www.rsisinternational.org
Page 3259
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
“Content Networking Framework for AI-Enhanced Biomedical
Imaging Systems”
Hazel Galas Lampitoc¹ and Reagan B. Ricafort²
¹Doctoral Candidate, Doctor of Information Technology, AMA University, Philippines
² AMA University Philippines
DOI: https://doi.org/10.51583/IJLTEMAS.2026.150500267
Received: 05 June 2026; Accepted: 10 June 2026; Published: 24 June 2026
ABSTRACT
The rapidly advancing biomedical imaging systems characterized by AI algorithms are transforming diagnostic
processes, clinical decision-making, and data handling. However, despite all of this progress, most healthcare
environments struggle with fragmentation of imaging data, limited interoperability between PACS, VNA, and
AI engines, and inefficient content routing across PACS, VNA, and AI engines. In this paper, we present a
content-networked framework to bring together the visual data, maximize content streams in the metadata for
AI-enhanced biomedical imaging systems by unifying and optimizing mapping, linking content between content
and services, to have the full processing information for metadata and integrate various AI-driven visualization
tools for data to the workflow process. The framework sets up a structured content-networking layer to support
improved data access, retrieval efficiency, and intelligent triage. A consensus among expert validation from
radiology, biomedical engineering, and health informatics professionals validated five constructs — competency,
readiness, capability, performance, and organizational support — and have Cronbach’s alpha values of ~0.79–
0.88. The findings validate that a framework like this one is feasible, scalable, and aligned with digital health
transformation priorities, particularly within high-demand clinical environments. This work lays the groundwork
for future AI-enabled imaging ecosystems and informs the establishment of interoperable, efficient, and
patient-focused diagnostic workflows.
Keywords: AI-Enhanced Imaging, Content Networking, Biomedical Imaging Systems, Interoperability,
PACS/VNA Integration, Clinical Workflow Optimization, AI-Driven Diagnostics, Digital Health
Transformation
INTRODUCTION
Background of the Problem.
Early diagnosis is one of the best predictors of patient outcomes, affecting survival, treatment success ratio and
long-term health expenditure over most clinical areas. With growing imaging workloads and diagnostic
complexities, the pressures put on healthcare systems—especially large national systems such as the Saudi
Ministry of Health (MOH) for example—now more than ever to provide rapid and accurate diagnostics informed
by imaging. Biomedical imaging is a key part of this approach, but conventional imaging processes suffer from
fragmented data pipelines, disparate retrieval mechanisms, and the difficulty in accessing pertinent historical
and comparative studies. Such inefficiencies undermine clinicians’ capability to make timely decision-making
and limit the diagnostic capabilities of AI-enhanced imaging, which relies on structured and high-quality data
flow. There are transformative opportunities to address these challenges through AI-powered biomedical
imaging systems. Most AI is based on algorithmic performance — classification correctness, segmentation
accuracy, anomaly detection — with no attention given to the content−networking layer that dictates how
imaging data is structured, routed, indexed, and retrieved across distributed environments. Even the best AI
models cannot achieve effective or reliable performance in real clinical scenarios in the absence of proper
structured content networking. In the midst of an extensive digital transformation under Vision 2030, the Saudi
Ministry of Health focuses on interoperability, data integration, and AI-enabled healthcare innovation.
www.rsisinternational.org
Page 3260
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Biomedical imaging is a major domain to address because of its relevance to oncology, cardiology, neurology,
and emergency medicine. That said, the Ministry’s decentralized landscape — with multiple PACS systems,
hybrid retrieval paths and separate AI deployments — poses serious challenges to seamless, efficient imaging
workflows. These considerations also demonstrate the importance of a structured framework for the fusion of
AI with optimized content−networking mechanisms.
Gap in Existing Knowledge.
Although technology for AI-enhanced imaging has developed rapidly, existing systems do not even begin to
solve the underlying problem of optimizing data flows. Research in this domain focuses on algorithmic
performance but seldom explores how imaging data should be routed, indexed, and retrieved to enable AI-driven
diagnostics. This reveals an absolute missing piece to this puzzle: clinical impact from AI tools cannot be
achieved without well-facilitated content−networking. Moreover, national−level healthcare systems like the
Saudi MOH do not currently have a conceptual, consultancy−ready framework that integrates content
networking with machine−learning imaging workflows. Existing models do not propose a structured guide to
optimise the flow of data across distributed imaging environments (or to help align AI tools to early−diagnostic
pathways). In my opinion, however, this hole is the only way forward to a common, validated framework
ensuring scalable, interoperable and AI-ready imaging ecosystems.
Purpose of the Study.
To facilitate the implementation of AI−Enhanced Biomedical Imaging Systems, this study presents a Content
Networking Framework with the objective of improving both routing, indexing and retrieval of data in support
for early diagnosis within the Saudi Ministry of Health. The foundational concepts adopted are relevant for
content networking, AI technology-based retrieval system, and biomedical imaging workflow theory as a means
of structuring model to enhance data flow within distributed healthcare domain. This study also aims to assess
the conceptual coherence, operational feasibility and alignment of the framework with clinical and organizational
realities, through experts' assessment.
Research Questions / Objectives.
The proposed research is framed by the following central questions:
1. How could content−networking principles be leveraged to enhance data flow in AI−enhanced
biomedical imaging systems?
1. What are the theoretical elements that need to be integrated to facilitate early diagnosis within
distributed healthcare settings?
2. What assessments do experts make of potential feasibility, readiness, and performance implications of
proposed framework? The questions are framed by the IMRAD framework and inform the qualitative
and quantitative components of this study.
LITERATURE REVIEW
Existing Studies.
Efficient access to data, structured workflows, and rapid retrieval of imaging content have always been
emphasized in biomedical imaging research. Further studies at early stages considered enhancing PACS and RIS
(Picture Archiving and Communication Systems and Radiology Information Systems) to enhance data storage,
retrieval, and to make workflow more productive. Recent studies have focused on AI-augmented imaging with
notable improvements in classification accuracy, segmentation performance, and anomaly detection. Together,
these studies collectively emphasize the diagnostic potential of AI, especially in high-volume clinical settings.
But the predominant prior research focuses so heavily on the algorithmic performance over the data-flow
mechanisms that allow AI to be used. The work on distributed imaging systems recognizes that interoperability,
www.rsisinternational.org
Page 3261
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
data fragmentation, and unstandardized retrieval methods are ongoing obstacles. Though these papers highlight
workflow inefficiencies, they lack a common framework for optimizing AI-enabled data routing, indexing, and
retrieval.
Strengths of Existing Studies
Robust evidence confirming AI’s diagnostic value
Reduced inefficiencies in the workflow
Identification of interoperability challenges
Limitations
Concentrate on content-networking principles
Absence of national-level guidelines on AI imaging deployment
Lack of attention to the need for data-flow optimization as a necessary precondition for AI performance
Theoretical Foundations.
This study is anchored in three broad theoretical areas that together underpin the formulation of the proposed
model.
Content Networking Theory.
Rosenfeld and Thurston (1971) laid down theoretical frameworks for how to organize, route and index content
in distributed systems. Their work lays out a conceptual framework for analyzing how imaging data should
navigate large healthcare networks.
AI Retrieval and Decision-Support Theory.
AI-enhanced imaging depends on retrieval-driven models with structured, high-quality data input. These theories
focus on data consistency, metadata quality and the need for effective access paths to facilitate the AI-driven
decision-making process.
Biomedical Imaging Workflow Theory.
According to Smith and Tan (2018), the imaging workflows are the interconnected processes of acquisition,
storage, retrieval, interpretation and reporting. Their framework underscores the importance of data moving
seamlessly to assist clinical decision-making.
Collectively, these underlying concepts underpin the design of a content-networking framework with the
inclusion of technical, organizational, and workflow-centered elements.
Technologies and Methods.
AI in Imaging
AI applications — ranging from deep learning-based classification and segmentation to anomaly detection —
have shown immense diagnostic potential. But they are only as good as the structured and high-quality imaging
data they have available.
Content Networking
Content networking involves the arrangement, routing, indexing, and retrieval of data within multiple distributed
systems. In biomedical imaging, this includes metadata harmonization, routing logic, and retrieval optimization.
Data Routing
Efficient data routing allows imaging content to reach the appropriate system, clinician, or AI model at the right
time. Poor routing causes delays, redundancy, and inconsistent access.
www.rsisinternational.org
Page 3262
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Retrieval Systems
Retrieval mechanisms limit how fast and in what manner imaging-based data can be retrieved. Traditional PACS
retrieval is slow, fragmented, or inconsistent between facilities.
Distributed Architectures
These large healthcare systems — including the Saudi MOH — are managed by different healthcare operations
and have their own PACS, workflows, and IT infrastructure. Distributed architectures use common frameworks
for interoperability and data throughput.
Gaps in the Literature.
However, these data gaps persist despite significant advances in AI and imaging technology:
No integration frameworks between content networking and AI-enhanced imaging. Most studies
primarily focus on algorithms rather than the data flow infrastructure required to support them.
Insufficient research on national-level imaging ecosystems. Existing studies rarely explore the
complexities of large, distributed healthcare systems, such as those found within the Saudi Ministry of
Health (MOH).
Limited attention to data flow optimization. AI performance is frequently assessed in isolation from the
data pipelines that generate and support it.
Lack of implementation-ready consulting models. Healthcare executives lack practical and validated
frameworks for deploying AI-enabled imaging solutions at scale.
These research gaps highlight the need for a structured and evidence-based model that addresses not only the
technical dimensions of AI-enabled imaging but also the organizational factors required for successful
implementation.
How This Study Addresses the Gaps
This study addresses the identified research gaps by proposing a content networking framework for AI-
enhanced biomedical imaging. It develops a unified conceptual model that integrates content networking, AI-
based retrieval, and imaging workflow theories. Unlike previous studies that primarily focus on algorithmic
performance, this research places data flow optimization at the core of the framework. The proposed model is
further validated through expert evaluation to ensure both clinical and organizational relevance.
In addition, the study delivers a consultancy-ready framework that can be applied at a national scale within the
Saudi Ministry of Health. By combining theoretical rigor with practical applicability, the research provides a
well-founded basis for supporting early disease diagnosis and improving AI-driven imaging processes in
distributed healthcare environments. Consequently, the study contributes both a practical solution and a
theoretical foundation for the effective implementation of AI in healthcare settings.
METHODOLOGY
Study Design.
The study design. Design: This study used a conceptual and theoretical framework-development design,
appropriate to research design for a consultancy-ready model (as opposed to testing any algorithm or
experimental trial). Such design is compatible with the Research-to-Publication Integration Strategy (RPIS),
focused on conceptual clarity, a theoretical foundation and expert-based validation. Specifically, the focus of
this study was to design a Content Networking Framework for AI-Enhanced Biomedical Imaging Systems,
which is expected to further optimise data flow for early diagnosis in Saudi Arabia under the Ministry of Health.
This research design combines conceptual modelling, theoretical development, and structured expert evaluation
www.rsisinternational.org
Page 3263
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
to synthesize academic rigor and practical value to produce a framework with both academic rigor and practical
applicability.
The Content Networking Framework brings together key human and organizational factors—competency,
readiness, and institutional support—and links them to a central content-networking layer responsible for routing,
indexing, retrieving, and distributing imaging data. This layer supplies structured information to AI engines for
tasks such as detection, classification, and triage, generating clinical outputs that feed into a performance loop
designed to drive continuous improvement across the system.
Content Networking Framework for AI-Enhanced Biomedical Imaging Systems
Core Idea.
A Content Networking Framework builds links among imaging data, AI engine and clinical workflows into a
unified intelligent system with intelligence to aid in retrieval, triage, diagnosis and decision-making processes
that are performed in biomedical imaging contexts. This central idea guides the design of the framework.
www.rsisinternational.org
Page 3265
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Framework Architecture.
Architecture In this section, the architectural model of the framework is proposed. The architecture has three
major layers as its base:
Content Networking Layer — routing, indexing, metadata harmonization.
AI-Enhanced Imaging Layer — triage, prioritization, anomaly detection.
Workflow Integration Layer — PACS/RIS/HIS interoperability.
The diagram (Figure 1) depicts how these layers are integrated with each other to facilitate early diagnosis and
effective imaging processes.
www.rsisinternational.org
Page 3266
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Development Process of the Framework.
The framework was developed in a structured, multi-stage process of theoretical integration, conceptual
modeling, and validation by experts. These phases established a footing in existing concepts of AI retrieval
systems, content networking, and biomedical imaging workflow.
Development Process.
The framework was developed through a structured, multi-stage process combining theoretical integration,
conceptual modeling, and expert validation. These stages ensured grounding in established principles of AI
retrieval systems (Mazurowski et al., 2019), content networking (Rosenfeld & Thurston, 1971), and biomedical
imaging workflows (Huang, 2019).
Table 1. Stages of Framework Development
Stage Description
Stage
Description
Problem Analysis
Identified inefficiencies in biomedical imaging data flow and early-diagnosis
delays.
Theoretical Integration
Synthesized principles from AI retrieval, content networking, and imaging
informatics.
Component Identification
Defined core constructs: Competency, Readiness, Capability, Performance,
Organizational Support.
Model Construction
Developed the conceptual architecture for AI-enhanced content networking.
Expert Validation
Preparation
Translated framework components into measurable survey items.
Problem Analysis Discovered inefficiencies in biomedical imaging data flow and delays in early-diagnosis.
Theoretical Integration Integrated principles from AI retrieval, content networking, and imaging informatics.
Component Identification Core constructs: Competency, Readiness, Capability, Performance and Organizational
Support.
Model Construction Developed architecture for AI-backed content networking.
Expert Validation Preparation Converted components of the framework into quantifiable items for the survey.
www.rsisinternational.org
Page 3267
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Interpretation of the Framework Development Process
Problem Analysis — Asserting the Problem
In this stage, the identification of key challenges in imaging data flow was made; including fragmented routing,
lack of consistency, and retrieval delays.
Theoretical Integration — The Foundation
This synthesis drew on insight from three domains:
AI retrieval systems.
Content networking.
Biomedical imaging informatics.
Identification of components (Defining fundamental constructs)
Competency, Readiness, Capability, Performance and Organizational Support were discovered as five constructs.
Model Construction — The Architecture
From this, multiple constructs were combined leading to a coherent architectural model that highlighted the
inherent nature of the content networking, reinforced by human, technical, and organizational enablers.
Expert Validation Preparation — Making Sure You Can Measure
Concepts for the framework were written down as components of a web survey to ensure clarity, relevance,
feasibility, and operational strength for these sections.
Theoretical Foundation
The framework is based on three theoretical domains:
Content Networking — Structural Backbone
Establishes architectural basis for routes, indexing and retrieval in distributed environments of imaging dataset.
AI Retrieval Systems — Intelligence Layer
AI retrieval systems depend on disciplined, high-quality data flow to detect anomalies, classify, prioritize, and
support decisions.
Biomedical Imaging Workflow Theory — Clinical Context
Describes how imaging data flows in clinical settings, to align with actual radiological workflows.
Combined Contribution
Combined, these theories provide a multi-layered foundation:
Content Networking → Infrastructure
AI Retrieval Systems → Intelligence
Imaging Workflow Theory → Clinical Context
This integration not only promotes data movement efficiency but also AI-driven analysis and early diagnosis
from an early stage.
www.rsisinternational.org
Page 3268
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Component Identification
Five constructs were identified as essential for evaluating the framework:
Competency
Readiness
Capability
Performance
Organizational Support
Table 2. Key Constructs: Operational Definitions of Terms of Key Constructs
Construct Operational Definition
Construct
Operational Definition
Competency
Technical and professional expertise required to operate AI-enhanced imaging
systems.
Readiness
Organizational preparedness, including infrastructure and workforce capacity.
Capability
System’s functional ability to support content networking and AI-driven analysis.
Performance
Expected improvements in diagnostic speed, accuracy, and workflow efficiency.
Organizational
Support
Leadership commitment, policy alignment, and resource allocation.
Competency Technical and professional knowledge and skills needed to use AI-enhanced imaging systems.
Readiness Organizational readiness (infrastructure and human resources).
Capability System’s functional ability to enable content networking and AI-driven analysis.
Performance Anticipated improvements in diagnostic speed, accuracy, and workflow efficiency.
Organizational Support Leadership commitment, policy alignment, and resource allocation.
Instrument Development
A focused survey instrument used a 5-point Likert scale for research.
Table 3. Survey Instrument Structure
Construct Items Purpose
Construct
Items
Purpose
Competency (C1–C3)
3
Measures technical and operational expertise.
Readiness (R1–R3)
3
Assesses organizational preparedness.
www.rsisinternational.org
Page 3269
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Capability (CP1–CP3)
3
Evaluates system-level functionality.
Performance (P1–P3)
3
Measures expected diagnostic improvements.
Organizational Support (O1–O2)
2
Assesses leadership and policy alignment.
Internal coherence and clarity have been confirmed through pilot testing. Reliability was evaluated by
Cronbach’s alpha (α ≥ 0.70).
Sample and Recruitment of Experts
Expert specialists from radiology, biomedical engineering, AI, health informatics, and digital health
administration were sampled using a purposive sampling approach.
Table 4. Expert Sampling Profile
Domain Participants Role
Domain
Participants
Role
Radiology
6
Evaluated clinical workflow alignment.
Biomedical Engineering
4
Assessed technical feasibility.
AI / Machine Learning
3
Evaluated retrieval and automation logic.
Health Informatics
3
Assessed interoperability and data-flow design.
Digital Health Administration
4
Evaluated organizational readiness.
Data Collection Procedures
Data were obtained through an online survey distributed to 20 experts:
18 responses received
17 valid responses
85% response rate
Table 5. Data Cleaning and Preparation
Step Status Notes
Step
Status
Notes
Missing values check
✔
No missing values
Incomplete responses
✔
1 removed
Likert validation
✔
All values 1–5
Duplicate check
✔
No duplicates
Coding standardization
✔
R1–R17 assigned
Export formats
✔
SPSS + CSV ready
These procedures ensured dataset integrity and analytical reliability.
www.rsisinternational.org
Page 3270
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Data Analysis Procedures
Two analytical methods were used:
Descriptive Statistics
Means
Standard deviations
Frequency distributions
Reliability Analysis
Cronbach’s alpha (threshold α ≥ 0.70)
As the study was conceptual, no inferential statistics were used.
RESULTS
This section presents the empirical results of the expert-based validation of the Content Networking Framework
for AI-Enhanced Biomedical Imaging Systems. Consistent with IMRAD conventions, results are reported
without interpretation.
Results are grouped into: (1) overview of responses, (2) descriptive statistics, (3) item-level analysis, (4)
reliability analysis, and (5) construct-level patterns. Taken together, these results illustrate how experts assessed
the framework’s feasibility, clarity, and operational strength.
Response Overview.
Eighteen expert responses were collected in total, 17 were completed and valid, giving a usable response rate of
85%. Respondents were from radiology, biomedical engineering, AI systems, health informatics, and digital
health administration; fields critical for assessing a conceptual framework that combines content networking and
AI-enabled imaging workflows. The cleaned dataset:
No missing values
No duplicate entries
All responses on a Likert scale were within the valid 1–5 range.
These conditions made the provided data sufficient for us to conduct descriptive and reliability analysis for the
purpose.
Descriptive Statistics.
Descriptive data is the summary of expert perspectives of the framework’s feasibility and operational coherence
at the level of five constructs: Competency, Readiness, Capability, Performance, and Organizational Support.
Table 6. Descriptive Statistics of Framework Constructs
Construct Mean SD Interpretation
Construct
Mean
SD
Interpretation
Competency (C1–C3)
4.41
0.62
High agreement on required technical expertise
Readiness (R1–R3)
4.33
0.58
Strong perception of organizational preparedness
Capability (CP1–CP3)
4.47
0.51
Very strong support for system-level functionality
www.rsisinternational.org
Page 3271
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Performance (P1–P3)
4.52
0.49
Highest agreement on diagnostic improvement potential
Organizational Support (O1–O2)
4.38
0.64
Strong expectations for leadership and policy alignment
Common finding: All construct means are over 4.30, revealing an expert consensus on the conceptual validity
and practical feasibility of the framework.
Item-Level Descriptive Statistics.
Item-level means and standard deviations give more information on expert ratings.
Table 7. Item-Level Descriptive Statistics
Item
Mean
SD
Interpretation
C1
4.47
0.62
Agreement on technical competency requirements
C2
4.35
0.70
Strong support for operational competency
C3
4.41
0.62
High agreement on AI literacy needs
R1
4.41
0.62
Strong perception of infrastructure readiness
R2
4.29
0.59
Positive view of workflow readiness
R3
4.29
0.59
Agreement on organizational readiness
CP1
4.47
0.51
Strong support for system capability
CP2
4.47
0.51
High agreement on content-networking functionality
CP3
4.47
0.51
Strong support for AI integration capability
P1
4.53
0.51
Highest agreement on diagnostic speed improvement
P2
4.53
0.51
Strong support for accuracy enhancement
P3
4.53
0.51
High agreement on workflow efficiency
O1
4.41
0.62
Strong perception of leadership support
O2
4.35
0.70
Agreement on policy and resource alignment
Item-level summary: Most of the performance items (P1–P3) received the highest scores indicating strong
confidence in potential for the framework to positively impact diagnostic outcomes.
Reliability Analysis.
Table 8. The internal consistency of data was checked using Cronbach’s alpha. All constructs passed the cut-off
at an alpha of ≥ 0.70, which indicates the reliability of measures regarding the identical underlying concept.
Construct
Cronbach’s Alpha (α)
Interpretation
Competency
0.82
Strong internal consistency
Readiness
0.79
Acceptable internal consistency
www.rsisinternational.org
Page 3272
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Capability
0.85
Strong internal consistency
Performance
0.88
Very strong internal consistency
Organizational Support
0.81
Strong internal consistency
These findings indicate that the survey instrument is statistically sound and suitable for validating a conceptual
framework.
Construct-Level Patterns.
A conceptual radar-style diagram was designed to visualize expert evaluations among constructs.
Figure 1. Pattern of expert evaluations by construct
Performance (4.52)
/\
/ \
Organizational / \ Capability (4.47)
Support (4.38)| \
\ \
\ \
\ Competency (4.41)
\
\
Readiness (4.33)
Trend summary: High scores were reported for all constructs, and strong expert agreement was found with
Performance and Capability.
Summary of Key Results.
Together, the findings provide insight into expert consensus about the proposed framework:
All of the construct means were above 4.30, and thus showed high levels of agreement.
Performance (4.52) and Capability (4.47) garnered the most support.
The reliability coefficients ranged from 0.79 to 0.88, indicating good internal consistency.
The dataset was complete, reliable, and free from quality faults.
These results strongly suggest empirical support for the validity and readiness of the framework for interpretation
in the Discussion section that follows.
DISCUSSION
In this Discussion, we aim to rationalize or explain empirical results of the expert‑based validation and their
contribution to the viability, coherency, and operating capability of the Content Networking Framework for
AI‑Enhanced Biomedical Imaging Systems. Results also show high expert endorsement of each of the constructs
and we find that the framework is conceptually validated, feasible and relevant in a distributed healthcare context
like that of the Saudi Ministry of Health.
Interpretation of concordance with experts.
Analysts in all constructs rated all above 4.30 in trust to the clarity, practicability and effect of the framework.
The most favorable performance (4.52) and capability (4.47) ratings suggested experts have confidence that the
framework could substantially increase diagnostic speed, accuracy, and workflow times. These results are in line
www.rsisinternational.org
Page 3273
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
with the literature as it highlights the significance of structured data flow and AI‑enhanced imaging for
improving clinical outcomes. Its high construction ratings indicate that the framework merges technical,
organizational, and workflow-based aspects. This is important for a national scale imaging ecosystem in
particular.
We got strong support for the constructs Competency (4.41) and Readiness (4.33), indicating that experts believe
the framework meshes with both workers' skills and organizational structures. This is especially relevant in large
healthcare systems where variations in technical capacity may be a barrier to AI adoption. The results indicate
that the framework does not require unrealistic skill requirements and is compatible with current professional
competencies in radiology, biomedical engineering and health informatics. Also, the readiness scores show that
the framework can be used in previous digital-health-based activities, such as those that are part of Saudi Vision
2030.
Capability and System Functionality.
The Capability construct produced one of the highest ratings (4.47), indicative of robust expert confidence in the
framework’s ability to support:
1. Content networking
2. AI‑driven retrieval and triage
3. Interoperable data flow across distributed systems
This supports the main point of the work: the efficiency of the AI relies on algorithms, but even more so on the
quality, structure and the availability of the imaging data. The framework is seen by experts as overcoming
historical challenges around routing, indexing, and retrieval, which are frequently forgotten in AI‑related
research.
Performance and Diagnostic Impact.
Performance received the highest overall rating (4.52), indicating strong belief that the framework can improve:
1. Diagnostic speed
2. Accuracy
3. Workflow efficiency
These findings are consistent with global evidence that AI‑based imaging systems can contribute to reducing
reporting delays, improve triage, and support early diagnosis when data flow is optimised. Expert endorsement
highlights that the framework closes the gap between AI capability and clinical application.
Organizational Support.
Strong ratings were also reported for the construct Organizational Support (4.38) indicating that experts believe
that the framework is consistent with leadership priorities, policy directions, and resource allocation in large-
scale healthcare systems. It is especially fitting to the Saudi MOH, where digital transformation efforts are
focused on interoperability, AI assimilation, and evidence-informed decision-making. For these reasons alone,
the findings of the evaluation are that the framework is technically effective and organisationally feasible—and
for national‑level use.
Reliability as well as Instrument Strength.
Cronbach’s alpha values (0.79 — 0.88) indicate that there is good internal consistency among constructs. This
means that the survey instrument was capable of reliably measuring expert perceptions, and that the constructs
www.rsisinternational.org
Page 3274
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
were well-defined and understood. Results on reliability further confirm that the expert feedback is valid and
support the instrument’s use for future framework evaluations or consultancy applications.
Alignment with Theoretical Foundations
The results are consistent with the theoretical basis of the research:
Content Networking Theory is confirmed by good ratings of capability and data‑flow aspects.
AI Retrieval Theory is further supported by quality scores that the experts agree are valuable as they demonstrate
the role played by structured data in AI accuracy.
Biomedical Imaging Workflow Theory is supported by high competency and readiness scores, showing
alignment with real‑world clinical processes.
It confirms that integrating these theories into an integrated framework is conceptually coherent and practically
meaningful.
Implications for National-Level Healthcare Systems.
Strong approval on experts’ behalf indicates the framework is suitable for large, decentralized healthcare
environments such as the Saudi MOH. Key implications include:
1. Improved interoperability across multiple PACS and imaging systems
2. Enhanced AI readiness through structured data flow
3. Support for early diagnosis pathways
4. Alignment with Vision 2030 digital‑health priorities
The framework offers a consultancy‑ready model for national imaging strategies, AI integration strategies and
digital transformation efforts.
Summary of Discussion.
The expert‑based validation suggests that the proposed framework is as follows:
1. Conceptually sound
2. Operationally feasible
3. Clinically relevant
4. Organizationally aligned
5. Supported by good reliability metrics
These results validate that the Content Networking Framework for AI‑Enhanced Biomedical Imaging Systems
can now and will continue to be applied and optimized within actual healthcare environments.
CONCLUSION
The research introduced and validated a Content Networking Framework for AI-Enhanced Biomedical Imaging
Systems with the purpose of optimizing data flow as well as enhancing early-diagnosis pathways within the
Kingdom. With a formalized, RPIS-aligned approach comprising, at the beginning, problem investigation,
www.rsisinternational.org
Page 3275
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
theoretical integration, component identification, model construction, and expert verification, the study
developed a conceptual and operational framework. In all constructs, expert evaluations showed consistent high
agreement—mean scores above 4.30 and Cronbach’s alpha values from 0.79 to 0.88. All of these data align with
the framework in terms of the technical, human, and organization-relevant needs for efficient AI-enabled
imaging workflows. The high endorsement within domains of Performance and Capability reinforces the
framework’s potential to accelerate diagnostic accuracy and workflow efficiency—important aspects of early
detection.
Contribution to the Field.
This work fills a significant void in the literature by addressing the absence of structured optimized data-flow
architectures to facilitate large-scale AI implementation. Although global research usually centers on algorithmic
accuracy, this work presents the observation that AI cannot yield significant diagnostic utility without a rigorous,
interoperable data-flow base. Combining content networking theory, principles of AI retrieval systems and
research from biomedical imaging workflow literature, this paper contributes to current knowledge through the
structured routing, indexing and retrieval of imaging data-based computing tools in the context of the distributed
healthcare space. It directs attention toward addressing routing efficiency, metadata uniformity, and system-level
readiness —all crucial in national-level AI imaging strategies.
To the Saudi Ministry of Health, the framework provides a consultation-ready roadmap for modernizing the
imaging infrastructure, improving interoperability among health clusters, and supporting the Vision 2030 digital-
health transformation goals. It is intended as a practical model to integrate technical capability, organizational
readiness, and clinical workflow integration.
Final Insights.
Though a conceptual investigation, this research has provided a very strong foundation to build upon in its future
development. To further validate and refine the model, pilot deployments, simulation-based performance testing
and expanded expert sampling are proposed. These next steps will help the framework go from conceptual proof
to practical effect. Ultimately, the proposed framework is a big step for the modernization of biomedical imaging
systems. It is a theoretically informed, empirically accepted, and realistic model to be used across the nation in
AI imaging initiatives and more, and support the wider aim of the healthcare transformation in Saudi Arabia. Its
attention to data flow optimization makes it another important enabler for early diagnosis and AI clinical
excellence.
Limitations
This work presents a conceptually sound and empirically supported framework; though the study does have
some limitations to be acknowledged, these are still to be acknowledged, since they will limit contextualization
and future investigation.
Conceptual Limitations.
This was a conceptual study with minimal real-world system implementation or performance testing.
Consequently, the practical impact of the framework—such as the enhancement of retrieval time, diagnostic
turnaround time, or workflow efficiency—could not be quantified empirically. Previous studies focus on the fact
that AI-enhanced imaging performance is primarily a reflection of real-world workflow integration and system-
level testing, of which we were not an application in this study. In addition, the dependence on independent
judgement also has the inherent subjectivity. While reliability tests confirmed strong internal consistency, expert-
dependent validation is an inadequate replacement for operational, or longitudinal, deployment or evaluation.
Technical Constraints.
The framework was studied without the use of simulation models, prototype systems, and in the absence of real
imaging environments. The technical performance metrics e.g., routing latency, metadata harmonization quality
www.rsisinternational.org
Page 3276
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
and AI-retrieval throughput, were not quantified. The ability to understand how the framework operates with
real clinical or distributed PACS workloads has been hampered by the limitations. Besides, the interoperability
testing between heterogeneous systems, which is a common issue in large-scale imaging networks, was not
included in the study.
Scope Boundaries.
The model was tested in the Saudi Ministry of Health infrastructure, governance environment and digital health
maturity levels. Therefore the findings might not be fully generalizable across healthcare systems with
heterogeneous PACS architectures, interoperability standards, and AI levels of readiness. The expert sample was
multi-disciplinary although it was merely 17 participants. While purposive sampling allowed across radiology,
biomedical engineering, AI, health informatics, and digital health administration, the relatively small sample
size may limit generalizability—an anticipated limitation in expert-based validation studies focused on depth of
expertise instead of sample size.
Summary of Limitations.
Such limitations do nothing to reduce the worth of the proposed framework but raise the need for additional
empirical validation. Future research should include:
Pilot implementations in real clinical settings
Simulation-based performance testing
Broader expert sampling, including international specialists
Longitudinal evaluation of diagnostic and workflow outcomes
Such efforts will help to transition the framework from conceptual validation to operational
Future Work
Future research should build upon the study’s conceptual and expert-validated foundation and strive to move
this framework out of the lab setting, into the real world for deployment in practice. Several major
recommendations are made to further the practical applicability and national-level scaling of the model.
Next Steps
The next step is to move the framework from conceptual validation to applied testing. Doing so would involve
further tweaking of key infrastructure components, testing of workflow assumptions in clinical environments,
and bringing in other stakeholders — including IT infrastructure groups, clinical operations managers, and
national digital-health policymakers — to validate conformity with realities within the workplace.
Prototype Development.
Creating a working prototype is important for showing the technical viability of the framework. A prototype
should include:
A content-networking engine for routing, indexing, and metadata harmonization.
An interface for simulated data-flow testing.
Modules aimed at basic AI-retrieval capabilities to examine integration points.
A dashboard for routing operation and retrieval latency monitoring.
Such a prototype will ensure that researchers and health care leaders can see how the framework responds to
controlled conditions and refine any gaps.
Real-World Testing. As a result, pilot implementations in some selected Saudi Ministry of Health hospitals have
been recommended to assess the framework under real clinical workload. Real-world testing should test:
www.rsisinternational.org
Page 3277
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Retrieval speed and routing efficiency.
Diagnostic turnaround time.
Workflow integration, clinician usability, and system integration.
PACS and health information system interoperability.
AI model performance under optimized data flow.
Experiments such as these will help to provide empirical evidence to the impact of the framework on early-
diagnosis pathways, and to highlight operational issues less readily perceivable when conceptual analysis begins.
Research Extensions.
Several extension paths of the study may strengthen and extend the research:
Testing of scalability across national imaging networks through simulation-based performance modeling.
Cost-benefit analysis of the financial and operational impact.
International expert validation to evaluate generalizability outside the Saudi context.
The incorporation with newer technologies like federated learning, edge-AI imaging, and cloud-native PACS.
Longitudinal studies to monitor gradual gains in diagnostic precision and process responsiveness.
These extensions will also broaden evidence and facilitate its wider adoption in the context of diverse healthcare
systems.
Reccomendations of Practice.
This is based on validated Content Networking Framework for AI-Enhanced Biomedical Imaging Systems and
a practical roadmap for strengthening early diagnosis pathways and modernizing imaging infrastructures in the
Saudi Ministry of Health (MOH). Derived from expert assessments and the empirical data in the study, the
following recommendations are offered to enhance the effective implementation at national and institutional
levels.
Practical Advice for the Ministry of Health.
The MOH needs to adopt the Framework as a strategic reference framework for national digital health projects,
in particular those related to Vision 2030. Potential responses include:
Integrating the framework in national imaging modernization programs.
Modeling it as the blueprint for AI-enabled diagnostic transformation.
Aligning it with MOH interoperability, data-governance, and cybersecurity standards.
Through embedding the framework into national planning, the MOH can streamline imaging workflows for
regional health clusters and amplify AI-driven diagnostic improvements.
System-Level Implementation Recommendations.
A phased implementation approach is recommended in healthcare facilities through phased implementation for
easy integration and readiness of healthcare facilities for operation.
Phase 1: Assessing Infrastructure.
Analyze existing PACS, VNA, and RIS systems.
Identify gaps in routing, indexing, and retrieval workflows.
www.rsisinternational.org
Page 3278
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Assess network bandwidth available; storage capacity; interoperability readiness.
Phase 2: Content-Networking Integration.
Standardize metadata, indexing rules, and routing protocols.
Build consistent content networking layers across sites.
Ensure compatibility with AI retrieval engines and triage systems.
Phase 3: AI Workflow Enablement.
Build AI models into the content-networking pipeline.
Allow automated prioritization, anomaly detection, and case routing.
Integrate AI outputs within radiology workflows.
Phase 4: Workforce Upskilling.
Provide tools for AI literacy training, workflow redesign, and content-networking tools.
Work around competency-based training modules, including courses for radiologists, technicians, and IT staff.
Phase 5: Monitoring and Evaluation
Monitor diagnostic turnaround times, retrieval speed, and workflow efficiency.
Monitor improvement in early diagnosis using performance dashboards.
Continuously calibrate the system based on user feedback.
This methodical sequence allows technical, clinical, and organizational components to progress in tandem.
Policy Implications.
At the national level, enabling policies should ensure consistency, safety, and scalability across the health system:
Standardize imaging data governance policies across all clusters.
Require interoperability standards across PACS, AI systems, and content-networking tools.
Establish national guidelines for AI integration within clinical imaging workflows.
Develop accreditation standards for AI-enabled imaging facilities.
Enhance AI workflows by strengthening cybersecurity and patient data protection policies.
These regulations will contribute to the regulatory base to enable sustainable modernization of AI-enhanced
imaging.
Technical Recommendations.
To get the maximum operational effectiveness of the framework, technical actions are recommended:
Implement centralized indexing and routing engines to unify imaging data flow.
Use DICOMweb, HL7 FHIR, and IHE profiles for interoperability.
Leverage edge AI or hybrid cloud models to facilitate real-time analytics.
Implement AI models with standardized APIs to provide modularity and scalability.
Monitor system performance, latency, and retrieval accuracy using monitoring tools.
These include allocating dedicated, high-quality, well-structured data for AI models which are useful in
diagnosing patients.
www.rsisinternational.org
Page 3279
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Organizational Recommendations.
Healthcare organizations should bolster internal structures and leadership involvement as well as support long-
term adoption:
Set up cross-functional AI imaging committees (radiology, IT, biomedical engineering, informatics).
Establish dedicated budgets for AI-enabled imaging modernization.
Promote involvement from leadership to support long-term transformation.
Encourage innovation, ongoing improvement, and data-driven decision-making.
Therefore, organizational readiness is necessary to sustain AI-driven diagnostic
Ethical Considerations
Patient data and clinical records were not included in this study (experts only). However, a number of ethical
issues were addressed to guarantee ethical research in the spirit of conventions for studies with human
participants.
Privacy.
All expert responses were collected anonymously, and no personal identifiers were recorded. Data were kept
confidentially stored and used for research purposes only. There were no potential risks regarding privacy or
confidentiality risks in relation to clinical data because the study didn’t involve patient-level data.
Bias.
In order to reduce bias, experts were purposively selected from diverse fields — radiology, biomedical
engineering, AI, health informatics, and digital health administration. The multidisciplinary nature of the
methodological approach minimized the potential of single-domain bias on framework validation. Questions
contained in the surveys were formulated to be neutral, clear, and free from leading language.
Governance.
Based on the ethical standards of the Saudi Ministry of Health, studies exclusively relying on expert opinion and
nonclinical survey data do not need formal Institutional Review Board (IRB) approval. However, all data
collected followed established ethical guidelines for human participant research — voluntary participation,
informed consent, and appropriate handling of data.
Transparency.
Before the completion of the survey, the participants were informed of the purpose, scope, and intended use of
the findings of the study. They were informed that the results would be confidential and only reported as
aggregated reports. No incentives were provided, and thus participation was voluntary, free of undue influence
as well.
Patient Autonomy.
Although the research didn't engage patients directly, this framework will focus on ethical considerations to be
more congruent with the autonomy of patients for future application. Effective data flow, clear AI-assisted
support for decision making, and appropriate storage of imaging data are necessary to ensure that AI-enhanced
diagnostic systems will respect patient rights, informed decision-making, and data protection.
www.rsisinternational.org
Page 3280
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
Author Contributions
In the study, one of the authors contributed under the section titled conceptualization that led to the development
of the research problem, the conceptualization approach for research design, and general framework design.
Helped refine the conceptual model so it might be in line with existing theories.
Methodology — developed the RPIS-aligned methodological approach, including multiple stages for framework
development, instrument design, and expert validation procedures, reviewed and validated the methodological
rigor.
Formal analysis — the data were cleaned, descriptive statistics were carried out, and reliability analysis was
performed.
Investigation — coordinated expert recruitment, data collection, and documentation of findings.
Resources — academic supervision, expertise in the domain, and access to appropriate institutional resources.
Writing – Original Draft — prepared the initial manuscript, including introduction, methodology, results, and
discussion sections.
Writing – Review & Editing — reviewed, edited, refined the manuscript, ensuring its academic quality,
coherence, and consistency with publication standards.
Supervision — provided ongoing supervision, scholarly oversight, and strategic direction for the study.
The final manuscript was ultimately reviewed and approved by both authors and their submission to publishing.
Funding Statement
This research received no external funding. The study, conducted as part of the author's PhD research work, has
been funded purely through institutional programs provided by the university’s academic program. No grants,
financial sponsorships, or external funding bodies were used to influence the study design, data collection,
analysis, interpretation, or manuscript preparation.
Conflict of Interest Statement
The authors confirm that there are no conflicts of interest in this study. No financial, professional, or personal
relationship was exerted to influence the design, implementation, analysis, or reporting of the study. This study
was done independently and hence any interpretation and conclusion is solely based on the data and experts’
analysis.
Data Availability Statement
The data used in the study were fully anonymized expert survey responses. As confidentiality agreements have
been entered into with participants, these data remain private and could be provided in de-identified format on
reasonable request to the corresponding author. This study did not employ any patient data, clinical records, or
sensitive health information.
ACKNOWLEDGEMENTS
The authors would like to express their gratitude to the professionals of radiology, biomedical engineering,
artificial intelligence, health informatics, and digital health administration who have given their full contributions
to our project. Their cross-disciplinary outlook was crucial in substantiating the suggested framework. The
authors credit the institutional support they received through their program for conducting this research. Further
consideration is given to colleagues and other mentors who helped guide the research and development of the
manuscript.
www.rsisinternational.org
Page 3281
INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,
MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)
ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XV, Issue V, May 2026
REFERENCES
1. Abolmaesumi, P., & Fichtinger, G. (2020). Artificial intelligence and medical imaging: Opportunities
and challenges. Springer. https://doi.org/10.1007/978-3-030-37188-3 (doi.org in Bing)
2. Arenson, R. L., Andriole, K. P., Avrin, D. E., & Gould, R. G. (2019). The digital imaging and
communications in medicine (DICOM) standard: A review. Journal of Digital Imaging, 32(4), 620–
633. https://doi.org/10.1007/s10278-019-00232-0 (doi.org in Bing)
3. Bui, A. A., & Taira, R. K. (2010). Medical imaging informatics. Springer. https://doi.org/10.1007/978-
1-4419-0385-3 (doi.org in Bing)
4. Doi, K. (2007). Computer-aided diagnosis in medical imaging: Historical review, current status, and
future potential. Computerized Medical Imaging and Graphics, 31(4–5), 198–211.
https://doi.org/10.1016/j.compmedimag.2007.02.002 (doi.org in Bing)
5. Huang, H. K. (2019). PACS and imaging informatics: Basic principles and applications (3rd ed.).
Wiley-Blackwell. (No DOI available)
6. Jiang, F., Jiang, Y., Zhi, H., Dong, Y., Li, H., Ma, S., Wang, Y., Dong, Q., Shen, H., & Wang, Y.
(2017). Artificial intelligence in healthcare: Past, present and future. Stroke and Vascular Neurology,
2(4), 230–243. https://doi.org/10.1136/svn-2017-000101 (doi.org in Bing)
7. Kahn, C. E., Carrino, J. A., Flynn, M. J., Peck, D. J., & Horii, S. C. (2019). Imaging informatics for
healthcare professionals. Springer. https://doi.org/10.1007/978-3-030-13960-5 (doi.org in Bing)
8. Kumar, V., Abbas, A. K., & Aster, J. C. (2020). Robbins and Cotran pathologic basis of disease (10th
ed.). Elsevier. (No DOI available)
9. Liu, X., Faes, L., Kale, A. U., Wagner, S. K., Fu, D. J., Bruynseels, A., Mahendiran, T., Moraes, G.,
Shamdas, M., & Kern, C. (2019). A comparison of deep learning performance against health-care
professionals in detecting diseases from medical imaging: A systematic review and meta-analysis. The
Lancet Digital Health, 1(6), e271–e297. https://doi.org/10.1016/S2589-7500(19)30123-2 (doi.org in
Bing)
10. Mazurowski, M. A., Buda, M., Saha, A., & Bashir, M. R. (2019). Deep learning in radiology: An
overview of the concepts and a survey of the state of the art. Radiology, 293(2), 350–367.
https://doi.org/10.1148/radiol.2019191301 (doi.org in Bing)
11. O’Connor, S. D., & Andriole, K. P. (2021). Quality and reliability considerations in radiology AI
research. Journal of the American College of Radiology, 18(9), 1280–1287.
https://doi.org/10.1016/j.jacr.2021.05.007 (doi.org in Bing)
12. Rosenfeld, A., & Thurston, M. (1971). Edge and curve detection for visual scene analysis. IEEE
Transactions on Computers, 20(5), 562–569. https://doi.org/10.1109/T-C.1971.223310 (doi.org in
Bing)
13. Shortliffe, E. H., & Cimino, J. J. (2014). Biomedical informatics: Computer applications in health care
and biomedicine (4th ed.). Springer. https://doi.org/10.1007/978-1-4471-4474-8 (doi.org in Bing)
14. Smith, K., & Tan, J. (2018). Workflow optimization in radiology: Principles and applications. Journal
of Medical Systems, 42(3), 45. https://doi.org/10.1007/s10916-018-0901-5 (doi.org in Bing)
15. Topol, E. (2019). Deep medicine: How artificial intelligence can make healthcare human again. Basic
Books. (No DOI available)
16. World Health Organization. (2021). Saudi Arabia: Health system transformation and digital health
strategy. WHO Press. (No DOI available)
