 0.35) (P = 0.66), but was significantly lower (P < 0.05) for the remaining PC-BCT reconstructions: 2.11 ± 0.92 (RM-2), 2.64 ± 0.80 (MIPs of RM-2), and 3.50 ± 1.23 (RM-1). Identical distribution qualifiers were assigned for PC-BCT and DBT in 18/20 participants, with excellent agreement (κ = 0.91), whereas identical morphologic qualifiers were assigned in only 5/20, with poor agreement (κ = 0.44). The median size of smallest discernible microcalcification particle was 0.2 versus 0.6 versus 1.1 mm in DBT versus RM-1 versus RM-2 (P < 0.001), likely due to blooming effects. Average glandular dose was 7.04 mGy (PC-BCT) versus 6.88 mGy (DBT) (P = 0.67). CONCLUSIONS: PC-BCT allows reliable detection of in-breast microcalcifications as long as they are not located in the posterior part of the breast and allows assessment of their distribution, but not of their individual morphology.
RESUMO
BACKGROUND: Quantitative techniques such as T2 and T1ρ mapping allow evaluating the cartilage and meniscus. We evaluated multi-interleaved X-prepared turbo-spin echo with intuitive relaxometry (MIXTURE) sequences with turbo spin-echo (TSE) contrast and additional parameter maps versus reference TSE sequences in an in situ model of human cartilage defects. METHODS: Standardized cartilage defects of 8, 5, and 3 mm in diameter were created in the lateral femora of ten human cadaveric knee specimens (81 ± 10 years old; nine males, one female). MIXTURE sequences providing proton density-weighted fat-saturated images and T2 maps or T1-weighted images and T1ρ maps as well as the corresponding two- and three-dimensional TSE reference sequences were acquired before and after defect creation (3-T scanner; knee coil). Defect delineability, bone texture, and cartilage relaxation times were quantified. Appropriate parametric or non-parametric tests were used. RESULTS: Overall, defect delineability and texture features were not significantly different between the MIXTURE and reference sequences (p ≤ 0.47). After defect creation, relaxation times significantly increased in the central femur (T2pre = 51 ± 4 ms [mean ± standard deviation] versus T2post = 56 ± 4 ms; p = 0.002) and all regions combined (T1ρpre = 40 ± 4 ms versus T1ρpost = 43 ± 4 ms; p = 0.004). CONCLUSIONS: MIXTURE permitted time-efficient simultaneous morphologic and quantitative joint assessment based on clinical image contrasts. While providing T2 or T1ρ maps in clinically feasible scan time, morphologic image features, i.e., cartilage defects and bone texture, were comparable between MIXTURE and reference sequences. RELEVANCE STATEMENT: Equally time-efficient and versatile, the MIXTURE sequence platform combines morphologic imaging using familiar contrasts, excellent image correspondence versus corresponding reference sequences and quantitative mapping information, thereby increasing the diagnostic value beyond mere morphology. KEY POINTS: ⢠Combined morphologic and quantitative MIXTURE sequences are based on three-dimensional TSE contrasts. ⢠MIXTURE sequences were studied in an in situ human cartilage defect model. ⢠Morphologic image features, i.e., defect delineabilty and bone texture, were investigated. ⢠Morphologic image features were similar between MIXTURE and reference sequences. ⢠MIXTURE allowed time-efficient simultaneous morphologic and quantitative knee joint assessment.
Assuntos
Cadáver , Cartilagem Articular , Articulação do Joelho , Imageamento por Ressonância Magnética , Humanos , Masculino , Imageamento por Ressonância Magnética/métodos , Feminino , Cartilagem Articular/diagnóstico por imagem , Articulação do Joelho/diagnóstico por imagem , Idoso de 80 Anos ou mais , IdosoRESUMO
BACKGROUND: Cardiovascular CT is required for planning transcatheter aortic valve implantation (TAVI). PURPOSE: To compare image quality, suitability for TAVI planning, and radiation dose of photon-counting CT (PCCT) with that of dual-source CT (DSCT). MATERIAL AND METHODS: Retrospective study on consecutive TAVI candidates with aortic valve stenosis who underwent contrast-enhanced aorto-ilio-femoral PCCT and/or DSCT between 01/2022 and 07/2023. Signal-to-noise (SNR) and contrast-to-noise ratio (CNR) were calculated by standardized ROI analysis. Image quality and suitability for TAVI planning were assessed by four independent expert readers (two cardiac radiologists, two cardiologists) on a 5-point-scale. CT dose index (CTDI) and dose-length-product (DLP) were used to calculate effective radiation dose (eRD). RESULTS: 300 patients (136 female, median age: 81 years, IQR: 76-84) underwent 302 CT examinations, with PCCT in 202, DSCT in 100; two patients underwent both. Although SNR and CNR were significantly lower in PCCT vs. DSCT images (33.0 ± 10.5 vs. 47.3 ± 16.4 and 47.3 ± 14.8 vs. 59.3 ± 21.9, P < .001, respectively), visual image quality was higher in PCCT vs. DSCT (4.8 vs. 3.3, P < .001), with moderate overall interreader agreement among radiologists and among cardiologists (κ = 0.60, respectively). Image quality was rated as "excellent" in 160/202 (79.2%) of PCCT vs. 5/100 (5%) of DSCT cases. Readers found images suitable to depict the aortic valve hinge points and to map the femoral access path in 99% of PCCT vs. 85% of DSCT (P < 0.01), with suitability ranked significantly higher in PCCT vs. DSCT (4.8 vs. 3.3, P < .001). Mean CTDI and DLP, and thus eRD, were significantly lower for PCCT vs. DSCT (22.4 vs. 62.9; 519.4 vs. 895.5, and 8.8 ± 4.5 mSv vs. 15.3 ± 5.8 mSv; all P < .001). CONCLUSION: PCCT improves image quality, effectively avoids non-diagnostic CT imaging for TAVI planning, and is associated with a lower radiation dose compared to state-of-the-art DSCT. Radiologists and cardiologists found PCCT images more suitable for TAVI planning.
RESUMO
Quantitative MRI techniques such as T2 and T1ρ mapping are beneficial in evaluating knee joint pathologies; however, long acquisition times limit their clinical adoption. MIXTURE (Multi-Interleaved X-prepared Turbo Spin-Echo with IntUitive RElaxometry) provides a versatile turbo spin-echo (TSE) platform for simultaneous morphologic and quantitative joint imaging. Two MIXTURE sequences were designed along clinical requirements: "MIX1", combining proton density (PD)-weighted fat-saturated (FS) images and T2 mapping (acquisition time: 4:59 min), and "MIX2", combining T1-weighted images and T1ρ mapping (6:38 min). MIXTURE sequences and their reference 2D and 3D TSE counterparts were acquired from ten human cadaveric knee joints at 3.0 T. Contrast, contrast-to-noise ratios, and coefficients of variation were comparatively evaluated using parametric tests. Clinical radiologists (n = 3) assessed diagnostic quality as a function of sequence and anatomic structure using five-point Likert scales and ordinal regression, with a significance level of α = 0.01. MIX1 and MIX2 had at least equal diagnostic quality compared to reference sequences of the same image weighting. Contrast, contrast-to-noise ratios, and coefficients of variation were largely similar for the PD-weighted FS and T1-weighted images. In clinically feasible scan times, MIXTURE sequences yield morphologic, TSE-based images of diagnostic quality and quantitative parameter maps with additional insights on soft tissue composition and ultrastructure.
RESUMO
OBJECTIVES: Large language models (LLMs) have shown potential in radiology, but their ability to aid radiologists in interpreting imaging studies remains unexplored. We investigated the effects of a state-of-the-art LLM (GPT-4) on the radiologists' diagnostic workflow. MATERIALS AND METHODS: In this retrospective study, six radiologists of different experience levels read 40 selected radiographic [n = 10], CT [n = 10], MRI [n = 10], and angiographic [n = 10] studies unassisted (session one) and assisted by GPT-4 (session two). Each imaging study was presented with demographic data, the chief complaint, and associated symptoms, and diagnoses were registered using an online survey tool. The impact of Artificial Intelligence (AI) on diagnostic accuracy, confidence, user experience, input prompts, and generated responses was assessed. False information was registered. Linear mixed-effect models were used to quantify the factors (fixed: experience, modality, AI assistance; random: radiologist) influencing diagnostic accuracy and confidence. RESULTS: When assessing if the correct diagnosis was among the top-3 differential diagnoses, diagnostic accuracy improved slightly from 181/240 (75.4%, unassisted) to 188/240 (78.3%, AI-assisted). Similar improvements were found when only the top differential diagnosis was considered. AI assistance was used in 77.5% of the readings. Three hundred nine prompts were generated, primarily involving differential diagnoses (59.1%) and imaging features of specific conditions (27.5%). Diagnostic confidence was significantly higher when readings were AI-assisted (p > 0.001). Twenty-three responses (7.4%) were classified as hallucinations, while two (0.6%) were misinterpretations. CONCLUSION: Integrating GPT-4 in the diagnostic process improved diagnostic accuracy slightly and diagnostic confidence significantly. Potentially harmful hallucinations and misinterpretations call for caution and highlight the need for further safeguarding measures. CLINICAL RELEVANCE STATEMENT: Using GPT-4 as a virtual assistant when reading images made six radiologists of different experience levels feel more confident and provide more accurate diagnoses; yet, GPT-4 gave factually incorrect and potentially harmful information in 7.4% of its responses.
RESUMO
BACKGROUND: To compare denoising diffusion probabilistic models (DDPM) and generative adversarial networks (GAN) for recovering contrast-enhanced breast magnetic resonance imaging (MRI) subtraction images from virtual low-dose subtraction images. METHODS: Retrospective, ethically approved study. DDPM- and GAN-reconstructed single-slice subtraction images of 50 breasts with enhancing lesions were compared to original ones at three dose levels (25%, 10%, 5%) using quantitative measures and radiologic evaluations. Two radiologists stated their preference based on the reconstruction quality and scored the lesion conspicuity as compared to the original, blinded to the model. Fifty lesion-free maximum intensity projections were evaluated for the presence of false-positives. Results were compared between models and dose levels, using generalized linear mixed models. RESULTS: At 5% dose, both radiologists preferred the GAN-generated images, whereas at 25% dose, both radiologists preferred the DDPM-generated images. Median lesion conspicuity scores did not differ between GAN and DDPM at 25% dose (5 versus 5, p = 1.000) and 10% dose (4 versus 4, p = 1.000). At 5% dose, both readers assigned higher conspicuity to the GAN than to the DDPM (3 versus 2, p = 0.007). In the lesion-free examinations, DDPM and GAN showed no differences in the false-positive rate at 5% (15% versus 22%), 10% (10% versus 6%), and 25% (6% versus 4%) (p = 1.000). CONCLUSIONS: Both GAN and DDPM yielded promising results in low-dose image reconstruction. However, neither of them showed superior results over the other model for all dose levels and evaluation metrics. Further development is needed to counteract false-positives. RELEVANCE STATEMENT: For MRI-based breast cancer screening, reducing the contrast agent dose is desirable. Diffusion probabilistic models and generative adversarial networks were capable of retrospectively enhancing the signal of low-dose images. Hence, they may supplement imaging with reduced doses in the future. KEY POINTS: ⢠Deep learning may help recover signal in low-dose contrast-enhanced breast MRI. ⢠Two models (DDPM and GAN) were trained at different dose levels. ⢠Radiologists preferred DDPM at 25%, and GAN images at 5% dose. ⢠Lesion conspicuity between DDPM and GAN was similar, except at 5% dose. ⢠GAN and DDPM yield promising results in low-dose image reconstruction.
Assuntos
Neoplasias da Mama , Meios de Contraste , Imageamento por Ressonância Magnética , Humanos , Feminino , Estudos Retrospectivos , Meios de Contraste/administração & dosagem , Neoplasias da Mama/diagnóstico por imagem , Imageamento por Ressonância Magnética/métodos , Pessoa de Meia-Idade , Modelos Estatísticos , Adulto , IdosoRESUMO
PURPOSE: Further acceleration of DWI in diagnostic radiology is desired but challenging mainly due to low SNR in high b-value images and associated bias in quantitative ADC values. Deep learning-based reconstruction and denoising may provide a solution to address this challenge. METHODS: The effects of SNR reduction on ADC bias and variability were investigated using a commercial diffusion phantom and numerical simulations. In the phantom, performance of different reconstruction methods, including conventional parallel (SENSE) imaging, compressed sensing (C-SENSE), and compressed SENSE acceleration with an artificial intelligence deep learning-based technique (C-SENSE AI), was compared at different acceleration factors and flip angles using ROI-based analysis. ADC bias was assessed by Lin's Concordance correlation coefficient (CCC) followed by bootstrapping to calculate confidence intervals (CI). ADC random measurement error (RME) was assessed by the mean coefficient of variation (CV¯) and non-parametric statistical tests. RESULTS: The simulations predicted increasingly negative bias and loss of precision towards lower SNR. These effects were confirmed in phantom measurements of increasing acceleration, for which CCC decreased from 0.947 to 0.279 and CV¯ increased from 0.043 to 0.439, and of decreasing flip angle, for which CCC decreased from 0.990 to 0.063 and CV¯ increased from 0.037 to 0.508. At high acceleration and low flip angle, C-SENSE AI reconstruction yielded best denoised ADC maps. For the lowest investigated flip angle, CCC = {0.630, 0.771 and 0.987} and CV¯={0.508, 0.426 and 0.254} were obtained for {SENSE, C-SENSE, C-SENSE AI}, the improvement by C-SENSE AI being significant as compared to the other methods (CV: p = 0.033 for C-SENSE AI vs. C-SENSE and p < 0.001 for C-SENSE AI vs. SENSE; CCC: non-overlapping CI between reconstruction methods). For the highest investigated acceleration factor, CCC = {0.479,0.926,0.960} and CV¯={0.519,0.119,0.118} were found, confirming the reduction of bias and RME by C-SENSE AI as compared to C-SENSE (by trend) and to SENSE (CV: p < 0.001; CCC: non-overlapping CI). CONCLUSION: ADC bias and random measurement error in DWI at low SNR, typically associated with scan acceleration, can be effectively reduced by deep-learning based C-SENSE AI reconstruction.
Assuntos
Aprendizado Profundo , Imagem de Difusão por Ressonância Magnética , Processamento de Imagem Assistida por Computador , Imagens de Fantasmas , Razão Sinal-Ruído , Imagem de Difusão por Ressonância Magnética/métodos , Processamento de Imagem Assistida por Computador/métodos , Humanos , Reprodutibilidade dos Testes , Algoritmos , Simulação por ComputadorRESUMO
Abbreviated MRI is an umbrella term, defined as a focused MRI examination tailored to answer a single specific clinical question. For abbreviated breast MRI, this question is: "Is there evidence of breast cancer?" Abbreviated MRI of the breast makes maximum use of the fact that the kinetics of breast cancers and of benign tissue differ most in the very early postcontrast phase; therefore, abbreviated breast MRI focuses on this period. The different published approaches to abbreviated MRI include the following three subtypes: (a) short protocols, consisting of a precontrast and either a single postcontrast acquisition (first postcontrast subtracted [FAST]) or a time-resolved series of postcontrast acquisitions with lower spatial resolution (ultrafast [UF]), obtained during the early postcontrast phase immediately after contrast agent injection; (b) abridged protocols, consisting of FAST or UF acquisitions plus selected additional pulse sequences; and (c) noncontrast protocols, where diffusion-weighted imaging replaces the contrast information. Abbreviated MRI was proposed to increase tolerability of and access to breast MRI as a screening tool. But its widening application now includes follow-up after breast cancer and even diagnostic assessment. This review defines the three subtypes of abbreviated MRI, highlighting the differences between the protocols and their clinical implications and summarizing the respective evidence on diagnostic accuracy and clinical utility.
Assuntos
Neoplasias da Mama , Imageamento por Ressonância Magnética , Humanos , Feminino , Imagem de Difusão por Ressonância Magnética , Mama/diagnóstico por imagem , Neoplasias da Mama/diagnóstico por imagem , CinéticaRESUMO
Increased attention has been given to MRI in radiation-free screening for malignant nodules in recent years. Our objective was to compare the performance of human readers and radiomic feature analysis based on stand-alone and complementary CT and MRI imaging in classifying pulmonary nodules. This single-center study comprises patients with CT findings of pulmonary nodules who underwent additional lung MRI and whose nodules were classified as benign/malignant by resection. For radiomic features analysis, 2D segmentation was performed for each lung nodule on axial CT, T2-weighted (T2w), and diffusion (DWI) images. The 105 extracted features were reduced by iterative backward selection. The performance of radiomics and human readers was compared by calculating accuracy with Clopper-Pearson confidence intervals. Fifty patients (mean age 63 +/- 10 years) with 66 pulmonary nodules (40 malignant) were evaluated. ACC values for radiomic features analysis vs. radiologists based on CT alone (0.68; 95%CI: 0.56, 0.79 vs. 0.59; 95%CI: 0.46, 0.71), T2w alone (0.65; 95%CI: 0.52, 0.77 vs. 0.68; 95%CI: 0.54, 0.78), DWI alone (0.61; 95%CI:0.48, 0.72 vs. 0.73; 95%CI: 0.60, 0.83), combined T2w/DWI (0.73; 95%CI: 0.60, 0.83 vs. 0.70; 95%CI: 0.57, 0.80), and combined CT/T2w/DWI (0.83; 95%CI: 0.72, 0.91 vs. 0.64; 95%CI: 0.51, 0.75) were calculated. This study is the first to show that by combining quantitative image information from CT, T2w, and DWI datasets, pulmonary nodule assessment through radiomics analysis is superior to using one modality alone, even exceeding human readers' performance.
RESUMO
BACKGROUND: Artificial intelligence (AI) models are increasingly used in the medical domain. However, as medical data is highly sensitive, special precautions to ensure its protection are required. The gold standard for privacy preservation is the introduction of differential privacy (DP) to model training. Prior work indicates that DP has negative implications on model accuracy and fairness, which are unacceptable in medicine and represent a main barrier to the widespread use of privacy-preserving techniques. In this work, we evaluated the effect of privacy-preserving training of AI models regarding accuracy and fairness compared to non-private training. METHODS: We used two datasets: (1) A large dataset (N = 193,311) of high quality clinical chest radiographs, and (2) a dataset (N = 1625) of 3D abdominal computed tomography (CT) images, with the task of classifying the presence of pancreatic ductal adenocarcinoma (PDAC). Both were retrospectively collected and manually labeled by experienced radiologists. We then compared non-private deep convolutional neural networks (CNNs) and privacy-preserving (DP) models with respect to privacy-utility trade-offs measured as area under the receiver operating characteristic curve (AUROC), and privacy-fairness trade-offs, measured as Pearson's r or Statistical Parity Difference. RESULTS: We find that, while the privacy-preserving training yields lower accuracy, it largely does not amplify discrimination against age, sex or co-morbidity. However, we find an indication that difficult diagnoses and subgroups suffer stronger performance hits in private training. CONCLUSIONS: Our study shows that - under the challenging realistic circumstances of a real-life clinical dataset - the privacy-preserving training of diagnostic deep learning models is possible with excellent diagnostic accuracy and fairness.
Artificial intelligence (AI), in which computers can learn to do tasks that normally require human intelligence, is particularly useful in medical imaging. However, AI should be used in a way that preserves patient privacy. We explored the balance between maintaining patient data privacy and AI performance in medical imaging. We use an approach called differential privacy to protect the privacy of patients' images. We show that, although training AI with differential privacy leads to a slight decrease in accuracy, it does not substantially increase bias against different age groups, genders, or patients with multiple health conditions. However, we notice that AI faces more challenges in accurately diagnosing complex cases and specific subgroups when trained under these privacy constraints. These findings highlight the importance of designing AI systems that are both privacy-conscious and capable of reliable diagnoses across patient groups.
RESUMO
A knowledge gap persists between machine learning (ML) developers (e.g., data scientists) and practitioners (e.g., clinicians), hampering the full utilization of ML for clinical data analysis. We investigated the potential of the ChatGPT Advanced Data Analysis (ADA), an extension of GPT-4, to bridge this gap and perform ML analyses efficiently. Real-world clinical datasets and study details from large trials across various medical specialties were presented to ChatGPT ADA without specific guidance. ChatGPT ADA autonomously developed state-of-the-art ML models based on the original study's training data to predict clinical outcomes such as cancer development, cancer progression, disease complications, or biomarkers such as pathogenic gene sequences. Following the re-implementation and optimization of the published models, the head-to-head comparison of the ChatGPT ADA-crafted ML models and their respective manually crafted counterparts revealed no significant differences in traditional performance metrics (p ≥ 0.072). Strikingly, the ChatGPT ADA-crafted ML models often outperformed their counterparts. In conclusion, ChatGPT ADA offers a promising avenue to democratize ML in medicine by simplifying complex data analyses, yet should enhance, not replace, specialized training and resources, to promote broader applications in medical research and practice.
Assuntos
Algoritmos , Neoplasias , Humanos , Benchmarking , Idioma , Aprendizado de MáquinaRESUMO
High-quality magnetic resonance (MR) imaging is essential for the precise assessment of the knee joint and plays a key role in the diagnostics, treatment and prognosis. Intact cartilage tissue is characterized by a smooth surface, uniform tissue thickness and an organized zonal structure, which are manifested as depth-dependent signal intensity variations. Cartilage pathologies are identifiable through alterations in signal intensity and morphology and should be communicated based on a precise terminology. Cartilage pathologies can show hyperintense and hypointense signal alterations. Cartilage defects are assessed based on their depth and should be described in terms of their location and extent. The following symptom constellations are of overarching clinical relevance in image reading and interpretation: symptom constellations associated with rapidly progressive forms of joint degeneration and unfavorable prognosis, accompanying symptom constellations mostly in connection with destabilizing meniscal lesions and subchondral insufficiency fractures (accelerated osteoarthritis) as well as symptoms beyond the "typical" degeneration, especially when a discrepancy is observed between (minor) structural changes and (major) synovitis and effusion (inflammatory arthropathy).
Assuntos
Cartilagem Articular , Osteoartrite do Joelho , Humanos , Osteoartrite do Joelho/complicações , Osteoartrite do Joelho/patologia , Cartilagem Articular/patologia , Progressão da Doença , Articulação do Joelho/patologia , Imageamento por Ressonância Magnética/métodosRESUMO
Magnetic resonance imaging (MRI) is the clinical method of choice for cartilage imaging in the context of degenerative and nondegenerative joint diseases. The MRI-based definitions of osteoarthritis rely on the detection of osteophytes, cartilage pathologies, bone marrow edema and meniscal lesions but currently a scientific consensus is lacking. In the clinical routine proton density-weighted, fat-suppressed 2D turbo spin echo sequences with echo times of 30-40â¯ms are predominantly used, which are sufficiently sensitive and specific for the assessment of cartilage. The additionally acquired T1-weighted sequences are primarily used for evaluating other intra-articular and periarticular structures. Diagnostically relevant artifacts include magic angle and chemical shift artifacts, which can lead to artificial signal enhancement in cartilage or incorrect representations of the subchondral lamina and its thickness. Although scientifically validated, high-resolution 3D gradient echo sequences (for cartilage segmentation) and compositional MR sequences (for quantification of physical tissue parameters) are currently reserved for scientific research questions. The future integration of artificial intelligence techniques in areas such as image reconstruction (to reduce scan times while maintaining image quality), image analysis (for automated identification of cartilage defects), and image postprocessing (for automated segmentation of cartilage in terms of volume and thickness) will significantly improve the diagnostic workflow and advance the field further.
Assuntos
Doenças das Cartilagens , Cartilagem Articular , Osteoartrite do Joelho , Humanos , Osteoartrite do Joelho/patologia , Cartilagem Articular/patologia , Inteligência Artificial , Doenças das Cartilagens/patologia , Imageamento por Ressonância Magnética/métodosRESUMO
Artificial intelligence (AI) has a multitude of applications in cancer research and oncology. However, the training of AI systems is impeded by the limited availability of large datasets due to data protection requirements and other regulatory obstacles. Federated and swarm learning represent possible solutions to this problem by collaboratively training AI models while avoiding data transfer. However, in these decentralized methods, weight updates are still transferred to the aggregation server for merging the models. This leaves the possibility for a breach of data privacy, for example by model inversion or membership inference attacks by untrusted servers. Somewhat-homomorphically-encrypted federated learning (SHEFL) is a solution to this problem because only encrypted weights are transferred, and model updates are performed in the encrypted space. Here, we demonstrate the first successful implementation of SHEFL in a range of clinically relevant tasks in cancer image analysis on multicentric datasets in radiology and histopathology. We show that SHEFL enables the training of AI models which outperform locally trained models and perform on par with models which are centrally trained. In the future, SHEFL can enable multiple institutions to co-train AI models without forsaking data governance and without ever transmitting any decryptable data to untrusted servers.
Assuntos
Neoplasias , Radiologia , Humanos , Inteligência Artificial , Aprendizagem , Neoplasias/diagnóstico por imagem , Processamento de Imagem Assistida por ComputadorRESUMO
Developing robust artificial intelligence (AI) models that generalize well to unseen datasets is challenging and usually requires large and variable datasets, preferably from multiple institutions. In federated learning (FL), a model is trained collaboratively at numerous sites that hold local datasets without exchanging them. So far, the impact of training strategy, i.e., local versus collaborative, on the diagnostic on-domain and off-domain performance of AI models interpreting chest radiographs has not been assessed. Consequently, using 610,000 chest radiographs from five institutions across the globe, we assessed diagnostic performance as a function of training strategy (i.e., local vs. collaborative), network architecture (i.e., convolutional vs. transformer-based), single versus cross-institutional performance (i.e., on-domain vs. off-domain), imaging finding (i.e., cardiomegaly, pleural effusion, pneumonia, atelectasis, consolidation, pneumothorax, and no abnormality), dataset size (i.e., from n = 18,000 to 213,921 radiographs), and dataset diversity. Large datasets not only showed minimal performance gains with FL but, in some instances, even exhibited decreases. In contrast, smaller datasets revealed marked improvements. Thus, on-domain performance was mainly driven by training data size. However, off-domain performance leaned more on training diversity. When trained collaboratively across diverse external institutions, AI models consistently surpassed models trained locally for off-domain tasks, emphasizing FL's potential in leveraging data diversity. In conclusion, FL can bolster diagnostic privacy, reproducibility, and off-domain reliability of AI models and, potentially, optimize healthcare outcomes.
Assuntos
Inteligência Artificial , Aprendizagem , Reprodutibilidade dos Testes , Generalização Psicológica , RadiografiaRESUMO
Large language models (LLMs) have shown potential in various applications, including clinical practice. However, their accuracy and utility in providing treatment recommendations for orthopedic conditions remain to be investigated. Thus, this pilot study aims to evaluate the validity of treatment recommendations generated by GPT-4 for common knee and shoulder orthopedic conditions using anonymized clinical MRI reports. A retrospective analysis was conducted using 20 anonymized clinical MRI reports, with varying severity and complexity. Treatment recommendations were elicited from GPT-4 and evaluated by two board-certified specialty-trained senior orthopedic surgeons. Their evaluation focused on semiquantitative gradings of accuracy and clinical utility and potential limitations of the LLM-generated recommendations. GPT-4 provided treatment recommendations for 20 patients (mean age, 50 years ± 19 [standard deviation]; 12 men) with acute and chronic knee and shoulder conditions. The LLM produced largely accurate and clinically useful recommendations. However, limited awareness of a patient's overall situation, a tendency to incorrectly appreciate treatment urgency, and largely schematic and unspecific treatment recommendations were observed and may reduce its clinical usefulness. In conclusion, LLM-based treatment recommendations are largely adequate and not prone to 'hallucinations', yet inadequate in particular situations. Critical guidance by healthcare professionals is obligatory, and independent use by patients is discouraged, given the dependency on precise data input.
Assuntos
Medicina , Doenças Musculoesqueléticas , Masculino , Humanos , Pessoa de Meia-Idade , Projetos Piloto , Estudos Retrospectivos , Idioma , Imageamento por Ressonância MagnéticaRESUMO
Background Clinicians consider both imaging and nonimaging data when diagnosing diseases; however, current machine learning approaches primarily consider data from a single modality. Purpose To develop a neural network architecture capable of integrating multimodal patient data and compare its performance to models incorporating a single modality for diagnosing up to 25 pathologic conditions. Materials and Methods In this retrospective study, imaging and nonimaging patient data were extracted from the Medical Information Mart for Intensive Care (MIMIC) database and an internal database comprised of chest radiographs and clinical parameters inpatients in the intensive care unit (ICU) (January 2008 to December 2020). The MIMIC and internal data sets were each split into training (n = 33 893, n = 28 809), validation (n = 740, n = 7203), and test (n = 1909, n = 9004) sets. A novel transformer-based neural network architecture was trained to diagnose up to 25 conditions using nonimaging data alone, imaging data alone, or multimodal data. Diagnostic performance was assessed using area under the receiver operating characteristic curve (AUC) analysis. Results The MIMIC and internal data sets included 36 542 patients (mean age, 63 years ± 17 [SD]; 20 567 male patients) and 45 016 patients (mean age, 66 years ± 16; 27 577 male patients), respectively. The multimodal model showed improved diagnostic performance for all pathologic conditions. For the MIMIC data set, the mean AUC was 0.77 (95% CI: 0.77, 0.78) when both chest radiographs and clinical parameters were used, compared with 0.70 (95% CI: 0.69, 0.71; P < .001) for only chest radiographs and 0.72 (95% CI: 0.72, 0.73; P < .001) for only clinical parameters. These findings were confirmed on the internal data set. Conclusion A model trained on imaging and nonimaging data outperformed models trained on only one type of data for diagnosing multiple diseases in patients in an ICU setting. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Kitamura and Topol in this issue.