Seminar Review: Caries Risk Assessment Tools

Suggested New Title (Question Format)

"Can We Predict Tooth Decay Before It Happens? A Critical Look at Caries Risk Assessment Tools"

• "Which Caries Risk Assessment Tool Should We Use - And Why?"
• "Are We Assessing Caries Risk the Right Way? Evaluating Tools and Their Clinical Relevance"
• "How Well Can We Predict Dental Caries? An Evidence-Based Appraisal of Risk Assessment Tools"

Slide-by-Slide Suggestions

Slide 1 - Title Slide

• Change the title as suggested above.
• Add your department name and institution.
• Remove "Seminar" as a floating text box - it looks unprofessional. Integrate it subtly (e.g., "MDS Seminar" as a subtitle).
• Add a date and a relevant background image (e.g., dental radiograph or a caries progression diagram).

Slide 2 - Contents

• Rename "Content" to "Overview" - sounds more academic.
• Group the content into sections for visual clarity:
  • Section 1: Background and rationale
  • Section 2: Individual tools (Cariogram, CAMBRA, ICDAS, etc.)
  • Section 3: Comparative evidence
  • Section 4: Emerging technologies and AI
  • Section 5: Conclusion
• This gives the audience a roadmap and makes your seminar feel structured.

Slide 3 - Introduction

• The introduction text is cut off ("to which the tooth is directly..."). Complete the definition of risk factor.
• Add a caries etiology diagram (Keyes triad or the expanded White spot model - bacteria, substrate, host, time).
• Split this slide into two slides:
  • Slide A: Definition of caries risk assessment + why it matters
  • Slide B: Definitions of risk factor, risk indicator, and risk marker (these are often confused and make for a good opening discussion point).

Slide 4 - Why Assess?

• Good content but presented as raw text. Convert to a 2-column visual:
  • Left: "Without CRA" (reactive treatment, missed high-risk patients, resource wastage)
  • Right: "With CRA" (targeted prevention, efficient resource use, better outcomes)
• Add a brief statistic here - e.g., prevalence of untreated caries globally or in India.

Slides 5-15 - Cariogram

• This section is too long and fragmented (10 slides). Condense to 4-5 slides:
  1. What is Cariogram + its origin (Bratthall, 1997)
  2. The pie chart graphic showing sectors (bacteria, diet, susceptibility, circumstances, green "chance to avoid caries")
  3. How to use it - scoring criteria (keep the scoring table slide)
  4. Interpretation of the green sector
  5. Advantages + Limitations (combine Slides 16 and 17)
• Slide 6 and Slide 14/15 appear to have only images with no text - make sure each image has a caption and a brief explanatory line.

Slides 18-22 - CAMBRA

• Good coverage. Add the CAMBRA form visual (the actual risk assessment form layout) if not already on Slide 20.
• Slide 22 has a flowchart that seems incomplete. Ensure it shows: Risk Assessment → Risk Stratification (Low/Moderate/High/Extreme) → Individualized Plan → Re-evaluation interval.
• Mention the 2023 update to CAMBRA guidelines if applicable.

Slides 23-24 - ICDAS

• The ICDAS score descriptions are cut off at Score 1. Include all 7 scores (0-6) with brief descriptions and ideally clinical photos.
• Add a comparison table: ICDAS vs conventional DMFT - why ICDAS is more sensitive for early lesion detection.
• Explain the ICDAS-II modification briefly.

Slides 25-27 - AAPD and ADA Forms

• Slide 25 text is cut off. Ensure the risk categorization (Low/Moderate/High) is fully described.
• Slide 27 has the ADA form title but no content visible. Either include the form image with annotation or describe the key risk indicators.
• Highlight the difference between AAPD (pediatric focus, age-stratified: 0-5 yrs vs ≥6 yrs) and ADA (adult-focused) forms.

Slide 28 - Dundee Model

• The text is cut off. Complete the description of the CHAID analysis and its results.
• Clarify how this model is different from or complementary to the other tools.
• If this is a niche tool, consider whether it deserves a full slide or can be briefly mentioned under "other tools."

Slides 29-30 - Traffic Light Matrix (TLM)

• Good concept. Add the actual color-coded matrix image if not already present.
• Mention that TLM is particularly useful for patient communication and motivation.

Slide 31 - CRAFT Approach

• Good addition. The text is cut off ("A green star is used to indicate 'sa..."). Complete it.
• Add the CRAFT chart image with the four parameters clearly labeled.

Slides 34-35 - CRA-RT and CARE Test

• CRA-RT: Good clinical pearl for community settings. Mention the 11 items briefly and the cut-off score concept.
• CARE test: The mechanism (salivary oligosaccharides) is well-explained. Add the sensitivity and specificity values if available from the original study.

Slides 36-37 - Recent Advances / AI

• Separate AI into its own dedicated slide - this is a major talking point in 2024-2026 and deserves more depth:
  • Machine learning models (Random Forest, CNN for radiograph analysis)
  • Apps (Cariometer, others)
  • Limitations of AI (data bias, lack of clinical validation, black-box problem)
• The Lactic acid impression slide content appears cut off. Complete it.

Slides 38-40 - Research / Literature Review

• Slide 40 is a duplicate of Slide 39 (same Agouropoulos 2019 study). Remove the duplicate.
• Add a summary comparison table across these research slides:

  Tool	Population	Validity	Preferred For
  CAMBRA	Adults	High	General practice
  Cariogram	Adults/Children	High	Research + clinical
  AAPD CRAF	Preschool	Moderate	Pediatric practice

• Add at least one Indian study on CRA tool applicability.

Slide 41 - Conclusion

• The conclusion text is cut off. Complete it.
• Structure the conclusion around your title question: "So, can we predict tooth decay before it happens? Yes - and here is what the evidence shows."
• End with a clear take-home message (e.g., "No single tool is perfect; the choice should be individualized based on patient age, setting, and clinical resources").

Slide 42 - References

• Remove the Lorem ipsum placeholder text from the second text box - this is a significant error that must be fixed before presentation.
• Use a consistent citation format (Vancouver or ADA style).
• Add 2-3 more recent references (2020-2025) to show current evidence.

General Presentation Improvements

Design and Formatting

1. Consistency: Use a uniform font (e.g., Calibri or Lato), font size (28-32pt for headings, 18-22pt for body), and color scheme throughout. Currently, text boxes appear inconsistent across slides.
2. Avoid walls of text: Every slide with a paragraph should be broken into 3-4 bullet points or a visual.
3. Add slide numbers to all slides.
4. Remove "Lorem ipsum" from Slide 42 immediately.
5. Complete all cut-off text - several shapes have content that is truncated.

Flow and Structure

1. Add transitions between sections - a brief "Section heading" slide (e.g., "SECTION 2: CAMBRA") helps the audience follow along.
2. The jump from individual tools directly to the research comparison section (Slide 38) is abrupt. Add a bridge slide: "Which Tool Performs Best? A Look at the Evidence."
3. Currently there are 43 slides for a seminar - this is on the longer side. Aim for 30-35 slides by merging the 10 Cariogram slides and removing the duplicate.

How to Present It

Opening (2-3 min)

• Start with a clinical scenario rather than a definition: "Imagine a 6-year-old child who comes in with no cavities. Should you give the same preventive advice as every other child, or can you identify - today - that this child will develop caries in the next 2 years?"
• This immediately makes the audience understand why CRA matters.

During the Presentation

• Engage the audience on each tool: "Has anyone used Cariogram in practice? What was your experience?"
• When presenting the Cariogram sections, do a live demo of the software (or show a pre-recorded short screen recording) - it is far more memorable than describing it.
• For ICDAS, show clinical photographs or a short image sequence of Score 0 through Score 6.
• When you get to the research comparison section (Slides 38-40), slow down - this is the most intellectually valuable part. Do not rush through it.

Critical Points to Emphasize

• The difference between risk factor vs risk indicator vs risk marker (often asked in viva).
• CAMBRA = ICDAS ≠ DMFT - DMFT only captures past disease, not future risk.
• The green sector of Cariogram = "chance to avoid caries" (not chance of getting it).
• The AAPD form is age-stratified - different forms for under 5 and over 6 years.

Closing (3-4 min)

• Return to your opening question and answer it directly using the evidence you presented.
• Mention that in Indian context, locally validated tools or adaptations are needed.
• Leave 5 minutes for questions - anticipate questions about AI in caries prediction and which tool is best for community dentistry settings.

Likely Viva/Discussion Questions to Prepare For

1. What is the difference between caries risk assessment and caries activity tests?
2. Which CRA tool would you use for a 3-year-old vs a 45-year-old?
3. What does the green sector in Cariogram represent exactly?
4. What are the limitations of CAMBRA in a developing country setting?
5. How does AI improve upon traditional CRA tools?
6. Has any CRA tool been validated for the Indian population?

Summary of Most Critical Fixes

What changed slide-by-slide:

Can AI Replace Caries Risk Assessment Tools?

What AI Can Do Well in Caries Risk

1. Radiographic Detection

• Sensitivity: 0.44–0.86
• Specificity: 0.85–0.98
• AUC: 0.84–0.98
• Accuracy: 0.73–0.98

2. Multi-Variable Pattern Recognition

3. Clinical Decision Support

Where AI Falls Short — The Hard Limits

1. External Validation Fails Badly

2. AI Operates in a "Black Box"

3. AI Cannot Assess What It Cannot Measure

• Assess salivary flow and buffering capacity
• Measure S. mutans or Lactobacilli counts
• Observe plaque quality and maturity
• Capture patient motivation, oral hygiene compliance, or dietary details in real time

4. Dataset Size and Diversity Problems

5. Ethical and Accountability Concerns

AI vs. Traditional CRA Tools — Head-to-Head

The Most Likely Future: AI + CRA Tools Working Together

1. AI reads the radiographs — automatically flags lesions on bitewings, classifies severity, and feeds this into the CRA tool as an input rather than making the clinician score it manually.
2. AI mines the EHR — extracts past caries history, medication lists, and dietary recall data and auto-populates CRA form fields.
3. CRA tools apply the clinical framework — the human clinician + validated tool integrates AI outputs with biological and behavioral factors to arrive at a final risk category.
4. AI monitors over time — tracks risk trend across visits and flags deteriorating patients for early intervention.

Bottom Line for Your Seminar

AI is a powerful tool within the CRA ecosystem, not a replacement for it. The evidence as of 2026 shows that AI performs excellently at specific narrow tasks (radiographic lesion detection, pattern mining in large datasets), but consistently fails the generalizability test when moved outside its training population. Traditional CRA tools — CAMBRA and Cariogram in particular — retain their superiority for holistic, individualized, explainable, clinician-guided risk assessment. The future is integration, not replacement.

• Reyes et al. (2022) Caries Research — PMID 35636386 — ML in caries diagnosis: promising but early stage
• Albano et al. (2024) BMC Oral Health — PMID 38402191 — AI for radiographic caries detection: good sensitivity/specificity
• Tirkkonen et al. (2026) BMC Oral Health — PMID 41546040 — External validation of ML models collapses across populations

AAPD Caries Risk Assessment Form — Scoring Criteria

How It Works (Important First)

"Use the circled responses to visualize the balance among risk factors, protective factors, and disease indicators. Use this balance or imbalance, together with clinical judgment, to assign a caries risk level." — AAPD Best Practice 2019

FORM 1 — Age 0–5 Years

Category 1: Biological / Social Factors

Note: For age 0–5, AAPD only distinguishes High vs. Low — there is no separate Moderate column in this age group form.

Category 2: Protective Factors

Category 3: Clinical Findings (Disease Indicators)

FORM 2 — Age ≥6 Years (Three-Column Form)

Category 1: Biological Factors

Category 2: Protective Factors

Category 3: Clinical Findings

Disease Indicators (Strongest Predictors — Separate from Clinical Findings)

Key rule: If ANY disease indicator is checked "Yes," the patient is almost certainly High Risk. Even if no disease indicators are present, a patient can still be High Risk if biological risk factors clearly outweigh protective factors.

Risk Determination Rules — Summary

A Practical Way to Think About It

RISK FACTORS                    PROTECTIVE FACTORS
(push towards High)             (push towards Low)
─────────────────               ──────────────────
Active maternal caries          Fluoridated water
Low SES                         Daily fluoride toothpaste
≥3 sugar exposures/day          Professional topical fluoride
No dental home                  Regular dental visits
Special needs                   Xylitol/MI Paste use
Visible plaque
S. mutans counts
White spot lesions
Recent restorations
─────────────────               ──────────────────
     DISEASE INDICATORS: If present → automatically HIGH

• Scale tips toward risk factors → High
• Scale is roughly balanced → Moderate
• Scale tips toward protective factors → Low