Lung cancer screening with CT: evaluation of radiologists and different computer assisted detection software (CAD) as first and second readers for lung nodule detection at different dose levels

OBJECTIVES To find the best pairing of first and second reader at highest sensitivity for detecting lung nodules with CT at various dose levels. MATERIALS AND METHODS An anthropomorphic lung phantom and artificial lung nodules were used to simulate screening CT-examination at standard dose (100 mAs, 120 kVp) and 8 different low dose levels, using 120, 100 and 80 kVp combined with 100, 50 and 25 mAs. At each dose level 40 phantoms were randomly filled with 75 solid and 25 ground glass nodules (5-12 mm). Two radiologists and 3 different computer aided detection softwares (CAD) were paired to find the highest sensitivity. RESULTS Sensitivities at standard dose were 92%, 90%, 84%, 79% and 73% for reader 1, 2, CAD1, CAD2, CAD3, respectively. Combined sensitivity for human readers 1 and 2 improved to 97%, (p1=0.063, p2=0.016). Highest sensitivities--between 97% and 99.0%--were achieved by combining any radiologist with any CAD at any dose level. Combining any two CADs, sensitivities between 85% and 88% were significantly lower than for radiologists combined with CAD (p<0.03). CONCLUSIONS Combination of a human observer with any of the tested CAD systems provide optimal sensitivity for lung nodule detection even at reduced dose at 25 mAs/80 kVp.

Response to neoadjuvant treatment of invasive ductal breast carcinomas including outcome evaluation: MRI analysis by an automatic CAD system in comparison to visual evaluation

BACKGROUND The aim of this study was to evaluate imaging-based response to standardized neoadjuvant chemotherapy (NACT) regimen by dynamic contrast-enhanced magnetic resonance mammography (DCE-MRM), whereas MR images were analyzed by an automatic computer-assisted diagnosis (CAD) system in comparison to visual evaluation. MRI findings were correlated with histopathologic response to NACT and also with the occurrence of metastases in a follow-up analysis. PATIENTS AND METHODS Fifty-four patients with invasive ductal breast carcinomas received two identical MRI examinations (before and after NACT; 1.5T, contrast medium gadoteric acid). Pre-therapeutic images were compared with post-therapeutic examinations by CAD and two blinded human observers, considering morphologic and dynamic MRI parameters as well as tumor size measurements. Imaging-assessed response to NACT was compared with histopathologically verified response. All clinical, histopathologic, and DCE-MRM parameters were correlated with the occurrence of distant metastases. RESULTS Initial and post-initial dynamic parameters significantly changed between pre- and post-therapeutic DCE-MRM. Visually evaluated DCE-MRM revealed sensitivity of 85.7%, specificity of 91.7%, and diagnostic accuracy of 87.0% in evaluating the response to NACT compared to histopathology. CAD analysis led to more false-negative findings (37.0%) compared to visual evaluation (11.1%), resulting in sensitivity of 52.4%, specificity of 100.0%, and diagnostic accuracy of 63.0%. The following dynamic MRI parameters showed significant associations to occurring metastases: Post-initial curve type before NACT (entire lesions, calculated by CAD) and post-initial curve type of the most enhancing tumor parts after NACT (calculated by CAD and manually). CONCLUSIONS In the accurate evaluation of response to neoadjuvant treatment, CAD systems can provide useful additional information due to the high specificity; however, they cannot replace visual imaging evaluation. Besides traditional prognostic factors, contrast medium-induced dynamic MRI parameters reveal significant associations to patient outcome, i.e. occurrence of distant metastases.

The performance of computer-aided detection when analyzing prior mammograms of newly detected breast cancers with special focus on the time interval from initial imaging to detection

PURPOSE: The clinical role of CAD systems to detect breast cancer, which have not been on cancer containing mammograms not detected by the radiologist was proven retrospectively. METHODS: All patients from 1992 to 2005 with a histologically verified malignant breast lesion and a mammogram at our department, were analyzed in retrospect focussing on the time of detection of the malignant lesion. All prior mammograms were analyzed by CAD (CADx, USA). The resulting CAD printout was matched with the cancer containing images yielding to the radiological diagnosis of breast cancer. CAD performance, sensitivity as well as the association of CAD and radiological features were analyzed. RESULTS: 278 mammograms fulfilled the inclusion criteria. 111 cases showed a retrospectively visible lesion (71 masses, 23 single microcalcification clusters, 16 masses with microcalcifications, in one case two microcalcification clusters). 54/87 masses and 34/41 microcalcifications were detected by CAD. Detection rates varied from 9/20 (ACR 1) to 5/7 (ACR 4) (45% vs. 71%). The detection of microcalcifications was not influenced by breast tissue density. CONCLUSION: CAD might be useful in an earlier detection of subtle breast cancer cases, which might remain otherwise undetected.

Classification of interstitial lung disease patterns using local DCT features and random forest.

Over the last decade, a plethora of computer-aided diagnosis (CAD) systems have been proposed aiming to improve the accuracy of the physicians in the diagnosis of interstitial lung diseases (ILD). In this study, we propose a scheme for the classification of HRCT image patches with ILD abnormalities as a basic component towards the quantification of the various ILD patterns in the lung. The feature extraction method relies on local spectral analysis using a DCT-based filter bank. After convolving the image with the filter bank, q-quantiles are computed for describing the distribution of local frequencies that characterize image texture. Then, the gray-level histogram values of the original image are added forming the final feature vector. The classification of the already described patches is done by a random forest (RF) classifier. The experimental results prove the superior performance and efficiency of the proposed approach compared against the state-of-the-art.

Performance of different polishing techniques for direct CAD/CAM ceramic restorations

This in vitro study evaluated the performance of three ceramic and two commonly used polishing methods on two CAD/CAM ceramics. Surface roughness and quality were compared. A glazed group (GLGR) of each ceramic material served as reference. One-hundred and twenty specimens of VITABLOCS Mark II (VITA) and 120 specimens of IPS Empress CAD (IPS) were roughened in a standardized manner. Twenty VITA and 20 IPS specimens were glazed (VITA Akzent Glaze/Empress Universal Glaze). Five polishing methods were investigated (n=20/group): 1) EVE Diacera W11DC-Set (EVE), 2) JOTA 9812-Set (JOTA), 3) OptraFine-System (OFI), 4) Sof-Lex 2382 discs (SOF) and 5) Brownie/Greenie/Occlubrush (BGO). Polishing quality was measured with a surface roughness meter (Ra and Rz values). The significance level was set at alpha=0.05. Kruskal Wallis tests and pairwise Wilcoxon rank sum tests with Bonferroni-Holm adjustment were used. Qualitative surface evaluation of representative specimens was done with SEM. On VITA ceramics, SOF produced lower Ra (p<0.00001) but higher Rz values than GLGR (p=0.003); EVE, JOTA, OFI and BGO yielded significantly higher Ra and Rz values than GLGR. On IPS ceramics, SOF and JOTA exhibited lower Ra values than GLGR (p<0.0001). Equivalent Ra but significantly higher Rz values occurred between GLGR and EVE, OFI or BGO. VITA and IPS exhibited the smoothest surfaces when polished with SOF. Nevertheless, ceramic polishing systems are still of interest to clinicians using CAD/CAM, as these methods are universally applicable and showed an increased durability compared to the investigated silicon polishers.

A systematic review of the clinical performance of CAD/CAM single-tooth restorations

PURPOSE: This systematic review sought to determine the long-term clinical survival rates of single-tooth restorations fabricated with computer-aided design/computer-assisted manufacture (CAD/CAM) technology, as well as the frequency of failures depending on the CAD/CAM system, the type of restoration, the selected material, and the luting agent. MATERIALS AND METHODS: An electronic search from 1985 to 2007 was performed using two databases: Medline/PubMed and Embase. Selected keywords and well-defined inclusion and exclusion criteria guided the search. All articles were first reviewed by title, then by abstract, and subsequently by a full text reading. Data were assessed and extracted by two independent examiners. The pooled results were statistically analyzed and the overall failure rate was calculated by assuming a Poisson-distributed number of events. In addition, reported failures were analyzed by CAD/CAM system, type of restoration, restorative material, and luting agent. RESULTS: From a total of 1,957 single-tooth restorations with a mean exposure time of 7.9 years and 170 failures, the failure rate was 1.75% per year, estimated per 100 restoration years (95% CI: 1.22% to 2.52%). The estimated total survival rate after 5 years of 91.6% (95% CI: 88.2% to 94.1%) was based on random-effects Poisson regression analysis. CONCLUSIONS: Long-term survival rates for CAD/CAM single-tooth Cerec 1, Cerec 2, and Celay restorations appear to be similar to conventional ones. No clinical studies or randomized clinical trials reporting on other CAD/CAM systems currently used in clinical practice and with follow-up reports of 3 or more years were found at the time of the search.

Surface Roughness and Microhardness of Two Recent CAD/CAM-Materials After Storage

Objectives: To investigate surface roughness and microhardness of two recent resin-ceramic materials for computer-aided design/computer-aided manufacturing (CAD/CAM) after polishing with three polishing systems. Surface roughness and microhardness were measured immediately after polishing and after six months storage including monthly artificial toothbrushing. Methods: Sixty specimens of Lava Ultimate (3M ESPE) and 60 specimens of VITA ENAMIC (VITA Zahnfabrik) were roughened in a standardized manner and polished with one of three polishing systems (n=20/group): Sof-Lex XT discs (SOFLEX; three-step (medium-superfine); 3M ESPE), VITA Polishing Set Clinical (VITA; two-step; VITA Zahnfabrik), or KENDA Unicus (KENDA; one-step; KENDA Dental). Surface roughness (Ra; μm) was measured with a profilometer and microhardness (Vickers; VHN) with a surface hardness indentation device. Ra and VHN were measured immediately after polishing and after six months storage (tap water, 37°C) including monthly artificial toothbrushing (500 cycles/month, toothpaste RDA ~70). Ra- and VHN-values were analysed with nonparametric ANOVA followed by Wilcoxon rank sum tests (α=0.05). Results: For Lava Ultimate, Ra (mean [standard deviation] before/after storage) remained the same when polished with SOFLEX (0.18 [0.09]/0.19 [0.10]; p=0.18), increased significantly with VITA (1.10 [0.44]/1.27 [0.39]; p=0.0001), and decreased significantly with KENDA (0.35 [0.07]/0.33 [0.08]; p=0.03). VHN (mean [standard deviation] before/after storage) decreased significantly regardless of polishing system (SOFLEX: 134.1 [5.6]/116.4 [3.6], VITA: 138.2 [10.5]/115.4 [5.9], KENDA: 135.1 [6.2]/116.7 [6.3]; all p<0.0001). For VITA ENAMIC, Ra (mean [standard deviation] before/after storage) increased significantly when polished with SOFLEX (0.37 [0.18]/0.41 [0.14]; p=0.01) and remained the same with VITA (1.32 [0.37]/1.31 [0.40]; p=0.58) and with KENDA (0.81 [0.35]/0.78 [0.32]; p=0.21). VHN (mean [standard deviation] before/after storage) remained the same regardless of polishing system (SOFLEX: 284.9 [24.6]/282.4 [31.8], VITA: 284.6 [28.5]/276.4 [25.8], KENDA: 292.6 [26.9]/282.9 [24.3]; p=0.42-1.00). Conclusion: Surface roughness and microhardness of Lava Ultimate was more affected by storage and artificial toothbrushing than was VITA ENAMIC.