Interinstitutional variability in testicular volumes and varicocele presence by ultrasound: surprising discrepancies and implications for clinical decision making.

OBJECTIVE

To determine the variability of testicular ultrasound (US) volumes and the clinical implications.

METHODS

A retrospective review identified 73 patients (62 adult and 11 adolescents) who had testicular US scans performed at both our institution (Urological Institute at Northeastern New York) and at an outside facility (outside hospital study). Testicular volume (TV) was measured using the Lambert formula: length × width × height × 0.71. TV difference (TVD) was calculated as percentage difference = 100 × (right TV - left TV)/left TV and serves as the intrapatient TVD. Patients with clinically palpable varicoceles were correlated to detection on US. Bland-Altman plots were generated to compare the limits of agreement for interinstitutional TVs. Interinstitutional agreement was measured using the Kappa (κ) coefficient.

RESULTS

Although the mean values for TVs were similar between institutions (19.1 vs 19.0 cm(3)), there was a wide range in the limits of agreement with a large standard deviation of difference (8.52 cm(3)). Interinstitutional agreement was poor regarding patients with a ≥ 20% TVD (κ = 0.09) and was even worse in patients with both varicoceles and ≥ 20% TVD (κ = 0.05). US agreement for verification of clinically detected varicoceles was good (κ = 0.67). If the variability found in this study was applied to identically sized testicles, an erroneous 20% TVD would be found in 38% of studies.

CONCLUSION

Interinstitutional or inter-ultrasonographer assessment of TV varies significantly. Clinicians should cautiously interpret the US-determined TV when making surgical decisions. Images, and not just reports, should be reviewed by clinicians.