The purpose of the study was to substantiate the capacities of X-ray recognition of the histological types of lung cancer and different tumor differentiation grades. The data available on 238 patients were studied; of them 169 patients were diagnosed as having primary lung cancer that was morphologically verified. In 154 patients, the tumor was not greater than 3.0 cm in diameter and belonged to the main histological types. X-ray study, analogue, digital tomography, X-ray surgery with percutaneous needle biopsy and rapid cytology were used. The diagnosis made by WHO ICD-10 C34.1-3 BDU (probable) and C34.1-3 M (valid) was a criterion for the quality of a accomplished diagnostic case. Of 13 traditional X-ray sings of lung cancer, 6 characteristics of a tumor image were taken for a following analysis. A number of signs correlate with the histological type of lung cancer in the symptom complexes that characterize the image of its histological types. Thus, "specula" are detectable in adenocarcinoma; a lobate pattern and minor focal clarification are also almost without exception characteristic. The differences between the images of squamous-cell carcinoma and adenocarcinoma were statistically significant (p < 0.05). The statistical models of lung cancer and its types were tested on 60 patients. Implementation of the optimized symptom complex positively changed the parameters of X-ray diagnosis: there were increases in the sensitivity from 68.6% to 80.0%, in the specificity from 24.3% to 65.0%; in the prognostic value of a positive result from 55.6% to 82.1%; in that of a negative result from 34.0% to 64.9%, and in the accuracy from 50.0% to 75.0%.
Download full-text PDF |
Source |
---|
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!