Skip to content

Latest commit

 

History

History
144 lines (104 loc) · 5.78 KB

treatments.md

File metadata and controls

144 lines (104 loc) · 5.78 KB

Treatment Interventions for Diabetic Wounds

Computational treatment study modeling 8 therapeutic interventions for diabetic foot ulcers. Each treatment is a TOML overlay applied on top of the diabetic wound profile, modifying specific biological parameters based on published mechanisms of action.

Available treatments

Treatment Targets Key references
anti_inflammatory M1/M2 transition, TNF-alpha, baseline inflammation Mirza & Koh 2011, Louiselle et al. 2021
hbo Tissue oxygenation, VEGF, angiogenesis, collagen Catrina et al. 2004, Thangarajah et al. 2009
npwt Blood flow, granulation, moisture, wound contraction Morykwas et al. 1997, Armstrong & Lavery 2005
doxycycline MMP inhibition, collagen/fibronectin preservation Lobmann et al. 2002, Smith et al. 1999
growth_factor Fibroblast activation, PDGF-BB mitogenesis Steed 2006, Smiell 1998
msc Multi-target paracrine (immune, vascular, matrix) Cao et al. 2017, Li et al. 2024
moisture TEWL reduction, hydration, exudate management Winter 1962, Junker et al. 2013
combination All four dysfunction axes simultaneously Multi-target rational combination

Running treatments

# Single treatment
./run.sh --study=diabetic-wound --skin=diabetic --treatment=hbo

# Full comparison study (all 8 treatments + baseline)
python3 scripts/study/treatment_study.py

# Specific treatments only
python3 scripts/study/treatment_study.py --treatments=hbo,msc,combination

# List available treatments
./run.sh --list-treatments

How treatments work

Treatments are TOML overlays in treatments/ applied after the diabetic profile and study config:

bdm.core.toml + modules/*/config.toml -> bdm.toml (merge)
  + profiles/diabetic.toml          (biology)
  + studies/diabetic-wound/preset.toml  (scenario)
  + treatments/hbo.toml             (intervention)

Each treatment modifies the diabetic dysfunction parameters back toward healthy values, proportional to the treatment's clinical effect size from published literature.

Treatment mechanisms

Anti-inflammatory (anti-TNF-alpha)

The core diabetic wound pathology is failed M1-to-M2 macrophage transition. TNF-alpha neutralization directly addresses this by:

  • Releasing the M1-to-M2 brake (m1_duration_factor: 3.0 -> 1.5)
  • Restoring M2 anti-inflammatory resolution (resolution_factor: 0.3 -> 0.7)
  • Reducing AGE/RAGE-driven baseline inflammation (0.001 -> 0.0002)
  • Restoring keratinocyte proliferation and migration

Hyperbaric oxygen (HBO)

HBO corrects the tissue hypoxia that drives much of diabetic wound dysfunction:

  • Restores HIF-1alpha/VEGF cycling (vegf_factor: 0.4 -> 1.0)
  • Improves microcirculation (perfusion basal: 0.7 -> 0.85)
  • Accelerates angiogenesis (angio_rate: 0.002 -> 0.008)
  • Partially rescues collagen synthesis via prolyl hydroxylase O2 supply

Negative pressure wound therapy (NPWT)

Multi-modal mechanical intervention:

  • 4x blood flow increase from sub-atmospheric pressure
  • Microdeformation stimulates cell proliferation
  • Macrodeformation assists wound edge advancement (inward_bias: 0.3 -> 0.5)
  • Sealed dressing controls moisture and clears inflammatory exudate

Doxycycline (sub-antimicrobial MMP inhibitor)

Targets the MMP/TIMP imbalance that destroys ECM in diabetic wounds:

  • Halves excess MMP activity (mmp_factor: 3.0 -> 1.5)
  • Restores TIMP balance (timp_factor: 0.5 -> 0.8)
  • Preserves collagen and fibronectin scaffolds
  • Mild anti-inflammatory effect via TACE inhibition

Growth factor (PDGF-BB / becaplermin)

FDA-approved recombinant growth factor for DFU:

  • Potent fibroblast mitogen (density_factor: 1.0 -> 1.8)
  • Rapid fibroblast activation (activation_factor: 2.0 -> 1.0)
  • Does not directly increase collagen synthesis (correct per literature)

Mesenchymal stem cell (MSC) therapy

Broadest-acting intervention via paracrine secretome:

  • M1-to-M2 macrophage reprogramming (IL-6, PGE2)
  • VEGF production (dominant paracrine factor, up to 100x from spheroids)
  • Fibroblast activation (PDGF, TGF-beta)
  • Keratinocyte proliferation/migration (EGF, bFGF, HGF)

Moisture dressings

Maintains optimal wound hydration:

  • 83% TEWL reduction (surface_loss_rate: 0.03 -> 0.005)
  • Hydrogel moisture donation (recovery_rate: 0.02 -> 0.06)
  • Exudate absorption clears inflammatory mediators

Combination therapy

Rational multi-target approach addressing all four diabetic dysfunction axes:

  1. Immune: anti-TNF-alpha (M1/M2 transition)
  2. Vascular: HBO (O2/VEGF/angiogenesis)
  3. ECM: doxycycline (MMP/collagen preservation)
  4. Moisture: advanced dressing (hydration)

Creating custom treatments

Create a new TOML file in treatments/:

# Treatment: my_therapy
# Literature: Author et al. Year
# Mechanism: what it does biologically

[skin.diabetic]
m1_duration_factor = 2.0    # override diabetic dysfunction
prolif_factor = 0.7          # partially restore proliferation

[skin.perfusion]
angio_rate = 0.005           # override vascular params

The treatment file only needs to contain the parameters it modifies. All other parameters remain at their diabetic profile values.

Output

The treatment study script produces:

  • output/treatment_study/metrics_<treatment>.csv contains full metrics for each run
  • output/treatment_study/treatment_comparison.csv contains the summary comparison table
  • Console output with closure rates, inflammation peaks, healing times

Biological validity

Treatment parameter mappings are derived from published quantitative data:

  • Clinical effect sizes (fold-changes, percentage improvements)
  • In vitro mechanistic studies (MMP inhibition %, fibroblast proliferation fold-change)
  • Animal model data (wound closure rates, histological measures)

See individual treatment TOML files for DOI-linked references.