Reverberant Wavefield Characterization and Directional Filtering for Penile Vibro-Elastography: A Pilot Feasibility Study in Post-Prostatectomy Erectile Dysfunction

Two common elastography methods gave physiologically reversed results in penile stiffness imaging; a filtering fix corrected them across all 6 patients

Journal: Physics in Medicine and Biology | Published: 2026-06-29 | Type: Pilot feasibility study | PMID: 42285148 Authors: Bui Ngoc Thang, Matt J Ziegelmann, Tobias S Kohler, Xiaoming Zhang (all Mayo Clinic, Rochester, MN) Funding/COI: Not disclosed

Summary

Penile vibro-elastography (PVE) measures corpus cavernosum stiffness by tracking how fast shear waves travel through tissue — stiffer tissue, faster waves. The problem: the corpus cavernosum is a cylinder, and cylindrical geometry creates standing waves (reverberant wavefields) that contaminate the signal and make standard estimation methods unreliable. This paper tests a filtering algorithm called LPVI-WBF (local phase velocity imaging with wavefield-based filtering) against two standard methods, and finds the standard methods routinely produce backwards results — reporting flaccid tissue as stiffer than erect tissue, which is physically impossible.

Claims

• In 59% of the 902 acquisitions across 6 patients, traveling wave energy was entirely absent — the signal was dominated by standing waves
• Mean standing wave fraction: 57 ± 17%, confirming reverberant contamination as the rule, not the exception
• LPVI-WBF correctly estimated erect shear wave speed as higher than flaccid (3.59 ± 0.48 m/s vs 3.25 ± 0.58 m/s, p<0.0001)
• Phase gradient (PG) and time-of-flight (TOF) methods produced the opposite result — erect < flaccid (p<0.0001), which contradicts known penile biomechanics
• LPVI-WBF dispersion exponent was near-zero (α=0.048, p=0.21), consistent with weakly viscoelastic tissue behavior
• PG (α=0.184) and TOF (α=0.321) showed significant frequency dependence, a signature of reverberant contamination rather than true tissue properties
• Simulations confirmed PG is accurate in clean wavefields but fails under reverberant conditions — which describes essentially all real penile acquisitions in this dataset

Study Quality

This is a pilot feasibility study, and it's honest about what that means. Six patients is not a sample from which any clinical conclusion can be drawn. What the study does establish is a technical claim: that the dominant wavefield physics in penile elastography are reverberant, not traveling, and that this renders two widely-used estimation methods systematically wrong in a predictable direction. That's a falsifiable, methodological claim, and the evidence for it — 902 acquisitions across four longitudinal visits with both phantom simulation and patient data — is internally consistent.

The simulation component is the backbone here. It tests PG and LPVI-WBF under controlled reverberant vs. clean conditions and reproduces the observed failure mode, giving the clinical findings a mechanistic explanation rather than just a correlation. The fact that all three frequencies (100, 150, 200 Hz) and both erectile states gave consistent results across all 6 patients strengthens the case that the reverberant contamination problem is structural, not patient-specific noise.

Red Flags

• N=6 is a proof-of-concept, not a validation study; no clinical interpretation of these stiffness values is warranted
• No phantom validation with ground-truth stiffness to confirm LPVI-WBF accuracy (the paper explicitly flags this as future work)
• Funding and conflict-of-interest not disclosed — notable given that this paper advocates for a specific algorithmic approach
• All patients had post-prostatectomy ED; whether these findings generalize to other ED etiologies or healthy penile tissue is unknown
• No histological or clinical correlation — it's unknown whether the stiffness differences LPVI-WBF detects map onto any outcome that matters (recovery, implant decision, etc.)
• LPVI-WBF's physiological consistency is demonstrated but its absolute accuracy is not established; producing the "right direction" result is necessary but not sufficient

Strengths

• The standing wave problem is clearly quantified, not just asserted (59% pure standing wave, 57% mean fraction)
• Simulations and in vivo findings are internally consistent and mutually reinforcing
• Longitudinal design (4 visits per patient) reduces the chance that results are session-specific artifacts
• The paper calls out exactly what validation is still needed rather than overclaiming

Verdict

This is a narrow, technically rigorous paper solving a specific signal-processing problem in a niche imaging modality. The finding that PG and TOF give physically backwards results in real penile elastography acquisitions is genuinely important if you are building or using PVE equipment — it means measurements from those methods in this anatomical context are likely meaningless or worse. LPVI-WBF looks promising as a fix, but "physiologically consistent direction" is a low bar; the absolute accuracy validation and the clinical-utility case still need to be made. Worth reading if you work in elastography or are evaluating PVE as a diagnostic tool. Worth ignoring if you're looking for clinical evidence about post-prostatectomy ED management.