Small peripheral lung carcinomas with five-year post-surgical follow-up: assessment by semi-automated volumetric measurement of tumour size, CT value and growth rate on TSCT.

OBJECTIVES: To retrospectively assess the utility of semi-automated measurements by stratification of CT values of tumour size, CT value and doubling time (DT) using thin-section computed tomography (CT) images. The post-surgical outcomes of favourable and problematic tumours (more advanced p stage than IA, post-surgical recurrence or mortality from lung cancer) were compared using the measured values. The computed DTs were compared with manually measured values. METHODS: The study subjects comprised 85 patients (aged 33-80 years, 48 women, 37 men), followed-up for more than 5 years postoperatively, with 89 lung lesions, including 17 atypical adenomatous hyperplasias and 72 lung cancers. DTs were determined in 45 lesions. RESULTS: For problematic lesions, whole tumour diameter and density were >18 mm and >-400 HU, respectively. The respective values for the tumour core (with CT values of -350 to 150 HU) were >15 mm and >-70 HU. Analysis of tumour core DTs showed interval tumour progression even if little progress was seen by standard tumour volume DT (TVDT). CONCLUSION: Software-based volumetric measurements by stratification of CT values provide valuable information on tumour core and help estimate tumour aggressiveness and interval tumour progression better than standard manually measured 2D-VDTs.

Identification of occult parechymal disease such as emphysema or airway disease using screening computed tomography.

RATIONALE: Chronic obstructive pulmonary disease (COPD) is a major public health problem. This study was performed to determine whether the low attenuation area (LAA) and visual score provided by low-dose computed tomography (CT) can be used to detect occult parenchymal disease, such as insidious COPD. METHODS: Each participant underwent low-dose CT scan and pulmonary function tests. The LAA% of the corresponding lung area was calculated. The cut-off level between the normal lung density area and LAA was defined as -960 HU, and the severity of emphysematous change (visual score) and LAA% were evaluated on three same chest CT slices obtained at full inspiration. RESULTS: Forty-eight of 2,247 individuals including 1058 non-smokers and 1189 smokers were diagnosed with COPD. Chest CT findings in individuals diagnosed with COPD showed centrilobular emphysema (50%), however, 17 of the subjects diagnosed with COPD had normal screening CT findings. Thirty-one subjects diagnosed with COPD showed a positive visual score, and 27 individuals with COPD showed LAA% of more than 30. Nine of 17 subjects with a negative visual score showed LAA% of more than 30. The visual score in smokers was significantly higher than that of non-smokers. The lung function in smokers was lower than that of non-smokers. Smokers also showed higher frequencies of chest CT abnormalities. CONCLUSION: Low-dose CT scans detected LAA and a positive visual score before COPD associated with an impaired lung function develops. Smokers with normal spirometry had a potential to develop an airflow obstruction accompanied with abnormal CT findings.

Radiological diagnosis of small pulmonary nodules detected on low-dose screening computed tomography.

BACKGROUND AND OBJECTIVE: Early detection and treatment of small malignant pulmonary lesions can improve survival; however, screening by CT detects many false positives. This study retrospectively evaluated a protocol for the diagnostic work-up of nodules detected by low-dose CT (LDCT) that are < or = 10 mm in diameter. METHODS: A health screening programme included LDCT. Lesions detected were allocated to one of four categories: negative, semi-negative, positive and semi-positive. Positive and semi-positive categories included non-calcified nodules without a polygonal shape, and these patients had an initial diagnostic HRCT and were then followed up using high-resolution CT (HRCT) at intervals determined by the characteristics of the lesion on screening LDCT and the initial diagnostic HRCT. RESULTS: There were 275 nodules detected on screening LDCT; 84 patients had lesions classified as positive and 99 as semi-positive. Thirteen nodules detected on screening LDCT were only determined to be polygonal and benign following the diagnostic HRCT. The sensitivity and specificity of the screening CT, when compared with diagnostic HRCT, for determining if nodules should be classified as positive were 100% and 97%. The sensitivity and specificity of the initial diagnostic HRCT for being able to predict lung cancer were 87.5% and 91.7% respectively. CONCLUSIONS: Following the detection of a pulmonary lesion on screening LDCT, a diagnostic HRCT is necessary to determine the timing of follow-up HRCT. Diagnostic HRCT is needed to rationalize the screening for lung cancer to reduce the frequency of unnecessary follow-up scans.

Long-term follow-up study of a population-based 1996-1998 mass screening programme for lung cancer using mobile low-dose spiral computed tomography.

Early diagnosis and treatment are important for improvement of the low survival rate of patients with lung cancer. The objective of this study was to evaluate the long-term survival rate of patients identified to have lung cancer by our population-based baseline and annual repeat low-radiation dose computed tomography (low-dose CT) screenings, conducted in 1996-1998. A total of 13,037 CT scans were obtained from 5480 subjects (2969 men, 2511 women) aged 40-74 years at the initial CT screening. Lung cancer was detected in 63 subjects (57 were detected by CT scans and underwent surgery; 1 was detected by sputum cytology and underwent surgery; 3 rejected treatment; and 2 were interval cases that developed symptoms prior to the next annual repeat CT screening). Follow-up study included review of medical records. Death certificates were examined to check for any deceased interval case among participants. Postoperative follow-up of the 50 survived patients ranged from 70 to 117 (median, 101) months. Eight patients died during follow-up (6 due to lung cancer from 20 to 67 months after surgery and 2 deaths unrelated to lung cancer, each 7 and 60 months following surgery). Three patients who rejected treatment died 14 months to 6 years after positive screening CT scans, and the 2 interval cases died at each 17 and 30 months, respectively, following negative screening CT scans. Survival was analysed in 59 patients with lung cancer detected by low-dose CT screening (excluding two patients; one was detected by sputum cytology and the other had mass lesion already noted on the chest radiograph of the previous year). The 10-year survival calculated by the Kaplan-Meier method was 83.1% (95% CI: 0.735-0.927) for death from all causes and 86.2% (95% CI: 0.773-0.951) for death from lung cancer. The survival rate was excellent for never-smokers, patients with BAC and adenocarcinoma/mixed types with non-solid CT density pattern, associated with Noguchi's type A or B and pathologic stage IA. A poorer prognosis was noted in smokers with adenocarcinomas/mixed types, associated with part-solid or solid CT density pattern and Noguchi's type C or D. All patients with non-solid tumours measuring 6-13.5mm at presentation are alive, patients with part-solid tumours, measuring 17mm or more, or solid tumours, measuring 13mm or more at presentation were associated with increased risk of lung cancer-related morbidity or mortality. The estimated rate of possible over-diagnosis was 13% in total and we failed to cure 17% of patients encountered in the programme. Low-dose CT screening substantially improves the 10-year survival for lung cancer with minimal use of invasive treatment procedures.

CT findings of early-stage small cell lung cancer in a low-dose CT screening programme.

The survival of patients with small cell lung cancer (SCLC) is related to T, N, M components, and early diagnosis and treatment of limited stage SCLC may improve survival. The objective of this study was to review the initial and annual repeat screening computed tomography (CT) images of all five patients with SCLC, encountered in our 1996-1998 population-based screening for lung cancer, to clarify any subtle, characteristic CT findings of early-stage small cell lung cancer. The medical records of the patients were reviewed to examine demographic and clinical features. We identified characteristic CT features of SCLC in the lung periphery, which were related to gross pathologic findings with longitudinal spread along the bronchial wall: a small spindle-shaped or pyramidal lesion was found as a subtle CT finding of SCLC, and irregularly shaped nodular lesions (vermiform, pine-cone-like or tandem-like nodular lesions) appeared at a more advanced stage. Tumour volume doubling time of the cases ranged from 38 days to 217 days. All five patients were male smokers: four patients underwent surgery and adjuvant chemotherapy; three of them remain alive, while the remaining patient, an interval case, died of lung cancer. One patient refused treatment and died of a cause other than lung cancer. Annual repeat CT screening was useful for detecting SCLC cases mostly at a curable stage, and information about CT features, presented here, should help physicians identify SCLC at an earlier-stage and lead to a more successful treatment of the disease.

Presence of local pleural adhesion in CT screening-detected small nodule in the lung periphery suggests noncancerous, inflammatory nature of the lesion.

PURPOSE: Differential diagnosis of small nodules in the lung periphery detected by low-dose chest CT screening is important before surgery. The aim of the study was to discriminate between benign and malignant lesions, identified in our preoperative imaging work-up examinations and confirmed during surgery, for nodules detected on CT screening. MATERIALS AND METHODS: This study is based on 106 patients (46 men and 60 women, median age: 61.5 years) with 123 CT screening-detected and histologically confirmed nodules smaller than 30 mm in the lung periphery identified between 2002 and 2005 at Azumi General Hospital, Japan. Lesions were classified into three groups according to histological findings: adenocarcinoma, atypical adenomatous hyperplasia (AAH) and inflammatory focal lesions. We examined the visceral pleura during surgery at a location close to lung nodules. RESULTS: The median diameter of resected lung nodules on high-resolution CT (HRCT) was 9.0 mm. Nodules were nonsolid in 42, partly solid in 51 and solid in 30. Histopathological diagnosis was lung cancer in 69, AAH in 21, other noninflammatory tumours in 6 and inflammatory lesions in 27. Fifty-four lesions were located in the subpleural zone. Eight of 123 nodules showed local pleural adhesions (LPA), while 2 were buried in extensive pleural adhesion. LPA was noted more frequently in inflammatory nodules than in cancer nodules (P<.01). CONCLUSION: The presence of LPA in close proximity to a small nodule is indicative of noncancerous lesion. This feature allows the discrimination of pulmonary peripheral inflammatory lesion from peripheral small cancer on chest low-dose CT screening.

[An octogenarian case of postoperative recurrent lung cancer responding to treatments with radiation plus gefitinib].

An 82-year-old woman, a never smoker, had a radical operation for CT screening revealed lung cancer in an other hospital in 1997. She was admitted to our hospital complaining of dry cough and dyspnea on effort in March 2004. She was diagnosed to have a local recurrence of lung cancer 6 years after the operation. After she underwent radiotherapy of the mediastinum (total 60 Gy) and daily administration of gefitinib for two weeks, the administration of gefitinib was continued every other day in the outpatient clinic. During follow-up, CYFRA gradually increased to 3.8 ng/mL, but then decreased to the normal range. The tumor response rate of metastasized lymph nodes of bronchial bifurcation reached 36%, and it was confirmed to be a partial response. Without harmful phenomena except skin eruptions, her quality of life was good with a performance status (PS) 0 at 85 years 4 months of age, 9 years 2 months after the resection, with 2 years 5 months of gefitinib administration. It will be useful as a treatment option for octogenarians having postoperative recurrent lung cancers with every other day administration of gefitinib.

A method for evaluating the performance of computer-aided detection of pulmonary nodules in lung cancer CT screening: detection limit for nodule size and density.

OBJECTIVE: We propose the application of virtual nodules to evaluate the performance of computer-aided detection (CAD) of lung nodules in cancer screening using low-dose CT. METHODS: The virtual nodules were generated based on the spatial resolution measured for a CT system used in an institution providing cancer screening and were fused into clinical lung images obtained at that institution, allowing site specificity. First, we validated virtual nodules as an alternative to artificial nodules inserted into a phantom. In addition, we compared the results of CAD analysis between the real nodules (n = 6) and the corresponding virtual nodules. Subsequently, virtual nodules of various sizes and contrasts between nodule density and background density (ΔCT) were inserted into clinical images (n = 10) and submitted for CAD analysis. RESULTS: In the validation study, 46 of 48 virtual nodules had the same CAD results as artificial nodules (kappa coefficient = 0.913). Real nodules and the corresponding virtual nodules showed the same CAD results. The detection limits of the tested CAD system were determined in terms of size and density of peripheral lung nodules; we demonstrated that a nodule with a 5-mm diameter was detected when the nodule had a ΔCT > 220 HU. CONCLUSION: Virtual nodules are effective in evaluating CAD performance using site-specific scan/reconstruction conditions. Advances in knowledge: Virtual nodules can be an effective means of evaluating site-specific CAD performance. The methodology for guiding the detection limit for nodule size/density might be a useful evaluation strategy.

Comparison of bronchoscopic diagnosis for peripheral pulmonary nodule under fluoroscopic guidance with CT guidance.

BACKGROUND: A new diagnostic procedure has been established for the selection of appropriate therapy for small lung lesions. We compared the sensitivity of real-time multi-slice computed tomography (MSCT) and X-ray television (TV) fluoroscopic guidance for performing bronchoscopy. METHODS: The first author performed and interpreted all bronchoscopies described in this study. The diagnosis of malignancy or benign was based on the results of histopathological examination, as well as clinical and imaging follow-up MSCT. We also compared the diagnostic yields of lesion size between MSCT and X-ray TV fluoroscopic guidance. RESULTS: Real-time MSCT and X-ray TV fluoroscopic guidance was conducted in 82 and 78 patients, respectively. The lesion size detected by real-time MSCT and X-ray TV fluoroscopic guidance was <10 mm (n=21, 14), 11-15 mm (n=24, 12), 16-20 mm (n=19,14), 21-25 mm (n=9, 12) and >26 mm (n=9, 26). The sensitivity of real-time MSCT- and X-ray TV fluoroscopic guidance was 62.2% and 52.6%, respectively. The sensitivity of real-time MSCT fluoroscopic guidance for histopathologic diagnosis of lesions less than 15 mm was higher than that of X-ray TV fluoroscopic guidance. While it was difficult to histopathologically diagnose small lung lesions less than 10mm in diameter, real-time MSCT fluoroscopic guidance offers a better chance of such diagnosis, irrespective of the size of the lesion, compared with X-ray TV fluoroscopic guidance. CONCLUSION: Real-time MSCT fluoroscopic guidance allows the bronchoscopist to accurately determine the location of the instruments in relation to the lesion in real time, thus helping to reduce the number of negative cases.

Clinical differences in the Global Initiative for Chronic Obstructive Lung Disease Stage 0.

This study was to examine the clinical differences between Stage 0 and normal subjects, using low-dose chest computed tomography (CT) and pulmonary function tests. Enrolled subjects performed as a health check for lung cancer screening including low-dose CT and pulmonary function tests. Subjects were divided into Stage 0, chronic obstructive pulmonary disease according to pulmonary function tests, and normal subjects. The severity of emphysema (visual score) was calculated on three low-dose CT slices. Low-dose CT and pulmonary function tests were performed in 1359 men and 888 women. The numbers and percentages of men and women smokers were 1076 (79.2%), and 107 (12.0%), respectively. A total of 722 individuals had one or more respiratory symptoms, such as cough (69.8%), sputum (75.8%), or shortness of breathing (0.83%). Of the 722 subjects, 71 (9.8%) individuals satisfied the criteria of chronic respiratory symptoms. Among the normal subjects, smoking caused differences in airflow limitation as a result of pulmonary function tests. The proportion of smokers and the visual score were significantly higher in Stage 0 than those in the normal subjects. The percentages of the maximal mid-expiratory flow (%MMF) and of the peak expiratory flow rate were significantly lower in Stage 0 than in the normal subjects. %MMF and the proportion of visual score were significantly lower in the smoking Stage 0 than in the nonsmoking Stage 0 subjects. Smoking would indicate early signs of emphysematous change between Stage 0 and normal subjects in comparison of pulmonary function tests and visual score of low-dose CT.

Improving radiologists' recommendations with computer-aided diagnosis for management of small nodules detected by CT.

RATIONALE AND OBJECTIVES: To evaluate how computer-aided diagnosis (CAD) can improve radiologists' recommendations for management of possible early lung cancers on CT. MATERIALS AND METHODS: Twenty-eight lung cancers and 28 benign lesions were employed. Each group of 28 lesions was classified into subgroups of two sizes (9 between 6 and 10 mm and 19 between 11 and 20 mm) and three patterns (8 with pure ground glass opacity [GGO], 12 with mixed GGO and 8 solid lesions). Sixteen radiologists participated in the observer study, first without and then with CAD. Radiologists' recommendations, including (1) follow-up in 12 months, (2) in 6 months, (3) in 3 months, or (4) biopsy, were compared at three levels of their malignancy probability ratings (low: 1%-33%; medium: 34%-66%; high: 67%-99%) for 896 observations (56 lesions by the 16 radiologists) in the two size subgroups and three patterns. RESULTS: The number of recommendations changed by radiologists by use of CAD was 163 (18%) among all 896 observations. Among these changed recommendations, the fraction showing a beneficial effect from CAD was 68% (111/163), and the fraction showing a beneficial effect regarding biopsy recommendations was 69% (48/70). With CAD, the radiologists' performance regarding biopsy recommendations was significantly improved for 43 lung cancers (31 changed to biopsy versus 12 changed away from biopsy; P = .003) and was also improved for 27 benign lesions (10 changed to biopsy versus 17 changed away from biopsy; P = .18). Most of the cancers with improved recommendations were solid lesions or mixed GGO and relatively large. CONCLUSION: CAD has the potential to improve the appropriateness of radiologists' recommendations for small malignant and benign lesions on CT scans.

Computer-aided diagnostic scheme for distinction between benign and malignant nodules in thoracic low-dose CT by use of massive training artificial neural network.

Low-dose helical computed tomography (LDCT) is being applied as a modality for lung cancer screening. It may be difficult, however, for radiologists to distinguish malignant from benign nodules in LDCT. Our purpose in this study was to develop a computer-aided diagnostic (CAD) scheme for distinction between benign and malignant nodules in LDCT scans by use of a massive training artificial neural network (MTANN). The MTANN is a trainable, highly nonlinear filter based on an artificial neural network. To distinguish malignant nodules from six different types of benign nodules, we developed multiple MTANNs (multi-MTANN) consisting of six expert MTANNs that are arranged in parallel. Each of the MTANNs was trained by use of input CT images and teaching images containing the estimate of the distribution for the "likelihood of being a malignant nodule," i.e., the teaching image for a malignant nodule contains a two-dimensional Gaussian distribution and that for a benign nodule contains zero. Each MTANN was trained independently with ten typical malignant nodules and ten benign nodules from each of the six types. The outputs of the six MTANNs were combined by use of an integration ANN such that the six types of benign nodules could be distinguished from malignant nodules. After training of the integration ANN, our scheme provided a value related to the "likelihood of malignancy" of a nodule, i.e., a higher value indicates a malignant nodule, and a lower value indicates a benign nodule. Our database consisted of 76 primary lung cancers in 73 patients and 413 benign nodules in 342 patients, which were obtained from a lung cancer screening program on 7847 screenees with LDCT for three years in Nagano, Japan. The performance of our scheme for distinction between benign and malignant nodules was evaluated by use of receiver operating characteristic (ROC) analysis. Our scheme achieved an Az (area under the ROC curve) value of 0.882 in a round-robin test. Our scheme correctly identified 100% (76/76) of malignant nodules as malignant, whereas 48% (200/413) of benign nodules were identified correctly as benign. Therefore, our scheme may be useful in assisting radiologists in the diagnosis of lung nodules in LDCT.

Evaluation of automated lung nodule detection on low-dose computed tomography scans from a lung cancer screening program(1).

RATIONALE AND OBJECTIVES: The purpose of this study was to evaluate the performance of a fully automated lung nodule detection method in a large database of low-dose computed tomography (CT) scans from a lung cancer screening program. Because nodules demonstrate a spectrum of radiologic appearances, the performance of the automated method was evaluated on the basis of nodule malignancy status, size, subtlety, and radiographic opacity. MATERIALS AND METHODS: A database of 393 thick-section (10 mm) low-dose CT scans was collected. Automated lung nodule detection proceeds in two phases: gray-level thresholding for the initial identification of nodule candidates, followed by the application of a rule-based classifier and linear discriminant analysis to distinguish between candidates that correspond to actual lung nodules and candidates that correspond to non-nodules. Free-response receiver operating characteristic analysis was used to evaluate the performance of the method based on a jackknife training/testing approach. RESULTS: An overall nodule detection sensitivity of 70% (330 of 470) was attained with an average of 1.6 false-positive detections per section. At the same false-positive rate, 83% (57 of 69) of the malignant lung nodules in the database were detected. When the method was trained specifically for malignant nodules, a sensitivity of 80% (55 of 69) was attained with 0.85 false-positives per section. CONCLUSION: We have evaluated an automated lung nodule detection method with a large number of low-dose CT scans from a lung cancer screening program. An overall sensitivity of 80% for malignant nodules was achieved with 0.85 false-positive detections per section. Such a computerized lung nodule detection method is expected to become an important part of CT-based lung cancer screening programs.

Subcentimeter large cell neuroendocrine carcinoma of the lung.

To our knowledge, no report exists of a subcentimeter size large cell neuroendocrine carcinoma (LCNEC) of the lung. A 75-year-old man participating in a low-dose CT screening program for lung cancer was found incidentally to have a partly-solid nodule in the right upper lung. After treatment with antibiotics, a repeat CT showed resolution of the nodule, but a new solid nodule measuring 9 x 9 mm was detected in the left lower lobe. The lesion showed marked enhancement on dynamic contrast-enhanced MRI. Video-assisted thoracic surgery and frozen section biopsy was suggestive of malignant lesion, resulting in extension of surgery to lobectomy with nodal dissection. The final diagnosis was stage IA-LCNEC. The estimated volume doubling time of the tumor was 30.1 days. These aggressive tumors may rarely have doubling times that overlap with benign processes.

Prognostic significance of high-resolution CT findings in small peripheral adenocarcinoma of the lung: a retrospective study on 64 patients.

OBJECTIVE: We studied the prognostic importance of high-resolution CT (HRCT) findings in lung adenocarcinomas. PATIENTS AND METHODS: HRCT findings (lesion size, percentage of ground-glass opacity (GGO) areas of lesion, and presence or absence of lobulation, coarse spiculation, air space, pleural tag, and multiplicity of lesion), clinical data (age and surgical method), and pathologic findings (tumor subtypes and presence or absence of nodal metastasis) in 64 consecutive patients with 64 peripheral adenocarcinomas of 20 mm or less (mean, 13 mm), including 36 women and 28 men with a mean age of 64 years were analyzed and correlated with survival of the patients using Kaplan-Meier method and stepwise Cox proportional hazards modeling. Follow-up periods of the patients ranged from 6 to 45 months (mean, 22 months). Tumors were classified into six subtypes (types A-F) according to tumor growth patterns defined by Noguchi et al. RESULTS: Six (9%) of the 64 patients died of lung cancer. In univariate analyses, a significant difference was noted for lesion size (P=0.043), the percentage of GGO areas (P=0.005), and tumor subtypes (P=0.006). Lesion size of <15 mm (n=35), a lesion with GGO areas of >57% (n=36), and type A (n=16) or type B adenocarcinomas (n=16) indicated a significantly better survival. In multivariate analyses using these three parameters as independent variables, the percentage of GGO areas was the only significant independent factor for survival (P=0.044, relative risk=0.95). CONCLUSION: GGO areas measured on HRCT may have an independent prognostic significance of small adenocarcinomas of the lung.

Discrimination of subtypes of small adenocarcinoma in the lung with thin-section CT.

We studied the usefulness of thin-section CT in discriminating two categories of adenocarcinoma in the lung. Thin-section CT findings, such as, lesion size, ground-glass opacity (GGO) areas of lesion and presence or absence of lobulation, coarse spiculation, air bronchogram, small air space, or pleural tag of lesion in 62 consecutive patients with 62 adenocarcinomas (35 type A or B tumors (Noguchi's classification) and 27 type C tumors) of < or =20 mm, including 36 women and 26 men with a mean age of 64 years were analyzed. We performed stepwise logistic modeling using all the CT findings as independent variables to estimate the significant factors for discriminating type C from type A or B tumor. Lesion size in type C tumors was significantly (P<0.001) greater than that in type A or B tumors. GGO areas in type C tumors were significantly (P<0.001) smaller than that in type A or B tumors. The prevalence of coarse spiculation, air bronchogram, and pleural tag in type C tumors was significantly greater (P=0.001, 0.010, and <0.001, respectively) than that in type A or B tumors. Logistic modeling revealed that the GGO area was the only significant factor for discriminating two categories (P<0.001). Using the percentage of GGO areas for predicting type C tumor, 40% or less showed the highest accuracy of 85% with 70% sensitivity and 97% specificity. GGO areas of 30% or less had no false-positive diagnosis (100% specificity) with 81% accuracy but its sensitivity was low (56%). Thin-section CT was useful in discriminating two categories of adenocarcinoma in the lung.

Alveolar shrinkage in bronchioloalveolar carcinoma without central fibrosis.

Preservation of alveolar architecture in small bronchioloalveolar carcinoma (BAC) is one of the most important factors for predicting their prognosis, alveolar shrinkage in BAC, however, has not been studied well. In ten cases of pure BAC without collapse or central fibrosis, we measured two two-dimensional (2D) parameters; side-length of the alveoli facing alveolar ducts and circumference of the alveoli forming complete circles. And we also examined three-dimensionally (3D) elastic fibers and myofibroblasts in the thick sections. In BAC, 2D parameters showed alveolar shrinkage. The elastic fibers forming the alveolar framework, including the alveolar orifice, were 3D sinuous and bent in BAC, and suggested that alveolar shrinkage cased by folding of the alveolar wall. Myofibroblasts lay transversely and longitudinally in the interstitium in BAC, intertwined with the elastic fibers. Proliferation of myofibroblasts may be of importance in alveolar wall-folding and alveolar shrinkage in BAC.

Selective enhancement filters for nodules, vessels, and airway walls in two- and three-dimensional CT scans.

Computer-aided diagnostic (CAD) schemes have been developed to assist radiologists in the early detection of lung cancer in radiographs and computed tomography (CT) images. In order to improve sensitivity for nodule detection, many researchers have employed a filter as a preprocessing step for enhancement of nodules. However, these filters enhance not only nodules, but also other anatomic structures such as ribs, blood vessels, and airway walls. Therefore, nodules are often detected together with a large number of false positives caused by these normal anatomic structures. In this study, we developed three selective enhancement filters for dot, line, and plane which can simultaneously enhance objects of a specific shape (for example, dot-like nodules) and suppress objects of other shapes (for example, line-like vessels). Therefore, as preprocessing steps, these filters would be useful for improving the sensitivity of nodule detection and for reducing the number of false positives. We applied our enhancement filters to synthesized images to demonstrate that they can selectively enhance a specific shape and suppress other shapes. We also applied our enhancement filters to real two-dimensional (2D) and three-dimensional (3D) CT images to show their effectiveness in the enhancement of specific objects in real medical images. We believe that the three enhancement filters developed in this study would be useful in the computerized detection of cancer in 2D and 3D medical images.

Massive training artificial neural network (MTANN) for reduction of false positives in computerized detection of lung nodules in low-dose computed tomography.

In this study, we investigated a pattern-recognition technique based on an artificial neural network (ANN), which is called a massive training artificial neural network (MTANN), for reduction of false positives in computerized detection of lung nodules in low-dose computed tomography (CT) images. The MTANN consists of a modified multilayer ANN, which is capable of operating on image data directly. The MTANN is trained by use of a large number of subregions extracted from input images together with the teacher images containing the distribution for the "likelihood of being a nodule." The output image is obtained by scanning an input image with the MTANN. The distinction between a nodule and a non-nodule is made by use of a score which is defined from the output image of the trained MTANN. In order to eliminate various types of non-nodules, we extended the capability of a single MTANN, and developed a multiple MTANN (Multi-MTANN). The Multi-MTANN consists of plural MTANNs that are arranged in parallel. Each MTANN is trained by using the same nodules, but with a different type of non-nodule. Each MTANN acts as an expert for a specific type of non-nodule, e.g., five different MTANNs were trained to distinguish nodules from various-sized vessels; four other MTANNs were applied to eliminate some other opacities. The outputs of the MTANNs were combined by using the logical AND operation such that each of the trained MTANNs eliminated none of the nodules, but removed the specific type of non-nodule with which the MTANN was trained, and thus removed various types of non-nodules. The Multi-MTANN consisting of nine MTANNs was trained with 10 typical nodules and 10 non-nodules representing each of nine different non-nodule types (90 training non-nodules overall) in a training set. The trained Multi-MTANN was applied to the reduction of false positives reported by our current computerized scheme for lung nodule detection based on a database of 63 low-dose CT scans (1765 sections), which contained 71 confirmed nodules including 66 biopsy-confirmed primary cancers, from a lung cancer screening program. The Multi-MTANN was applied to 58 true positives (nodules from 54 patients) and 1726 false positives (non-nodules) reported by our current scheme in a validation test; these were different from the training set. The results indicated that 83% (1424/1726) of non-nodules were removed with a reduction of one true positive (nodule), i.e., a classification sensitivity of 98.3% (57 of 58 nodules). By using the Multi-MTANN, the false-positive rate of our current scheme was improved from 0.98 to 0.18 false positives per section (from 27.4 to 4.8 per patient) at an overall sensitivity of 80.3% (57/71).

Computerized scheme for determination of the likelihood measure of malignancy for pulmonary nodules on low-dose CT images.

An automated computerized scheme has been developed for determination of the likelihood measure of malignancy of pulmonary nodules on low-dose helical CT (LDCT) images. Our database consisted of 76 primary lung cancers (147 slices) and 413 benign nodules (576 slices). With this automated computerized scheme, the location of a nodule was first indicated by a radiologist. The outline of the nodule was segmented automatically by use of a dynamic programming technique. Various objective features on the nodules were determined by use of outline analysis and image analysis, and the likelihood measure of malignancy was determined by use of linear discriminant analysis (LDA). The effect of many different combinations of features and the performance of LDA in distinguishing benign nodules from malignant ones were evaluated by means of receiver operating characteristic (ROC) analysis. The Az value (area under the ROC curve) obtained by the computerized scheme in distinguishing benign nodules from malignant ones was 0.828 when a single slice was employed for each of the nodules. However, the Az value was improved to 0.846 when multiple slices were used for determination of the likelihood measure of malignancy. The Az values obtained by the computerized scheme on LDCT images were significantly greater than the Az value of 0.70, which was obtained from our previous observer studies by radiologists in distinguishing benign nodules from malignant ones on LDCT images. The automated computerized scheme for determination of the likelihood measure of malignancy would be useful in assisting radiologists to distinguish between benign and malignant pulmonary nodules on LDCT images.