Since the first pioneering report of hyperpolarized [1-¹³C]pyruvate magnetic resonance imaging (MRI) of the Warburg effect in prostate cancer patients, clinical dissemination of the technique has been rapid; close to 10 sites worldwide now possess a polarizer fit for the clinic, and more than 30 clinical trials, predominantly for oncological applications, are already registered on the US and European clinical trials databases. Hyperpolarized ¹³C probes to study pathophysiological processes beyond the Warburg effect, including tricarboxylic acid cycle metabolism, intra-cellular pH and cellular necrosis have also been demonstrated in the preclinical arena and are pending clinical translation, and the simultaneous injection of multiple co-polarized agents is opening the door to high-sensitivity, multi-functional molecular MRI with a single dose. Here, we review the biomedical applications to date of the two polarization methods that have been used for in vivo hyperpolarized ¹³C molecular MRI; namely, dissolution dynamic nuclear polarization and parahydrogen-induced polarization. The basic concept of hyperpolarization and the fundamental theory underpinning these two key ¹³C hyperpolarization methods, along with recent technological advances that have facilitated biomedical realization, are also covered.

T₂-fluid-attenuated inversion recovery images (FLAIR) mismatch sign is now known to be a specific yet insensitive image feature for IDH-mutant, 1p19q non-codeleted astrocytoma. The current study revealed that lesion presenting T₂-FLAIR mismatch exhibited extremely long T₁- and T₂-relaxation time while T₂-FLAIR matched lesions showed low to moderate values. On the other hand, IDH-wildtype tumors presented noticeably short T₁- and T₂-relaxation time. These different relaxation time characteristics seemed to render T₂-FLAIR mismatch sign of becoming such a unique and specific image feature for IDH-mutant, 1p19q non-codeleted astrocytoma.

Purpose: Increased use of deep convolutional neural networks (DCNNs) in medical imaging diagnosis requires determinate evaluation of diagnostic performance. We performed the fundamental investigation of diagnostic performance of DCNNs using the detection task of brain metastasis.

Methods: We retrospectively investigated AlexNet and GoogLeNet using 3117 positive and 37961 negative MRI images with and without metastasis regarding (1) diagnostic biases, (2) the optimal K number of K-fold cross validations (K-CVs), (3) the optimal positive versus negative image ratio, (4) the accuracy improvement curves, (5) the accuracy range prediction by the bootstrap method, and (6) metastatic lesion detection by regions with CNNs (R-CNNs).

Results: Respectively, AlexNet and GoogLeNet had (1) 50 ± 4.6% and 50 ± 4.9% of the maximal mean ± 95% confidence intervals (95% CIs) measured with equal-sized negative versus negative image datasets and positive versus positive image datasets, (2) no less than 10 and 4 of K number in K-CVs fell within the respective maximum biases of 4.6% or 4.9%, (3) 74% of the highest accuracy with equal positive versus negative image ratio dataset and 91% of that with four times of negative-to-positive image ratio dataset, (4) the accuracy improvement curves increasing from 69% to 74% and 73% to 88% as positive versus negative pairs of the training images increased from 500 to 2495, (5) at least nine and six out of 10-CV result sets essential to predict the accuracy ranges by the bootstrap method, and (6) 50% and 45% of metastatic lesion detection accuracies by R-CNNs.

Conclusions: Our research presented methodological fundamentals to evaluate diagnostic features in the visual recognition of DCNNs. Our series will help to conduct the accuracy investigation of computer diagnosis in medical imaging.

Purpose: This study aimed to evaluate the effect of chemical exchange saturation transfer (CEST) on the ischemic regions in hypoxic-ischemic encephalopathy (HIE) in comparison with diffusion-weighted imaging (DWI) and magnetic resonance spectroscopy (MRS) using a 7T-MRI.

Methods: We used neonatal rats (n = 8), aged 8 days, to clarify the progression of HIE. The rat model of HIE was developed by ligating and severing the left common carotid artery, followed by 45 minutes of recovery, and 60 minutes of hypoxia (8% O₂/92% N₂; 34°C). At 0–2 and 24 hours after the onset of HIE, CEST imaging, DWI, and MRS were performed with a 7T-MRI. The magnetization transfer ratio (MTR) asymmetry curves and four MTR asymmetry maps at 0.5, 1.0, 2.0, and 3.5 ppm were calculated using the CEST images. Fractional anisotropy (FA) and apparent diffusion coefficient (ADC) maps were calculated by DWI, and brain metabolites were assessed by MRS.

Results: In the ischemic regions of neonatal rats, FA was significantly increased at 0–2 hours and decreased at 24 hours after the onset of HIE. ADC in the ipsilateral side was significantly lower than that of contralateral side. All rats with HIE showed hypointense areas on MTR asymmetry maps (2.0 and 3.5 ppm), that did not correspond with the hyperintense areas on DWI. In addition, a significant increase in lactate levels was observed at 0–2 and 24 hours after the onset of HIE.

Conclusion: CEST MTR maps did not correspond with the hyperintense areas on DWI at 0–2 and 24 hours after the onset of HIE. The change of multi offset CEST signal may be primarily related to the brain metabolites and pH alterations, such as that caused by lactate, after the onset of HIE.

Purpose: Idiopathic normal pressure hydrocephalus (iNPH) and Alzheimer’s disease (AD) are geriatric diseases and common causes of dementia. Recently, many studies on the segmentation, disease detection, or classification of MRI using deep learning have been conducted. The aim of this study was to differentiate iNPH and AD using a residual extraction approach in the deep learning method.

Methods: Twenty-three patients with iNPH, 23 patients with AD and 23 healthy controls were included in this study. All patients and volunteers underwent brain MRI with a 3T unit, and we used only whole-brain three-dimensional (3D) T₁-weighted images. We designed a fully automated, end-to-end 3D deep learning classifier to differentiate iNPH, AD and control. We evaluated the performance of our model using a leave-one-out cross-validation test. We also evaluated the validity of the result by visualizing important areas in the process of differentiating AD and iNPH on the original input image using the Gradient-weighted Class Activation Mapping (Grad-CAM) technique.

Results: Twenty-one out of 23 iNPH cases, 19 out of 23 AD cases and 22 out of 23 controls were correctly diagnosed. The accuracy was 0.90. In the Grad-CAM heat map, brain parenchyma surrounding the lateral ventricle was highlighted in about half of the iNPH cases that were successfully diagnosed. The medial temporal lobe or inferior horn of the lateral ventricle was highlighted in many successfully diagnosed cases of AD. About half of the successful cases showed nonspecific heat maps.

Conclusion: Residual extraction approach in a deep learning method achieved a high accuracy for the differential diagnosis of iNPH, AD, and healthy controls trained with a small number of cases.