RADIOLOGICAL AND PATHOLOGICAL CORRELATION OF LUNG NODULES IN A BACKGROUND OF METASTATIC DISEASE

Abstract

Background: Computed tomography (CT) imaging has improved the chances of detecting small indeterminate (<1 cm) lung nodules. The determination of the underlying malignant or benign nature of a lung nodule poses a great diagnostic challenge and depends on a number of factors, including the radiographic appearance of nodule, the presence of non-pulmonary metastases, characteristics of growth and histological criteria. Methods: The medical records of 89 patients admitted to our specialist cancer centre between 2008 and 2013 were reviewed. Patients of all age groups and tumour category were included in the study. Clinical data of these patients were collected and the following parameters were analysed: Radiographic diagnosis, location, size, laterality and number of nodules and histological impression. The radiological findings were then correlated with histopathological findings. The nodules were sub-classified into groups on the basis of size (A = 0–0.5 cm; B = 0.5–0.9 cm; C = 1.0–1.5 cm and D = >1.5 cm). Results: CT scan reports of 89 patients with lung nodules were reviewed. On radiology, 73/89 (82%) were reported to be malignant nodule. Histopathological review of the biopsies of these 89 nodules confirmed malignancy in 50/89 (56.2%) patients. CT scan was found to be highly sensitive (94%, 95% confidence interval [CI]: 83.43–98.68%) but with a very low specificity (33.3%, 95% CI: 19.10–50.22%). CT scan was found to have a higher negative predictive value (81.2%, 95% CI: 54.34–95.73%) and a lower positive predictive value 64.4% (95% CI: 52.31–75.25%) when correlated with histopathological findings. Pathology of these nodules included metastatic sarcoma (27/89; 30.3%) and carcinoma (18/89; 20.2%). The frequency of the biopsy-proven malignant nodules on the right side was 26/45 (57.8%) and on the left side was 24/44 (54.5%) (P = 0.832). Malignant nodules were more frequent in lower lobes (28/43, 65.1%) than in upper lobes (14/32, 43.8%). These two sites combined accounted for 84% of all malignant nodules. There was a significant correlation between nodule size and likelihood of underlying malignancy. The overall prevalence of malignancy in the larger nodules (C and D) was much higher (23/30 and 76.7%) compared to the smaller sized (A and B) nodules (27/58 and 46.8%), P < 0.05.Conclusion: CT scan is a useful tool in the initial clinical assessment of indeterminate lung nodules with high sensitivity (94%) and a high negative predictive value (81.2%).Key words: Computed tomography, fibrosis, indeterminate lung nodule, infection, lung nodule, malignancy, metastases