Thin-slice Scans Research Articles

Screening for lung cancer using thin-slice low-dose computed tomography in southwestern China: a population-based real-world study.

Lung cancer is one of the most common malignant tumors threatening human life and health. At present, low-dose computed tomography (LDCT) screening for the high-risk population to achieve early diagnosis and treatment of lung cancer has become the first choice recommended by many authoritative international medical organizations. To further optimize the lung cancer screening method, we conducted a real-world study of LDCT lung cancer screening in a large sample of a healthy physical examination population, comparing differences in lung nodules and lung cancer detection between thin and thick-slice LDCT scanning. A total of 29 296 subjects who underwent low-dose thick-slice CT scanning (5 mm thickness) from January 2015 to December 2015 and 28 058 subjects who underwent low-dose thin-slice CT scanning (1 mm thickness) from January 2018 to December 2018 in West China Hospital were included. The positive detection rate, detection rate of lung cancer, pathological stage of lung cancer, and mortality rate of lung cancer were analyzed and compared between the two groups. The positive rate of LDCT screening in the thin-slice scanning group was significantly higher than that in the thick-slice scanning group (20.1% vs. 14.4%, p < 0.001). In addition, the lung cancer detection rate in the thin-slice LDCT screening positive group was significantly higher than that in the thick-slice scanning group (78.0% vs. 52.9%, p < 0.001). The screening positive rate of low-dose thin-slice CT scanning is higher and more early-stage lung cancer (IA1 stage) can be detected in the screen-positive group.

English

BACKGROUND Multidetector computed tomography (MDCT) is the imaging of choice and the most commonly used investigation in head and neck lesions, because of its fast and readily available nature. Our study focused on contrast enhanced multidetector CT profile of neck masses in determining their nature among different demographic profile. It provides volumetric helical data, optimal multiplanar and 3D reconstructions. METHODS This is a retrospective observational study carried out in the Department of Radiology, Gauhati Medical College and Hospital, Guwahati, among 60 cases of clinically diagnosed neck masses, from December 2011 to June 2013. All patients were selected from out-patient departments (OPD) and indoor wards of various departments (mostly from ENT department). Few cases were also taken from Dr. B. Baruah Cancer Institute. Patients from both sexes and all ages were included as part of the study. Chi square test/Fischer exact test have been used to find the significance of association of CT scan findings with the final diagnosis. Diagnostic statistics such as sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy have been used to find the correlation of CT scan with the final diagnosis. RESULTS Thin slice scanning with thinner reconstructions, maximum intensity production (MIP), shaded surface display (SSD), multiplanar reformation (MPR) and curved reformatted images are the advantages of MDCT. CT has 100 % accuracy in bony involvement detection and 96 % accuracy in neck lesion diagnosis. Therefore, CT can accurately localise and characterise the neck lesions. CONCLUSIONS The neck lesion location and its characteristics determination has significantly improved because of computed tomography. Bone erosions and expansion are accurately determined by a CT scan. CT scan can accurately delineate the disease that provides a definite pre-operative diagnosis, planning radiotherapy ports and post therapy follow up. KEYWORDS Bone Invasion, Lymph Nodes, Neck Space Extensions, Multidetector-CT

Attention-based full slice brain CT image diagnosis with explanations

Diagnose brain diseases by brain CT images is one of the most common ways. But, it usually takes several (>7) years to train a professional doctor because it is very challenging to diagnose brain diseases correctly. The study on automated assistance of brain CT diagnosis is still limited. In this paper, we research the challenges of this task and propose a method by simulating human doctor diagnosis habits. Our method analyzes a full slice of brain CT images, instead of every single one, to take into account continuous changes of the whole brain structure, simulate the way the doctor diagnoses. To avoid redundancies in a thin slice scan, we propose a redundancy removal and data augmentation method that can both reduce computation complexity and improve performance without information loss. Doctors make a diagnosis by observing several key images and key points in them. Our method achieved this by two steps of attention mechanisms. It can highlight the images and key points that have significant impacts on the prediction and explain the results. We evaluated our method on two public datasets CQ500 and RSNA, which achieved 0.9262 and 0.8650 F1 score respectively. Moreover, an experienced doctor (with 29 years of experience) verified the promising clinical application value of the proposed method through manual experiments.

Deep Learning Approach Applied to Prediction of Bone Age Based on Computed Tomography Orthopedic Image Processing

Objective: Deep learning and neural network models are new research directions in the field of machine learning and artificial intelligence. Deep learning has made breakthroughs in image recognition and speech recognition applications, and has also shown unique advantages in face recognition and information retrieval, and has been widely used. Methods: Thin-layer computed tomography (CT) scan and multiplanar reconstruction (MPR) and volume reconstruction (VR) techniques were used to perform CT thin-slice scan and volume of the bilateral clavicle sternum at 548 number of l5~25 years old. Reproduction (volume rendering, VR) and three-dimensional image recombination, measuring and calculating the longest diameter of the sternal end of the bilateral clavicle, the longest diameter of the metaphysis and its length ratio, the area of the epiphysis, the area of the metaphysis and its area ratio, etc. We establish a mathematical model of bone age inference, and then substitute 50 training samples into the mathematical model to test the accuracy of the model. Results: There was a statistically significant difference between the male and female sex ratios in the same age group (P &lt; 0.05). The established mathematical model shows that the developmental law of the sternal skeletal bone is highly correlated with the biological age. The accuracy of all models is 70.5% (±1.0 years old) and 82.5% (±1.5 years). Skeletal X-ray images show different gradation changes in black and white, with black-and-white contrast and hierarchical image features. Based on the advantages of deep learning in image recognition, we combine it with bone age assessment research to build a forensic bone age automation. Conclusions: This paper harnesses the basic concepts of deep learning and its network structure, and expounds the research progress of deep learning in image recognition in different research fields at home and abroad in recent years, as well as the advantages and application prospects of deep learning in bone age assessment.

Preliminary study of medialization thyroplasty for unilateral vocal fold immobility

Objective:To explore the feasibility and effectiveness of medialization thyroplasty for the treatment of unilateral vocal fold immobility （UVFI）. Method:Eight UVFI patients were performed medialization thyroplasty under local anesthesia. We made a window in the thyroid cartilage under local anesthesia, then insert the preformed silicone implant. The paralyzed vocal fold was medialized to make the glottis closed. Fibrolaryngoscope video recording, subjective voice analysis and CT thin slice scan of larynx were done before and after surgery to evaluate closure of vocal cords, improvement of voice and position of implantation. Result:The preoperative and postoperative voice handicap index 30（VHI-30） of the 8 patients were 91.5（64.5, 97.5） and 22.5（5.0, 47.5） respectively, which showed statistical difference（P<0.05）. GRBAS results: The postoperative G, R, B, A were all smaller than preoperative ones, which showed statistical difference（P<0.05）; the pre and postoperative S was both 0. The fibrolaryhgoscope recording showed the preoperative and postoperative score of incomplete glottis closure was 1.0（1.0, 1.0） and 4.0（2.5, 5.0） respectively, which showed statistical difference（P<0.05）. Postoperative laryngeal CT showed significant vocal cord medialization on the affected side. Aspiration was significantly improved in 4 patients who were suffered from this symptom before the surgery. No complication occurred with the 8 patients during 5 to 48 months follow up. Conclusion:Medialization thyroplasty can effectively improve vocalization and quality of life in patients with UVFI.

Analysis of the advantages of 3D printing in the surgical treatment of multiple rib fractures: 5 cases report

BackgroundRib fractures account for a fairly high proportion of chest injuries, ranging from 55 to 80%. The most common mechanisms of injury include: traffic accident, extrusion and falls from significant heights. Besides, the surgical treatment of multiple rib fractures has been accepted by more and more medical professionals. We reported 5 clinical cases of patients with multiple rib fractures undergoing open reduction and internal fixation using 3D printing technology.Case presentationRetrospective analysis of 5 clinical cases of multiple rib fractures from January 2017 to August 2018 in our hospital. A preoperative CT thin slice scan was used to reconstruct the 3D model according to the scanning results, and 3D printing technology was adopted to prepare the rib model. Preoperative reconstruction of the rib’s normal shape and lock plate for the shaped ribs was created according to reconstructed model. For multiple fractures especially patients with severely deformed rib shape, it is suggested to intraoperative shape directly to the metal bone plate fixed on the ribs on both ends of the fracture line, in order to establish a basic support frame. The other various fracture section can be fixed on the lock plate respectively. Postoperative chest radiographs of the 5 patients showed that the internal fixations were in good and natural shape. The thoracic contour was well formed and symmetrically with the contralateral side.ConclusionMaking the rib model and the pre-shaped titanium alloy rib locking plate using 3D printing technology, provided a more minimally invasive and precisely individualized treatment for some rib fracture operations.

Design and basic research on accuracy of a novel individualized three-dimensional printed navigation template in atlantoaxial pedicle screw placement.

ObjectiveTo design and evaluate the accuracy of a novel navigation template suitable for posterior cervical screw placement surgery by using 3D printing technology to improve the existing guiding template design.MethodsThe researchers (including spine surgeons and technicians) used CT to perform thin-slice scanning on 12 cases of normal upper cervical vertebral specimens and defined the screw channels that were completely located in the pedicle without penetrating the cortex as ideal screw channels, then designed the ideal channel of the upper cervical vertebral (atlantoaxial) pedicle screw by computer software which was regarded as the preset values, and recorded the screw entrance point, transverse angle and sagittal angle of the ideal channel. Then, researchers designed the novel navigation templates for placement pedicle screw according to the ideal screw channel preset values and manufactured them with one for every single vertebra by 3D printer. A senior spine surgeon performed the posterior surgery to implant pedicle screw on the specimens by the novel navigation templates, then performed CT thin-slice scanning on the specimens again after removing the screws, and reconstructed the actual screws channel by computer software, recorded the screw entrance point, transverse angle and sagittal angle of the actual channels which were defined as the actual values and evaluated them according to Kawaguchi’s pedicle screw evaluation standard finally. The differences between the preoperative preset values of ideal screw channel and the postoperative actual values of actual screw channel were compared by a nonparametric paired rank test.Results48 screws were placed on 12 cases of upper cervical vertebral specimens in total. It showed that the grade 0, I, II, III channels in this study were 47, 1, 0, 0, respectively. The grade 0 channels accounted for 97.92% of the total number of channels. There was no significant difference with regard to the screw entrance point, the transverse angle, and the sagittal angle between the preoperative preset values of ideal screw channels and the postoperative actual values of actual screw channels.ConclusionTo implant pedicle screw assisted with the novel individually navigation template designed by 3D printed in the posterior cervical surgery can improve accuracy of pedicle screw placement and safety of the surgery.

Population specificity in the estimation of skeletal age and sex: case studies using a Western Australian population

ABSTRACTThe expertise of the forensic anthropologist is grounded in the study of skeletal morphology relative to normal state; the latter is variously defined according to the complex interrelationship between biological function, genetics and intra-specific variation (amongst others). In relation to sex estimation, there are multiple dimorphic attributes in the adult skeleton that facilitate accurate sex estimations. It is, however, well established that there are population specific influences in not only skeletal growth, but also the expression and magnitude of sexual dimorphism (notwithstanding relative consistency in ‘universally dimorphic traits’ that relate to biological functioning – e.g., pelvic shape in relation to female functional requirements for childbirth).The recognition of the need for statistical quantification of forensic assessment is mandated by judicial requirement. Application of reference standards ‘foreign’ to that of the local jurisdiction, however, introduces an unknown variation and by association possible misclassifications. The present paper aims to empirically demonstrate the effect of the latter in a Western Australian population using multivariate classification statistics. We report here the analysis of a sample comprising three-dimensionally reconstructed cranial (n = 400) and pelvic (n = 400) multi-detector computed tomography (MDCT) thin-slice scans. In the morphometric analysis of cranial dimorphism, we demonstrate an unacceptably large sex bias; foreign standards thus frequently misclassify females.

Automatic lung nodule detection using multi-scale dot nodule-enhancement filter and weighted support vector machines in chest computed tomography.

A novel CAD scheme for automated lung nodule detection is proposed to assist radiologists with the detection of lung cancer on CT scans. The proposed scheme is composed of four major steps: (1) lung volume segmentation, (2) nodule candidate extraction and grouping, (3) false positives reduction for the non-vessel tree group, and (4) classification for the vessel tree group. Lung segmentation is performed first. Then, 3D labeling technology is used to divide nodule candidates into two groups. For the non-vessel tree group, nodule candidates are classified as true nodules at the false positive reduction stage if the candidates survive the rule-based classifier and are not screened out by the dot filter. For the vessel tree group, nodule candidates are extracted using dot filter. Next, RSFS feature selection is used to select the most discriminating features for classification. Finally, WSVM with an undersampling approach is adopted to discriminate true nodules from vessel bifurcations in vessel tree group. The proposed method was evaluated on 154 thin-slice scans with 204 nodules in the LIDC database. The performance of the proposed CAD scheme yielded a high sensitivity (87.81%) while maintaining a low false rate (1.057 FPs/scan). The experimental results indicate the performance of our method may be better than the existing methods.

Automatic lung nodule detection using a 3D deep convolutional neural network combined with a multi-scale prediction strategy in chest CTs

ObjectiveA novel computer-aided detection (CAD) scheme for lung nodule detection using a 3D deep convolutional neural network combined with a multi-scale prediction strategy is proposed to assist radiologists by providing a second opinion on accurate lung nodule detection, which is a crucial step in early diagnosis of lung cancer. MethodA 3D deep convolutional neural network (CNN) with multi-scale prediction was used to detect lung nodules after the lungs were segmented from chest CT scans, with a comprehensive method utilized. Compared with a 2D CNN, a 3D CNN can utilize richer spatial 3D contextual information and generate more discriminative features after being trained with 3D samples to fully represent lung nodules. Furthermore, a multi-scale lung nodule prediction strategy, including multi-scale cube prediction and cube clustering, is also proposed to detect extremely small nodules. ResultThe proposed method was evaluated on 888 thin-slice scans with 1186 nodules in the LUNA16 database. All results were obtained via 10-fold cross-validation. Three options of the proposed scheme are provided for selection according to the actual needs. The sensitivity of the proposed scheme with the primary option reached 87.94% and 92.93% at one and four false positives per scan, respectively. Meanwhile, the competition performance metric (CPM) score is very satisfying (0.7967). ConclusionThe experimental results demonstrate the outstanding detection performance of the proposed nodule detection scheme. In addition, the proposed scheme can be extended to other medical image recognition fields.

Preoperative evaluation of liver volume in living donor liver transplantation.

OBJECTIVE:The aim of the present study was to retrospectively evaluate the difference between the preoperative estimated volume and the actual intraoperative graft volume determined in donor right hepatectomies and to evaluate the possible effect of age, gender, and body mass index on the difference.METHODS:A total of 225 donor hepatectomies performed at the center between 2016 and 2017 were evaluated for the study. Left hepatectomies and left lateral segmentectomies were excluded from the analysis. As a result, 174 donor right hepatectomies were included in the study. Volumetric analysis was performed with dynamic hepatic computed tomography (CT), including non-contrast analysis, followed by non-ionic, contrast-enhanced arterial, portal, and hepatic-phase, thin-slice scanning. Volumetric analysis was performed based on the CT images using automatic volume calculating software.RESULTS:The mean preoperatively estimated graft volume was 800±112 g and the mean intraoperatively measured actual graft volume was 750±131 g. There was a statistically significant difference (p=0.003). Age and body mass index had a significant impact on the discrepancy between the predicted and actual graft volume, while gender did not.CONCLUSION:A thorough preoperative evaluation of the donor graft volume should be performed in order to prevent donor morbidity and mortality, as well as small-for-size and large-for-size phenomena in the implanted grafts. Physicians working in the field of transplantation should be aware of the fact that a difference of 10% between the predicted and the actual graft volume is usually encountered.

Preparation of simulate craniocerebral models via three dimensional printing technique

Three dimensional (3D) printing technique was used to prepare the simulate craniocerebral models, which were applied to preoperative planning and surgical simulation. The image data was collected from PACS system. Image data of skull bone, brain tissue and tumors, cerebral arteries and aneurysms, and functional regions and relative neural tracts of the brain were extracted from thin slice scan (slice thickness 0.5 mm) of computed tomography (CT), magnetic resonance imaging (MRI, slice thickness 1mm), computed tomography angiography (CTA), and functional magnetic resonance imaging (fMRI) data, respectively. MIMICS software was applied to reconstruct colored virtual models by identifying and differentiating tissues according to their gray scales. Then the colored virtual models were submitted to 3D printer which produced life-sized craniocerebral models for surgical planning and surgical simulation. 3D printing craniocerebral models allowed neurosurgeons to perform complex procedures in specific clinical cases though detailed surgical planning. It offered great convenience for evaluating the size of spatial fissure of sellar region before surgery, which helped to optimize surgical approach planning. These 3D models also provided detailed information about the location of aneurysms and their parent arteries, which helped surgeons to choose appropriate aneurismal clips, as well as perform surgical simulation. The models further gave clear indications of depth and extent of tumors and their relationship to eloquent cortical areas and adjacent neural tracts, which were able to avoid surgical damaging of important neural structures. As a novel and promising technique, the application of 3D printing craniocerebral models could improve the surgical planning by converting virtual visualization into real life-sized models.It also contributes to functional anatomy study.

The diagnostic value of multislice computed tomography in children with congenital lobar emphysema

Objective: To investigate the diagnostic value of multislice computed tomography (MSCT) in children with congenital lobar emphysema (CLE). Materials & methods: Seven children with CLE diagnosed by surgery pathology had undergone thin-slice scanning of MSCT, multiplanar reconstruction and volume rendering technique. The diagnostic value of MSCT in children with CLE was evaluated. Results: Among seven patients with CLE, three cases were located in the left upper lobe, two cases in the right upper lobe and two cases in the right middle-upper lobe. All cases were manifested by variable degrees of bronchus stenosis in the corresponding lung segment or lobe, increased lung radiolucence and volume, and reduced lung markings. In this group, there was one case, respectively, associated with left posterior mediastinal bronchogenic cyst, bronchiectasis in the right lung, a lung cyst in the right upper lobe, angiodysplasia in the right upper lung and congenital funnel chest; two cases of multiple lung bullae in the right lung, three cases of infection in both lungs and four cases of mediastinal hernia. Conclusion: MSCT and multidimensional reconstruction is a noninvasive diagnostic method that displays the location, degree and accompanying anomalies of CLE in three dimensions. It has significance for clinicians in that it reduces both misdiagnosis and the time taken to receive treatment.

Study on inferior petrosal sinus and its confluence pattern with relevant veins by MSCT

To explore the anatomic route of inferior petrosal sinus (IPS) after going out of the cranium and its confluence patterns with internal jugular vein (IJV), anterior condylar vein (ACV) and lateral condylar vein (LCV), and to supply knowledge about typing of IPS-IJV junction, so as to provide reference evidence for evaluation of transvenous access route in the diagnosis and treatment of skull base and cavernous sinus lesions. In 120 patients, the IPS route and its confluence with IJV, ACV and LCV were shown by multi-planar reconstruction (MPR) and curve multi-planar reconstruction (CMPR). Combining with continuous thin slice, the IPS-IJV junction was further subdivided according to the level of IPS confluence with IJV and whether there is an anastomosis with sigmoid sinus (SS). Furthermore, the IPS length, venous diameter at IPS-IJV junction and IPS-SS communicating branch were determined and compared. Inferior petrosal sinus directly draining into jugular bulb (JB) or/and draining into JB after communication with SS was found in 28 sides (11.7%, pattern A); IPS draining into IJV at the level of exterior opening of hypoglossal canal in 114 sides (47.5%, pattern B); IPS draining into IJV in a lower extracranial level in 83 sides (34.6%, pattern C); IPS with multiple junctions draining into IJV near the jugular foramen in 12 sides (5.0%, pattern D); IPS directly draining into VVP in 1 side (0.4%, pattern E); IPS being absent in 1 side (0.4%, pattern F). IPS draining into VVP via ACV was seen in 218 sides, IPS draining into VVP via LCV in 100 sides and IPS directly draining into IJV in 14 sides. After going out of the cranium, IPS goes along with IJV for a relatively long distance in some cases. The IPS extracranial length over 40 mm was found in ten sides and the lowest level of IPS route was at the inferior margin of the fourth cervical vertebra. The venous diameter at the IPS-IJV confluence was 0.8-5.7 mm (mean 2.51 mm) and it was significantly larger on the right side than on the left (P = 0.01). However, there was no remarkable difference between male patients and female ones. Continuous thin-slice scanning by multislice spiral computed tomography in combination with MPR and CMPR can clearly show IPS route and its confluence with relevant veins, and determine the feasibility of procedures via IPS. Therefore, it can be used as an effective method for preoperative evaluation of IPS for diagnosis and treatment of skull base and cavernous sinus lesions by IJV access route.

The technique of radical pancreaticoduodenectomy for malignant tumor in pancreatic head with pressed superior mesenteric blood vessels or portal vein

To investigate the technique of radical pancreaticoduodenectomy for malignant tumor in pancreatic head with pressed superior mesenteric blood vessel or portal vein. From March 2005 to March 2007, thin slice scan and vessel-reconstruction of 56 patients of malignant tumor in pancreatic head with pressed superior mesenteric blood vessels or portal vein were carried out using multidetector spiral CT to evaluate whether peripheral vessels of pancreatic tumor were invaded and whether the tumor was resectable. During the operation, 3 vascular blocking bands for superior mesenteric vein, portal vein and spleen vein or 4 vascular blocking bands (additional one for inferior mesenteric vein) were preset. Under the cross and traction between superior mesenteric vein and superior mesenteric artery, resected the uncinate process of pancreas thoroughly. Using those methods, radical pancreaticoduodenectomy for 56 patients above-mentioned were successfully accomplished. The accuracy for preoperative judging by using multidetector spiral CT whether the peripheral vessels of pancreatic cancer were invaded and whether the tumor was resectable was 98% and 100% separately. Thirty-seven of 56 patients, whose superior mesenteric blood vessels or portal veins were pressed by the tumor of pancreatic head, were operated using 3 vascular blocking bands and 2 patients using 4 vascular blocking bands, followed by suturing the bleeding points of the superior mesenteric vein with 5-0 vascular suture Proline. One patient's superior mesenteric vein was partially resected and restored. The operations cost 5-8 h each and the blood loss was 200-600 ml. There were no operative or postoperative hemorrhage or pancreatic juice leakage. According to the follow-up up to now, 2 patients died of multiple live tumor metastases 7 and 9 months separately after operation, the other 54 patients were still alive. Thin slice scan and vessel-reconstruction using multidetector spiral CT can accurately judge whether the blood vessels near the pancreatic tumor were invaded and whether the tumor was resectable, using 3 vascular blocking bands or 4 vascular blocking bands and cross, traction of the superior mesenteric blood vessels, operator can easily accomplish the radical pancreaticoduodenectomy of malignant tumor in pancreatic head with pressed superior mesenteric blood vessels and portal vein, which was not resectable or need combined resection of the blood vessels in the traditional opinion.

Imaging diagnosis of two unusual forms of gallstone ileus

ileus caused by gallstone has become more and more common recently due to popular usage of computed tomography (CT) and magnetic resonance imaging (MRI). In the past, imaging diagnosis relied primarily on conventional radiography and contrast-enhanced fluoroscopy. 1 CT has subsequently been shown to be quite useful for this diagnosis. 2,3 Recently, MR cholangiopancreatography (MRCP) has emerged as a valuable noninvasive technique for evaluation of biliary disease. Here we described the CT, MRCP and gastroenterography findings in two patients with unusual form of gallstone ileus (GI), one is Bouveret's syndrome and the other is negative gallstone-induced ileus. Few reports about GI with angiografin and MRCP findings of Bouveret's syndrome has been published. 3-8

Percutaneous the Diagnostic Value of Spiral CT in Gallbladder Carcinoma(40 cases report)

目的 探讨螺旋CT扫描对胆囊癌的诊断价值.方法对上腹部行螺旋CT扫描、胆囊区薄层扫描及胆囊MPR(多平面重建)重建,分析胆囊癌CT表现.结果本文40例胆囊癌的诊断与手术病理或细胞学都符合,病例类型以腺癌最为常见,占70%.结论CT扫描对胆囊癌的诊断较容易,但对壁厚型胆囊癌有一定的局限性,且早期胆囊癌诊断率不高,须结合临床表现及CT表现来进行胆囊癌与胆囊炎,胆囊癌与肝癌的鉴别诊断。

False positive reduction for lung nodule CAD using support vector machines and genetic algorithms

In this paper, we propose a machine learning approach to reduce false positive lung nodules identified in multi-slice CT scans by CAD algorithms. From a pool of features computed from the thin-slice scans, a genetic algorithm is used to determine an optimal feature subset for training a classifier that will eliminate as many of the false positives as possible while retaining the true nodules. We use support vector machines as the classifier for its superior performance. The experiment was conducted on a database of 66 true nodules and 123 false ones. From 15 features calculated for each nodule our approach selected 9 as the optimal feature subset size and the resulting classifier trained with a selected set of 9 features was able to achieve 98.5% sensitivity and 82.9% specificity using leave-one-out cross validation.

Automated lung segmentation in X-ray computed tomography: development and evaluation of a heuristic threshold-based scheme 1

Rationale and Objectives. To develop and evaluate a reliable, fully-automated lung segmentation scheme for application in X-ray computed tomography. Materials and Methods. The automated scheme was heuristically developed using a slice-based, pixel-value threshold and two sets of classification rules. Features used in the rules include size, circularity, and location. The segmentation scheme operates slice-by-slice and performs three key operations: (1) image preprocessing to remove background pixels, (2) computation and application of a pixel-value threshold to identify lung tissue, and (3) refinement of the initial segmented regions to prune incorrectly detected airways and separate fused right and left lungs. Results. The performance of the automated segmentation scheme was evaluated using 101 computed tomography cases (91 thick slice, 10 thin slice scans). The 91 thick cases were pre- and post-surgery from 50 patients and were not independent. The automated scheme successfully segmented 94.0% of the 2,969 thick slice images and 97.6% of the 1,161 thin slice images. The mean difference of the total lung volumes calculated by the automated scheme and functional residual capacity plus 60% inspiratory capacity was −24.7 ± 508.1 mL. The mean differences of the total lung volumes calculated by the automated scheme and an established, commonly used semi-automated scheme were 95.2 ± 52.5 mL and −27.7 ± 66.9 mL for the thick and thin slice cases, respectively. Conclusion. This simple, fully-automated lung segmentation scheme provides an objective tool to facilitate lung segmentation from computed tomography scans.

Helical CT reconstruction with longitudinal filtration.

Each of the currently used helical reconstruction algorithms represents one tradeoff selection of the image noise or mA requirement and resultant artifacts versus the width of the section profile. For many applications, the performance tradeoffs different from those existed are desired. In this paper, we present a general framework of incorporating the longitudinal (z) filtration in helical CT reconstruction, which leads to a new class of helical reconstruction algorithms with tunable performance characteristics to suit various applications. The z-filtering algorithm enables us to derive the performance curve of image noise or mA requirement versus the width of the section profile, whereas the existing helical reconstruction algorithms only represent limited discrete samples of the curve. We demonstrate that using a new helical reconstruction algorithm derived from this general framework, a substantial reduction in image noise or mA requirement can be achieved with a slight increase of FWHM of the section profile. Thus for many applications requiring both compact section profile and low image noise or mA requirement, the new algorithm represents a better tradeoff than both the 180 degrees and 360 degrees interpolation algorithms. The new algorithm also offers the following potential clinical benefits: (1) longer x-ray tube life; (2) longer scan duration with less x-ray tube cooling delays--longer volume coverage and/or longer thin-slice scan; (3) less radiation to patient; (4) less image noise; and (5) less helical artifacts. Initial clinical results are also presented.