Shih-Chung Benedict Lo

Lung Nodules: Improved Detection with Software That Suppresses the Rib and Clavicle on Chest Radiographs

PURPOSE To demonstrate possible superiority in the performance of a radiologist who is tasked with detecting actionable nodules and aided by the bone suppression and soft-tissue visualization algorithm of a new software program that produces a modified image by suppressing the ribs and clavicles, filtering noise, and equalizing the contrast in the area of the lungs. MATERIALS AND METHODS The study and use of anonymized and deidentified data received approval from the MedStar-Georgetown University Oncology Institutional Review Board. Informed consent was obtained from 15 study radiologists. The study radiologists participated as observers in a reader study of 368 patients in an approximately 2:1 cancer-free-to-cancer ratio. The localized receiver operating characteristic (LROC) method was used for analyses. Images were rerandomized for each radiologist. Each patient image was sequentially read, first with the standard radiograph and then with the software-aided image. Normal studies were confirmed with computed tomography (CT), follow-up, and/or panel consensus. RESULTS Each reader and the combined scores of the 15 readers showed improvement. The area under the combined LROC curve increased significantly from 0.460 unaided to 0.558 aided by visualization software (P = .0001). When measured according to the readers indication that a case should be sent or not sent for CT or biopsy, sensitivity for cancer detection increased from 49.5% unaided to 66.3% aided by software (P < .0001); specificity decreased from 96.1% to 91.8% (P = .004). Seventy-four percent of the aided detections occurred in cancers with 70% or greater overlap of the bone and the nodule. CONCLUSION The radiologists using visualization software significantly increased their detection of lung cancers and benign nodules.

Potential for unnecessary patient exposure from the use of storage phosphor imaging systems

It is possible to use storage phosphor radiography (SR) devices in a manner that results in excess exposure to the patient without the operators knowledge. Because these SR systems have an automatic correction for the final optical density (OD) of the image, the technologist and radiologist will not be able to use excessive blackness of the image as a sign of overexposure. Tests reported here demonstrate that it is possible to obtain images of a chest phantom that appear acceptable with a 32 times difference in exposure (maximal exposure .86 R). It is possible to obtain exposures of a pelvis phantom that appear acceptable up to the tube limit of our machine (4.8 R). Tests of the Fuji AC-1 demonstrate that it will accept a much wider range of exposures than the AGFA ADC prototype which permits only a 7 times difference in exposure before the image is degraded.

Convolution neural-network-based detection of lung structures

Chest radiography is one of the most primary and widely used techniques in diagnostic imaging. Nowadays with the advent of digital radiology, the digital medical image processing techniques for digital chest radiographs have attracted considerable attention, and several studies on the computer-aided diagnosis (CADx) as well as on the conventional image processing techniques for chest radiographs have been reported. In the automatic diagnostic process for chest radiographs, it is important to outline the areas of the lungs, the heart, and the diaphragm. This is because the original chest radiograph is composed of important anatomic structures and, without knowing exact positions of the organs, the automatic diagnosis may result in unexpected detections. The automatic extraction of an anatomical structure from digital chest radiographs can be a useful tool for (1) the evaluation of heart size, (2) automatic detection of interstitial lung diseases, (3) automatic detection of lung nodules, and (4) data compression, etc. Based on the clearly defined boundaries of heart area, rib spaces, rib positions, and rib cage extracted, one should be able to use this information to facilitate the tasks of the CADx on chest radiographs. In this paper, we present an automatic scheme for the detection of lung field from chest radiographs by using a shift-invariant convolution neural network. A novel algorithm for smoothing boundaries of lungs is also presented.

Classification of lung nodules in diagnostic CT: an approach based on 3D vascular features, nodule density distribution, and shape features

We have developed various segmentation and analysis methods for the quantification of lung nodules in thoracic CT. Our methods include the enhancement of lung structures followed by a series of segmentation methods to extract the nodule and to form 3D configuration at an area of interest. The vascular index, aspect ratio, circularity, irregularity, extent, compactness, and convexity were also computed as shape features for quantifying the nodule boundary. The density distribution of the nodule was modeled based on its internal homogeneity and/or heterogeneity. We also used several density related features including entropy, difference entropy as well as other first and second order moments. We have collected 48 cases of lung nodules scanned by thin-slice diagnostic CT. Of these cases, 24 are benign and 24 are malignant. A jackknife experiment was performed using a standard back-propagation neural network as the classifier. The LABROC result showed that the Az of this preliminary study is 0.89.

Phase contrast digital mammography using molybdenum x-ray: clinical implications in detectability improvement

We have applied phase imaging on digital mammography to investigate adequate contrast of printed images for digital phase contrast mammography using a practical molybdenum X-ray tube. Phase contrast mammography procedures were performed with defined air gap (e.g., 0.6 m) configuration using customized mammography equipment and a computed radiography (CR) system. Magnified (x2) phase contrast images acquired with 0.0875mm per pixel were mapped onto the laser imager resolution at 0.04375mm per pixel for printing life-size object on wet processing silver halide recording film. For contact mammography of screen-film system, we used MinR2000 system as a baseline method. ACR 156 phantom printed images with contrasts of 2.8, 3.7, 4.9, 5.7 and 6.7 were evaluated by five radiologists. The ACR scores for the life-size image based on the 2 times magnified phase contrast image acquired by the computed radiography were higher than the scores of MinR2000 image, when the contrast of printed images for both methods was 3.7. The ACR scores were lower in the low contrast images (i.e., 2.8) than its higher contrast counterparts (i.e., >= 3.7) for all techniques used. The detectability improvement should be due to higher spatial resolution and lower noise in the phase contrast images.

Mammographic mass detection by stochastic modeling and a multimodular neural network

In this paper, we have developed a combined method utilizing morphological operations, a finite generalized Gaussian mixture (FGGM) modeling, and a contextual Bayesian relaxation labeling technique (CBRL) to enhance and extract suspicious masses. A feature space is constructed based on multiple feature extraction from the regions of interest (ROIs). Finally, a multi-modular neural network (MMNN) is employed to distinguish true masses from non-masses. We have applied these methods to test our mammogram database. The true masses in the database were identified by a radiologist with biopsy reports. The results demonstrated that all the areas of suspicious masses in mammograms were extracted in the prescan step using the proposed segmentation procedure. We found that 6 - 15 suspected masses per mammogram were detected and required further evaluation. We also found that the MMNN can reduce the number of suspicious masses with a sensitivity of 84% at 1 - 2 false positive (FP) per mammogram based on the database containing 46 mammograms (23 of them have biopsy proven masses). In conclusion, the experimental results indicate that morphological filtering combined with FGGM model-based segmentation is an effective way to extract mammographic suspicious mass patterns. Compared with conventional neural networks, the probabilistic MMNN can lead to a more efficient learning algorithm and can provide more understanding in the analysis of the distribution patterns of multiple features extracted from the suspicious masses.

Classification of false-positive findings on computer-aided detection of breast microcalcifications

False positive detections of microcalcifications by computer aided diagnosis (CADx) systems are a distraction to the radiologist and raise questions as to the eventual clinical utility of computer aided diagnosis systems. We have carefully analyzed the mammographic findings that appear in the locations of CADx detections and have counted and classified them.

Enhanced lung cancer detection in temporal subtraction chest radiography using directional edge filtering techniques

We have developed a series of directional edge enhancement and edge extraction methods that can accurately segment posterior and anterior ribs in chest radiography. These methods can also separate the lower and upper edges of ribs. The edges were first enhanced by two sets of proximate parabola curve models for left and right sides of the image. We used a directional edge filtering technique to remove low signals and noises on the edge enhanced image in the multiresolution domain. Finally, we employed a rib curve projection and reasoning method to reconstruct the rib edges and remove false edges for the upper and lower bound of the rib edges independently. A two-step registration, corresponding to global and local matching, is applied for current and prior images assisted by their corresponding edge images. The subtraction images were then processed by a rule-based CAD system. The FROC results were compared to that obtained by the original image using a CAD system consisting of rule-based and convolution neural network processing. The majority of lung cancer in temporal subtraction images were lit-up. The FROC results were significantly improved using the subtraction image with the rule-based CAD.

Computer-aided detection of lung cancer on chest radiographs: algorithm performance vs. radiologists' performance by size of cancer

Our goal was to perform a pre-clinical test of the performance of a new pre-commercial system for detection of primary early-stage lung cancer on chest radiographs developed by Deus Technologies, LLC. The RapidScreenTM RS 2000 System integrates state of the art technical development in this field.

Computerized detection and classification of microcalcifications on mammograms

We are developing computer-aided diagnosis algorithms to assist radiologists in detection and classification of microcalcifications on mammograms. A digitized mammogram was processed with a difference-image technique and signal segmentation methods to identify suspicious signals. False-positive detections were reduced by using morphological features as well as a convolution neural network. A regional clustering technique was applied to the remaining signals to identify clinically significant clustered microcalcifications. For the development of a malignant/benign classifier, the microcalcifications were extracted from the digital images by computerized segmentation techniques. A number of visibility descriptors and shape descriptors were developed to describe the features of the microcalcifications. Linear discriminant analysis and receiver operating characteristic (ROC) methodology were used to classify the benign and malignant microcalcifications. For detection of microcalcifications, the computer reached a true-positive (TP) rate of 100% at 0.1 false-positive (FP) clusters per image for obvious microcalcifications, a TP rate of 93% at 1 FP clusters per image for average subtle microcalcifications, and a TP rate of 87% at 1.5 FP clusters per image for very subtle microcalcifications. For classification of microcalcifications, preliminary results indicated that an area under the ROC curve (Az) of 0.91 and 0.89 could be achieved during training, and an Az of 0.82 and 0.87 during jackknife testing for obvious and subtle clusters, respectively. When all cases were combined, the Az was 0.87 and 0.84, respectively, for training and jackknife testing.