RESUMEN
To address the diagnostic performance of scratch-imprint cytology (SIC), in this study we compared intraoperative diagnoses of pulmonary lesions between SIC and frozen section histology (FSH) for accuracy with respect to the final pathological diagnosis. We histologically divided 206 pulmonary lesions (resected surgically) into two groups (benign and malignant) and compared each intraoperative diagnosis by SIC and FSH with the final pathological diagnoses. We also examined the radiological existence of pure ground-glass opacity (GGO) nodules in each group. The diagnostic sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 91.5%, 100%, 100%, 63.6%, and 92.6%, respectively for SIC, and 98.2%, 100%, 100%, 92.1% and 98.5%, respectively, for FSH. Thus, we concluded that diagnosis by SIC is reliable for malignancy, but not for benign lesions. All pure GGO nodules (19; 9.2%) were noninfectious and malignant with a high accuracy of FSH diagnosis (100%), in comparison with those of low accuracy with a SIC diagnosis (57.9%). SIC can be an appropriate intraoperative diagnostic tool where multiple cytotechnologists observe intraoperative SIC preparations scratched evenly across the whole lesion including the peripheral area of the mass.
Asunto(s)
Citodiagnóstico/métodos , Enfermedades Pulmonares/diagnóstico , Secciones por Congelación/métodos , Humanos , Sensibilidad y EspecificidadRESUMEN
We aimed to propose a biosafety algorithm for the protection of pathology staff during intraoperative examinations of pulmonary lesions when working with cytological imprints and/or frozen sections for the intraoperative diagnosis of pulmonary lesions. We examined 148 pulmonary surgical tissues obtained intraoperatively for imprint cytology (IC) and for frozen sectioning and compared the diagnoses against the final pathological diagnoses. We analyzed concordance and non-concordance rates and then used the data to produce a biosafety algorithm. The diagnostic sensitivity, specificity, positive predictive value, negative predictive value and accuracy of scratch-IC were 91%, 100%, 100%, 50% and 92%, respectively, and those of frozen sectioning were 99%, 100%, 100%, 96% and 99%, respectively. Our data indicate that frozen sectioning is unnecessary if scratch-IC yields a 'malignant' diagnosis but recommended with a 'benign' diagnosis. When a scratch-IC preparation deemed inadequate for a diagnosis or an abscess, the pathologist must consult the surgeon concerning the possibility of granuloma with caseous necrosis and should ask the surgeon to be prepared for a frozen section. If granuloma with caseous necrosis is found in the frozen section, the pathologist must immediately communicate the information to entire staff and perform a PCR test before making a permanent section.