Nodule Detection Performance Research Articles

High performance lung nodule detection schemes in CT using local and global information

A key issue in current computer-aided diagnostic (CAD) schemes for nodule detection in CT is the large number of false positives, because current schemes use only global three-dimensional (3D) information to detect nodules and discard useful local two-dimensional (2D) information. Thus, the authors integrated local and global information to markedly improve the performance levels of CAD schemes. Our database was obtained from the standard CT lung nodule database created by the Lung Image Database Consortium (LIDC). It consisted of 85 CT scans with 111 nodules of 3 mm or larger in diameter. The 111 nodules were confirmed by at least two of the four radiologists participated in the LIDC. Twenty-six nodules were missed by two of the four radiologists and were thus very difficult to detect. The authors developed five CAD schemes for nodule detection in CT using global 3D information (3D scheme), local 2D information (2D scheme), and both local and global information (2D + 3D scheme, 2D - 3D scheme, and 3D - 2D scheme). The 3D scheme, which was developed previously, used only global 3D information and discarded local 2D information, as other CAD schemes did. The 2D scheme used a uniform viewpoint reformation technique to decompose a 3D nodule candidate into a set of 2D reformatted images generated from representative viewpoints, and selected and used "effective" 2D reformatted images to remove false positives. The 2D + 3D scheme, 2D - 3D scheme, and 3D - 2D scheme used complementary local and global information in different ways to further improve the performance of lung nodule detection. The authors employed a leave-one-scan-out testing method for evaluation of the performance levels of the five CAD schemes. At the sensitivities of 85%, 80%, and 75%, the existing 3D scheme reported 17.3, 7.4, and 2.8 false positives per scan, respectively; the 2D scheme improved the detection performance and reduced the numbers of false positives to 7.6, 2.5, and 1.3 per scan; the 2D + 3D scheme further reduced those to 2.7, 1.9, and 0.6 per scan; the 2D - 3D scheme reduced those to 7.6, 2.1, and 0.8 per scan; and the 3D - 2D scheme reduced those to 17.3, 1.6, and 1.0 per scan. The local 2D information appears to be more useful than the global 3D information for nodule detection, particularly, when it is integrated with 3D information.

The Effects of Ambient Lighting in Chest Radiology Reading Rooms

Under typical dark chest radiography reading room conditions, a radiologist's pupils contract and dilate as their visual focus intermittently shifts between the high luminance monitor and the darker background wall, resulting in increased visual fatigue and degradation of diagnostic performance. A controlled increase of ambient lighting may minimize these visual adjustments and potentially improve comfort and accuracy. This study was designed to determine the effect of a controlled increase of ambient lighting on chest radiologist nodule detection performance. Four chest radiologists read 100 radiographs (50 normal and 50 containing a subtle nodule) under low (E=1 lx) and elevated (E=50 lx) ambient lighting levels on a DICOM-calibrated, medical-grade liquid crystal display. Radiologists were asked to identify nodule locations and rate their detection confidence. A receiver operating characteristic (ROC) analysis of radiologist results was performed and area under ROC curve (AUC) values calculated for each ambient lighting level. Additionally, radiologist selection times under both illuminance conditions were determined. Average AUC values did not significantly differ (p>0.05) between ambient lighting levels (estimated mean difference=-0.03; 95% CI, (-0.08, 0.03)). Average selection times decreased or remained constant with increased illuminance. The most considerable decreases occurred for false positive identification times (35.4±18.8 to 26.2±14.9 s) and true positive identification times (29.7±18.3 to 24.5±15.5 s). No performance differences were statistically significant. Study findings suggest that a controlled increase of ambient lighting within darkly lit chest radiology reading rooms, to a level more suitable for performance of common radiological tasks, does not appear to have a statistically significant effect on nodule detection performance.

Influence of liquid crystal displays (LCDs) with different resolutions on the detection of pulmonary nodules: An observer performance study

Purpose To evaluate the influence of monochrome LCDs with different resolutions on the detection performance of pulmonary nodules. Materials and methods 187 chest DR images were selected from our hospital's picture archiving and communication system (PACS), including 111 normal cases and 76 cases with solitary pulmonary nodules. Those positive images were divided into two groups, A and B, according to the diameter of nodules. Three high-, three mid-, and three low-experienced radiologists interpreted those images with a 2 megapixel (MP), a 3 MP and a 5 MP monochrome LCD independently. Regarding the presence of the nodule, five-point confidence level rating scale was used: definite absence, probable absence, indetermination, probable presence and definite presence. The observers were requested to rank each image on the given display according to the presence of pulmonary nodule. Receiver operation characteristic (ROC) analysis was used to interpret the data. Results When detecting the pulmonary nodules of Group A on 2 MP-, 3 MP- and 5 MP-LCD, the areas under ROC curves (AUCs) were 0.641, 0.683 and 0.732 for high-experienced radiologists, 0.633, 0.679 and 0.716 for mid-experienced radiologists, 0.620, 0.652 and 0.719 for low-experienced radiologists, respectively; when detecting the pulmonary nodules of Group B, the AUCs were 0.811, 0.830 and 0.842 for high-experienced radiologists, 0.771, 0.821 and 0.837 for mid-experienced radiologists, 0.759, 0.770 and 0.829 for low-experienced radiologists, respectively. However, there were no significant differences among different reading modalities. Conclusion For detecting pulmonary nodules, it is equivalent of observer performance among 2 MP-, 3 MP- and 5 MP-LCD.

Compare Display Schemes for Lung Nodule CT Screening

This study investigated the relative efficiencies of a stereographic display and two monoscopic display schemes for detecting lung nodules in chest computed tomography (CT). The ultimate goal was to determine whether stereoscopic display provides advantages for visualization and interpretation of three-dimensional (3D) medical image datasets. A retrospective study that compared lung nodule detection performances achieved using three different schemes for displaying 3D CT data was conducted. The display modes included slice-by-slice, orthogonal maximum intensity projection (MIP), and stereoscopic display. One hundred lung-cancer screening CT examinations containing 647 nodules were interpreted by eight radiologists, in each of the display modes. Reading times and displayed slab thickness versus time were recorded, as well as the probability, location, and size for each detected nodule. Nodule detection performance was analyzed using the receiver operating characteristic method. The stereo display mode provided higher detection performance with a shorter interpretation time, as compared to the other display modes tested in the study, although the difference was not statistically significant. The analysis also showed that there was no difference in the patterns of displayed slab thickness versus time between the stereo and MIP display modes. Most radiologists preferred reading the 3D data at a slab thickness that corresponded to five CT slices. Our results indicate that stereo display has the potential to improve radiologists' performance for detecting lung nodules in CT datasets. The experience gained in conducting the study also strongly suggests that further benefits can be achieved through providing readers with additional functionality.

Computer-Aided Pulmonary Nodule Detection – Performance of Two CAD Systems at Different CT Dose Levels

To evaluate the impact of dose reduction on the performance of computer-aided lung nodule detection systems (CAD) of two manufacturers by comparing respective CAD results on ultra-low-dose computed tomography (ULD-CT) and standard dose CT (SD-CT). Multi-slice computed tomography (MSCT) data sets of 26 patients (13 male and 13 female, patients 31 - 74 years old) were retrospectively selected for CAD analysis. Indication for CT examination was staging of a known primary malignancy or suspected pulmonary malignancy. CT images were consecutively acquired at 5 mAs (ULD-CT) and 75 mAs (SD-CT) with 120 kV tube voltage (1 mm slice thickness). The standard of reference was determined by three experienced readers in consensus. CAD reading algorithms (pre-commercial CAD system, Philips, Netherlands: CAD-1; LungCARE, Siemens, Germany: CAD-2) were applied to the CT data sets. Consensus reading identified 253 nodules on SD-CT and ULD-CT. Nodules ranged in diameter between 2 and 41 mm (mean diameter 4.8 mm). Detection rates were recorded with 72 % and 62 % (CAD-1 vs. CAD-2) for SD-CT and with 73 % and 56 % for ULD-CT. Median false positive rates per patient were calculated with 6 and 5 (CAD-1 vs. CAD-2) for SD-CT and with 8 and 3 for ULD-CT. After separate statistical analysis of nodules with diameters of 5 mm and greater, the detection rates increased to 83 % and 61 % for SD-CT and to 89 % and 67 % for ULD-CT (CAD-1 vs. CAD-2). For both CAD systems there were no significant differences between the detection rates for standard and ultra-low-dose data sets (p > 0.05). Dose reduction of the underlying CT scan did not significantly influence nodule detection performance of the tested CAD systems.

Assessment of Radiologist Performance in the Detection of Lung Nodules: Dependence on the Definition of “Truth”

Studies that evaluate the lung nodule detection performance of radiologists or computerized methods depend on an initial inventory of the nodules within the thoracic images (the "truth"). The purpose of this study was to analyze (1) variability in the "truth" defined by different combinations of experienced thoracic radiologists and (2) variability in the performance of other experienced thoracic radiologists based on these definitions of "truth" in the context of lung nodule detection in computed tomographic (CT) scans. Twenty-five thoracic CT scans were reviewed by four thoracic radiologists, who independently marked lesions they considered to be nodules >or=3 mm in maximum diameter. Panel "truth" sets of nodules were then derived from the nodules marked by different combinations of two and three of these four radiologists. The nodule detection performance of the other radiologists was evaluated based on these panel "truth" sets. The number of "true" nodules in the different panel "truth" sets ranged from 15 to 89 (mean 49.8 +/- 25.6). The mean radiologist nodule detection sensitivities across radiologists and panel "truth" sets for different panel "truth" conditions ranged from 51.0 to 83.2%; mean false-positive rates ranged from 0.33 to 1.39 per case. Substantial variability exists across radiologists in the task of lung nodule identification in CT scans. The definition of "truth" on which lung nodule detection studies are based must be carefully considered, because even experienced thoracic radiologists may not perform well when measured against the "truth" established by other experienced thoracic radiologists.

Characterization of Radiologists’ Search Strategies for Lung Nodule Detection: Slice-Based Versus Volumetric Displays

The goal of this study was to assess whether radiologists' search paths for lung nodule detection in chest computed tomography (CT) between different rendering and display schemes have reliable properties that can be exploited as an indicator of ergonomic efficiency for the purpose of comparing different display paradigms. Eight radiologists retrospectively viewed 30 lung cancer screening CT exams, containing a total of 91 nodules, in each of three display modes [i.e., slice-by-slice, orthogonal maximum intensity projection (MIP) and stereoscopic] for the purpose of detecting and classifying lung nodules. Radiologists' search patterns in the axial direction were recorded and analyzed along with the location, size, and shape for each detected feature, and the likelihood that the feature is an actual nodule. Nodule detection performance was analyzed by employing free-response receiver operating characteristic methods. Search paths were clearly different between slice-by-slice displays and volumetric displays but, aside from training and novelty effects, not between MIP and stereographic displays. Novelty and training effects were associated with the stereographic display mode, as evidenced by differences between the beginning and end of the study. The stereo display provided higher detection and classification performance with less interpretation time compared to other display modes tested in the study; however, the differences were not statistically significant. Our preliminary results indicate a potential role for the use of radiologists' search paths in evaluating the relative ergonomic efficiencies of different display paradigms, but systematic training and practice is necessary to eliminate training curve and novelty effects before search strategies can be meaningfully compared.

Computer-aided Detection of Solid Lung Nodules in Lossy Compressed Multidetector Computed Tomography Chest Exams

To assess the effect of three-dimensional (3D) lossy image compression of multidetector computed tomography chest scans on computer-aided detection (CAD) of solid lung nodules greater than 4 mm in size. A total of 120 cases, acquired with 1.25-mm collimation, were collected from 5 different sites, of which 66/120 were low-dose cases. Two chest radiologists established that 37 cases had no actionable lung nodules; the remaining 83 cases contained 169 nodules (range 3.8-35.0 mm, mean 5.8 mm +/- 3.0 [SD]). All cases were compressed using the 3D Set Partitioning in Hierarchical Trees algorithm to 24:1, 48:1, and 96:1 levels. A study of the effect of compression on computer-aided detection (CAD) sensitivity was performed at operating points of 2.5 false marks (FM), 5 FM, and 10 FM per case using McNemar's test. Logistic regression models were used to evaluate the impact on CAD sensitivity by compression level on nodule and image characteristics. Compared with no compression, there was no significant degradation in CAD sensitivity found at any of the studied compression levels and operating points. However, between compression levels, there was marginal association with sensitivity. Specifically, 24:1 level was significantly better than 96:1 at all operating points, and occasionally better than no compression at 10 FM/case. Based on multivariate analysis, nodule location was found to be a significant predictor (P = .01) with a lower sensitivity associated with juxtapleural nodules. Nodule size, dose, reconstruction filter, and contrast medium were not significant predictors. CAD detection performance of solid lung nodules did not suffer until 48:1 compression.

Influence of rib structure on detection of subtle lung nodules

Purpose To access the influence of anatomic noise on the detectability of subtle lung nodules depicted on chest radiographs. Material and methods From normal chest radiography images, 132 square regions were extracted, of which the centers were on the upper margin of a rib, the inside of a rib, the lower margin of a rib, and the central region between two adjoining ribs. Simulated nodules were digitally superimposed at the centers of these extracted square images. Twelve radiologists viewed 50 soft-copy images consisting of these 792 processed images, including the noise-added images. The observer's confidence level for the square images containing single nodules was used as an index of observer performance. Results Results indicated statistically reliable effects of the relationship between rib structures and nodule positions on the detection performance ( P < 0.001). The nodule detectability on the images with a center located between two adjoining ribs was significantly the best, whereas it was significantly the worst on the noise-added images with a center located between two adjoining ribs. Conclusion The rib structures overlying a subtle lung nodule on chest X-ray images have a detrimental effect on nodule detection performance as anatomic noise, regardless of the nodule location on ribs.

The influence of liquid crystal display (LCD) monitors on observer performance for the detection of nodular lesions on chest radiographs

To access the influence of liquid crystal display (LCD) monitors on the detectability of nodular lesions depicted on chest radiographs by comparing them with a high-resolution cathode ray tube (CRT) monitor. Ten radiologists interpreted 247 soft-copy images on LCD monitors with pixel arrays of 1,024x1,280, 1,200x1,600, 1,536x2,048 and 2,048x2,560, and a CRT monitor with a pixel array of 2,048x2,560, and were asked to indicate their individual confidence levels regarding the presence of a nodule. These images were chest radiographs with and without a lung nodule from the "Standard Digital Image Database" created by the Japanese Society of Radiological Technology. The luminance distributions of all monitors were adjusted to the same, and the ambient illumination was 200 lux. Observer performance was analyzed in terms of the receiver-operating characteristics. No significant statistical differences in nodule detection performance were found among the four LCD monitors and the CRT monitor. The nodule detection performance on the LCD monitors with a spatial resolution higher than a matrix size of 1,024x1,280 was found to be equivalent to that on the high-resolution CRT monitor.

Computed-aided diagnosis system for detection of liver tumor in CT liver image

AIM: To evaluate prospectively the influence of pulmonary nodule characteristics on detection performances of a computer-aided diagnosis (CAD) tool and experienced chest radiologists using multislice CT (MSCT).MATERIALS AND METHODS: MSCT scans of 20 consecutive patients were evaluated by a CAD system and two independent chest radiologists for presence of pulmonary nodules. Nodule size, position, margin, matrix characteristics, vascular and pleural attachments and reader confidence were recorded and data compared with an independent standard of reference. Statistical analysis for predictors influencing nodule detection or reader performance included chi-squared, retrograde stepwise conditional logistic regression with odds ratios and nodule detection proportion estimates (DPE), and ROC analysis.RESULTS: For 135 nodules, detection rates for CAD and readers were 76.3, 52.6 and 52.6%, respectively; false-positive rates were 0.55, 0.25 and 0.15 per examination, respectively. In consensus with CAD the reader detection rate increased to 93.3%, and the false-positive rate dropped to 0.1/scan. DPEs for nodules ≤5 mm were significantly higher for ICAD than for the readers (p<0.05). Absence of vascular attachment was the only significant predictor of nodule detection by CAD (p=0.0006–0.008). There were no predictors of nodule detection for reader consensus with CAD. In contrast, vascular attachment predicted nodule detection by the readers (p=0.0001–0.003). Reader sensitivity was higher for nodules with vascular attachment than for unattached nodules (sensitivities 0.768 and 0.369; 95% confidence intervals=0.651–0.861 and 0.253–0.498, respectively).CONCLUSION: CAD increases nodule detection rates, decreases false-positive rates and compensates for deficient reader performance in detection of smallest lesions and of nodules without vascular attachment.

Computer-aided detection of pulmonary nodules: influence of nodule characteristics on detection performance

AIM: To evaluate prospectively the influence of pulmonary nodule characteristics on detection performances of a computer-aided diagnosis (CAD) tool and experienced chest radiologists using multislice CT (MSCT). MATERIALS AND METHODS: MSCT scans of 20 consecutive patients were evaluated by a CAD system and two independent chest radiologists for presence of pulmonary nodules. Nodule size, position, margin, matrix characteristics, vascular and pleural attachments and reader confidence were recorded and data compared with an independent standard of reference. Statistical analysis for predictors influencing nodule detection or reader performance included chi-squared, retrograde stepwise conditional logistic regression with odds ratios and nodule detection proportion estimates (DPE), and ROC analysis. RESULTS: For 135 nodules, detection rates for CAD and readers were 76.3, 52.6 and 52.6%, respectively; false-positive rates were 0.55, 0.25 and 0.15 per examination, respectively. In consensus with CAD the reader detection rate increased to 93.3%, and the false-positive rate dropped to 0.1/scan. DPEs for nodules ≤5 mm were significantly higher for ICAD than for the readers ( p<0.05). Absence of vascular attachment was the only significant predictor of nodule detection by CAD ( p=0.0006–0.008). There were no predictors of nodule detection for reader consensus with CAD. In contrast, vascular attachment predicted nodule detection by the readers ( p=0.0001–0.003). Reader sensitivity was higher for nodules with vascular attachment than for unattached nodules (sensitivities 0.768 and 0.369; 95% confidence intervals=0.651–0.861 and 0.253–0.498, respectively). CONCLUSION: CAD increases nodule detection rates, decreases false-positive rates and compensates for deficient reader performance in detection of smallest lesions and of nodules without vascular attachment.

Influence of monitor luminance change on observer performance for detection of abnormalities depicted on chest radiographs.

To investigate how changes in luminance affect the detection accuracy of radiologists viewing chest radiograph images on high-resolution CRT monitors. Thirteen radiologists performed a detection task for 11 chest radiograph images with simulated nodules on a monitor with 11 luminance conditions (the maximum luminance ranges from 157.4-369.0 candela/m2) simulating CRT degraded by long-term usage, under the ambient illumination of 200 lux; the observation order was always from the darkest to the brightest. There was a statistically reliable effect of the 11 monitor display conditions on the detection of nodules (P < 0.001). In the conditions in which the maximum luminance of the CRT was 60.7% or below that of the standard display luminance, the correctly detected nodule number reliably deteriorated. The luminance change in CRT monitor display under long-term usage will have a detrimental effect on nodule detection performance in chest radiograph images.

A Perceptually Based Method for Enhancing Pulmonary Nodule Recognition

Feedback of chest areas receiving prolonged gaze durations significantly increases nodule detection performance. Why feedback circling enhances performance when other cueing methods produce equivocal results was assessed. Chest and noise images with nodule targets were used to determine: what type of cue is most effective; whether circling influences the way the eye samples the target; whether circling limits processing of distracting information outside its boundary. Circling improves performance more than cues with less complete boundaries and increases the accuracy and frequency with which nodules are fixated. Outside distractors were detected less often with than without the circle present. Circling isolates the abnormal region from the rest of the image, making disembedding and integration of nodule features more likely and insulates this region from distractors. The facilitative effects of circling are generalizable to other images in which low contrast targets are embedded in noisy backgrounds.