An advanced approach to reconstruct CT images from limited-angle projections, reducing radiation dose and tube load.

The process of reconstructing CT scan images from limited angle projections is critical and requires strict adherence to the ALARA principle. This principle is designed to minimize radiation exposure while maintaining image quality. Our study utilized filter back-projection (FBP) and algebraic iterative reconstruction (IR) algorithms to reconstruct brain CT images from 200 projection lines and a 100 × 100 matrix size. By combining the results of a MATLAB function with the insights of a radiologist, we can produce high-quality images that decrease radiation dose and tube load. Our findings reveal that the algebraic method is superior to the filter back-projection in preserving image quality when utilizing limited-angle projections.


1.Introduction
Radiation is a type of energy that can penetrate and interact with matter.There are two categories of radiation: ionizing and non-ionizing.Ionizing radiation can further be classified into two types: electromagnetic (EM) radiation and particle radiation.Electromagnetic radiation, such as X-rays and gamma rays, has high energy that can easily and non-invasively penetrate biological tissue.This means it is categorized as ionizing radiation and is primarily used for diagnostic and treatment purposes, replacing surgery in minimally invasive procedures.[1].
Since the second revolution in the field of diagnostic imaging, the invention of the Computed Tomography (CT) scan in 1972, it has become a primary diagnostic modality [2] .Over time, the CT application has developed from spiral scanning in 1980 to the advent of multidetector-row technology in 1990 [3].The applications of CT have continued to increase dramatically with significant improvements in quality, resolution, accuracy, and speed [4].Consequently, it has become possible to replace surgical diagnostic procedures with the CT scan, which is a rapid, accurate diagnostic modality that can be performed quickly and safely.This is due to its fast scanning speed and isotropic spatial resolution at 0.3-0.4mm.[5] .
CT scans offer valuable diagnostic information and aid in patient management, but they also come with potential risks.Research suggests that CT scans can increase the risk of malignancy, and they contribute significantly to a patient's exposure to medical radiation.As a result, reducing radiation dose in CT scans is a critical area of focus in the medical field.
When it comes to CT imaging for diagnostic purposes, the radiation dose plays a critical role in determining the quality and accuracy of the images produced.It's essential to maintain image quality while minimizing radiation exposure.By understanding the relationship between image quality and CT dose reduction, we can determine the optimal balance between the two.To achieve this balance, there are two approaches.The first is to establish the appropriate target image qualityfor the specific diagnostic task at hand.The second approach is to enhance certain imagequality factors, such asreducing noise, which allows us to decrease radiation dose or improve CNR, SNR, and other aspects that contribute to superior image quality [6].
Significant progress has been made in minimizing radiation exposure while improving dose efficiency [7].To achieve this goal, efforts have been focused on optimizing various components of the CT system, including the collimator, beam-shaping filter, and detector.Additionally, advancements have been made in scan range, automatic exposure control, dual energy CT, and logarithms.Among these, the most significant is Iterative Reconstruction (IR) logarithms [6][8]The purpose of this study is to use MatLab 2016 to generate CT images from limited projection reconstruction at specific angles while adhering to the ALARA principle, which aims to minimize radiation exposure while preserving image quality to the greatest extent possible.

Material and Method
Within this research endeavor, algebraic functions in MATLAB 2016 were utilized on an HP CORE i5 PC to construct new CT images and plots in diagnostic quality.Data was imported from Siemens and Philips CT scanners in the form of CT DICOM images onto the researchers' personal computer.The MATLAB software was subsequently employed to form a 100x100 pixel matrix (Figure 2) from the original image (Figure 1).This matrix was then applied to create restricted reconstruction images through the application of two distinct reconstruction algorithms: Filter Back-Projection and algebraic iterative reconstruction (IR) algorithms.A series of CT images were generated using varying angles of limited projection, including 45 degrees, 90 degrees, and 180 degrees, from the original images.Each set comprised 200 projection lines and a matrix size of 100 x 100 pixels.Both aforementioned reconstruction algorithms were algorithms were employed to produce the final images, which were then evaluated against specific quality criteria.

3.Result and Discussion
Figures 3 and 4 provide a comprehensive comparison of two distinct methods for reconstructing from limited angles.These visual aids demonstrate how each method performs with varying numbers of angles.Thoroughly examining the data presented in these figures will allow for a more comprehensive understanding of the advantages and drawbacks of each approach.The second image (B) was created using 90 angles projections at the same matrix size in MATLAB.Through a meticulous examination of quality standards by a proficient radiologist and an in-depth analysis of MATLAB results, it has been determined that generating CT images using limited angle projections is a viable option.This innovative technique can effectively minimize patient radiation exposure and tube loading.
In the realm of reconstructing images from limited angle projections while preserving their quality, algebraic algebraic iterative reconstruction (IR) algorithms outperforms filter back-projection (FBP).FBP fails to provide any meaningful insights when working with angles of 45, 90, and 180 degrees.In contrast, the algebraic method significantly improves image quality, which was confirmed by Yu and Zeng's study in 2015.
It's important to note that using a 100x100-pixel matrix size may not provide accurate diagnoses due to the limited capabilities of the researcher's computer.The matrix size is a crucial factor in image quality, as demonstrated by Haney Alsleem's research.A larger matrix size preserves spatial resolution and enhances image quality, although there may be an increase in image noise compared to a 512 matrix size, as noted in Hata et al.'s 2018 study.

4.Conclusion
This study marks a significant milestone in our country's efforts to generate brain CT images from limited angle projections and assess the quality criteria for algebraic and FBP methods.The findings indicate that producing diagnostic images from limited angle projections can be accomplished through the use of more sophisticated computers that can handle larger matrix sizes -a promising breakthrough.The data gathered during this study is a valuable source for researchers who are striving to reduce radiation exposure to patients during CT scans while maintaining image quality and minimizing tube loading.

Figure 1 .
Figure 1.Original brain CT scan with IDose2 reconstruction and 256 matrix size.

Figure 2 .
Figure 2.shows the new CT scan image with a 100 × 100 matrix size created using MATLAB.

Figure 3 .
Figure 3. shows two images of a brain CT scan created using Filter Back-Projection (FBP).The first image (A) was created using 45 angles projections at a matrix size of 100 x 100 in MATLAB.

Figure 4 .
Figure 4. a comparison between the algebraic iterative reconstruction (IR) algorithms and FBP methods.(A) A brain CT image reconstructed from a Limited Angel with a 100 × 100 matrix size and 200 lines of projection using the algebraic iterative reconstruction (IR) algorithms from 90 degree via MATLAB.(B) shows the FBP reconstructed image from 90 angles with a 100 × 100 matrix size.