Quantitative preoperative peroneal vessel assessment in fibula free flap surgery

Authors: Rothchild E, Saini N, Smith IT, Yom J, Ricci JA.

Affiliation: Albert Einstein College of Medicine; Donald & Barbara Zucker School of Medicine at Hofstra/Northwell (Radiology; Plastic Surgery).

Journal: Journal of Reconstructive Microsurgery , October 2025

DOI: 10.1055/a-2717-4909.

Key takeaways

• In 117 FFF patients, higher peroneal Bollinger scores independently predicted any 90-day complication (adjusted OR 2.23).

• Each doubling of Bollinger score increased complication risk by 123% and LOS by 27%.

• LLACS (lower-limb arterial calcification score) alone did not predict overall complications; signal only for fistula on univariate analysis.

• Combining Bollinger + LLACS identified a high-risk group with 8.36× higher odds of complications.

Background

Atherosclerosis and arteriosclerosis in donor vessels may compromise free-flap outcomes; yet, current CTA assessments are largely qualitative.

Objective

Test whether quantitative CTA-derived metrics, the peroneal Bollinger score (luminal stenosis) and LLACS (arterial wall calcification), stratify perioperative risk in fibula free flap (FFF) reconstruction.

Methods

• Design/setting/LOE: Retrospective cohort, university-affiliated tertiary center; Aug 2021–Mar 2023; outcomes within 90 days. Level III.

• n: 117 consecutive FFF patients (mean age 56.8; mean BMI 27.0).

• Imaging protocol: Routine lower-extremity CTA (single-arterial or triple-phase); 3-mm reformats; scores calculated on PACS.

• Scoring:

  • Bollinger: segmental luminal stenosis on contrast phase, peroneal segment only.

    • Scored 0–15 quantifying luminal stenosis severity/extent; higher = worse donor vessel quality.

  • LLACS: Agatston calcification on non-contrast crural segment.

• Endpoints: Any complication (infection, hematoma, seroma, fistula, dehiscence, flap failure); length of stay (LOS).

• Stats: Scores log2(score+1) transformed; logistic regression for complications; Poisson for LOS; ROC cutoffs; combined risk strata (low/moderate/high). Adjusted for BMI, gender, smoking, ASA, CCI.

Results

• Complications: 42/117 (35.9%); infection 23.9%; dehiscence 8.5%; fistula 4.3%; hematoma 3.4%; flap failure 3.4%. Median LOS 11 days.

• Primary outcome: Increasing Bollinger score → any complication (unadj OR 2.10; adj OR 2.23, p=0.044).

• Flap failure (signal): Unadjusted OR 3.38 (p=0.035); multivariate not estimable due to low events.

• LOS: Higher Bollinger associated with longer LOS (β=1.27, p<0.001).

• ROC cutoffs: Bollinger 0.5 (OR 3.95 for any complication); LLACS 8.4 (NS).

• LLACS alone: No independent association with overall complications; trend for fistula (unadj OR 1.25; adjusted borderline).

• Combined risk matrix: High-risk (both above cutoffs) → OR 8.36 vs low-risk for any complication.

Conclusion

Quantitative CTA scoring—especially the peroneal Bollinger score—adds actionable risk stratification for FFF; pairing it with LLACS further isolates a very high-risk cohort.

Strengths & limitations

• Strengths: First quantitative, vessel-specific assessment in FFF; standardized CTA protocol; multivariable modeling; creation of a pragmatic risk matrix.

• Limitations: Retrospective, single-center; 90-day horizon only; exclusion of severe peroneal occlusions (selection bias); modest event counts limited failure modeling.  

Critiques and questions

• Generalizability & thresholds: Cutoffs (Bollinger ≥0.5; LLACS ≥8.4) have modest AUCs and wide CIs; they should be prospectively validated across centers with standardized CTA phases/recons. The highest complication-specific OR is risk for hematoma, which has questionable relationship to atherosclerotic occlusion.

• Causality vs correlation: Bollinger is vessel-specific (plausible pathophysiology), but residual confounding (e.g., smoking intensity, diabetes severity) may persist despite adjustment. Event counts limit failure modeling.

• Implementation: Reporting Bollinger routinely requires minimal extra time if radiology templates include a peroneal segment score, but may be difficult for surgeons alone to implement.