Methodological Determinants of Grayscale Pixel Intensity in Ram Testicular Ultrasonography: Effects of
ROI
Placement and Imaging Plane Across Age Groups and Two Reproductive Periods
Bülent Bülbül, Şükrü Doğan, Mustafa Yiğit Nizam, Necdet Akay ABSTRACT
Testicular ultrasonography is increasingly used in ram breeding soundness evaluations, yet methodological standardization of quantitative grayscale image analysis remains limited. The primary objective of this study was to determine the effects of region‐of‐interest (ROI) placement and imaging plane on testicular parenchymal numerical pixel values (NPVs) under a fixed ultrasonographic acquisition protocol, in Güney Karaman rams. A secondary objective was to compare NPV among age groups and between two predefined reproductive‐period sampling points, August and March. Twenty‐nine healthy, fertility‐proven rams were assigned to three age groups: 1.5 ( n = 8), 3.1 ( n = 12), and 4.6 years ( n = 9). B‐mode ultrasonography was performed with a 7.5‐MHz linear transducer during the breeding (August) and non‐breeding season (March). From each testis, 10 standardized regions of interest were analysed (six longitudinal and four transversal, with five above and five below the mediastinum testis). ImageJ analysis yielded 1160 spot measurements. A linear mixed model with random intercepts for animal and testis ID accounted for the hierarchical repeated‐measures structure. NPVs increased with age, were higher in August than in March (71.65 ± 1.75 vs. 60.89 ± 1.75, p < 0.001), and were markedly higher below than above the mediastinum testis (89.50 ± 1.75 vs. 43.03 ± 1.75, p < 0.001). Transversal images yielded higher NPVs than longitudinal images (68.61 ± 1.77 vs. 63.93 ± 1.74, p < 0.001). Random‐effects analysis showed that 31%, 7%, and 62% of total variance were attributable to between‐animal, between‐testis, and spot‐level variation, respectively. These findings demonstrate that ram testicular echogenicity is influenced by biological and methodological factors, highlighting the need for standardized multi‐site sampling protocols.