Increasing incidence of primary malignant brain tumors: influence of diagnostic methods.

Background: The incidence of brain cancer has increased dramatically over the last decades in most developed countries. Whether these trends can be attributed to improved diagnosis is not clear.

Purpose: To determine the effect of new imaging technology on increased rates of brain cancer, we assessed the level of detection for neurological disorders when computed tomography (CT) and magnetic resonance imaging (MRI) results were not available.

Methods: A neurologist performed a blind review of hospital charts from 356 randomly selected patients, hospitalized between 1985 and 1989 for neurological disorders, including brain cancer. All prediagnosis information except CT and MRI results was used as a basis for diagnostic re-evaluation. Also, a random sample of 151 brain cancer patients diagnosed between 1960 and 1965 was selected for a description of diagnostic methods used during that period.

Results: A comparison between the original diagnoses and the re-evaluations for patients in the 1985-1989 sample indicated that there was, among the diseases selected, a 24% misclassification when CT scans and MRI were not available. In particular, 20% of brain tumors were undetected (95% confidence interval = 15%-25%), and 10% of non-tumor disorders were inaccurately labeled as brain tumors in the absence of these tests. The repeatability of the re-evaluations was 86%.

Conclusions: Among elderly North Americans, at least twofold increases in brain cancer incidence were observed over the last two decades. Since our findings show that CT scans and MRI are responsible for the detection of about 20% of brain tumors, we conclude that other factors also are responsible for the observed trends.