Morphology filter bank for extracting nodular and linear patterns in medical images

Automated Detection of Cerebral Microbleeds on Two-dimensional Gradient-recalled Echo T2* Weighted Images Using a Morphology Filter Bank and Convolutional Neural Network

PURPOSE

We present a novel algorithm for the automated detection of cerebral microbleeds (CMBs) on 2D gradient-recalled echo T2* weighted images (T2*WIs). This approach combines a morphology filter bank with a convolutional neural network (CNN) to improve the efficiency of CMB detection. A technical evaluation was performed to ascertain the algorithm's accuracy.

METHODS

In this retrospective study, 60 patients with CMBs on T2*WIs were included. The gold standard was set by three neuroradiologists based on the Microbleed Anatomic Rating Scale guidelines. Images with CMBs were extracted from the training dataset comprising 30 cases using a morphology filter bank, and false positives (FPs) were removed based on the threshold of size and signal intensity. The extracted images were used to train the CNN (Vgg16). To determine the effectiveness of the morphology filter bank, the outcomes of the following two methods for detecting CMBs from the 30-case test dataset were compared: (a) employing the morphology filter bank and additional FP removal and (b) comprehensive detection without filters. The trained CNN processed both sets of initial CMB candidates, and the final CMB candidates were compared with the gold standard. The sensitivity and FPs per patient of both methods were compared.

RESULTS

After CNN processing, the morphology-filter-bank-based method had a 95.0% sensitivity with 4.37 FPs per patient. In contrast, the comprehensive method had a 97.5% sensitivity with 25.87 FPs per patient.

CONCLUSION

Through effective CMB candidate refinement with a morphology filter bank and FP removal with a CNN, we achieved a high CMB detection rate and low FP count. Combining a CNN and morphology filter bank may facilitate the accurate automated detection of CMBs on T2*WIs.

Simultaneous Arterial and Venous Imaging Using 3D Quantitative Parameter Mapping

PURPOSE

To increase the number of images that can be acquired in MR examinations using quantitative parameters, we developed a method for obtaining arterial and venous images with mapping of proton density (PD), RF inhomogeneity (B1), longitudinal relaxation time (T1), apparent transverse relaxation time (T2*), and magnetic susceptibility through calculation, all with the same spatial resolution.

METHODS

The proposed method uses partially RF-spoiled gradient echo sequences to obtain 3D images of a subject with multiple scan parameters. The PD, B1, T1, T2*, and magnetic susceptibility maps are estimated using the quantification method we previously developed. Arterial images are obtained by adding images using optimized weights to emphasize the arteries. A morphology filter is used to obtain venous images from the magnetic susceptibility maps. For evaluation, images obtained from four out of five healthy volunteers were used to optimize the weights used in the arterial-image calculation, and the optimized weights were applied to the images from the fifth volunteer to obtain an arterial image. Arterial images of the five volunteers were calculated using the leave-one-out method, and the contrast between the arterial and background regions defined using the reference time-of-flight (TOF) method was evaluated using the area under the receiver operation characteristic curve (AUC). The contrast between venous and background regions defined by a reference quantitative susceptibility mapping (QSM) method was also evaluated for the venous image.

RESULTS

The AUC to discriminate blood vessels and background using the proposed method was 0.905 for the arterial image and 0.920 for the venous image.

CONCLUSION

The results indicate that the arterial images and venous images have high signal intensity at the same region as determined from the reference TOF and QSM methods, demonstrating the possibility of acquiring vasculature images with quantitative parameter mapping through calculation in an integrated manner.