Differential perfusion patterns of perforator and random flaps assessed by indocyanine green imaging

Authors: AlQhtani A Z, Lee N, Kim H, Eom J S, Han H H

Affiliation: Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea

Journal: Journal of Reconstructive Microsurgery, Sept 17, 2025

PMID: 40962273

Key takeaways

• Optimal ICGA timing differs by flap type in a rat model.

• Perforator flaps: best necrosis prediction at ~50 seconds post-injection. Random flaps: ~150 seconds (p = 0.0028).

• Mean necrosis: 47.3% (perforator) vs 55.1% (random).

• Findings support flap-specific ICGA protocols; translation to humans requires validation.

Background

ICG angiography (ICGA) is widely used for flap perfusion assessment, but optimal interpretation timing is inconsistent across studies and may vary by flap type.

Objective

Identify flap-specific ICGA timing windows that best predict eventual tissue necrosis for perforator and random pattern flaps in a controlled preclinical model.

Methods

• Design/setting/LOE: Preclinical randomized rat study; single-lab experiment (Level of Evidence not applicable/animal study).

• Sample: Sixteen male Sprague–Dawley rats randomized: perforator (n=8) vs random flap (n=8). Deaths during experiment: 1 perforator, 2 random → analyzed ≈13 animals.

• Flap models: Perforator: 6×4 cm² lower abdominal flap based on superficial epigastric artery; Random: 3×10 cm² dorsal flap without dominant perforator.

• ICG protocol: 0.25 mg ICG via femoral vein; near-infrared images every 10 s (0–60 s) then every 30 s to 4 min. Hypoperfusion = fluorescence <30% of peak.

• Outcome measures: (1) ICGA hypoperfused area (time-resolved), (2) gross necrosis on postoperative day 7; compared alignment over time.

• Statistics: Mann–Whitney U where applicable; interval-censored survival with EMICM; generalized log-rank; proportional hazards regression; α<0.05.

  
  

Results

• Alignment window (ICGA vs final necrosis):

  • Perforator flaps: necrosis matched ICGA boundary between 10–50 s, most often 30–40 s.

  • Random flaps: alignment occurred 30–150 s, most commonly 120–150 s.

• Best single time points: 50 s (perforator) vs 150 s (random) for necrosis prediction; generalized log-rank p = 0.0028; proportional hazards p = 0.0032.

• Necrosis burden (mean ± SD): 47.3% ±14.3 (perforator) vs 55.1% ±16.1 (random).

• Attrition: 3 intra-study deaths (1 perforator, 2 random) excluded; remaining animals analyzed for timing correlation.

Conclusions

ICGA interpretation should be flap-specific: early time points (~50 s) best predict necrosis in perforator flaps, whereas later windows (~150 s) are optimal in random flaps; adopting tailored timing may reduce false positives and improve intraoperative decisions.

Strengths: Randomization to flap type; predefined imaging cadence; objective intensity threshold (<30% peak); time-to-event modeling aligns ICGA signal with hard endpoint (necrosis day 7).

Limitations: Animal model; non-identical flap sizes (6×4 vs 3×10 cm²) may affect perfusion dynamics; small N with attrition; gross necrosis without histopathology; device-/platform-specific considerations.

Clinical relevance

For intraoperative ICGA in human surgery, consider earlier reads for perforator flaps and extended observation for random (e.g., mastectomy) flaps. Differences in human physiology and ICGA platforms mean these exact seconds are not directly transferrable; the concept—tailor timing to flap vascular topology—is the actionable takeaway.

Questions/critiques

Potential biases/confounders: Different flap surface areas and tissue thickness may alter perfusion kinetics; lack of temperature/vasoactive control reporting; gross (not histologic) necrosis assessment; small cohorts reduce precision of “best time” estimates.

External validity: Human flaps vary (vasopressors, comorbidities, radiation, thickness) and clinical systems may acquire later time windows; translation needs prospective clinical validation by flap class.