RESUMEN
Hepatocellular carcinoma frequently recurs after surgery, necessitating personalized clinical approaches based on tumor avatar models. However, location-dependent oxygen concentrations resulting from the dual hepatic vascular supply drive the inherent heterogeneity of the tumor microenvironment, which presents challenges in developing an avatar model. In this study, tissue samples from 12 patients with hepatocellular carcinoma are cultured directly on a chip and separated based on preference of oxygen concentration. Establishing a dual gradient system with drug perfusion perpendicular to the oxygen gradient enables the simultaneous separation of cells and evaluation of drug responsiveness. The results are further cross-validated by implanting the chips into mice at various oxygen levels using a patient-derived xenograft model. Hepatocellular carcinoma cells exposed to hypoxia exhibit invasive and recurrent characteristics that mirror clinical outcomes. This chip provides valuable insights into treatment prognosis by identifying the dominant hepatocellular carcinoma type in each patient, potentially guiding personalized therapeutic interventions.
Asunto(s)
Carcinoma Hepatocelular , Neoplasias Hepáticas , Oxígeno , Microambiente Tumoral , Carcinoma Hepatocelular/patología , Carcinoma Hepatocelular/metabolismo , Humanos , Neoplasias Hepáticas/patología , Neoplasias Hepáticas/metabolismo , Animales , Ratones , Oxígeno/metabolismo , Línea Celular Tumoral , Masculino , Femenino , Ensayos Antitumor por Modelo de Xenoinjerto , Persona de Mediana Edad , Dispositivos Laboratorio en un ChipRESUMEN
Recent advances in contrastive learning have significantly improved the performance of deep learning models. In contrastive learning of medical images, dealing with positive representation is sometimes difficult because some strong augmentation techniques can disrupt contrastive learning owing to the subtle differences between other standardized CXRs compared to augmented positive pairs; therefore, additional efforts are required. In this study, we propose intermediate feature approximation (IFA) loss, which improves the performance of contrastive convolutional neural networks by focusing more on positive representations of CXRs without additional augmentations. The IFA loss encourages the feature maps of a query image and its positive pair to resemble each other by maximizing the cosine similarity between the intermediate feature outputs of the original data and the positive pairs. Therefore, we used the InfoNCE loss, which is commonly used loss to address negative representations, and the IFA loss, which addresses positive representations, together to improve the contrastive network. We evaluated the performance of the network using various downstream tasks, including classification, object detection, and a generative adversarial network (GAN) inversion task. The downstream task results demonstrated that IFA loss can improve the performance of effectively overcoming data imbalance and data scarcity; furthermore, it can serve as a perceptual loss encoder for GAN inversion. In addition, we have made our model publicly available to facilitate access and encourage further research and collaboration in the field.
Asunto(s)
Aprendizaje Profundo , Redes Neurales de la Computación , Humanos , Radiografía TorácicaRESUMEN
RATIONALE AND OBJECTIVES: The effect of different computed tomography (CT) reconstruction kernels on the quantification of interstitial lung disease (ILD) has not been clearly demonstrated. The study aimed to investigate the effect of reconstruction kernels on the quantification of ILD on CT and determine whether deep learning-based kernel conversion can reduce the variability of automated quantification results between different CT kernels. MATERIALS AND METHODS: Patients with ILD or interstitial lung abnormality who underwent noncontrast high-resolution CT between June 2022 and September 2022 were retrospectively included. Images were reconstructed with three different kernels: B30f, B50f, and B60f. B60f was regarded as the reference standard for quantification, and B30f and B50f images were converted to B60f images using a deep learning-based algorithm. Each disease pattern of ILD and the fibrotic score were quantified using commercial software. The effect of kernel conversion on measurement variability was estimated using intraclass correlation coefficient (ICC) and Bland-Altman method. RESULTS: A total of 194 patients were included in the study. Application of different kernels induced differences in the quantified extent of each pattern. Reticular opacity and honeycombing were underestimated on B30f images and overestimated on B50f images. After kernel conversion, measurement variability was reduced (mean difference, from -2.0 to 3.9 to -0.3 to 0.4%, and 95% limits of agreement [LOA], from [-5.0, 12.7] to [-2.7, 2.1]). The fibrotic score for converted B60f from B50f images was almost equivalent to the original B60f (ICC, 1.000; mean difference, 0.0; and 95% LOA [-0.4, 0.4]). CONCLUSION: Quantitative CT analysis of ILD was affected by the application of different kernels, but deep learning-based kernel conversion effectively reduced measurement variability, improving the reproducibility of quantification.
Asunto(s)
Aprendizaje Profundo , Enfermedades Pulmonares Intersticiales , Humanos , Reproducibilidad de los Resultados , Estudios Retrospectivos , Tomografía Computarizada por Rayos X/métodos , Enfermedades Pulmonares Intersticiales/diagnóstico por imagen , Pulmón/diagnóstico por imagenRESUMEN
Training deep learning models on medical images heavily depends on experts' expensive and laborious manual labels. In addition, these images, labels, and even models themselves are not widely publicly accessible and suffer from various kinds of bias and imbalances. In this paper, chest X-ray pre-trained model via self-supervised contrastive learning (CheSS) was proposed to learn models with various representations in chest radiographs (CXRs). Our contribution is a publicly accessible pretrained model trained with a 4.8-M CXR dataset using self-supervised learning with a contrastive learning and its validation with various kinds of downstream tasks including classification on the 6-class diseases in internal dataset, diseases classification in CheXpert, bone suppression, and nodule generation. When compared to a scratch model, on the 6-class classification test dataset, we achieved 28.5% increase in accuracy. On the CheXpert dataset, we achieved 1.3% increase in mean area under the receiver operating characteristic curve on the full dataset and 11.4% increase only using 1% data in stress test manner. On bone suppression with perceptual loss, we achieved improvement in peak signal to noise ratio from 34.99 to 37.77, structural similarity index measure from 0.976 to 0.977, and root-square-mean error from 4.410 to 3.301 when compared to ImageNet pretrained model. Finally, on nodule generation, we achieved improvement in Fréchet inception distance from 24.06 to 17.07. Our study showed the decent transferability of CheSS weights. CheSS weights can help researchers overcome data imbalance, data shortage, and inaccessibility of medical image datasets. CheSS weight is available at https://github.com/mi2rl/CheSS .
Asunto(s)
Rayos X , Humanos , Curva ROC , Radiografía , Relación Señal-RuidoRESUMEN
Hematoxylin and eosin (H&E) staining is the gold standard modality for diagnosis in medicine. However, the dosage ratio of hematoxylin to eosin in H&E staining has not been standardized yet. Additionally, H&E stains fade out at various speeds. Therefore, the staining quality could differ among each image, and stain normalization is a critical preprocessing approach for training deep learning (DL) models, especially in long-term and/or multicenter digital pathology studies. However, conventional methods for stain normalization have some significant drawbacks, such as collapsing in the structure and/or texture of tissue. In addition, conventional methods must require a reference patch or slide. Meanwhile, DL-based methods have a risk of overfitting and/or grid artifacts. We developed a score-based diffusion model of colorization for stain normalization. However, mistransfer, in which the model confuses hematoxylin with eosin, can occur using a score-based diffusion model due to its high diversity nature. To overcome this mistransfer, we propose a stain separation method using sparse non-negative matrix factorization (SNMF), which can decompose pathology slide into Hematoxylin and Eosin to normalize each stain component. Furthermore, inpainting with overlapped moving window patches was used to prevent grid artifacts of whole slide image normalization. Our method can normalize the whole slide pathology images through this stain normalization pipeline with decent performance.
Asunto(s)
Algoritmos , Colorantes , Colorantes/química , Hematoxilina , Eosina Amarillenta-(YS) , Coloración y EtiquetadoRESUMEN
Deep learning-based applications have great potential to enhance the quality of medical services. The power of deep learning depends on open databases and innovation. Radiologists can act as important mediators between deep learning and medicine by simultaneously playing pioneering and gatekeeping roles. The application of deep learning technology in medicine is sometimes restricted by ethical or legal issues, including patient privacy and confidentiality, data ownership, and limitations in patient agreement. In this paper, we present an open platform, MI2RLNet, for sharing source code and various pre-trained weights for models to use in downstream tasks, including education, application, and transfer learning, to encourage deep learning research in radiology. In addition, we describe how to use this open platform in the GitHub environment. Our source code and models may contribute to further deep learning research in radiology, which may facilitate applications in medicine and healthcare, especially in medical imaging, in the near future. All code is available at https://github.com/mi2rl/MI2RLNet.