We were interested to read the article by Sajnani and King in the July 2012 issue of the AJO-DO . The authors retrospectively tried to predict impaction of a maxillary canine using measurements made on a panoramic radiograph. As the authors point out in their conclusion, such a prediction is possible at 8 years of age by using geometric measurements on panoramic radiographs.
As a rule of thumb, diagnosis of an outcome can be made cross-sectionally; however, 2 separate prospective cohort data sets should be used for prediction. Moreover, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), likelihood ratio positive (true positive/false positive), and likelihood ratio negative (false negative/true negative) as well as odds ratio (true results/false results, preferably more than 50) are among the tests to evaluate the validity (diagnostic accuracy) of a single test compared with a gold standard. Area under the receiver operating characteristic curve (AUC) is also reported for diagnostic rather than prognostic values of a model. Why did the authors not use these well-known tests?
They reported a clinically discernible difference of 4 mm at the age of 8 years and beyond between the mean distance of the tip of the impacted canine group and that of the antimere group from the occlusal plane ( P <0.05). Reporting such differences has nothing to do with either diagnosis or prediction of a clinical outcome.