The Integration of Deep Learning in Radiotherapy: Exploring Challenges, Opportunities, and Future Directions through an Umbrella Review

This study investigates, through a narrative review, the transformative impact of deep learning (DL) in the field of radiotherapy, particularly in light of the accelerated developments prompted by the COVID-19 pandemic. The proposed approach was based on an umbrella review following a standard narrative checklist and a qualification process. The selection process identified 19 systematic review studies. Through an analysis of current research, the study highlights the revolutionary potential of DL algorithms in optimizing treatment planning, image analysis, and patient outcome prediction in radiotherapy. It underscores the necessity of further exploration into specific research areas to unlock the full capabilities of DL technology. Moreover, the study emphasizes the intricate interplay between digital radiology and radiotherapy, revealing how advancements in one field can significantly influence the other. This interdependence is crucial for addressing complex challenges and advancing the integration of cutting-edge technologies into clinical practice. Collaborative efforts among researchers, clinicians, and regulatory bodies are deemed essential to effectively navigate the evolving landscape of DL in radiotherapy. By fostering interdisciplinary collaborations and conducting thorough investigations, stakeholders can fully leverage the transformative power of DL to enhance patient care and refine therapeutic strategies. Ultimately, this promises to usher in a new era of personalized and optimized radiotherapy treatment for improved patient outcomes.

Comparative morphofunctional analysis of axial skeleton excluding the skull of primates based on the anatomical, radiographic, and tomographic description of the black-striped capuchin (Sapajus libidinosus Spix, 1823)

Comparative anatomy the basis for studies of evolution, and radiographic and tomographic aspects, as auxiliary methods in the investigation of anatomical particularities, reinforce evolutionary research. Therefore, the aim of this study was to describe the vertebrae, sternum, and ribs of the capuchin monkey (Sapajus libidinosus) by means of anatomical dissection and radiographic and tomographic images. To this purpose, four cadavers were used in the anatomical analysis and five living animals for the imaging exams. The bones were described and compared with data from other primates species found in literature. Student's t-test for independent samples was performed. The vertebral column of the comprises seven cervical, 13 or 14 thoracic, five or six lumbar, two or three sacral, and 23 or 24 caudal vertebrae. The atlas is characterized by three foramina on the wing. The seventh cervical vertebra had a transverse foramen in one specimen. The anticlinal vertebra is always the penultimate thoracic one, the ninth pair of ribs is always the last sternal pair, and the last two are buoyant. The sternal was composed of five or six sternebrae. The lumbar vertebrae showed a bifurcated spinous process. Three different sacral morphologies were observed. The structures identified macroscopically could be well determined through radiographic and tomographic images. S. libidinosus presented anatomical characteristics more similar to those of man and of platirrinos monkeys. The knowledge obtained by macroscopic anatomy and tomographic and radiological exams contributes significantly to comparative evolutionary studies.

The Regulation of Artificial Intelligence in Digital Radiology in the Scientific Literature: A Narrative Review of Reviews

Today, there is growing interest in artificial intelligence (AI) in the field of digital radiology (DR). This is also due to the push that has been applied in this sector due to the pandemic. Many studies are devoted to the challenges of integration in the health domain. One of the most important challenges is that of regulations. This study conducted a narrative review of reviews on the international approach to the regulation of AI in DR. The design of the study was based on: (I) An overview on Scopus and Pubmed (II) A qualification and eligibility process based on a standardized checklist and a scoring system. The results have highlighted an international approach to the regulation of these systems classified as "software as medical devices (SaMD)" arranged into: ethical issues, international regulatory framework, and bottlenecks of the legal issues. Several recommendations emerge from the analysis. They are all based on fundamental pillars: (a) The need to overcome a differentiated approach between countries. (b) The need for greater transparency and publicity of information both for SaMDs as a whole and for the algorithms and test patterns. (c) The need for an interdisciplinary approach that avoids bias (including demographic) in algorithms and test data. (d) The need to reduce some limits/gaps of the scientific literature production that do not cover the international approach.

Osteology applied to image diagnosis of the forelimb of the black-striped capuchin (Sapajus libidinosus Spix, 1823)

Macroscopic bone analysis and evaluation through imaging methods are essential in the recognition of natural and altered structures. Therefore, this study aimed at describing the osteology the thoracic limb of Sapajus libidinosus in bone pieces, identifying them in radiographic and tomographic images. For this, four cadavers were used in the macroscopic analysis and five animals for the imaging exams, of which four were euthanized and added to the macroscopic stage. For imaging exams, the animals were kept anesthetized. All bones were documented, structures described, and compared with literature data from human and nonhuman primates. There was no statistical difference between males and females regarding the length of the forelimb bones. Most of the bone structures of the scapula were well identified in the imaging methods, being more restricted in the ventrodorsal projection. The clavicle presented very limited visualization. The humerus, as well as the radius and ulna, were not well portrayed in their proximal and distal epiphyses by radiography. However, they were well identified on tomography. All structures described in the macroscopic image of the carpal and metacarpal bones could be identified through radiography and tomography, and the radiographic examination of this region is an excellent method for identifying fractures. The glenoid notch of the scapula was not visualized by any imaging method. S. libidinosus presented anatomical characteristics more similar to those of neotropical primates and man, being a great indicator of an experimental model for studies in these species.

Anatomy applied to image diagnosis of the hind limb in the black-striped capuchin (Sapajus libidinosus Spix, 1823)

The knowledge of anatomy and imaging exams emerges as an important tool in the study of evolutionary processes of a species, in the elaboration of diagnosis, and the successful choice of the appropriate clinical and surgical procedures. Therefore, this study aims to describe the osteology of the hind limb of Sapajus libidinosus by means of gross, radiographic, and tomographic images. Four cadavers were used in the macroscopic analysis and five animals for the imaging exams, of which four were eventually euthanized and added to the macroscopic study. For imaging exams, they were kept anesthetized. All bones of the hind limb were documented, their structures were described, and compared with data in the literature from human and nonhuman primates. We have performed Student's t test for independent samples. There was no statistical difference between the sexes regarding the length of the hind limb bones. The coxal bone was largely well described using imaging methods. A small penile bone was present at the tip of the penis and it could be identified by all analysis methods. The femur, as well as the tibia and fibula, were not well portrayed in their proximal and distal epiphyses by radiography (Rx). However, they were well identified on tomography. No third trochanter was observed in the femur and the patella had a triangular shape. All the structures described by gross anatomy of the tarsus and metatarsus could be identified by Rx and tomography. More subtle structures, such as the popliteal notch on the tibia, and the gluteal tuberosity pectineal line and facies aspera on the coxal bone, were not identified by medical imaging. S. libidinosus presented anatomical characteristics that were similar to those of larger New World and Old World monkeys, including man. This suggests it's value as an experimental model for studies in recent primates.

The Artificial Intelligence in Digital Radiology: Part 1: The Challenges, Acceptance and Consensus

Artificial intelligence is having important developments in the world of digital radiology also thanks to the boost given to the research sector by the COVID-19 pandemic. In the last two years, there was an important development of studies focused on both challenges and acceptance and consensus in the field of Artificial Intelligence. The challenges and acceptance and consensus are two strategic aspects in the development and integration of technologies in the health domain. The study conducted two narrative reviews by means of two parallel points of view to take stock both on the ongoing challenges and on initiatives conducted to face the acceptance and consensus in this area. The methodology of the review was based on: (I) search of PubMed and Scopus and (II) an eligibility assessment, using parameters with 5 levels of score. The results have: (a) highlighted and categorized the important challenges in place. (b) Illustrated the different types of studies conducted through original questionnaires. The study suggests for future research based on questionnaires a better calibration and inclusion of the challenges in place together with validation and administration paths at an international level.

A deep learning approach to dental restoration classification from bitewing and periapical radiographs

OBJECTIVE

The aim of this study was to examine the success of deep learning-based convolutional neural networks (CNN) in the detection and differentiation of amalgam, composite resin, and metal-ceramic restorations from bitewing and periapical radiographs.

METHOD AND MATERIALS

Five hundred and fifty bitewing and periapical radiographs were used. Eighty percent of the images were used for training, and 20% were left for testing. Twenty percent of the images allocated for training were then used for validation during learning. The image classification model was based on the application of CNN. The model used Resnet34 architecture, which is pre-trained on the ImageNet dataset. Average sensitivity, receiver operating characteristic (ROC) curve, and area under the curve (AUC) were calculated for performance evaluation of the model.

RESULTS

The model training loss was 0.13, and the validation loss was 0.63. The independent test group result was 0.67. Amalgam AUC was 0.95, composite AUC was 0.95, and metal-ceramic AUC was 1.00. The average AUC was 0.97. The false positive rate in the validation set was 18, the false negative rate was 18, the true positive rate was 60, and the true negative rate was 138. The true positive rate was 0.82 for amalgam, 0.75 for composite, and 0.73 for metal-ceramic.

CONCLUSION

Deep learning-based CNNs from periapical and bitewing radiographs appear to be a promising technique for the detection and differentiation of restorations.

Lessons from the COVID-19 Pandemic on the Use of Artificial Intelligence in Digital Radiology: The Submission of a Survey to Investigate the Opinion of Insiders

The development of artificial intelligence (AI) during the COVID-19 pandemic is there for all to see, and has undoubtedly mainly concerned the activities of digital radiology. Nevertheless, the strong perception in the research and clinical application environment is that AI in radiology is like a hammer in search of a nail. Notable developments and opportunities do not seem to be combined, now, in the time of the COVID-19 pandemic, with a stable, effective, and concrete use in clinical routine; the use of AI often seems limited to use in research applications. This study considers the future perceived integration of AI with digital radiology after the COVID-19 pandemic and proposes a methodology that, by means of a wide interaction of the involved actors, allows a positioning exercise for acceptance evaluation using a general purpose electronic survey. The methodology was tested on a first category of professionals, the medical radiology technicians (MRT), and allowed to (i) collect their impressions on the issue in a structured way, and (ii) collect their suggestions and their comments in order to create a specific tool for this professional figure to be used in scientific societies. This study is useful for the stakeholders in the field, and yielded several noteworthy observations, among them (iii) the perception of great development in thoracic radiography and CT, but a loss of opportunity in integration with non-radiological technologies; (iv) the belief that it is appropriate to invest in training and infrastructure dedicated to AI; and (v) the widespread idea that AI can become a strong complementary tool to human activity. From a general point of view, the study is a clear invitation to face the last yard of AI in digital radiology, a last yard that depends a lot on the opinion and the ability to accept these technologies by the operators of digital radiology.

Detection of Pulmonary Nodule Growth with Chest Tomosynthesis: A Human Observer Study Using Simulated Nodules

RATIONALE AND OBJECTIVES

Chest tomosynthesis has been suggested as a suitable alternative to CT for follow-up of pulmonary nodules. The aim of the present study was to investigate the possibility of detecting pulmonary nodule growth using chest tomosynthesis.

MATERIALS AND METHODS

Simulated nodules with volumes of approximately 100 mm³ and 300 mm³ as well as additional versions with increasing volumes were created. The nodules were inserted into images from pairs of chest tomosynthesis examinations, simulating cases where the nodule had either remained stable in size or increased in size between the two imaging occasions. Nodule volume growths ranging from 11% to 252% were included. A simulated dose reduction was applied to a subset of the cases. Cases differing in terms of nodule size, dose level, and nodule position relative to the plane of image reconstruction were included. Observers rated their confidence that the nodules were stable in size or not. The rating data for the nodules that were stable in size was compared to the rating data for the nodules simulated to have increased in size using ROC analysis.

RESULTS

Area under the curve values ranging from 0.65 to 1 were found. The lowest area under the curve values were found when there was a mismatch in nodule position relative to the reconstructed image plane between the two examinations. Nodule size and dose level affected the results.

CONCLUSION

The study indicates that chest tomosynthesis can be used to detect pulmonary nodule growth. Nodule size, dose level, and mismatch in position relative to the image reconstruction plane in the baseline and follow-up examination may affect the precision.

Characterization and validation of the thorax phantom Lungman for dose assessment in chest radiography optimization studies

This work concerns the validation of the Kyoto-Kagaku thorax anthropomorphic phantom Lungman for use in chest radiography optimization. The equivalence in terms of polymethyl methacrylate (PMMA) was established for the lung and mediastinum regions of the phantom. Patient chest examination data acquired under automatic exposure control were collated over a 2-year period for a standard x-ray room. Parameters surveyed included exposure index, air kerma area product, and exposure time, which were compared with Lungman values. Finally, a voxel model was developed by segmenting computed tomography images of the phantom and implemented in PENELOPE/penEasy Monte Carlo code to compare phantom tissue-equivalent materials with materials from ICRP Publication 89 in terms of organ dose. PMMA equivalence varied depending on tube voltage, from 9.5 to 10.0 cm and from 13.5 to 13.7 cm, for the lungs and mediastinum regions, respectively. For the survey, close agreement was found between the phantom and the patients' median values (deviations lay between 8% and 14%). Differences in lung doses, an important organ for optimization in chest radiography, were below 13% when comparing the use of phantom tissue-equivalent materials versus ICRP materials. The study confirms the value of the Lungman for chest optimization studies.

Comparison of the diagnostic accuracy of direct digital radiography system, filtered images, and subtraction radiography

Background:

To compare the diagnostic accuracy of three different imaging systems: Direct digital radiography system (DDR-CMOS), four types of filtered images, and a priori and a posteriori registration of digital subtraction radiography (DSR) in the diagnosis of proximal defects.

Materials and Methods:

The teeth were arranged in pairs in 10 blocks of vinyl polysiloxane, and proximal defects were performed with drills of 0.25, 0.5, and 1 mm diameter. Kodak RVG 6100 sensor was used to capture the images. A posteriori DSR registrations were done with Regeemy 0.2.43 and subtraction with Image Tool 3.0. Filtered images were obtained with Kodak Dental Imaging 6.1 software. Images (n = 360) were evaluated by three raters, all experts in dental radiology.

Results:

Sensitivity and specificity of the area under the receiver operator characteristic (ROC) curve (Az) were higher for DSR images with all three drills (Az = 0.896, 0.979, and 1.000 for drills 0.25, 0.5, and 1 mm, respectively). The highest values were found for 1-mm drills and the lowest for 0.25-mm drills, with negative filter having the lowest values of all (Az = 0.631).

Conclusion:

The best method of diagnosis was by using a DSR. The negative filter obtained the worst results. Larger drills showed the highest sensitivity and specificity values of the area under the ROC curve.

Computers in imaging and health care: now and in the future

Early picture archiving and communication systems (PACS) were characterized by the use of very expensive hardware devices, cumbersome display stations, duplication of database content, lack of interfaces to other clinical information systems, and immaturity in their understanding of the folder manager concepts and workflow reengineering. They were implemented historically at large academic medical centers by biomedical engineers and imaging informaticists. PACS were nonstandard, home-grown projects with mixed clinical acceptance. However, they clearly showed the great potential for PACS and filmless medical imaging. Filmless radiology is a reality today. The advent of efficient softcopy display of images provides a means for dealing with the ever-increasing number of studies and number of images per study. Computer power has increased, and archival storage cost has decreased to the extent that the economics of PACS is justifiable with respect to film. Network bandwidths have increased to allow large studies of many megabytes to arrive at display stations within seconds of examination completion. PACS vendors have recognized the need for efficient workflow and have built systems with intelligence in the management of patient data. Close integration with the hospital information system (HIS)-radiology information system (RIS) is critical for system functionality. Successful implementation of PACS requires integration or interoperation with hospital and radiology information systems. Besides the economic advantages, secure rapid access to all clinical information on patients, including imaging studies, anytime and anywhere, enhances the quality of patient care, although it is difficult to quantify. Medical image management systems are maturing, providing access outside of the radiology department to images and clinical information throughout the hospital or the enterprise via the Internet. Small and medium-sized community hospitals, private practices, and outpatient centers in rural areas will begin realizing the benefits of PACS already realized by the large tertiary care academic medical centers and research institutions. Hand-held devices and the Worldwide Web are going to change the way people communicate and do business. The impact on health care will be huge, including radiology. Computer-aided diagnosis, decision support tools, virtual imaging, and guidance systems will transform our practice as value-added applications utilizing the technologies pushed by PACS development efforts. Outcomes data and the electronic medical record (EMR) will drive our interactions with referring physicians and we expect the radiologist to become the informaticist, a new version of the medical management consultant.

Characterization of monochrome CRT display systems in the field

This article presents a review of image quality assessment methods for monochrome CRTs in the field as opposed to the laboratory. The review includes image quality programs at the University of Washington, the University of Texas at Houston, the University of Michigan, and the University of Arizona. CRT manufacturers and display-board suppliers also are concerned with image quality, particularly with respect to the life time of the CRT. The programs show that the need for image quality assessment for CRTs in the clinic is recognized. Although several experimental programs are in place, there is no universally accepted program. In fact, the clinical consequences of degraded monitor performance are not even well known and must be established. The existing programs mainly are based on the most comprehensive test pattern, the SMPTE pattern. The programs permit assessment of maximum luminance, display function, dynamic range, and contrast. They do not permit assessment of spatial resolution. There is no easy method to determine the spatial resolution in the field as precisely as desired simply because there are no visual aids (test patterns) to reliably determine loss of spatial resolution and signal-to-noise ratio using human observers. This report also presents initial and encouraging data obtained at the University of Arizona with a CCD camera. This CCD camera has the potential to be developed into an important tool for practical CRT evaluation for the clinic.