Avaliação da influência de dados clínicos na formação das hipóteses de diagnóstico de lesões intraósseas em radiografias panorâmicas

Abstract

Diagnosis is a fundamental step in clinical practice. Its construction is based on the collection of information during the anamnesis, physical examination and in the complementary examination. The latter plays a fundamental role in concluding the diagnosis and establishing the treatment plan. The present investigation used panoramic radiography as a complementary exam and aimed to evaluate the influence of clinical data of patients in the elaboration of hypotheses for diagnostic imaging. The sample consisted of 41 panoramic radiographs with mandatory presence of intraosseous lesions with conclusive histopathological diagnosis. Three evaluators, experienced radiologists, evaluated the images in two moments: with and without clinical information, to verify their influence on the diagnostic hypotheses raised. The evaluators were responsible to determine, in each evaluation, the observed pathology group, three diagnostic hypotheses for the lesion, their confidence degree in the diagnosis using a scale from 1 to 5 and indicate whether other information from the case were desired for the final diagnosis. Data were collected and submitted to statistical analysis (α = 0.05). To compare the two assessments regarding the group of acquired pathologies, McNemar's statistical test was used. To analyze the final diagnosis of the lesion, Wilcoxon test was used, assigning a score of 3 if the evaluator's diagnosiscoincidedwith the histopathologicaldiagnosisin the first hypothesis,score 2 in the secondhypothesis,score 1 in the third hypothesis, and score 0 in the case of absence of the correct hypothesis. For the comparison of the confidence level between the evaluators, the Wilcoxon test was also used. Regarding the additional information desired by the specialists, a descriptive analysis was carried out. The results revealed that no statistically significant differences (p=0.166) were observed for the groups of indicated path ologies. As for the diagnostic hypotheses suggested by the evaluators, with and without clinical data, a higher median value was observed in the second evaluation with clinical data (p=0.023). The evaluators' confidence level did not differ (p=0.081), and different data were indicated as desirable for the construction of diagnostic hypotheses. It was concluded that clinical data contribute for the radiology specialist for a better performance in the diagnostic hypothesis of intraosseous lesions, possibly favoring the patient's treatment plan.