Abstract / Summary
Abstract AIM: To determine whether serum leptin concentration and body mass index (BMI) differ between patients with ocular or ocular adnexal tumours and tumour-free controls, and to quantify the precision of any null finding. METHODS: A case–control study was conducted at the reconstruction and oncology clinic of a tertiary university hospital in Medan, Indonesia, from December 2025. Thirty consecutive patients with an ocular or adnexal tumour and 25 controls attending the same clinic for refractive error were enrolled. Fasting serum leptin was measured by enzyme-linked immunosorbent assay, sampled after excision in the malignant group, and BMI was interpreted using Asia-Pacific thresholds. Cases were analysed both as a single group and divided into malignant and benign lesions. Every comparison was accompanied by an effect size with a 95% confidence interval (CI), and logistic regression was adjusted for age and sex. The smallest detectable difference was calculated at the achieved sample size. RESULTS: Median serum leptin was 17.7 ng/mL (interquartile range 14.8–31.5) in cases and 24.0 ng/mL (12.7–61.0) in controls, a difference in the opposite direction to that predicted and not statistically significant (Hodges-Lehmann −6.39 ng/mL, 95% CI −19.0 to +4.7; P=0.208). Discrimination was absent, with an area under the receiver operating characteristic curve of 0.40 (95% CI 0.25–0.56). Separating the 10 malignant from the 20 benign lesions did not change the result (Kruskal-Wallis P=0.437; malignant versus control P=0.281). After adjustment for age and sex, leptin was not associated with malignancy (odds ratio 0.886 per 10 ng/mL, 95% CI 0.654–1.201), whereas age was (odds ratio 1.162 per year, 95% CI 1.018–1.326, P=0.026). BMI showed the same pattern. Leptin correlated with BMI in every subset examined (all participants ρ=0.634, cases ρ=0.672, controls ρ=0.663; all P<0.001), matching published population estimates. At this sample size the smallest detectable difference was a 2.48-fold ratio of geometric means. CONCLUSION: In this cohort, serum leptin and BMI do not distinguish patients with ocular tumours from controls. The expected leptin–BMI relationship is preserved, indicating that the null result reflects the biology sampled rather than assay failure. Because sampling in the malignant group followed excision, these data address leptin after treatment rather than as a preoperative marker, and they provide no basis for adding serum leptin to the assessment of ocular masses. Detecting a 1.5-fold difference would require 138 participants per group sampled before intervention.