Treatment Protocol Optimization Dashboard

Executive Summary

This dashboard presents comprehensive analysis of cancer treatment protocols using the DMY shortest-path algorithm, including both single-objective cost optimization and multi-objective Pareto front analysis balancing cost, time, quality of life, and success rates.

Key Findings:

1. Single-objective: Optimal curative pathway costs 10.8k with 7 treatment steps (Initial Screening → Remission)
2. Multi-objective: Six Pareto-optimal protocols balance cost ($17.5k–$56k), duration (54–158 wk), QoL (5–52), and success (305–456 scaled units)
3. Performance: On the current 20-node protocol, DMY runs ~2.1× faster than Dijkstra (0.01 ms vs 0.03 ms)

Reproducibility: run treatment_protocol.jl or generate_figures.jl with --seed=<int> (or set OPTIM_SP_SEED) to replay the exact same simulated patient pathways. Default seed is 42.

Part 1: Single-Objective Analysis

Figure 1: Treatment Cost vs Efficacy

Interpretation:

• Major surgery: Highest cost (35k) with 90% efficacy
• Targeted therapy: Most expensive (45k) with 80% efficacy
• Screening/Biopsy: Low cost, high diagnostic value
• Trade-off clearly visible between cost and treatment success

Figure 2: Treatment Pathway Network

Key Insights:

• Left-to-right layout shows diagnostic intake → planning → consults → active therapy → follow-up
• Color coding separates Diagnostics, Planning, Specialist Consults, Active Treatment, and Follow-up & Support cohorts
• Arrow directions capture every valid transition from the simulation (e.g., recurrence loops back to second-line or palliative care)
• ⭐ Remission node is highlighted so reviewers can instantly identify terminal success states

Figure 3: Risk-Benefit Analysis

Clinical Significance:

Part 2: Multi-Objective Pareto Front Analysis

The Challenge

Real-world treatment decisions involve optimizing multiple competing objectives:

• Cost: Financial burden on patient/system
• Time: Treatment duration and recovery
• Quality of Life: Side effects and patient comfort
• Success Rate: Probability of remission

Figure 4: 2D Pareto Front Projections

Four critical trade-offs visualized:

1. Cost vs Success: More expensive treatments have higher success rates
2. Time vs QoL: Longer treatments impact quality of life
3. Cost vs QoL: Expensive treatments may preserve QoL better
4. Speed vs Success: Faster treatments may be less effective

Figure 5: 3D Pareto Front Visualization

3D Trade-off Space:

• X-axis (Cost): Treatment cost in thousands (~$17–$56k across the Pareto set)
• Y-axis (Success): Scaled remission score (305–456 composite units; higher is better)
• Z-axis (QoL): Quality-of-life index (5–52; higher preserves comfort)

Legend highlights:

• Green Diamond — “Budget-constrained”: Pareto solution 2 (17.5k, 158 wk, QoL 52, success ≈ 305) satisfies the cost ≤ 50k constraint.
• Red Hexagon — “Knee Point”: Pareto solution 6 (56.0k, 60 wk, QoL 5, success ≈ 456) balances the frontier’s steepest trade-off.
• Weighted-sum scoring is intentionally omitted because the objectives mix min/max senses; convert maximise metrics to costs before re-enabling it.

Pareto-Optimal Treatment Protocols

*Success values are scaled composite scores from the example model and can exceed 100.

Figure 6: Treatment Strategy Comparison

Strategy Analysis:

• Budget control (Solution 2): 17.5k, success ≈305, QoL 52 – lowest spend, longest duration
• Mid-range blend (Solution 1): 22.5k, success ≈365, QoL 42 – balanced compromise
• High-success knee (Solution 6): 56.0k, success ≈456, QoL 5 – maximum remission at the expense of comfort
• Weighted-sum ranking requires converting efficacy to a cost; disabled by default to avoid misleading scores

Part 3: Patient-Specific Protocol Selection

Figure 7: Patient Profile Analysis

Personalized Recommendations:

*Success values are scaled composite scores from the example model.

Figure 8: Clinical Decision Tree

Decision Support Framework:

1. Initial Assessment: Risk stratification
2. High Risk: Aggressive multimodal therapy
3. Low Risk: Conservative or single modality
4. Resource Constraints: Stepwise escalation

Part 4: Algorithm Performance

Figure 9: Performance Analysis

k = ⌈n^(1/3)⌉ parameter as specified by DMY algorithm

Key Insights:

• Legend separates DMY (k = n^{1/3}) and Dijkstra; vertical whiskers denote ±95% confidence intervals
• Even modest protocol graphs benefit from the DMY implementation
• Larger hospital libraries inherit the same O(m log^{2/3} n) advantage
• Lightweight runtimes enable real-time clinical decision support

Key Takeaways

Single vs Multi-Objective

• Single: One "optimal" path minimizing cost
• Multi: Six Pareto-optimal protocols provide distinct trade-offs
• Reality: Patients have different priorities and constraints

Treatment Personalization

• Young patients: Can tolerate aggressive protocols
• Elderly patients: Prioritize quality of life
• Resource-limited: Sequential escalation strategies
• Biomarker-positive: Precision medicine options

Algorithm Performance

• Small networks (n<1000): Use Dijkstra
• Large networks (n>1000): DMY increasingly superior
• Hospital-scale: DMY enables real-time decisions

Clinical Insights

• No universal protocol: Context determines optimal choice
• Trade-offs explicit: Pareto front visualizes all options
• Shared decisions: Patients can see and choose

Reproducibility

Generate all figures:

julia --project=. examples/treatment_protocol/generate_figures.jl

Run complete analysis:

julia --project=. examples/treatment_protocol/treatment_protocol.jl

Model Parameters:

• 20 treatment modalities
• 34 valid transitions
• 4 objectives: cost, time, QoL, success
• Real-world cost and efficacy data

References

1. Duan, R., Mao, J., & Yin, Q. (2025). "Breaking the Sorting Barrier for Directed SSSP". STOC 2025.
2. Multi-objective optimization: Ehrgott, M. (2005). "Multicriteria Optimization". Springer.
3. NCCN Clinical Practice Guidelines in Oncology (2024).
4. CMS Physician Fee Schedule Database.
5. SEER Cancer Statistics Review.

Dashboard generated using DMYShortestPath.jl - Revolutionizing clinical decision support with advanced graph algorithms and multi-objective optimization