Step 1: Anchor on the truth standard. The disease is defined by the gold standard, which labels 100 individuals as truly diabetic. Sensitivity asks: of these 100, how many did the screening test catch?
Step 2: Read off true positives. The screening test flagged 90 positives, all assumed to be among the truly diseased, so true positives $a = 90$.
Step 3: Account for the missed cases. The screening test failed to detect $100 - 90 = 10$ genuinely diseased people; these are the false negatives $c = 10$.
Step 4: Plug into the relation. \[\text{Sensitivity} = \frac{a}{a+c} = \frac{90}{90+10} = \frac{90}{100}\] The denominator is always the total truly diseased (100), confirming option (a) and ruling out the variants that misuse the count.
\[\boxed{90/100}\]