A self-tuned graph-based framework for localization and grading prostate cancer lesions: An initial evaluation based on multiparametric magnetic resonance imaging

Multimodal Image Classification by Multiview Latent Pattern Extraction, Selection, and Correlation

The large amount of data available in the modern big data era opens new opportunities to expand our knowledge by integrating information from heterogeneous sources. Multiview learning has recently achieved tremendous success in deriving complementary information from multiple data modalities. This article proposes a framework called multiview latent space projection (MVLSP) to integrate features extracted from multiple sources in a discriminative way to facilitate binary and multiclass classifications. Our approach is associated with three innovations. First, most existing multiview learning algorithms promote pairwise consistency between two views and do not have a natural extension to applications with more than two views. MVLSP finds optimum mappings from a common latent space to match the feature space in each of the views. As the matching is performed on a view-by-view basis, the framework can be readily extended to multiview applications. Second, feature selection in the common latent space can be readily achieved by adding a class view, which matches the latent space representations of training samples with their corresponding labels. Then, high-order view correlations are extracted by considering feature-label correlations. Third, a technique is proposed to optimize the integration of different latent patterns based on their correlations. The experimental results on the prostate image dataset demonstrate the effectiveness of the proposed method.

Pulmonary artery stiffness in fetuses with tetralogy of Fallot

PURPOSE

This study evaluated the elastic characteristics of the pulmonary trunk and distal branches in fetuses diagnosed with tetralogy of Fallot (TOF) using Doppler echocardiography.

METHODS

Data on 42 fetuses diagnosed with TOF and 84 gestational age-matched normal fetuses were prospectively collected from the Second Xiangya Hospital of Central South University between August 2022 and January 2023. The severity of TOF was classified into three categories based on the z-score of the pulmonary annulus diameter: mild (z-score ≥-2), moderate (-4

RESULTS

Compared with the normal group, the MPA-PAS in fetuses with TOF was significantly higher, while the DLPA-PAS and DRPA-PAS were notably lower (all P<0.05). The MPA-PAS of fetuses with severe TOF was higher than that of those with mild and moderate TOF (all P<0.05). However, there were no significant differences in the DLPA-PAS or DRPA-PAS among fetuses with mild, moderate, and severe TOF (all P>0.05).

CONCLUSION

Fetuses diagnosed with TOF exhibited increased vascular stiffness in the MPA and reduced stiffness in the distal pulmonary artery (PA). Larger-scale follow-up studies are required to elucidate the relationships between these changes in vascular stiffness and PA development in patients with TOF.

Collapse

Key Words

• Fetus
• Pulmonary arterial elasticity
• Pulmonary arterial stiffness
• Tetralogy of Fallot
• Ultrasonic cardiography

Collapse

MESH Headings Collapse

Grants Collapse

Affiliation(s)

Yushan Liu Department of Ultrasound Diagnosis, Second Xiangya Hospital, Central South University, Changsha, China
Ran Xu Department of Urology, Second Xiangya Hospital, Central South University, Changsha, China
Dan Zhou Department of Ultrasound Diagnosis, Second Xiangya Hospital, Central South University, Changsha, China
Yang Yang Department of Ultrasound Diagnosis, Second Xiangya Hospital, Central South University, Changsha, China
Ganqiong Xu Department of Ultrasound Diagnosis, Second Xiangya Hospital, Central South University, Changsha, China
Shi Zeng Department of Ultrasound Diagnosis, Second Xiangya Hospital, Central South University, Changsha, China

Collapse

Brain tissues have single-voxel signatures in multi-spectral MRI

Since the seminal works by Brodmann and contemporaries, it is well-known that different brain regions exhibit unique cytoarchitectonic and myeloarchitectonic features. Transferring the approach of classifying brain tissues - and other tissues - based on their intrinsic features to the realm of magnetic resonance (MR) is a longstanding endeavor. In the 1990s, atlas-based segmentation replaced earlier multi-spectral classification approaches because of the large overlap between the class distributions. Here, we explored the feasibility of performing global brain classification based on intrinsic MR features, and used several technological advances: ultra-high field MRI, q-space trajectory diffusion imaging revealing voxel-intrinsic diffusion properties, chemical exchange saturation transfer and semi-solid magnetization transfer imaging as a marker of myelination and neurochemistry, and current neural network architectures to analyze the data. In particular, we used the raw image data as well to increase the number of input features. We found that a global brain classification of roughly 97 brain regions was feasible with gross classification accuracy of 60%; and that mapping from voxel-intrinsic MR data to the brain region to which the data belongs is possible. This indicates the presence of unique MR signals of different brain regions, similar to their cytoarchitectonic and myeloarchitectonic fingerprints.

Algorithms applied to spatially registered multi-parametric MRI for prostate tumor volume measurement

Background

Prostate tumor volume correlates with critical components of cancer staging such as Gleason score (GS) grade, predicted disease progression, and metastasis. Therefore, non-invasive tumor volume measurement may elevate clinical management. Radiology assessments of multi-parametric MRI (MP-MRI) commonly visually examine individual images to determine possible tumor presence. This study combines registered MP-MRI into a single image that display normal tissue and possible lesions. This study tests and exploits the vector nature of spatially registered MP-MRI by using supervised target detection algorithms (STDA) and color display and psychovisual analysis (CIELAB) to non-invasively estimate prostate tumor volume.

Methods

MRI, including T1, T2, diffusion [apparent diffusion coefficient (ADC)], dynamic contrast enhanced (DCE) images, were resampled, rescaled, translated, and stitched to form spatially registered Multi-parametric cubes. The multi-parametric or multi-spectral signatures (7-component or T1, T2, ADC, etc.) that characterize the prostate tumors were inserted into target detection algorithms with conical decision surfaces (adaptive cosine estimator, ACE). Various detection thresholds were applied to discriminate tumor from normal tissue. In addition, tumor appeared as yellow in color images that were created by assigning red to washout from DCE, green to high B from diffusion, and blue to autonomous diffusion image. The yellow voxels in the three-channel hypercube were visually identified by a reader and recording voxels that exceed a threshold in the b* component of the CIELAB algorithm. The number of reported tumor voxels were converted to volume based on spatial resolution and slice separation. The tumor volume measurements were quantitatively validated by comparing the tumor volume computations to the pathologist's assessment of the histology of sectioned whole mount prostates from 26 consecutive patients with prostate adenocarcinoma who underwent radical prostatectomy. This study analyzed tumors exceeding 1 cc and that also took up contrast material (18 patients).

Results

High correlation coefficients for tumor volume measurements using supervised target detection and color analysis vs. histology from wholemount prostatectomy were computed (R=0.83 and 0.91, respectively). A linear fit for tumor volume measurements using for supervised target detection and color analysis vs. tumor measurements from radical prostatectomy (after correcting for shrinkage from the radical prostatectomy) results in a slope of 1.02 and 3.02, respectively. A polynomial fit for the color analysis to the histology found (R=0.95). Voxels exceeding a threshold in the b* part of the CIELAB algorithm yielded correlation coefficients (0.71, 0.80) offsets (0.01 cc, -0.63 cc) and slopes (1.99, 0.89) against the wholemount prostatectomy and color analysis, respectively.

Conclusions

Supervised target detection and color display and analysis applied to registered MP-MRI non-invasively estimates prostate tumor volumes >1 cc and displaying angiogenesis.