RESUMO
OBJECTIVES: The interpretation of mammograms requires many years of training and experience. Currently, training in mammography, like the rest of diagnostic radiology, is through institutional libraries, books, and experience accumulated over time. We explore whether artificial Intelligence (AI)-generated images can help in simulation education and result in measurable improvement in performance of residents in training. METHODS: We developed a generative adversarial network (GAN) that was capable of generating mammography images with varying characteristics, such as size and density, and created a tool with which a user could control these characteristics. The tool allowed the user (a radiology resident) to realistically insert cancers within different regions of the mammogram. We then provided this tool to residents in training. Residents were randomized into a practice group and a non-practice group, and the difference in performance before and after practice with such a tool (in comparison to no intervention in the non-practice group) was assessed. RESULTS: Fifty residents participated in the study, 27 underwent simulation training, and 23 did not. There was a significant improvement in the sensitivity (7.43 percent, significant at p-value = 0.03), negative predictive value (5.05 percent, significant at p-value = 0.008) and accuracy (6.49 percent, significant at p-value = 0.01) among residents in the detection of cancer on mammograms after simulation training. CONCLUSION: Our study shows the value of simulation training in diagnostic radiology and explores the potential of generative AI to enable such simulation training. CLINICAL RELEVANCE STATEMENT: Using generative artificial intelligence, simulation training modules can be developed that can help residents in training by providing them with a visual impression of a variety of different cases. KEY POINTS: Generative networks can produce diagnostic imaging with specific characteristics, potentially useful for training residents. Training with generating images improved residents' mammographic diagnostic abilities. Development of a game-like interface that exploits these networks can result in improvement in performance over a short training period.
RESUMO
While breast carcinoma is the most feared pathology in women with breast lumps, infections continue to be an important aetiology, especially in countries with low to middle socio-economic status. The breast infections or mastitis can present as acute painful breast or recurrent episodes of breast lumps with or without pain. The common causes include puerperal, non-puerperal, and idiopathic mastitis whereas uncommon causes like tuberculosis, filariasis, hydatid and other parasitic infections are still seen in developing countries. Imaging with digital mammography may be difficult due to pain or inadequate due to increased breast density. Ultrasound serves as the modality of choice for detailed assessment in these patients. Since the imaging features are often overlapping with malignancy, biopsy is almost always indicated. However, there are certain imaging findings that may point to the diagnosis of mastitis and can help in accurate radiologic-pathologic correlation. This article aims to illustrate the varied clinico-radiological features of patients with tropical breast infections.
Assuntos
Neoplasias da Mama , Mastite , Humanos , Feminino , Mastite/diagnóstico por imagem , Neoplasias da Mama/patologia , Mamografia/métodos , Biópsia , Diagnóstico Diferencial , Dor/diagnósticoRESUMO
PURPOSE: To evaluate the role of conventional diffusion weighted imaging, diffusion kurtosis imaging (DKI), and intravoxel incoherent motion (IVIM) in distinguishing benign from malignant adnexal masses. METHODS: 38 patients with 45 adnexal masses were enrolled in this prospective study and assessed with multiparametric MRI, including the IVIM-DKI sequence, on a 3 T MRI system. The mean apparent diffusion coefficient (ADC) from conventional DWI, the apparent diffusion coefficient derived from DKI (Dapp), the apparent kurtosis coefficient (Kapp), true diffusion coefficient (Dt), perfusion fraction (f) and pseudo-diffusion coefficient (Dp) were measured. RESULTS: The mean ADC, Dapp, and Dt were significantly higher in benign adnexal masses than in malignant adnexal masses (p < 0.001). f and Dp were also significantly higher in benign adnexal masses, with p values of 0.026 and 0.002, respectively. Kapp was higher in malignant masses (p < 0.001). Among mean ADC, Dapp, and Dt, mean ADC had the highest area under the curve (AUC) of 0.885. However, no statistically significant differences were observed between the ROCs of various diffusion parameters. CONCLUSION: The mean ADC, Dapp, and Kapp are useful parameters in discriminating between benign and malignant adnexal masses. Dt derived from IVIM also helps in distinguishing benign and malignant adnexal masses; however, no incremental role of IVIM and DKI over ADC could be identified in our study.
Assuntos
Doenças dos Anexos , Imagem de Difusão por Ressonância Magnética , Humanos , Feminino , Imagem de Difusão por Ressonância Magnética/métodos , Pessoa de Meia-Idade , Adulto , Estudos Prospectivos , Doenças dos Anexos/diagnóstico por imagem , Diagnóstico Diferencial , Idoso , Ultrassonografia/métodos , Interpretação de Imagem Assistida por Computador/métodos , Sensibilidade e Especificidade , AdolescenteRESUMO
The objective of this research was to study the contrast enhancement patterns of the different molecular subtypes of breast cancer on contrast-enhanced ultrasound (CEUS) using both qualitative and quantitative parameters. This prospective study included females with a single breast mass which was histopathologically proven carcinoma. B mode ultrasound (USG) and CEUS were performed in all patients during baseline assessment. Qualitative CEUS assessment encompassed enhancement pattern, presence of fill-in and washout. Quantitative assessment included measurement of peak enhancement, time to peak; area under the curve and mean transit time. A p-value < 0.05 was considered statistically significant for differentiating the subtypes. The included thirty masses were categorised into two subtypes-triple negative breast cancer (TNBC) (36.7%) and non-TNBC (63.3%) subtypes. With B-mode USG, a statistically significant difference was observed between the two groups with respect to their shape and margins. TNBC lesions showed an oval shape, circumscribed margins and peripheral nodular enhancement on CEUS with the absence of fill-in even in the delayed phase (p-value - 0.04). The two subtypes did not significantly differ in terms of quantitative perfusion parameters. The various subtypes of breast cancer therefore possess distinct contrast enhancement patterns. CEUS potentially allows differentiation amongst these molecular subtypes that may aid in radiology-pathology (rad-path) correlation and follow up of the patients.