Fractal and Spectral Dimensions as Determinants of Thermal Ablation Outcomes in Cancer Tissues

This study demonstrates that incorporating tissue fractal geometry and spectral dimensions into a fractal-fractional bio-heat model successfully explains the clinical variability in thermal ablation outcomes, revealing that topological connectivity is a key determinant of coagulation zone expansion and ablative efficacy in malignant neoplasms.

The Gist

The Big Picture: Why Cancer Treatments Sometimes Miss the Mark

Imagine you are trying to cook a steak. You have a perfect recipe (the standard medical model) that tells you exactly how long to leave the pan on and how much heat to apply. But sometimes, the steak comes out burnt on the outside and raw in the middle, or sometimes it's perfectly cooked one day and undercooked the next, even though you used the exact same settings.

This is the problem doctors face with Thermal Ablation, a cancer treatment that "cooks" tumors to kill them. Despite using precise machines, the size of the dead (necrotic) zone varies wildly. Sometimes the tumor isn't fully destroyed, leading to it growing back.

This paper argues that the reason for this unpredictability isn't just bad equipment or human error. It's because cancer tissue isn't a uniform block of meat; it's a complex, tangled, fractal maze.

---

The Core Concept: The "Fractal" Tissue

To understand the paper, you need to understand two new ways of looking at tissue:

1. Fractal Dimension (The "Roughness"): Think of a coastline. If you measure it with a long ruler, it looks short. If you measure it with a tiny ruler, following every little cove and rock, it becomes much longer. Biological tissues are like this—they are "rough" and self-similar at different scales. The paper says that as cancer gets worse, this "roughness" (Fractal Dimension) increases.
2. Spectral Dimension (The "Connectivity"): This is the paper's big discovery. Imagine a city.
   • High Connectivity: A city with a grid of streets where you can get from A to B in many ways. Heat (or traffic) flows easily.
   • Low Connectivity: A city with dead ends, blocked roads, and isolated neighborhoods. Heat gets stuck in one area and can't spread.
   • The authors call this the Spectral Dimension. It measures how well-connected the tissue is, not just how rough it looks.

The Experiment: Cooking in a Maze

The researchers built a super-computer simulation to test what happens when you try to "cook" a tumor.

• The Old Way (Fourier Diffusion): Imagine heat spreading through a block of butter. It spreads evenly and predictably. This is what current medical models assume.
• The New Way (Fractal-Fractional Model): Imagine trying to spread heat through a spaghetti bowl or a sponge. The heat gets trapped in the nooks and crannies, moves slowly in some directions, and jumps quickly in others. It also has "memory"—the tissue remembers it was hot and resists getting hotter again (a phenomenon called thermotolerance).

They added a "smart controller" (like a thermostat) to the simulation. In real life, doctors use these to keep the temperature at the probe tip from getting too hot (to avoid burning the skin) while trying to cook the tumor.

The Key Findings

1. The "Trap" Effect

When the tissue has low connectivity (low Spectral Dimension), the heat gets trapped.

• Analogy: Imagine pouring water into a sponge with very tight, disconnected holes. The water stays right where you poured it and doesn't spread out to wet the whole sponge.
• Result: The "cooked" zone (the dead tumor area) ends up being smaller than the doctors expect, even if they use the same amount of energy.

2. The Mystery of Liver Metastases

Doctors have noticed a strange pattern: When they treat primary liver cancer (cancer that started in the liver), they get a good result. But when they treat liver metastases (cancer that spread to the liver from the colon or elsewhere), the treatment often fails to kill as much of the tumor.

Why?
The paper explains this using the "Sponge" analogy:

• Primary Liver Cancer: The tissue is like a loose, open sponge. Heat spreads well.
• Metastases: These tumors often grow with a thick, hard, fibrous shell (like a dense, compacted brick wall). This makes the internal "roads" very blocked.
• The Math: The model shows that metastases have a much lower Spectral Dimension (lower connectivity). The heat hits a dead end and can't reach the edges of the tumor. This explains why the "cooked" zone is smaller for metastases.

3. The "Uncertainty" Factor

The paper found that the biggest source of error in predicting treatment size comes from not knowing the Spectral Dimension.

• If you don't know how "connected" the tissue is, you can't predict how big the dead zone will be.
• Interestingly, the uncertainty is highest in healthy tissue (which is more variable) and lower in advanced tumors (which are more consistently "rough"). This explains why it's hard to predict how much healthy tissue will be affected, but slightly easier to predict the tumor's reaction once it's fully formed.

The Takeaway: A New Way to Cook

The authors are proposing a shift in how we think about cancer treatment:

• Old View: "Apply X amount of heat for Y minutes."
• New View: "We need to map the connectivity of the tumor first."

Just as a chef needs to know if they are cooking a dense steak or a fluffy soufflé, doctors need to know if the tumor is a "loose sponge" or a "compact brick." If a tumor has low connectivity (like a metastasis), the doctor might need to change the strategy—perhaps applying heat for longer, using a different method, or accepting that the "cooked" zone will be smaller than the machine's standard settings suggest.

Summary in One Sentence

This paper proves that the "shape" and "connectivity" of a tumor's internal structure (its fractal nature) act like a hidden traffic jam for heat, explaining why some cancers are harder to cook than others and suggesting that future treatments must account for this invisible maze to be successful.

Technical Summary

1. Problem Statement

Clinical thermal ablation (e.g., radiofrequency or microwave ablation) for cancer treatment suffers from significant unpredictability in lesion size and shape. Classical bio-heat models (based on Fourier diffusion and the Pennes equation) fail to fully explain this variability, with applied energy accounting for as little as 25% of the variance in effective ablation volume.
Key unresolved issues include:

• Anomalous Transport: Biological tissues exhibit structural heterogeneity (fractality) and memory effects (thermotolerance), causing heat transfer to deviate from standard diffusion.
• Clinical Discrepancies: There is a documented reduction in ablative efficacy for liver metastases (LM) compared to primary hepatocellular carcinomas (HCC), which current models cannot physically explain.
• Missing Parameters: While the fractal dimension ($D_f$) is a known biomarker for tumor grade, the spectral dimension ($d_s$), which governs topological connectivity and diffusion pathways, has not been integrated into thermal ablation modeling.

2. Methodology

The authors developed a realistic fractal-fractional bio-heat model to simulate thermal ablation under state-of-the-art clinical conditions.

A. Mathematical Framework

The core of the study is a modified Pennes bio-heat equation incorporating:

1. Temporal Fractional Derivative (Memory Effect): A Caputo fractional derivative of order $\alpha$ ($0 < \alpha < 2$) replaces the standard time derivative to model thermotolerance and non-Fourier heat propagation.
   • $\alpha < 1$: Sub-diffusive (damped, slow response).
   • $\alpha = 1$: Classical diffusion.
   • $\alpha > 1$: Super-diffusive (oscillatory, rapid transport).
2. Spatial Fractal Geometry: The Laplacian operator is modified to account for the fractal dimension ($D_f$) and spectral dimension ($d_s$). This introduces a scale-dependent effective thermal conductivity $\lambda_{eff}(r) = \lambda_0 (r/\ell)^{-\theta}$, where $\theta = (2D_f/d_s) - 2$.
3. Non-linear Blood Perfusion: A dynamic model based on the degree of stasis (DS) and Arrhenius kinetics accounts for vasodilation (increased perfusion at 41–45°C) followed by vascular collapse at higher temperatures.
4. Cell Viability: Thermal damage is calculated using the Arrhenius kinetic model, defining the coagulation zone ($\Omega \geq 4.6$, ~99% cell death) and the periablation zone ($0.6 \leq \Omega < 2.1$).

B. Simulation Setup

• Control Strategy: The model employs a Proportional-Integral (PI) controller to modulate the heat source ($Q_s$) in real-time, maintaining a target temperature ($T_{lim} = 90^\circ$C) at the applicator tip. This mimics closed-loop clinical systems.
• Numerical Implementation: Solved using a fractional Predictor-Corrector Finite Difference Method (FDM) in Python.
• Parameters: Simulations varied the fractional order ($\alpha$), fractal dimension ($D_f \in [1.5, 1.9]$), and spectral dimension ($d_s$). Sensitivity analysis was performed to quantify the impact of these parameters on the coagulation radius ($r_c$).

3. Key Contributions

1. Integration of Spectral Dimension: The study is the first to explicitly link the spectral dimension ($d_s$)—a measure of topological connectivity—to thermal ablation outcomes, identifying it as a critical driver of clinical variability.
2. Closed-Loop Fractal Modeling: Unlike previous open-loop models, this work integrates fractal-fractional physics with PI-controlled power delivery, revealing how feedback control interacts with tissue topology.
3. Physical Explanation of Clinical Anomalies: The model provides a rigorous physical mechanism explaining why liver metastases yield smaller ablation zones than primary HCCs, attributing it to differences in topological connectivity ($d_s$) rather than just geometry.

4. Key Results

A. Impact of Anomalous Transport and Control

• PI Control Efficacy: The PI controller successfully maintained the target temperature across all regimes ($\alpha$), but the spatial expansion of the coagulation zone became highly dependent on tissue topology once the controller engaged.
• Sub-diffusive vs. Super-diffusive:
  • In sub-diffusive regimes ($\alpha < 1$), heat trapping occurs, leading to sustained thermal damage even after cooling.
  • In super-diffusive regimes ($\alpha > 1$), heat dissipates rapidly, causing sharp drops in temperature once the source is off.
• Geometric vs. Topological Effects:
  • Fractal Dimension ($D_f$): Lower $D_f$ (less complex structure) leads to larger thermal expansion.
  • Spectral Dimension ($d_s$): Lower $d_s$ (reduced connectivity) significantly hinders heat diffusion, shrinking the ablation radius.

B. Sensitivity and Uncertainty Analysis

• Topological Uncertainty: The uncertainty in the final ablation size is driven primarily by the indeterminacy of $d_s$.
• Tumor Evolution: Uncertainty is highest in healthy tissue (low $D_f$) and decreases as tumors evolve (higher $D_f$). This aligns with clinical data showing primary HCCs (higher $D_f$) have more predictable ablation outcomes than metastases (lower $D_f$).
• Sensitivity Index: The coagulation radius is highly non-linearly sensitive to both $D_f$ and $d_s$ (Sensitivity Index $S > 1$).

C. Explaining the HCC vs. Metastasis Discrepancy

The model successfully reproduced the clinically observed ~15% reduction in ablation radius for liver metastases compared to HCC.

• Mechanism: To reproduce this, the spectral dimension of metastases ($d_s^{LM}$) must be strictly lower than that of HCC ($d_s^{HCC}$).
• Biological Correlation: This lower $d_s$ corresponds to the desmoplastic growth pattern of metastases, characterized by dense fibrotic rims and rigid extracellular matrices that disrupt topological continuity and act as barriers to heat propagation.

5. Significance and Future Outlook

• Paradigm Shift: The paper argues that thermal dosimetry must move beyond simple geometric descriptors to include topological parameters ($d_s$).
• Clinical Implications:
  • Predictive Planning: Understanding $d_s$ could allow clinicians to predict ablation outcomes more accurately, particularly for metastatic tumors.
  • Treatment Optimization: Strategies could be tailored to the specific topological connectivity of the tumor (e.g., adjusting power duration for low-connectivity metastases).
• Future Work: The authors propose using inverse modeling to experimentally determine $d_s$ in biological tissues, potentially via imaging or inverse thermal analysis, to enable "fractal-aware" treatment planning.

In conclusion, this study demonstrates that the spectral dimension is a fundamental determinant of thermal ablation efficacy. By accounting for the fractal and topological nature of cancer tissues, the proposed model resolves long-standing clinical variabilities and offers a pathway toward more precise, personalized cancer therapies.