ABSTRACT
A completely automated system for the identification of pleural nodules in low-dose and thin-slice computed tomography (CT) of the lung has been developed. The directional-gradient concentration method has been applied to the pleura surface and combined with a morphological opening-based procedure to generate a list of nodule candidates. Each nodule candidate is characterized by 12 morphological and textural features, which are analyzed by a rule-based filter and a neural classifier. This detection system has been developed and validated on a dataset of 42 annotated CT scans. The k-fold cross validation has been used to evaluate the neural classifier performance. The system performance variability due to different ground truth agreement levels is discussed. In particular, the poor 44% sensitivity obtained on the ground truth with agreement level 1 (nodules annotated by only one radiologist) with six FP per scan grows up to the 72% if the underlying ground truth is changed to the agreement level 2 (nodules annotated by two radiologists).
Subject(s)
Diagnosis, Computer-Assisted/methods , Lung Neoplasms/diagnostic imaging , Lung Neoplasms/diagnosis , Solitary Pulmonary Nodule/diagnostic imaging , Solitary Pulmonary Nodule/diagnosis , Tomography, X-Ray Computed/methods , Databases, Factual , Diagnosis, Computer-Assisted/statistics & numerical data , False Positive Reactions , Humans , Imaging, Three-Dimensional , Pattern Recognition, Automated , ROC Curve , Radiographic Image Interpretation, Computer-Assisted/methods , Tomography, X-Ray Computed/statistics & numerical dataABSTRACT
A computer-aided detection (CAD) system for the identification of small pulmonary nodules in low-dose and thin-slice CT scans has been developed. The automated procedure for selecting the nodule candidates is mainly based on a filter enhancing spherical-shaped objects. A neural approach based on the classification of each single voxel of a nodule candidate has been purposely developed and implemented to reduce the amount of false-positive findings per scan. The CAD system has been trained to be sensitive to small internal and sub-pleural pulmonary nodules collected in a database of low-dose and thin-slice CT scans. The system performance has been evaluated on a data set of 39 CT containing 75 internal and 27 sub-pleural nodules. The FROC curve obtained on this data set shows high values of sensitivity to lung nodules (80-85% range) at an acceptable level of false positive findings per patient (10-13 FP/scan).
Subject(s)
Artificial Intelligence , Diagnosis, Computer-Assisted , Image Processing, Computer-Assisted , Imaging, Three-Dimensional , Lung Neoplasms/diagnostic imaging , Solitary Pulmonary Nodule/diagnostic imaging , Tomography, Spiral Computed , Algorithms , Humans , Italy , Lung/diagnostic imaging , Mass Screening , Neural Networks, Computer , Pattern Recognition, Automated , ROC Curve , Radiation Dosage , Sensitivity and Specificity , SoftwareABSTRACT
The protective action of 1,2,3,4-tetrahydro-9-aminoacridine (THA) against the long-lasting inactivation of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) brought about by diisopropylfluorophosphate (DFP) and physostigmine, as well as by neostigmine in the case of AChE only, was evaluated by a dilution technique using Electrophorus electricus AChE and horse serum BuChE as target enzymes. In parallel experiments, the ability of physostigmine itself to protect these enzymes from DFP was evaluated and compared with that of THA. THA pretreatment was seen to prevent in a dose-dependent manner the inhibition of both AChE and BuChE. However, it was appreciably more potent towards AChE than towards BuChE. THA mean EC50 values for protecting AChE against 10, 40 and 100 microM DFP were 0.04, 0.16 and 0.45 microM, respectively; against 1 microM physostigmine the value was 1.8 microM and against 1.2 microM neostigmine it was 3.0 microM. The THA mean EC50 value for protecting BuChE against 3 microM physostigmine was 0.55 microM and the values for protecting against 3, 10 and 40 microM DFP were 1.5, 3 and greater than 10 microM, respectively. The protective action of THA was time independent: recovery of the maximal enzymic activity was immediate upon dilution. Unlike THA, the protective action of physostigmine developed progressively after dilution and was maximal within 3-4 (AChE) or 6-8 hr (BuChE). Under our experimental conditions, 0.3 microM physostigmine protected approximately 70% of AChE from 40 microM DFP and 5 microM physostigmine protected 9 and 47% of BuChE from 40 and 3 microM DFP, respectively. The results of this work suggest that THA exerts its protective action by shielding the active site of AChE and BuChE from the attack of the inactivating agents on account of its higher enzymic affinity, whereas the protective action of physostigmine against DFP takes advantage also of the carbamylation of the enzyme. These results are in line with the hypothesis that protection of AChE is the primary mechanism responsible for the antidotal action of THA against organophosphorus poisoning.
