Accuracy of orchidometry in boys with varicocele

Abstract

Orchidometric evaluation of the testis has been proposed as a cost-effective alternative to measurement of the testis with high-frequency linear ultrasound, which may be costly in terms of hospital resources and patient time. It is known from animal experiments, autopsy series, and small clinical studies that, under ideal conditions, orchidometry may approximate ultrasound measurement. However, little is known of the effectiveness of orchidometry in the clinical setting in a large sample of adolescents with varicocele. We sought to analyze the performance characteristics of Rochester orchidometry and its agreement with ultrasound testis volumes in boys with varicocele. Our institutional varicocele database was analyzed from March 2000 to May 2013, including all boys with Rochester orchidometric measurement and ultrasound-based volume measurement performed on the same day. The Lambert formula (L*W*H*0.71) was used to calculate ultrasound volumes. Seven-hundred and twenty measurements were included: 360 of the left testis, and 360 of the right testis. Each subject was included once; in the event of serial measurements the earliest measurement was analyzed. Bland-Altman plots with 95% limits of agreement were used to compare orchidometry and ultrasound measurements. Analysis was performed with JMP, v11 Pro. Age at exam ranged from 11.2 to 18.5 years (median 15.8). With respect to varicocele grade, 183 (50.8%) were grade III, 113 (31.4%) were grade II, 42 (11.7%) were grade I, 12 (0.3%) were bilateral, and 10 (0.3%) were ungraded. Mean ultrasound left testis volume was 13.6 cc (SD 6.6) and mean right testis volume was 15.1 cc (SD 6.9). Eleven surgeons performed the orchidometric measurements; one surgeon performed 71% of the exams. Mean overestimation on the right was 2.0 cc (SD 4.2) and on the left was 1.9 cc (SD 4.1); each was highly statistically significant (p < 0.0001, paired t test). Error was correlated with testis size, implying a greater degree of overestimation with increasing volume (p < 0.01, Pearson's correlation 0.09). Amount of volume overestimation and variability was not significantly different for right and left testis. Sensitivity and specificity of Rochester orchidometry to detect a testis volume differential (TVD) of 20% were 33% (95% CI 23-42%) and 96% (95% CI 92-97%), respectively. Testis size, varicocele grade, or examining surgeon had no effect on sensitivity or specificity. We have shown in a large series of boys with adolescent varicocele that in clinical practice there is a modest degree of overestimation of testis volume on average (1.9-2.0 cc), although there is a large range of volume estimation, such that the 95% confidence interval ranges are quite wide, from approximately 6 cc lower than the true volume to 10 cc greater than the true volume. Furthermore, the low sensitivity (33%) of orchidometry for 20% testis volume differential renders this a suboptimal screening tool for this clinical parameter, which has been shown to be associated with semen analysis outcomes. Knowledge of the performance characteristics of orchidometry is similarly important for research, as factors such as the prevalence of testis volume differential are then dependent on the modality of measurement. Lastly, that this was conducted over a long time course with several surgeons involved suggests that these data reflect real-world application of orchidometry. Appropriate caution should be exercised when relying solely on orchidometric evaluation of the testis. Rochester orchidometry in general appears to overestimate testis size, and there is wide variability in the estimation. In clinical practice, the sensitivity of Rochester orchidometry is modest in detecting a 20% testis volume differential; this difference would be missed in approximately two out of three of boys screened with orchidometry alone.