Establishment Of A Diagnostic Score System To Predict Benign Or Malignancy In Solitary Pulmonary Nodules

Objective:To evaluate the clinical features, radiological findings and risk factorsaffecting the diagnosis of solitary pulmonary nodules (SPN) by literature reviewand statistical analysis, and to set up a diagnostic score system to predict benignor malignancy and a processes about diagnosis and treatment in solitarypulmonary nodules.Methods:1. Factors study: A retrospective study was performed in our hospitalinvolving121patients with definite pathological diagnosis of SPN from January2013to2014October.These patients were included66males and55femalesand divided into A, B groups. Of all these patients,76patients were used to setup the diagnostic score system (A group),45patients were used for furthervalidation (B group). Clinical features, radiological findings and risk factorsstudied here were totally16features, including persistent dry cough withtreatment failure, blood sputum, persistent unrelieved chest pain, speculation,lobulation, age, smoking index, family history of malignancy etc.2. Establish a diagnostic score system: According to the literature review andclinical practice, we gave a score to every features based on its incidence andimportance in the diagnosis of SPN malignancy. The univariate and multivariateanalysis were performed respectively to find the independent predictors whichwere consistence with the score system. Then the threshold was found throughthe receiver-operating characteristic curve (ROC curve) and the diagnostic scoresystem was built. 3. Verify the scoring system and establish the flow chart of diagnosis andtreatment: Use group B patients to verify scoring system sensitivity, specificity,accuracy, positive predictive value, negative predictive value. A flow chart wasdraw according to our results and ACCP guideline for handling SPN in ourclinical practice.Results:1. According to the literature review and clinical practice, we give a score(3/2/1) to each feature based on its incidence and importance in the diagnosis ofmalignancy. Logistic univariate analysis was used to determine the correlationof every single feature with the diagnosis of SPN malignancy. For more riskfactors included, the statistical significance was defined by P value<0.1.Features with P values <0.1included lobulation, spiculation, nodularuneven density, vacuole sign, blood sputum, traction of pleural, air bronchogram,smoking index. Further we found that SPN malignancy was significantlycorrelated with lobulation (OR1.908), speculation (OR2.289), nodular unevendensity (OR2.867), vacuole sign (OR8.340) and blood sputum (OR17.661) bymultivariate logistic regression analysis.2. Five independent predictors came from multivariate analysis and76patients were used to create ROC curve and set up our score system. The areaunder the curve (AUC) we got was0.891(P <0.05,95%CI0.8190.963). Thesensitivity and specificity of this ROC curve were87.5%and80.6%,respectively. According to the Youden index (0.681),3.5was considered as thethreshold of our diagnostic score system for determining SPN malignancy. If thetotal score <3.5, there was a higher possibility of benign nodules than that ofmalignant nodules; if the total score≥3.5, the malignant SPN would beconcerned.3. Validation of the above diagnostic score system was performed in theremaining45patients. The88.5%sensitivity,78.9%specificity and the84.4% accuracy were obtained. The positive predictive value was85.2%, the negativepredictive value was83.3%. A flow chart was draw according to our results andACCP guideline for handling SPN in our clinical practice.Conclusions:1. Lobulation, spiculation, nodular uneven density, vacuole sign and bloodysputum were significantly correlated with the diagnosis of SPN malignancy aftergiven a score.2. The diagnostic score system including the five independent predictors andthe threshold of3.5have a better diagnostic capability in determining SPNmalignancy.3. Combined with ACCP guidelines for the diagnosis and treatment of SPN,clinicians can make a reasonable decision to deal with SPN, eventuallyachieving early diagnosis, early treatment, and avoiding overtreatment.