Background/Objectives: This study aimed to investigate the accuracy of Tumor, Node, Metastasis (TNM) classification based on radiology reports using GPT3.5-turbo (GPT3.5) and the utility of multilingual large language models (LLMs) in both Japanese and English. Methods: Utilizing GPT3.5, we developed a system to automatically generate TNM classifications from chest computed tomography reports for lung cancer and evaluate its performance. We statistically analyzed the impact of providing full or partial TNM definitions in both languages using a generalized linear mixed model. Results: The highest accuracy was attained with full TNM definitions and radiology reports in English (M = 94%, N = 80%, T = 47%, and TNM combined = 36%). Providing definitions for each of the T, N, and M factors statistically improved their respective accuracies (T: odds ratio [OR] = 2.35, p < 0.001; N: OR = 1.94, p < 0.01; M: OR = 2.50, p < 0.001). Japanese reports exhibited decreased N and M accuracies (N accuracy: OR = 0.74 and M accuracy: OR = 0.21). Conclusions: This study underscores the potential of multilingual LLMs for automatic TNM classification in radiology reports. Even without additional model training, performance improvements were evident with the provided TNM definitions, indicating LLMs’ relevance in radiology contexts.

With the increasing use of postmortem imaging, deep learning (DL)-based automated analysis may assist in the detection of intracranial hemorrhages. However, limited postmortem data complicate model training. This study aims to assess the accuracy of DL models in detecting intracranial hemorrhages in postmortem head computed tomography (CT) scans using transfer learning. A total of 75,000 labeled head CT images from the Radiological Society of North America Intracranial Hemorrhage Detection Challenge serve as the training data for the 15 DL models. Each model is fine-tuned via transfer learning. A total of 134 postmortem cases with hemorrhage status confirmed by autopsy serve as the external test set. Model performance is evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, training time, inference time, and number of parameters. Spearman’s rank correlation coefficients are calculated for these metrics. DenseNet201 achieves the highest AUC (0.907), with the AUCs of the 15 models ranging from 0.862 to 0.907. A longer inference time moderately correlates with higher AUC (Spearman’s ρ = 0.586, p = 0.022), whereas the number of parameters is not positively correlated with performance (ρ = −0.472, p = 0.076). The sensitivity and specificity are 0.828 and 0.871, respectively. Transfer learning using a large non-postmortem dataset enables accurate intracranial hemorrhage detection using postmortem CT, potentially reducing the autopsy workload. The results demonstrate that models with fewer parameters often perform comparably to more complex models, emphasizing the need to balance accuracy with computational efficiency.

Objective: In positron emission tomography (PET)/magnetic resonance imaging (MRI), attenuation correction (AC) for PET of the head is achieved by MRI data to generate pseudo-computed tomography (CT) images. However, for the torso, AC becomes more challenging due to the complexity of separating bone components. Additionally, generating accurate MRI-based CT using deep learning poses significant difficulties for the chest, primarily because perfectly paired MRI and CT training data are hard to obtain owing to respiratory motion and body movements. We previously demonstrated that MRI-to-CT conversion can be achieved without deformation, even using unsupervised learning for zero echo time (ZTE) MRI and CT data from different individuals. Building on this foundation, our study aims to apply this approach to AC in chest PET/MRI and assess their quantitative accuracy, reproducibility, and external validity. Methods: The datasets used included (1) training dataset (unpaired ZTE MRI and CT of PET/CT, n = 360 and 500, respectively); (2) test dataset (paired PET/MRI and PET/CT, n = 25 and 25, respectively); (3) repeatability assessment dataset (repeated PET/MRI, n = 15 × 2 scans for the same patient); and (4) external validation dataset (paired MRI component of PET/MRI and CT, n = 30 and 30, respectively, acquired at another institution). Unpaired training data were used to train the deep learning model of pseudo-CT generation from ZTE. The accuracy, repeatability, and reproducibility of the PET/MRI scans using ZTE- and deep learning-based AC (MRAC _ZTE ) were evaluated based on the similarity of the histograms and the mean standardized uptake value (SUVmean) of physiological background of bone and liver. Results: The histogram correlation coefficients between MRAC _ZTE and the AC map based on the CT (CTAC) for the spine were significantly higher than those between conventional AC (MRAC _Dixon ) and CTAC. Additionally, bone SUVmean obtained using MRAC _ZTE showed reduced bias relative to CTAC compared with MRAC _Dixon . This method proved to be reproducible on each patient level and robust against external validation. Conclusions: Unsupervised learning with unpaired ZTE and CT data enabled pseudo-CT generation with bone components that closely matched CT-based attenuation maps. Integration into MR-based attenuation correction resulted in stable physiological uptake measurements in chest PET/MRI, supporting the feasibility of this approach.

With the increasing use of postmortem imaging, deep learning (DL)-based automated analysis may assist in detecting intracranial hemorrhages. However, limited postmortem data complicate model training. This study aims to assess the accuracy of DL models in detecting intracranial hemorrhages in postmortem head computed tomography (CT) scans using transfer learning. 75,000 labeled head CT images from the Radiological Society of North America Intracranial Hemorrhage Detection Challenge serve as training data for 15 DL models. Each model is fine-tuned via transfer learning. A total of 124 postmortem cases with hemorrhage status confirmed by autopsy serve as the external test set. The model performance is evaluated using the area under the receiver operating characteristic curve (AUC), sensitivity, specificity, training time, inference time, and number of parameters. Spearman’s rank correlation coefficients are calculated for these metrics. DenseNet201 achieves the highest AUC (0.907), with the AUCs of the 15 models ranging from 0.862 to 0.907. Longer inference time moderately correlates with higher AUC (Spearman’s ρ = 0.586, p = 0.022), whereas the number of parameters is not positively correlated with performance (ρ = -0.472, p = 0.076). The sensitivity and specificity are 0.906 and 0.986, respectively. Transfer learning using a large non-postmortem dataset enables accurate intracranial hemorrhage detection on postmortem CT, potentially reducing autopsy workload. The results demonstrate that models with fewer parameters often perform comparably to more complex models, emphasizing the need to balance accuracy with computational efficiency.

The integrated positron emission tomography/magnetic resonance imaging (PET/MRI) scanner facilitates the simultaneous acquisition of metabolic information via PET and morphological information with high soft-tissue contrast using MRI. Although PET/MRI facilitates the capture of high-accuracy fusion images, its major drawback can be attributed to the difficulty encountered when performing attenuation correction, which is necessary for quantitative PET evaluation. The combined PET/MRI scanning requires the generation of attenuation-correction maps from MRI owing to no direct relationship between the gamma-ray attenuation information and MRIs. While MRI-based bone-tissue segmentation can be readily performed for the head and pelvis regions, the realization of accurate bone segmentation via chest CT generation remains a challenging task. This can be attributed to the respiratory and cardiac motions occurring in the chest as well as its anatomically complicated structure and relatively thin bone cortex. This paper presents a means to minimise the anatomical structural changes without human annotation by adding structural constraints using a modality-independent neighbourhood descriptor (MIND) to a generative adversarial network (GAN) that can transform unpaired images. The results obtained in this study revealed the proposed U-GAT-IT + MIND approach to outperform all other competing approaches. The findings of this study hint towards possibility of synthesising clinically acceptable CT images from chest MRI without human annotation, thereby minimising the changes in the anatomical structure.