Publications by authors named "Tomoki Imokawa"
Article Synopsis
- Brain perfusion SPECT is a crucial imaging technique that identifies neurodegenerative changes early on, playing an essential role in dementia evaluation and diagnosis.
- It highlights different patterns of hypoperfusion associated with various types of dementia, such as Alzheimer's disease and dementia with Lewy bodies, which helps differentiate between them and predicts disease progression.
- The method is shifting towards using biomarkers like brain perfusion SPECT for diagnosing dementia, guiding treatment with disease-modifying drugs based on individual patient profiles.
Article Synopsis
- A study evaluated the use of deep learning to automatically segment abnormal uptakes in dedicated breast positron emission tomography (dbPET) images to enhance detection of small breast cancers.
- The researchers trained a U-Net-based model on a dataset that included augmented collage images to address the issue of limited training data, resulting in high Dice scores and classification accuracies.
- The findings suggest that deep learning is effective for dbPET segmentation, and the inclusion of collage images significantly improves model performance, especially for cases without breast cancer.
This study aimed to evaluate the ability of the pix2pix generative adversarial network (GAN) to improve the image quality of low-count dedicated breast positron emission tomography (dbPET). Pairs of full- and low-count dbPET images were collected from 49 breasts. An image synthesis model was constructed using pix2pix GAN for each acquisition time with training (3776 pairs from 16 breasts) and validation data (1652 pairs from 7 breasts).
View Article and Find Full Text PDFArticle Synopsis
- - This study aimed to evaluate how well deep learning (DL), specifically using convolutional neural networks (CNNs), can identify normal versus metastatic axillary lymph nodes in ultrasound images compared to radiologists' assessments.
- - Researchers used 300 normal and 328 metastatic lymph node images to train a DL model (Xception) and tested it on 100 images, measuring its accuracy through sensitivity, specificity, and the area under the receiver operating characteristic curve (AUC).
- - The DL model achieved high accuracy, with 94% sensitivity and 88% specificity, and its performance was comparable to an experienced radiologist, showing potential to enhance diagnostic skills for less experienced residents.
Article Synopsis
- This study evaluated different deep learning (DL) models to predict breast cancer using dedicated breast positron emission tomography (dbPET) images from women diagnosed with breast cancer or non-breast cancer.
- A Xception-based DL model was trained and tested on dbPET images, showing high sensitivity (93%) and specificity (93%) compared to expert radiologists.
- The findings suggest that the DL model's diagnostic performance is comparable to that of radiology experts, indicating its potential for practical application in breast cancer detection.
Article Synopsis
- Deep learning (DL) has shown great potential in enhancing image processing, but its application in PET/CT scans for breast cancer has not been extensively researched.
- This study used a dataset of 400 images from 200 breast cancer (BC) and 200 non-BC patients to develop two DL models, one using only 0-degree PET maximum-intensity projection (MIP) images and the other incorporating four different MIP angles.
- The four-angle DL model achieved a diagnostic sensitivity of 96% and specificity of 80%, outperforming the traditional methods utilized by radiologists, suggesting that such DL models could significantly assist in future breast cancer diagnostics.
Xanthogranulomatous pancreatitis (XGP) is extremely rare, with only 31 cases reported in the English literature to date. We reviewed previously reported 17 articles about XGP and report an additional case of XGP. This is the first case of XGP with xanthogranulomatous cholecystitis accompanied by intraductal papillary mucinous carcinoma (IPMC) in an 80-year-old woman.
View Article and Find Full Text PDF