Diagnostic accuracy of salivary scintigraphic indices in xerostomic populations.
PURPOSE: Three decades of work to enhance the diagnostic accuracy of salivary scintigraphy have generated various plausible decision criteria. This study evaluates four commonly cited numeric indices in studies of xerostomic populations and how accurately they identify Sjögren's syndrome, chronic sialadenitis, radiation sialadenitis, and drug effects and distinguish each from the other. METHODS: Stimulated dynamic salivary scintigraphy was performed on 295 xerostomic patients and on 31 controls. The nonparametric area under the receiver operating characteristic curves expressed the diagnostic accuracy of the following scintigraphic indices: the parotid:submandibular ratio of unstimulated glandular activity, the peak:baseline uptake ratio, its time of occurrence, and the stimulated excretion fraction. RESULTS: The stimulated excretion fraction distinguished Sjögren's syndrome and radiation sialadenitis from healthy states with respective accuracies of 0.78 and 0.90. The maximum diagnostic payoff in Sjögren's syndrome occurred at a cutoff of 73%, yielding a 73% rate of test sensitivity and a 73% rate of specificity. The other three indices were not useful. Even the stimulated excretion fraction performed indifferently or poorly in most other diagnostic tasks. CONCLUSIONS: In the scintigraphic examination of xerostomic and healthy populations, an acceptable diagnostic utility of the stimulated excretion fraction was evident only in Sjögren's syndrome and radiation sialadenitis. When presented with differential diagnostic alternatives not involving radiation sialadenitis, none of the four numeric indices performed acceptably.