A Responsible Deterministic Framework for CT Radiation Governance V2.0

PAGER-CT Framework: Radiation Governance & Lifecycle Stewardship

Created by Dr. Sharad Maheshwari MD - imagingsimplified@gmail.com - Founder of BeResponsibleAI

Navigation Menu

PAGER-CT Framework

Institute for Responsible Healthcare AI

🏠 Welcome to the PAGER-CT Interactive Platform

CT optimization is not solely a scanner parameter problem. It is a structured intersection of upstream clinical justification, human workflow factors, clinical uncertainty management, and medical physics. This exhibit presents a testable, deterministic governance architecture where every rule separates Principle, Proposed Default, Documented Exceptions, and Local Validation.

🗺️ Understanding the Project Flow

📑 Step 1

Executive Summary

The core thesis: Moving from naive dose minimization to deterministic, task-specific stewardship.

🗺️ Step 2

Master System Map

A unified visual doctrine mapping the continuous lifecycle stewardship loop.

🏥 Step 3

Clinical Reality

The qualitative systemic forces driving utilization: throughput, medicolegal asymmetry, and reassurance.

🛡️ Step 4

J-Gate: Justification

Upstream governance checkpoint with indication-specific defaults and documented clinical exceptions.

📋 Step 5

Operational Factors

Mapping all modifiable pre-scan and console parameters into the governance architecture.

🧩 Step 6

PAGER-CT Domains

Testable governance across Phenotype & Size, Anatomy, Geometry, Exam, and Reconstruction.

⚛️ Step 7

Physics & SSDE

Conditional scout-centering pathway, separated fixed-mAs vs. clinical kV displays, and NPS texture validation.

⚖️ Step 8

ALADA & False Economy

Adapting ALADA/ALADAIP concepts to whole-body CT, unquantified sufficiency schematics, and 3 SSDE safeguards.

🌐 Step 9

Global Guidelines

Strategic alignment with ACR, AAPM Reports 204/220/293, ICRP 103/135, EuroSafe, and DIMITRA.

🏛️ Step 10

RATSe Governance & KPIs

The canonical 5-pillar RATSe overlay, failure mode safeguards, and evidence-tagged KPI targets.

🔍 Step 11

Evidence & Claim Register

Complete traceability classifying every number, equation, default rule, and schematic.

A Responsible Deterministic Framework for CT Radiation Governance

CT radiation optimization is both an upstream clinical stewardship challenge and a technical physics discipline. Because a substantial share of avoidable radiation exposure is associated with ordering, repeat imaging, and multiphase protocoling decisions prior to scan acquisition (Hendee et al., Radiology 2010;257:240–245; ICRP Publication 103), pre-scan justification governance is an essential partner to console-level dose management.

⚕️ Core Thesis & Testable Stance

CT provides indispensable clinical value: rapid acquisition, high spatial resolution, emergency triage certainty, and low operator dependency. Accordingly, PAGER-CT is NOT anti-CT and NOT anti-emergency imaging.

To remain both clinically safe and scientifically falsifiable, every governance rule in PAGER-CT is structured around four explicit elements: (1) Governing Principle, (2) Proposed Operational Default, (3) Documented Clinical Exceptions, and (4) Local Validation Protocol.

🏢 The Layered Governance Architecture

🛡️

J-Gate

Justification & Phase Defaults

⚙️

PAGER-CT

5-Domain Execution Layer

🏛️

RATSe

5-Pillar Governance Overlay

📈

Stewardship

Override Audit & Size-Binned DRLs

⚖️ How PAGER-CT Differs from Unilateral Models

Dimension	Naive Dose Minimization (Misapplied ALARA)	Ungoverned Black-Box AI	PAGER-CT Governance Framework
Optimization Philosophy	Unconditional dose reduction without task-specific noise ceilings	Opaque post-processing denoising without upstream justification	Indication-oriented, patient-specific sufficiency (proposed adaptation of ALADA/ALADAIP concepts with ICRP justification/optimization)
Size-Adjusted Metric	CTDIvol (mGy — Unadjusted 16/32 cm Phantom Output)	Proprietary Vendor Quality Indices	SSDE (mGy — Size-scaled CTDIvol estimating central-section patient dose per AAPM 204/220/293, paired with CTDIvol & DLP)
Governance Scope	Scanner Console Parameters Only	Image Reconstruction Only	End-to-End: J-Gate Justification + Exception Auditing + 5-Domain Execution
Explainability & Validation	Static Lookup Charts	Non-traceable Neural Weights	Explicit Principle → Default → Exception → Local Validation Register

🗺️ The PAGER-CT Master System Map

A unified doctrine mapping the flow from clinical uncertainty to deterministic physics optimization, governed by a continuous lifecycle stewardship loop.

Continuous Stewardship Audit ↺

▶

🏥 Clinical Pressures

ED Throughput • Medicolegal Asymmetry • Reassurance & Serial Follow-Up

▼

🛡️ J-Gate (Justification Layer)

Clinical Necessity • Prior Imaging Review • Indication-Specific Phase Defaults & Exceptions

▼

⚙️ PAGER-CT (Operational Layer)

P — Phenotype & Size (SSDE, mGy)

A — Anatomy

G — Geometry

E — Examination

R — Reconstruction

▼

🏛️ RATSe (Canonical 5-Pillar Governance Overlay)

Responsibility • Accountability • Transparency • Safety & Security • ethics, Equity & Environment

▼

📈 Lifecycle Stewardship Loop

J-Gate Override Audits • Size-Binned SSDE (mGy) & DLP Benchmarks • Task Adequacy Review

🏥 The Clinical Reality & Urgency

A practical governance framework must account for the real-world clinical and operational pressures that shape imaging utilization. Avoidable radiation exposure rarely stems from a simple lack of physics knowledge; it frequently emerges from systemic workflow and decision-making pressures upstream of the scanner table (Hendee et al., Radiology 2010;257:240–245).

⏱️ High-Volume Utilization & Throughput

In acute care and emergency settings, CT serves as a rapid uncertainty-resolution tool. A cross-sectional scan takes minutes and accelerates disposition, whereas operator-dependent modalities or observation carry operational delays.

⚖️ Medicolegal Asymmetry & Defensive Ordering

Retrospective hindsight bias reframes diagnostic uncertainty as omission. Following adverse clinical outcomes, scrutiny frequently centers on why a scan or additional contrast phase was not performed, encouraging defensive multiphase escalation.

🧘 Reassurance Imaging & Serial Follow-Up

Cross-sectional imaging is often ordered to achieve rapid clinical closure. In recurrent presentations (e.g., chronic stone formers) or stable surveillance, repeat or multiphase exams may occur without documented interval clinical change.

Qualitative Taxonomy of Ordering Pressures

Conceptual Schematic

Schematic — Not Empirical Data: Equal-quadrant conceptual taxonomy illustrating the four interacting domains that J-Gate governance balances. No empirical percentage split is implied.

🛡️ J-Gate: Justification Governance Layer

Because ordering and protocoling decisions upstream of acquisition strongly influence patient exposure (Hendee et al., Radiology 2010;257:240–245; ICRP Publication 103), justification must precede parameter optimization. Rather than enforcing rigid blanket prohibitions, J-Gate establishes indication-specific protocol defaults with documented clinical exceptions and an immediate emergency fast-track pathway.

Core Governance Variables

• 1. Clinical Question & Scope: What specific pathology is suspected, and will the exam alter management?
• 2. Prior Exam Adequacy: Triggers review (not automatic cancellation) when recent diagnostic imaging is on file.
• 3. Indication-Specific Phase Default: Matches contrast phases to the suspected pathology, with explicit exception pathways.
• 4. Emergency Fast-Track: Unconditional zero-delay bypass for unstable, trauma, or stroke patients.

Interactive Deterministic Rule & Exception Simulator

Proposed Governance Logic

Select a clinical scenario to inspect the 4-part governance structure: Principle → Proposed Default → Documented Exceptions → Validation Pathway.

SCENARIO: Recurrent Flank Pain / Known Urolithiasis (Prior CT < 12 Months) • PRINCIPLE: Avoid routine multiphase or standard-dose repeat scanning in uncomplicated recurrent stone formers. • TRIGGER_REVIEW: Prior CT < 12 months flags clinical review of prior stone burden and adequacy (does not auto-cancel). • PROPOSED_DEFAULT: Ultrasound first or Low-Dose Non-Contrast CT KUB (single phase) for uncomplicated colic. • DOCUMENTED_EXCEPTIONS: Escalate to IV contrast / multiphase (e.g., CT Urography) if fever/leukocytosis (suspected pyonephrosis/abscess), unexplained hematuria, solitary kidney, or atypical vascular pain. • VALIDATION_PATHWAY: Log exception reason in RIS; audit low-dose stone CT diagnostic adequacy and urology intervention rates.

🧬 Governing Phase Creep Through Defaults & Exceptions

Unnecessary multiphase scanning (e.g., ordering non-contrast + arterial + portal venous + delayed phases by habit for routine follow-up) multiplies per-exam CTDI_vol (mGy), DLP (mGy·cm), and SSDE (mGy). J-Gate addresses this not by banning multiphase protocols, but by setting indication-specific single-phase defaults (such as single portal venous phase for routine restaging of non-hypervascular malignancies) while preserving explicit, pre-authorized clinical exceptions for hypervascular tumors (HCC, neuroendocrine, RCC, melanoma), suspected active hemorrhage, or urothelial evaluation.

📋 Operational Mapping of Modifiable Factors

Every modifiable pre-scan and acquisition variable maps into the layered J-Gate, PAGER-CT, and RATSe architecture—ensuring each technical and workflow factor has an explicit governance home.

🛡️

J-Gate (Justification Governance Layer)

Clinical & Repeat Imaging Review

• Emergency vs. elective / Baseline vs. follow-up
• Alternative non-ionizing modality (US / MRI)
• Prior diagnostic CT review trigger
• Documented interval clinical change

Phase Defaults & Exceptions

• Indication-specific single-phase defaults
• Documented multiphase exception criteria
• Focused anatomic z-coverage vs. pan-scan
• Dual-energy virtual non-contrast where validated

Fast-Track Override Pathways

• Life-threatening polytrauma / ICU instability
• Acute intracranial hemorrhage / stroke code
• Active arterial hemorrhage / aortic syndrome

P — Phenotype & Size Domain

• Age & pediatric vs. adult sizing
• Sex & pregnancy verification
• Water-equivalent / effective diameter (D_w / D_eff for SSDE in mGy)
• Body habitus extremes (Pediatric, Cachectic, Bariatric)
• Breath-hold cooperation & motion risk

A — Anatomy Domain

• Anatomic region attenuation profile
• Regional radiosensitive organs (lens, thyroid, breast, gonads)
• Diagnostic task contrast mechanism
• Task-specific noise tolerance
• Low-contrast vs. high-contrast detectability

G — Geometry Domain

• Vertical & lateral isocenter alignment (Bowtie/AEC)
• Scout projection (AP, PA, Lateral) & table-height interaction
• Z-axis scan length discipline (primary DLP driver)
• Arm elevation outside abdominal/thoracic beam
• Prosthesis/metal positioning awareness

E — Examination Domain

• Tube potential (kV) & tube current modulation (AEC)
• Automated kV selection & reference quality indices
• Contrast bolus tracking & timing accuracy
• Pitch, rotation time & collimation selection
• Prevention of contrast-timing failure rescans

R — Reconstruction Domain

• Iterative reconstruction (IR) blending validation
• Deep learning image reconstruction (DLIR) validation
• Slice thickness & reconstruction kernel matching
• Noise Power Spectrum (NPS) texture preservation
• Metal Artifact Reduction (MAR) algorithms
• Task-specific low-contrast detectability verification

🏛️

RATSe Canonical 5-Pillar Overlay & Human Factors

Canonical RATSe Pillars

• R — Responsibility: Indication-justified imaging
• A — Accountability: Auditable override & exception logs
• T — Transparency: Traceable evidence & rule logic
• S — Safety & Security: Anti-underdosing & MAR safeguards
• e — Ethics, Equity & Environment: Habitus fairness & sustainable scanner utilization

Workflow & Human Factors

• Technologist centering & positioning feedback
• ED throughput pressure mitigation
• Standardized nomenclature (RadLex / AAPM)
• Radiologist–physicist–technologist protocol review

Decision Support & Risk Mitigation

• Automated prior-exam review prompts in CDS/RIS
• 3D camera automated table-height centering
• Bariatric photon-starvation safeguards
• Motion & bolus-timing failure root-cause review

Core Architectural Insight

PAGER-CT is NOT merely a scanner console lookup table. It is a testable governance architecture integrating patient habitus, regional anatomy, projection geometry, examination physics, image reconstruction validation, and institutional accountability.

🧩 PAGER-CT: Operational Core Domains

Once an exam passes J-Gate justification, PAGER-CT governs how the CT is executed safely across five operational domains—pairing each domain's principle with a proposed default and local validation requirement.

P — Phenotype & Size Domain

Accounts for patient age, physical habitus, and water-equivalent attenuation diameter (D_w). Adapts kV/mAs selection for pediatric vs. bariatric patients and evaluates cooperation/motion risk.

Dose Metric Anchor: SSDE (mGy) — estimate of average absorbed dose to the central section of the patient's scan via size-scaled phantom CTDI_vol (AAPM 204/220/293); requires specifying reference phantom (16/32 cm) and method (D_eff/D_w).

A — Anatomy Domain

Accounts for the intrinsic contrast mechanism of the target organ system and regional radiosensitive tissues (lens, thyroid, breast, gonads) within or adjacent to the primary beam.

Governance Default: Match target noise index / reference mAs to anatomic & task contrast requirements.

G — Geometry Domain

High-impact operational domain. Governs isocenter alignment, scout projection geometry (AP/PA/Lateral), arm positioning, and strict z-axis coverage boundaries.

Proposed Default: Target vertical centering within ±2 cm of gantry isocenter prior to scout acquisition.

E — Examination Domain

Governs indication-driven kV/AEC modulation, contrast bolus timing, pitch, and phase discipline. Implements indication-specific single-phase defaults with documented clinical exceptions.

Objective: Indication-specific spectral & temporal execution validated per scanner.

R — Reconstruction Domain

Governs post-projection image generation (Filtered Back Projection, Iterative Reconstruction, Deep Learning Reconstruction, and MAR). Because reconstruction strength, dose, Noise Power Spectrum (NPS) texture, and lesion detectability interact in vendor-specific ways, protocols require task-specific validation.

Validation Requirement: Any IR/DLIR blending cap or minimum photon-flux floor for low-contrast parenchymal tasks (e.g., hepatic metastases, acute stroke) must be locally validated per scanner model, reconstruction algorithm, and clinical task.

⚛️ Operational Physics & SSDE Integration

Rigorous CT stewardship requires accurate medical physics: understanding how vertical miscentering interacts conditionally with scout projection and Bowtie filtration, separating fixed-mAs tube potential (kV) scaling from clinical iso-noise AEC behavior, and validating reconstruction noise texture.

Geometry Domain: Conditional Scout Miscentering & Bowtie Filter Pathway

Published Literature Conditional Causal Schematic

Automatic Exposure Control (AEC) estimates patient attenuation and projected width from the localizer scout (topogram). Crucially, the magnitude and direction of tube-current (mAs) error from vertical miscentering are not universal constants—they depend on scout projection geometry (AP overhead tube vs. PA under-table tube vs. lateral), table attenuation in the scout, and vendor AEC implementation:

Pathway A: Patient Shifted Closer to Active Scout Tube

(e.g., Table too HIGH in an AP overhead-tube scout, OR table too LOW in a PA under-table scout)

Geometric Mechanism: Fan-beam divergence magnifies the projected silhouette on the detector.
AEC Consequence: Size-based AEC interprets the wider shadow as a larger patient and escalates tube current (mAs) and dose.

Target State: Patient Centered at Gantry Isocenter

(Proposed Operational Default: Vertical offset within ±2 cm of isocenter)

Geometric Mechanism: True 1:1 projected width at isocenter; thickest central anatomy aligns with the center of the physical Bowtie Filter.
AEC Consequence: Accurate mAs modulation and uniform noise distribution.

Pathway B: Patient Shifted Farther from Active Scout Tube

(e.g., Table too LOW in an AP overhead-tube scout, OR table too HIGH in a PA under-table scout)

Geometric Mechanism: Fan-beam geometry minifies the projected silhouette.
AEC Consequence: AEC under-estimates patient size and reduces mAs, increasing quantum noise and risking non-diagnostic image quality.

📚 Published Study-Specific Findings (Not Universal Correction Factors): • Kaasalainen et al. (AJR 2014;203:123–130, PMID: 24951205): Using a PA scout with ±6 cm vertical offsets, the adult phantom's relative dose increased by up to 38% at the lowest table position and decreased by up to 23% at the highest table position; lateral scouts showed much smaller dose variation.
• Li et al. (Med Phys 2007;34:3814–3820) & AP Scout Geometry: With an overhead AP scout, raising the table above isocenter (closer to the overhead focal spot) magnifies the scout and increases dose, whereas lowering the table reduces AEC output; table attenuation entering the scout also modifies AEC response.

🛡️ Bowtie Filter Penalty & Testable Governance Stance: • Direction-Independent Bowtie Penalty: Regardless of AP or PA scout direction, any vertical offset misaligns the thickest central patient anatomy with the thinner periphery of the physical Bowtie filter, increasing surface dose and image noise.
• Proposed Default & Local Validation: Target table centering within ±2 cm of isocenter (assisted by 3D camera centering where available) and audit centering compliance stratified by scanner model and scout projection.

Examination & Phenotype Domains: Tube Potential (kV) Physics vs. Clinical AEC Optimization

Derived Physics Model Published Principle

To avoid conflating fixed-parameter physics with clinical Automatic Exposure Control (AEC) performance, PAGER-CT separates the fixed-mAs radiation output relationship from clinical iso-noise iodine optimization:

Display 1: Fixed-mAs Relative Output (Approximate Power-Law) D / D₁₂₀ ≈ (kV / 120)^2.6

At strictly constant mAs, scanner output (CTDI_vol) varies approximately with the 2.5 to 3.0 power of tube potential (depending on beam filtration and phantom size). The bars below show the derived output ratio for an illustrative exponent of 2.6—not a clinical AEC dose prediction.

Derived Fixed-mAs Values: 70 kV = 0.25 | 80 kV = 0.35 | 100 kV = 0.62 | 120 kV = 1.00 (Reference).

Display 2: Clinical Optimization Under AEC Iodine K-Edge: 33.2 keV

Lowering tube potential (e.g., from 120 kV to 80 or 100 kV) shifts the mean X-ray energy spectrum closer to the K-edge of Iodine (33.2 keV), substantially increasing photoelectric attenuation and iodine HU conspicuity.

1. Why Constant-mAs ≠ Clinical Dose: At fixed mAs, lowering kV increases iodine HU contrast but also increases quantum noise due to reduced photon penetration.

2. Clinical Iso-Noise AEC Modulation: In clinical protocols (e.g., automated kV/mAs selection), the scanner increases mA partially at lower kV to maintain target noise. Net dose savings and contrast-to-noise efficiency (CNR²/CTDI_vol) depend on patient water-equivalent diameter (D_w), contrast concentration, scanner filtration, and clinical task (highest benefit in CT angiography and pediatric/small-adult habitus).

3. Bariatric Safeguard: In large-habitus patients, low kV risks hitting the X-ray tube current (mA) ceiling and causing photon starvation—requiring escalation to 120–140 kV or tin-filtered/dual-energy modes.

Validation Rule: Evaluate net CTDI_vol (mGy), SSDE (mGy), and task CNR across patient D_w strata on each local scanner before locking automated kV tables.

Reconstruction Domain: Noise Power Spectrum (NPS) & Task-Specific Protocol Validation

Task-Specific Validation Requirement

Iterative Reconstruction (IR) and Deep Learning Image Reconstruction (DLIR) reduce pixel noise magnitude (standard deviation of HU). However, strong non-linear regularization can shift the Noise Power Spectrum (NPS) toward lower spatial frequencies and alter task-specific Modulation Transfer Function (MTF_task)—producing an over-smoothed or blotchy texture that can impair human observer detectability for small, low-contrast lesions despite favorable simple SNR/CNR numbers.

⚖️ Governing Principle (Task-Coupled Reconstruction): Reconstruction strength, radiation output, noise texture, and diagnostic performance must be evaluated together. High-intrinsic-contrast tasks (urolithiasis, CT angiography, bone, pulmonary nodules) tolerate stronger denoising and lower photon flux than subtle low-contrast parenchymal tasks (hepatic metastases, pancreatic masses, acute ischemic stroke).

🔬 Local Scanner & Protocol Validation Requirement: Because there is no universal reconstruction-strength cap or minimum photon-flux value that applies across all vendors, scanner models, and DLIR/IR algorithms, any strength ceiling or minimum mAs/CTDI_vol floor must be defined and validated locally for the specific scanner, algorithm, and clinical task using phantom low-contrast detectability and radiologist reader sign-off.

⚖️ Diagnostic Sufficiency (ALADA / ALADAIP Adaptation) & SSDE Safeguards

The objective of clinical CT optimization is not unconditional dose reduction, but achieving the lowest exposure that reliably preserves task-specific diagnostic confidence.

🏛️

Distinguishing ICRP ALARA, Dental CBCT ALADA/ALADAIP, and PAGER-CT

Provenance & Scope

In ICRP guidance (ICRP Publication 103), the principle of optimization (ALARA — As Low As Reasonably Achievable) is always conditioned on maintaining diagnostic efficacy. The clinical hazard is not the ICRP principle itself, but the naive misapplication of dose minimization without adequate attention to task-specific diagnostic image quality.

📌 Canonical PAGER-CT Positioning Statement:

"PAGER-CT adopts an indication-oriented, patient-specific approach to diagnostic sufficiency, drawing on ALADA/ALADAIP concepts while retaining established ICRP justification and optimization principles."

Historical Provenance vs. Proposed Adaptation:
• Original Context (Dental & Pediatric CBCT): ALADA (As Low As Diagnostically Acceptable) emerged in NCRP discussions (Bushberg JT, Health Phys 2014;106:122–130) and dental cone-beam CT literature (Jaju PP, Jaju SP. Imaging Sci Dent 2015;45(4):263–265). The European DIMITRA Research Group (Oenning AC et al., Pediatr Radiol 2018;48(3):308–316) expanded this for pediatric dental CBCT into ALADAIP (As Low As Diagnostically Acceptable being Indication-oriented and Patient-specific; timeline reviewed in Braz Oral Res 2025;39:e14, PMID: 39936713; PMCID: PMC11808696).
• PAGER-CT Scope: Applying ALADA/ALADAIP to multi-detector and photon-counting whole-body clinical CT is presented here as a proposed operational governance adaptation rather than an already-standardized whole-body regulatory metric.

Task-Specific Sufficiency Knee-Point

Schematic — Not Empirical Data

Required photon flux depends on intrinsic lesion contrast. Low-contrast hepatic metastasis detection requires lower noise than high-contrast ureteral stone detection.

Unquantified Conceptual Schematic: Illustrates the qualitative plateau of diagnostic confidence above the task-sufficiency threshold. Axes carry no empirical percentage values.

The False Economy of Underdosing

Schematic — Not Empirical Data

PAGER-CT enforces the Anti-Underdosing Principle: reducing exposure below task sufficiency degrades diagnostic confidence and increases the likelihood of indeterminate reports or repeat imaging.

Unquantified Conceptual Schematic: Illustrates the inverse qualitative trade-off between severe underdosing and indeterminate/rescan risk. Not an empirical probability curve.

The CT Dose Metrics Hierarchy & Mandatory SSDE Safeguards

Metric	Standard Unit	Physical Meaning & Governance Role in PAGER-CT
CTDIvol	mGy	Standardized scanner radiation output measured in a 16 cm or 32 cm cylindrical PMMA phantom. Quantifies scanner output—does not account for individual patient attenuation diameter.
DLP	mGy·cm	Dose-Length Product (CTDIvol × scan length). Primary operational metric for auditing z-axis scan length discipline and multiphase utilization.
SSDE	mGy	Size-Specific Dose Estimate (AAPM Reports 204, 220 & 293). Estimate of the average absorbed dose to the central section of the patient's scan volume (in mGy), calculated by scaling phantom CTDIvol by size-dependent conversion factors (fsize) derived from effective diameter (Deff, Report 204), water-equivalent diameter (Dw, Report 220), or head CT conversion factors (Report 293).
Organ Dose / Effective Dose	mGy / mSv	Monte Carlo tissue-specific absorbed dose (mGy) and ICRP tissue-weighted reference population risk index (mSv). Used for population-level epidemiology/risk modeling rather than individual console optimization.

🛡️ Three Mandatory Methodological Safeguards for SSDE Reporting & Benchmarking

1. Specify Reference Phantom & Method: Always specify the reference PMMA phantom (32 cm for adult body CT; 16 cm for head/pediatric standards per protocol) and whether size was determined geometrically (D_eff, AAPM 204) or by X-ray attenuation (D_w, AAPM 220). Values from different methods are not directly interchangeable without specification.

2. Stratify by Body Region, Indication & Size Bin: Never apply a single SSDE figure as a universal benchmark across anatomical regions, clinical tasks (e.g., low-dose stone vs. multiphase liver), or patient-size strata (cf. size-binned DRL methodology in Kanal et al., Radiology 2017;284:120–133).

3. Do Not Equate SSDE with Organ Risk or Cumulative Harm: SSDE is a central-section physical absorbed-dose index (mGy). It is not individual organ dose, effective dose (mSv), biological risk, or a validated metric for summing cumulative dose across heterogeneous serial exams.

🌐 Global Guidelines & Strategic Integration

PAGER-CT does not replace established radiation protection initiatives—it integrates their principles into an auditable, workflow-embedded governance architecture.

📚 Integrated Reference Frameworks

🦉

ACR Image Wisely® & Image Gently®

Operationalizes American College of Radiology campaigns for adult and pediatric dose stewardship by embedding Phenotype & Size checks, indication-specific phase defaults, and size-binned SSDE (mGy) tracking into departmental protocol governance.

📊

AAPM Reports 204 (2011), 220 (2014) & 293 (2019)

Implements American Association of Physicists in Medicine standards for Size-Specific Dose Estimate (SSDE in mGy): body SSDE using geometric effective diameter D_eff (Report 204), water-equivalent attenuation diameter D_w (Report 220), and head CT SSDE (Report 293), paired with U.S. size-binned diagnostic reference levels and achievable doses (Kanal et al., Radiology 2017;284:120–133).

🌍

ICRP Publications 103 & 135 and ESR EuroSafe Imaging

Upholds the International Commission on Radiological Protection pillars of Justification (via J-Gate), Optimization (via PAGER-CT), and Diagnostic Reference Levels (ICRP Publication 135), supporting the European Society of Radiology EuroSafe Imaging Call for Action.

🎯

Proposed Whole-Body CT Adaptation of ALADA (2014–2015) & ALADAIP (2017–2018)

Adapts the NCRP/CBCT ALADA concept (Bushberg 2014; Jaju PP & Jaju SP, Imaging Sci Dent 2015;45:263–265) and the European DIMITRA project's pediatric CBCT ALADAIP concept (Oenning AC et al., Pediatr Radiol 2018;48:308–316; reviewed in Braz Oral Res 2025;39:e14, PMC11808696) into a proposed 5-domain governance model for clinical multi-detector and photon-counting CT.

🏛️ RATSe Governance Overlay, Failure Modes & KPIs

While PAGER-CT governs technical acquisition and reconstruction validation, the RATSe Governance Overlay provides the canonical five-pillar institutional accountability framework to audit upstream J-Gate overrides, prevent protocol drift, and manage clinical edge cases.

The Canonical 5 Pillars of RATSe

Standardized 5-letter governance expansion used uniformly across the PAGER-CT architecture and slide deck

R

Responsibility

Every exposure must be clinically justified via J-Gate, indication-driven, and tailored to the patient's physical habitus and clinical urgency.

A

Accountability

All major acquisition overrides (phase exceptions, manual kV/mAs escalation, repeat scans, J-Gate bypasses) are logged for periodic multidisciplinary review.

T

Transparency

Deterministic protocol defaults, documented clinical exceptions, dose metrics (CTDI_vol, DLP, SSDE), and evidence status remain transparent and traceable.

S

Safety & Security

Enforces the Anti-Underdosing Principle, photon-starvation safeguards in obesity, metal artifact governance, and motion-arresting protocols.

e

Ethics, Equity & Environment

Ensures equitable protocol quality across pediatric, adult, and bariatric habitus, while reducing avoidable rescans, contrast waste, and scanner energy/tube wear.

🎯 Why PAGER-CT Goes Beyond a Standard DRL Audit: Bridging Dose Indices & Clinical Stewardship

Because SSDE (mGy), CTDI_vol (mGy), and DLP (mGy·cm) are physical dose indices rather than direct measures of biological risk or diagnostic sufficiency, a conventional Diagnostic Reference Level (DRL) audit only evaluates scanner output after the decision to scan has already been made. PAGER-CT bridges this gap by leading with upstream J-Gate Justification and Exception Auditing (tracking avoided exams, indication-specific phase defaults, and documented clinical exceptions) alongside downstream diagnostic adequacy safeguards (centering compliance and technical rescan monitoring).

Explicit Failure Mode Governance

Obesity & Photon Starvation ▼

Mechanism: Large water-equivalent diameter (D_w) strongly attenuates low-energy photons. If low kV is used, the X-ray tube hits its maximum mA ceiling, causing severe photon starvation and streak artifacts.

Governance Action: Disable low-kV selection above scanner-validated D_w thresholds; escalate to 120–140 kV with longer rotation time or dual-source mode, accepting higher CTDI_vol (mGy) to ensure a diagnostic exam on the first pass.

Metal Implants & AEC Over-Ranging ▼

Mechanism: High-Z orthopedic hardware causes severe beam hardening and photon starvation. Standard AEC reads the dense metal on the scout and spikes tube current across the anatomic zone.

Governance Action: Enable high-kV / tin-filtered or Dual-Energy virtual monochromatic (110–140 keV) acquisition with Metal Artifact Reduction (MAR) reconstruction, and cap AEC mAs spikes over metal zones.

Pediatric / Uncooperative Motion vs. Dose ▼

Mechanism: Respiratory or gross patient motion blurs spatial resolution, often forcing a repeat acquisition that doubles total radiation and contrast exposure.

Governance Action: Prioritize temporal resolution (fastest gantry rotation, high-pitch helical mode) over aggressive mAs reduction. A single motion-free scan is always superior to a failed ultra-low-dose scan followed by a repeat.

Operational Stewardship Dashboard

Hypothetical Governance Mockup

💡 Evidence Status Disclosure: All KPI values below are explicitly tagged as [Illustrative Synthetic Value] or [Proposed Local Target — Pending Validation] to demonstrate how a department operationalizes PAGER-CT auditing. None represent empirical clinical trial results.

112 Synthetic Example

J-Gate Exceptions Logged

Monthly upstream audit of phase & repeat-scan exception rationale

88% Target: ≥85%

Centering Compliance

Proposed local target: ≥85% within ±2 cm of gantry isocenter

14% Synthetic Example

Unindicated Multiphase Reduction

Illustrative YoY metric tracking adherence to single-phase defaults

4.2% Proposed Ceiling

Technical Rescan Ceiling

Proposed local ceiling tracking motion, bolus-timing & underdosing repeats

Stratum-Specific Adult Abd/Pelvis SSDE Mockup Proposed Local Target

Scope Specification: Single-phase routine adult Abd/Pelvis CT, 29–33 cm D_w bin (32 cm PMMA phantom, AAPM 220). Contextualized against 15 mGy published U.S. 75th-percentile reference for the 29–33 cm bin (Kanal et al., Radiology 2017;284:120–133).

12.4 mGy

Lifecycle Stewardship Integration

Ultimately, PAGER-CT with J-Gate and RATSe forms a testable, end-to-end clinical governance architecture—connecting upstream clinical justification and indication-specific phase defaults, through the conditional physics of patient centering, spectral kV selection, and task-validated reconstruction, to transparent longitudinal auditing.

Institute for Responsible Healthcare AI

🔍 Evidence & Claim-to-Source Register

In accordance with the Transparency pillar of RATSe, every quantitative value, physical relationship, clinical rule, and visual schematic displayed in this platform is classified below into one of four explicit evidence categories, alongside its exact source and operational scope.

1. Published Literature

Peer-reviewed clinical/phantom study or consensus report (AAPM, ICRP, NCRP) with specified experimental scope.

2. Derived Physics Model

Calculated directly from an explicit physical equation with stated boundary conditions (e.g., fixed mAs).

3. Proposed Local Target

Testable institutional default or governance threshold proposed by PAGER-CT, pending local clinical validation.

4. Conceptual Schematic

Unquantified visual diagram or synthetic dashboard mockup illustrating system workflow; zero empirical interpretation.

Claim / Displayed Figure	Evidence Status	Primary Source / Equation	Applicable Scope & Qualification
Upstream ordering, repeat imaging & multiphase utilization drive avoidable CT exposure	Published Literature	Hendee WR et al., Radiology 2010;257(1):240–245; ICRP Pub 103 (2007)	Supports pre-scan justification governance (J-Gate) alongside console parameter optimization.
Vertical miscentering (±6 cm) altered adult phantom dose by +38% (lowest table) to −23% (highest table) on PA scout	Published Literature	Kaasalainen T et al., AJR 2014;203(1):123–130 (PMID: 24951205); Li J et al., Med Phys 2007;34:3814–3820	Study- and scanner-specific phantom results. Direction reverses between PA (under-table) and AP (overhead) scouts; Bowtie mismatch increases surface dose/noise in both directions.
Isocenter centering tolerance within ±2 cm and ≥85% departmental compliance target	Proposed Local Target	PAGER-CT Geometry Domain Governance Rule	Operational quality target pending local validation per scanner model and 3D-camera availability (88% dashboard value is an illustrative synthetic mockup).
Fixed-mAs relative CTDIvol ratios: 0.25 (70 kV), 0.35 (80 kV), 0.62 (100 kV), 1.00 (120 kV)	Derived Physics Model	D / D120 = (kV / 120)2.6 at constant mAs; Iodine K-edge = 33.2 keV	Valid only for fixed mAs with an approximate exponent of 2.6. Under clinical iso-noise AEC, the scanner increases mA at lower kV, so net dose savings must be evaluated per patient Dw and scanner.
SSDE (mGy) definition & conversion methodology (Deff, Dw, Head CT)	Published Literature	AAPM Reports 204 (2011), 220 (2014), and 293 (2019)	Estimates average absorbed dose (mGy) to the central section of the scan volume. Requires specifying reference phantom (16/32 cm) and method; not organ dose, effective dose (mSv), or cumulative risk.
15 mGy Adult Abdomen-Pelvis SSDE reference vs. 12.4 mGy local mockup	Published Ref Mockup Target	Kanal KM et al., Radiology 2017;284(1):120–133 (U.S. ACR Dose Index Registry)	15 mGy is the published U.S. 75th-percentile DRL for single-phase adult abdomen-pelvis CT in the median 29–33 cm size bin (12 mGy achievable dose). 12.4 mGy is a hypothetical local dashboard example for that specific stratum.
ALADA (2014–2015) and ALADAIP (2017–2018) provenance vs. whole-body CT adaptation	Published Provenance Proposed Adaptation	Bushberg 2014; Jaju PP & Jaju SP, Imaging Sci Dent 2015;45:263–265; Oenning AC et al., Pediatr Radiol 2018;48:308–316; Braz Oral Res 2025;39:e14 (PMC11808696)	Originally developed for radiation protection and dental/pediatric CBCT; extended in PAGER-CT as a proposed indication-oriented, patient-specific operational framework for clinical CT.
Task-specific validation of IR/DLIR reconstruction strength and low-contrast detectability	Proposed Validation Rule	PAGER-CT Reconstruction Domain Governance Rule	Replaces rigid universal IR/DLIR caps with a requirement to validate reconstruction blending and minimum photon flux per scanner, algorithm, and low-contrast task.
Ordering Pressures Taxonomy, ALADA Sufficiency Curve, False Economy Curve, and KPI Dashboard Counts (112, 14%, 4.2%)	Conceptual Schematic	PAGER-CT Exhibit Visual Models	Unquantified conceptual schematics and synthetic governance dashboard examples; no empirical percentages or clinical probability estimates are claimed.