| The differentiation of the character of solitary pulmonary consolidatory lesions is an old and unsolvemental problem in the radiologic field. There are many methods which have their advantages and disadvantages on differentiation. A lot of new problems have been found in the process of their application. They are described in many articles in literatures. The history of methods which are designed to make use of lesions' morphology and sign in conventional X-ray or CT to differentiate the character of lesion is long.The fact has proved: A benign lesion will be misdiagnosed as a malignant one if only a few signs or morphologies which are more seen in malignant lesions in CT have been considered. It will result in unnecessary surgical treatment. Combination of all kinds of their morphologies and signs in conventional X-ray and CT have become one of the methods to differentiate the character of solitary pulmonary consolidatory lesion.We have found some attempts on differentiating the character of the lesions in foreign literatures. Foreign learners designed a method called "artificial neural network" .They drew some character of lesions such as size, dense, lobulation,spiculation and some clinical character which were considered as a whole. These data were put into the computer. With the help of computer, we acquired the probability of malignancy of the lesion by means of some arithmetical formula.There are a few researches in domestic literatures now. On the basis of many studies, Majority authors in our country have come into a general agreement that we can improve the diagnostic accuracy of the character of the lesions if all kinds of signs and morphologies of the lesions in conventional X-ray or CT have been considered as a whole body.1.PURPOSE:To improve the diagnostic accuracy of malignant solitary pulmonary consolidatory lesions, we designed a method called "evaluated with malignant class of the lesions" .We drew a study group to test the method .Its clinical significance is also discussed.2.MATERIAL AND METHODS:The study group was drawn from patients between 7,2000 and 7,2002 for solitary pulmonary consolidatory lesions. Forty-three patients who underwent CT examination among them have been consecutive collected as our study group finally. There are twenty-seven malignant lesions and sixteen benign lesions in study group.we designed a method called "evaluated with malignant class of the lesions " at the basis of seven morphologies and signs which were more seen in CT imagings of the malignant lesions.We invited three radiologists who blinded to the final diagnosis examined all CT imagings of the study group according to seven signs and morphologies of the lesions which are given first. According to method we designed, we evaluated the lesions' malignant class by means of the result of the examining. We also calculated the corrective rate of screening the malignant lesions from the totle study group while adoping different malignant class as threshold respectively.3.RESULT:The study result proved: Although seven morphologies and signs we selected in this study are more seen in the malignant group, we can find they exist in CT imagings of the benign group to some extent. The statistical result displays an interesting phenomena that the higher malignant class we evaluated, the higher diagnostic accuracy of the malignant lesions.4.CONCLUSION:According to the statistical result, we conclude that differentiate the character of lesions with our designed method may be helpful to improve the diagnostic accuracy of malignant lesions in CT. Considering the number of the sample which are collected to test is small, we need more investigation to test the validity of the method. |