RESUMEN
Purpose: We developed a method to visualize the image distortion induced by nonlinear noise reduction algorithms in computed tomography (CT) systems. Approach: Nonlinear distortion was defined as the induced residual when testing a reconstruction algorithm by the criteria for a linear system. Two types of images were developed: a nonlinear distortion of an object (NLDobject) image and a nonlinear distortion of noise (NLDnoise) image to visualize the nonlinear distortion induced by an algorithm. Calculation of the images requires access to the sinogram data, which is seldomly fully provided. Hence, an approximation of the NLDobject image was estimated. Using simulated CT acquisitions, four noise levels were added onto forward projected sinograms of a typical CT image; these were noise reduced using a median filter with the simultaneous iterative reconstruction technique or a total variation filter with the conjugate gradient least-squares algorithm. The linear reconstruction technique filtered back-projection was also analyzed for comparison. Results: Structures in the NLDobject image indicated contrast and resolution reduction of the nonlinear denoising. Although the approximated NLDobject image represented the original NLDobject image well, it had a higher random uncertainty. The NLDnoise image for the median filter indicated both stochastic variations and structures reminding of the object while for the total variation filter only stochastic variations were indicated. Conclusions: The developed images visualize nonlinear distortions of denoising algorithms. The object may be distorted by the noise and vice versa. Analyzing the distortion correlated to the object is more critical than analyzing a distortion of stochastic variations. The absence of nonlinear distortion may measure the robustness of the denoising algorithm.
RESUMEN
BACKGROUND: With the World Health Organization's (WHO) publication of the 2021-2030 neglected tropical diseases (NTDs) roadmap, the current gap in global diagnostics became painfully apparent. Improving existing diagnostic standards with state-of-the-art technology and artificial intelligence has the potential to close this gap. METHODOLOGY/PRINCIPAL FINDINGS: We prototyped an artificial intelligence-based digital pathology (AI-DP) device to explore automated scanning and detection of helminth eggs in stool prepared with the Kato-Katz (KK) technique, the current diagnostic standard for diagnosing soil-transmitted helminths (STHs; Ascaris lumbricoides, Trichuris trichiura and hookworms) and Schistosoma mansoni (SCH) infections. First, we embedded a prototype whole slide imaging scanner into field studies in Cambodia, Ethiopia, Kenya and Tanzania. With the scanner, over 300 KK stool thick smears were scanned, resulting in total of 7,780 field-of-view (FOV) images containing 16,990 annotated helminth eggs (Ascaris: 8,600; Trichuris: 4,083; hookworms: 3,623; SCH: 684). Around 90% of the annotated eggs were used to train a deep learning-based object detection model. From an unseen test set of 752 FOV images containing 1,671 manually verified STH and SCH eggs (the remaining 10% of annotated eggs), our trained object detection model extracted and classified helminth eggs from co-infected FOV images in KK stool thick smears, achieving a weighted average precision (± standard deviation) of 94.9% ± 0.8% and a weighted average recall of 96.1% ± 2.1% across all four helminth egg species. CONCLUSIONS/SIGNIFICANCE: We present a proof-of-concept for an AI-DP device for automated scanning and detection of helminth eggs in KK stool thick smears. We identified obstacles that need to be addressed before the diagnostic performance can be evaluated against the target product profiles for both STH and SCH. Given that these obstacles are primarily associated with the required hardware and scanning methodology, opposed to the feasibility of AI-based results, we are hopeful that this research can support the 2030 NTDs road map and eventually other poverty-related diseases for which microscopy is the diagnostic standard.
Asunto(s)
Helmintiasis , Helmintos , Ancylostomatoidea , Animales , Inteligencia Artificial , Ascaris lumbricoides , Heces/parasitología , Helmintiasis/diagnóstico , Helmintiasis/parasitología , Enfermedades Desatendidas/diagnóstico , Schistosoma mansoni , Suelo/parasitología , TrichurisRESUMEN
Denoising reconstruction techniques can introduce nonlinear properties into computed tomography (CT) systems. These nonlinear algorithms introduce distortion which affects the assessment of the resolution of the system. The purpose of the present study was to decouple and investigate amplitude modulation and waveform distortion in reconstruction algorithms in CT. The methodology developed by Wells, J. R. and Dobbins, J. T. III [Frequency response and distortion properties of nonlinear image processing algorithms and the importance of imaging context. Med. Phys. 40, 091906 (2013)] was adapted to CT reconstruction algorithms. The CT simulating program ASTRA Toolbox© for MATLAB™ was used for the reconstruction of the sinusoidal wave functions. Filtered back projection and the simultaneous iterative reconstruction technique were investigated with simple nonlinear mechanisms: a median filter and a non-negative constraint, respectively. The native reconstruction algorithms were not free from nonlinear waveform distortion, however, none of the metrics showed any dependence on the contrast-to-noise ratio (CNR). Furthermore, the algorithms including nonlinear mechanisms showed a clear and specific CNR dependence, indicating the necessity for distortion analysis in nonlinear CT reconstruction.
Asunto(s)
Algoritmos , Tomografía Computarizada por Rayos X , Procesamiento de Imagen Asistido por Computador , Fantasmas de Imagen , Dosis de Radiación , Interpretación de Imagen Radiográfica Asistida por ComputadorRESUMEN
The purpose of this study was to investigate the effect of adaptive statistical iterative reconstruction (ASiR) on the visualisation of anatomical structures and diagnostic image quality in paediatric cerebral computed tomography (CT) examinations. Forty paediatric patients undergoing routine cerebral CT were included in the study. The raw data from CT scans were reconstructed into stacks of 5 mm thick axial images at various levels of ASiR. Three paediatric radiologists rated six questions related to the visualisation of anatomical structures and one question on diagnostic image quality, in a blinded randomised visual grading study. The evaluated anatomical structures demonstrated enhanced visibility with increasing level of ASiR, apart from the cerebrospinal fluid space around the brain. In this study, 60 % ASiR was found to be the optimal level of ASiR for paediatric cerebral CT examinations. This shows that the commonly used 30 % ASiR may not always be the optimal level.
Asunto(s)
Encéfalo/anatomía & histología , Encéfalo/diagnóstico por imagen , Modelos Estadísticos , Exposición a la Radiación/prevención & control , Interpretación de Imagen Radiográfica Asistida por Computador/métodos , Tomografía Computarizada por Rayos X/métodos , Adolescente , Niño , Preescolar , Simulación por Computador , Interpretación Estadística de Datos , Femenino , Humanos , Lactante , Recién Nacido , Masculino , Variaciones Dependientes del Observador , Dosis de Radiación , Exposición a la Radiación/análisis , Protección Radiológica/métodos , Intensificación de Imagen Radiográfica/métodos , Reproducibilidad de los Resultados , Sensibilidad y Especificidad , Relación Señal-Ruido , Método Simple CiegoRESUMEN
The purpose of this study was to investigate the effect of different combinations of convolution kernel and the level of Adaptive Statistical iterative Reconstruction (ASiR™) on diagnostic image quality as well as visualisation of anatomical structures in paediatric abdominal computed tomography (CT) examinations. Thirty-five paediatric patients with abdominal pain with non-specified pathology undergoing abdominal CT were included in the study. Transaxial stacks of 5-mm-thick images were retrospectively reconstructed at various ASiR levels, in combination with three convolution kernels. Four paediatric radiologists rated the diagnostic image quality and the delineation of six anatomical structures in a blinded randomised visual grading study. Image quality at a given ASiR level was found to be dependent on the kernel, and a more edge-enhancing kernel benefitted from a higher ASiR level. An ASiR level of 70 % together with the Soft™ or Standard™ kernel was suggested to be the optimal combination for paediatric abdominal CT examinations.