dSPIC: a deep SPECT image classification network for automated multi-disease, multi-lesion diagnosis

Background Functional imaging especially the SPECT bone scintigraphy has been accepted as the effective clinical tool for diagnosis, treatment, evaluation, and prevention of various diseases including metastasis. However, SPECT imaging is brightly characterized by poor resolution, low signal-to-noise ratio, as well as the high sensitivity and low specificity because of the visually similar characteristics of lesions between diseases on imaging findings. Methods Focusing on the automated diagnosis of diseases with whole-body SPECT scintigraphic images, in this work, a self-defined convolutional neural network is developed to survey the presence or absence of diseases of concern. The data preprocessing mainly including data augmentation is first conducted to cope with the problem of limited samples of SPECT images by applying the geometric transformation operations and generative adversarial network techniques on the original SPECT imaging data. An end-to-end deep SPECT image classification network named dSPIC is developed to extract the optimal features from images and then to classify these images into classes, including metastasis, arthritis, and normal, where there may be multiple diseases existing in a single image. Results A group of real-world data of whole-body SPECT images is used to evaluate the self-defined network, obtaining a best (worst) value of 0.7747 (0.6910), 0.7883 (0.7407), 0.7863 (0.6956), 0.8820 (0.8273) and 0.7860 (0.7230) for accuracy, precision, sensitivity, specificity, and F-1 score, respectively, on the testing samples from the original and augmented datasets. Conclusions The prominent classification performance in contrast to other related deep classifiers including the classical AlexNet network demonstrates that the built deep network dSPIC is workable and promising for the multi-disease, multi-lesion classification task of whole-body SPECT bone scintigraphy images.

Synthesis and Evaluation of Novel Norfloxacin Isonitrile 99mTc Complexes as Potential Bacterial Infection Imaging Agents

To develop potential technetium-99m single-photon emission computed tomography (SPECT) imaging agents for bacterial infection imaging, the novel norfloxacin isonitrile derivatives CN4NF and CN5NF were synthesized and radiolabeled with a [99mTc][Tc(I)]+ core to obtain [99mTc]Tc-CN4NF and [99mTc]Tc-CN5NF. These compounds were produced in high radiolabeling yields and showed hydrophilicity and good stability in vitro. The bacterial binding assay indicated that [99mTc]Tc-CN4NF and [99mTc]Tc-CN5NF were specific to bacteria. Compared with [99mTc]Tc-CN4NF, biodistribution studies of [99mTc]Tc-CN5NF showed a higher uptake in bacteria-infected tissues than in turpentine-induced abscesses, indicating that [99mTc]Tc-CN5NF could distinguish bacterial infection from sterile inflammation. In addition, [99mTc]Tc-CN5NF had higher abscess/blood and abscess/muscle ratios. SPECT image of [99mTc]Tc-CN5NF showed that there was a clear accumulation in the infection site, suggesting that it could be a potential bacterial infection imaging radiotracer.

Fast algorithm for the determination of the optimum filter for bone SPECT image reconstruction

In SPECT, the reconstructed images are strongly affected by poisson noise, poor spatial resolution and bad contrast due to the radioactivity disintegration and procedures acquisition. In this paper, we propose an algorithm to improve the traditional FBP reconstruction and to choose the most suitable technique for bone SPECT image denoising. The proposed approach is composed of two steps. The first one consists of denoising the acquired sinograms using successively eight currently used filters in nuclear medicine: Wiener, Metz, Hamming, Hann, Shepp-Logan, Parzen, Butterworth and Gaussian combined with Butterworth filters. The second step is a simultaneous reconstruction of the axial slices using a new 3D FBP algorithm for each filter. A comparative study of these filters is tested and evaluated on a dataset containing thirty one bone SPECT image. The results show that the difference between these filters is statistically significantly different from each other (p<0.05) and the 3D FBP with the combination between Butterworth and Gaussian provide the best performance. The selected method is compared to three denoising methods. These methods are tested on a Shepp Logan phantom and bone SPECT images. Experimental results show that the 3D FBP reconstruction with the pre-processing combination (Gaussian (Std=0.3) + Butterworth (fc=0.47, ordre=3)) filter is more accurate and robust compared to other methods. It provides the highest performance in term of contrast, SNR, CNR ensuring a shorter processing time. It accelerates the reconstruction, reduces noise and artifacts while preserving detailed features. This approach could be considered as a valuable candidate to enhance the quality of the reconstructed bone SPECT image.

The usefulness of SwiftScan technology for bone scintigraphy using a novel anthropomorphic phantom

The aim of this study was to demonstrate the usefulness of SwiftScan with a low-energy high-resolution and sensitivity (LEHRS) collimator for bone scintigraphy using a novel bone phantom simulating the human body. SwiftScan planar image of lateral view was acquired in clinical condition; thereafter, each planar image of different blend ratio (0–80%) of Crality 2D processing were created. SwiftScan planar images with reduced acquisition time by 25–75% were created by Poisson’s resampling processing. SwiftScan single photon emission computed tomography (SPECT) was acquired with step-and-shoot and continuous mode, and SPECT images were reconstructed using a three-dimensional ordered subset expectation maximization incorporating attenuation, scatter and spatial resolution corrections. SwiftScan planar image showed a high contrast to noise ratio (CNR) and low percent of the coefficient of variance (%CV) compared with conventional planar image. The CNR of the tumor parts in SwiftScan SPECT was higher than that of the conventional SPECT image of step and shoot acquisition, while the %CV showed the lowest value in all systems. In conclusion, SwiftScan planar and SPECT images were able to reduce the image noise compared with planar and SPECT image with a low-energy high-resolution collimator, so that SwiftScan planar and SPECT images could be obtained a high CNR. Furthermore, the SwiftScan planar image was able to reduce the acquisition time by 25% when the blend ratio of Clarity 2D processing set to more than 40%.