The diagnosis of Hashimoto's thyroiditis

Computer methods were used to estimate the usefulness of several clinical signs and laboratory tests in the diagnosis of chronic lymphocytic (Hashimoto's) thyroiditis. Information was drawn from the records of 217 patients with this disease seen at two hospitals in Southern California. The parameters studied included the physical characteristics of the thyroid gland; the appearance of the radioisotope thyroid scan; the response to a perchlorate discharge test; the serum antithyroglobulin antibody titer; the serum TSH concentration measured by radioimmunoassay; the thyroid radioiodine uptake response to exogenous TSH stimulation; and the serum PBI-T4I difference. Of the above, the PBI-T4I difference was deleted, being of limited value. The TSH stimulation test and serum TSH measurement were considered as alternative ways to evaluate thyroid reserve. Therefore, five diagnostic markers remained, all useful but not definitive. Computer and rule-of-thumb methods (two or more of the five markers positive) were tested for successful diagnosis of 145 patients with Hashimoto's thyroiditis proven by pathological examination of biopsied tissue; 23 of these were patients not used in the original pool of data. As a rule of thumb, it was found that Hashimoto's thyroiditis is the likely diagnosis if two or more out of the five useful markers are in its favor. For best results, four and preferably all of the five criteria should be tested. If this is done, the expected diagnostic accuracy for patients with Hashimoto's disease is by the rule-of-thumb method 67% correct, 21% indecisive, and 12% false negative; by computer methods it is 88% correct, 4% indecisive, and 8% false negative. By each method the number of false positive diagnoses was equivalent to 25% of the total number of patients with this disease. The false positive results nearly all occurred in patients with goiter associated with a defect in thyroid hormone synthesis.