Search

To evaluate the performance of a deep learning-based computer-aided diagnosis (CAD) system at detecting pulmonary nodules on CT by comparing radiologists’ readings with and without CAD. A total of 120 chest CT images were randomly selected from patients with suspected lung cancer. The gold standard of nodules ≥ 3 mm was established by a panel of three expert radiologists. Two less experienced radiologists read the images without and afterward with CAD system. Their reading times were recorded. The radiologists’ sensitivity increased from 20.9% to 38.0% with the introduction of CAD. The positive predictive value (PPV) decreased from 70.5% to 61.8%, and the F1-score increased from 32.2% to 47.0%. The sensitivity significantly increased from 13.7% to 32.4% for small nodules (3–6 mm) and from 33.3% to 47.6% for medium nodules (6–10 mm). CAD alone showed a sensitivity of 70.3%, a PPV of 57.9%, and an F1-score of 63.5%. Reading time decreased by 11.3% with the use of CAD. CAD improved the less experienced radiologists’ sensitivity in detecting pulmonary nodules of all sizes, especially including a significant improvement in the detection of clinically important-sized medium nodules (6–10 mm) as well as small nodules (3–6 mm) and reduced their reading time.

Objective To report clinical features of bone metastases (BM) from head and neck squamous cell carcinoma (HNSCC). Methods Among 772 patients with HNSCC diagnosed at our hospital over 9 years, 30 patients (3.9%) had clinical evidence of BM (24 men and 6 women; mean age: 63 years). We assessed the time interval from the primary diagnosis to BM development, symptoms attributable to BM, presence of distant metastases to other organs, number of BM, sites of BM, morphologic changes on computed tomography (CT) images, treatment for BM, and overall survival (OS). Results BM at the initial stage were found in 9 patients with HNSCC (30%), and in 21 patients (70%) with HNSCC during the course of the disease. In the later patients, the median time interval from the primary diagnosis was 11.5 months. Nineteen patients (63%) did not have BM-related symptoms, 6 (20%) had pain, 3 (10%) had neurologic symptoms resulting from vertebral or skull metastases, and 2 (7%) had hypercalcemia. Seventeen patients (57%) showed bone-exclusive metastases, and 13 (43%) had distant metastases in other organs. Eleven patients (37%) had monostotic metastases (solitary BM), and 19 patients (63%) had polyostotic metastases (multiple BM). When combined, 9 patients (30%) showed bone-exclusive and monostotic metastases. The most commonly affected site was the thoracolumbar spine, accounting for 34% of total BM, followed by the pelvis (24%), shoulder and thorax (21%), and the extremities (17%). Notably, metastases to bones above the clavicle (craniofacial bones and cervical spine) accounted for only 3% of all bone lesions. CT images showed variable morphologic patterns with osteolytic type in 17 patients (57%), intertrabecular in 7 (23%), osteoblastic in 4 (13%), and mixed in 2 (7%). Systematic chemotherapy for BM was performed in 19 patients and radiotherapy in 18. The median survival time for patients with bone-exclusive and monostotic metastases was significantly longer than that for patients with multi-organ metastases or polyostotic metastases at 18.2 months vs. 5.7 months (p = 0.02). Neither chemotherapy nor radiotherapy extended OS. Conclusion Thirty percent of BM cases from HNSCC showed bone-exclusive and monostotic metastases. These patients tended to show a more favorable prognosis than patients with multi-organ metastases or polyostotic metastases.

Drug deposition in the peripheral airways has been reported to differ between pressurized metered dose inhalers (pMDIs) and dry powder inhalers (DPIs). This study was conducted to determine the drug deposition fraction from three inhaled corticosteroids/long-acting β2-agonists (ICS/LABA) combination products for inhalation in the peripheral airways. An airway deposition model was constructed using computed tomography (CT) images obtained in patients with asthma. Functional respiratory imaging was used to determine the airway deposition of Flutiform® (Kyorin Pharmaceutical), Symbicort® (AstraZeneca and Astellas Pharma), and Relvar® (GlaxoSmithKline K.K.) without actual drug inhalation based on airway images from CT and computational fluid dynamics using drug particle size. The mean drug deposition fractions and standard deviations (SDs) in the peripheral airways for Flutiform, Symbicort, and Relvar were 30.33% ± 1.89%, 18.95% ± 2.03%, and 6.58% ± 0.98% for ICS, and 29.17% ± 1.81%, 17.24% ± 1.87%, and 11.30% ± 1.42% for LABA, respectively, showing that significantly higher proportions of ICS and LABA were deposited to the peripheral airways from Flutiform than from Symbicort and Relvar (p

The aim of this study was to evaluate the decreasing of cardiac motion artifact and whether the extent of ground-glass attenuation of idiopathic pulmonary fibrosis (IPF) was accurately assessed by electrocardiography (ECG)-triggered high-resolution computed tomography (HRCT) by 0.5-s/rotation multidetector-row CT (MDCT).ECG-triggered HRCT were scanned at the end-diastolic phase by a MDCT scanner with the following scan parameters; axial four-slice mode, 0.5 mm collimation, 0.5-s/rotation, 120 kVp, 200 mA/rotation, high-frequency algorithm, and half reconstruction. In 42 patients with IPF, both conventional HRCT (ECG gating(-), full reconstruction) and ECG-triggered HRCT were performed at the same levels (10-mm intervals) with the above scan parameters. The correlation between percent diffusion of carbon monoxide of the lung (%DLCO) and the mean extent of ground-glass attenuation on both conventional HRCT and ECG-triggered HRCT was evaluated with the Spearman rank correlation coefficient test.The correlation between %DLCO and the mean extent of ground-glass attenuation on ECG-triggered HRCT (observer A: r = -0.790, P0.0001; observer B: r = -0.710, P0.0001) was superior to that on conventional HRCT (observer A: r = -0.395, P0.05; observer B: r = -0.577, P = 0.002) for both observers.ECG-triggered HRCT by 0.5 s/rotation MDCT can reduce the cardiac motion artifact and is useful for evaluating the extent of ground-glass attenuation of IPF.

OBJECTIVE. The purpose of this study was to evaluate the radiographic and CT findings of inhalational talc pneumoconiosis.CONCLUSION. Large opacities of talc pneumoconiosis progress more often than do small opacities. The CT findings of talc pneumoconiosis overlap those of silicosis and asbestosis.

To determine the effect of changing pitch and collimation on depiction of pulmonary metastases on scans of 5-mm section thickness obtained with multi-detector row computed tomography (CT) compared with those obtained with single-detector row CT.In five autopsy lungs, 1,013 metastatic 0.5-30.0-mm nodules were detected at helical CT with 1-mm collimation and histopathologically diagnosed as metastases. Each nodule was numbered, and its localization was recorded as the standard for subsequent studies. Four types of scans of 5-mm section thickness were obtained with multi-detector row CT and four sets of helical pitch and table speed, respectively, as follows: set A, 3:1 and 7.5 mm per rotation; set B, 6:1 and 15 mm per rotation; set C, 6:1 and 30 mm per rotation; set D, conventional and 5-mm interval. Conventional helical CT scans (set E) were obtained with 5-mm collimation at single-detector row CT. Two independent observers evaluated the five sets of CT scans.Acquisition times for sets A-D, respectively, were 1.9, 3.8, 7.5, and 1.5 times faster than they were for set E. The mean numbers of detected nodules were 671 (66%) in set A, 661 (65%) in set B, 678 (67%) in set C, 654 (65%) in set D, and 656 (65%) in set E; there was no significant difference in the number of detected nodules among the five sets (P =.997, McNemar test and Bonferroni equation).Regardless of varying pitch or detector collimation, multi- and single-detector row CT scans obtained with 5-mm section thickness have almost the same ability to depict pulmonary metastases and are equivalent.

The purpose of this study was to evaluate the accuracy of temporal subtraction with a commercially available computer-assisted diagnosis system for the detection of multifocal hazy pulmonary opacities on chest radiographs, which are sometimes difficult to detect directly on chest radiographs.Thirty healthy patients and 30 patients with new multifocal hazy pulmonary opacities that were confirmed by serial chest CT examinations were evaluated with and without temporal subtraction images. Chest radiographs were taken from either film-screen or digital radiography images and were digitized with a spatial resolution of 0.171 mm per pixel. Temporal subtraction images were produced by an iterative image-warping technique. We designed an observer performance study in which observers (six chest radiologists and four residents) indicated their confidence level for the presence or absence of hazy pulmonary opacities on two sets of images, current and previous radiographs only (set A), and current and previous radiographs with temporal subtraction images (set B). Receiver operating characteristic curves were generated.For chest radiologists, observer performance with set B (with temporal subtraction images; A(z) = 0.947) was superior to that with set A (without temporal subtraction images; A(z) = 0.916) (p0.05). For residents, no statistically significant difference was found between sets A and B.The temporal subtraction technique clearly improves diagnostic accuracy for the detection of multifocal hazy pulmonary opacities on chest radiographs, especially when the observers are experienced chest radiologists who have sufficient skill to evaluate the patient's condition as revealed on the images.

The purpose of our study was to correlate the high-resolution CT findings of small peripheral adenocarcinoma of the lung with underlying histopathology and to evaluate the prognostic implications of the CT findings.The high-resolution CT findings of small peripheral adenocarcinoma of the lung in 224 patients were analyzed by two independent observers for location, size, marginal characteristics, and extent of ground-glass opacity and necrosis. The pathologic specimens were reviewed by an experienced lung pathologist.One hundred and thirty-two patients had bronchioloalveolar carcinoma and 92 had adenocarcinoma. The extent of ground-glass opacity was greater in bronchioloalveolar carcinomas (mean +/- SD, 29% +/- 31.6%) than in other adenocarcinomas (8% +/- 13.3%) (p0.001). The extent of ground-glass opacity was significantly greater in patients without recurrence (p = 0.020) and those without nodal (p = 0.017) or distant (p = 0.007) metastases than in patients with nodal or distant metastases or in whom the carcinoma had recurred.The extent of ground-glass opacity in a nodule is greater in bronchioloalveolar carcinomas than in other adenocarcinomas. Greater extent of ground-glass opacity also correlates with improved prognosis.

The purpose of this study was to describe the high-resolution computed tomography (HRCT) findings of pulmonary involvement in primary Sjögren's syndrome. The study included 60 patients who met the diagnostic criteria for primary Sjögren's syndrome. The authors retrospectively reviewed the presence, extent, and distribution of various HRCT findings. Results showed that the most common HRCT findings were areas with ground-glass attenuation (92%), followed by subpleural small nodules (78%), non-septal linear opacity (75%), interlobular septal thickening (55%), bronchiectasis (38%), and cysts (30%).

In this paper, we present the results of time study of CAD (Computer-Aided Diagnosis) on temporal subtraction for ground-glass shadow on chest radiographs. Our results showed that the average reading time with the CAD system was significantly longer than without employing the CAD system. However, diagnostic accuracy with temporal subtraction images was slightly improved than without temporal subtraction (with; Az=0.96, without; Az=0.95). Consequently, we could interpret that the radiologist's performance was enhanced by CAD on temporal subtraction, even with longer reading time.

Purpose: The purpose of this work was to evaluate the CT features of thymoma and to determine the most helpful findings in differentiating invasive from noninvasive thymoma. Method: The CT scans from 27 patients with invasive thymoma and 23 with noninvasive thymoma were independently assessed by two observers without knowledge of their invasiveness. The presence and distribution of various CT findings were independently analyzed. Results: Invasive thymomas were more likely to have lobulated (16/27, 59%) or irregular (6/27, 22%) contours than noninvasive thymomas (8/23, 35% and 1.5/23, 6%, respectively) (p < 0.05). Invasive thymomas had a higher prevalence of low attenuation areas within the tumor (16/27, 60%) than noninvasive thymomas (5/23, 22%) (p < 0.001) as well as foci of calcification (14.5/27, 54% vs. 6/23, 26%; p < 0.01). Conclusion: The presence of lobulated or irregular contour, areas of low attenuation, and multifocal calcification is suggestive of invasive thymoma.

A respiratory gating technique was developed to allow computed tomography-guided needle biopsy of small pulmonary nodules. Twenty-three pulmonary nodules less than 15 mm in diameter underwent biopsy with the use of this technique. There were 14 true-positive, eight true-negative, and one false-negative result (diagnostic accuracy, 96%). The diagnostic accuracy for small nodules without this technique in a historical control was 69% (P

RATIONALE AND OBJECTIVES To investigate whether expiratory high-resolution computed tomography (HRCT) is more useful than inspiratory HRCT for the detection of early-phase diffuse alveolar damage. METHODS Eleven anesthetized rabbits were scanned with both inspiratory and expiratory HRCT every 30 minutes during mechanical ventilation. Ten rabbits were killed after the detection of pulmonary abnormalities on both inspiratory and expiratory HRCT. The remaining rabbit was killed when the pulmonary abnormalities appeared only on expiratory HRCT. RESULTS In four cases (36%), the abnormal findings were detected earlier on expiratory HRCT than on inspiratory HRCT. In seven cases (64%), the abnormalities appeared simultaneously on inspiratory and expiratory HRCT. In all 11 cases, the histopathological changes of areas with abnormal CT findings corresponded to the exudative or proliferative phase of diffuse alveolar damage. CONCLUSIONS Expiratory HRCT has the potential to detect the abnormalities of diffuse alveolar damage earlier than inspiratory HRCT.

The purpose of this study was to determine whether acute parenchymal lung diseases can be differentiated on the basis of the pattern and distribution of abnormalities revealed on high-resolution CT.High-resolution CT scans of 90 patients with acute parenchymal lung diseases (19 with bacterial pneumonia, 13 with mycoplasmal pneumonia, 21 with acute interstitial pneumonia, 18 with hypersensitivity pneumonitis, 10 with acute eosinophilic pneumonia, and nine with pulmonary hemorrhage) were independently assessed by two observers who had no knowledge of clinical or pathologic data. The observers recorded abnormalities, their first-choice diagnosis, and their degree of confidence in their first-choice diagnosis.The two observers made a correct first-choice diagnosis in an average of 55 (61%) of 90 cases. Correct first-choice diagnosis was made in 50% of cases of bacterial pneumonia, 62% of mycoplasmal pneumonia, 90% of acute interstitial pneumonia, 72% of hypersensitivity pneumonitis, 30% of acute eosinophilic pneumonia, and 28% of pulmonary hemorrhage. CT findings allowed distinction between infectious and noninfectious causes in 81 (90%) of 90 cases.High-resolution CT is helpful in the differential diagnosis of infectious from noninfectious acute parenchymal lung disease. However, high-resolution CT is of limited value in making a specific diagnosis.

Radiation-induced thyroid dysfunction is considered a late effect. We prospectively assessed acute reactions of the thyroid to external neck irradiation.This study included 22 patients in whom the thyroid was incidentally exposed to therapeutic doses of radiation. Thyroid function tests included measurements of serum thyroid stimulating hormone (TSH), free and total triiodothyronine (T3) and thyroxine (T4), thyroglobulin, and antithyroid antibodies. These tests were performed before radiotherapy (baseline values), after approximately 40 Gy had been administrated, 2 weeks after the end of radiotherapy, and 3 and 6 months after the beginning of radiotherapy.Mean serum levels of TSH were 1.53, 0.55, 0.78, 2.14, and 7.57 microU/ml before radiotherapy, after 40 Gy irradiation, 2 weeks after the end of radiotherapy, and 3 and 6 months after radiotherapy, respectively. Thus, levels of TSH exhibited two phases: a significant decrease during radiotherapy (thyrotoxic phase) and an increase after radiotherapy (hypothyroid phase) (baseline vs. 40 Gy: p0.0001, baseline vs. 6 months: p = 0.003). Increases of thyroid hormones were subtle during radiotherapy.We believe that radiation promotes release of excessive amounts of thyroid hormones during radiotherapy owing to suppression of TSH secretion. In addition to the late damage (hypothyroidism), thyrotoxicosis occurs when the thyroid gland receives a therapeutic doses of external radiation.

Purpose: To identify the normal performance of left ventricular (LV) regional contraction using cine MR imaging with presaturation myocardial tagging. Material and Methods: Sixteen normal volunteers were examined on a 1.5 T MR system with tagging cine sequences. Tags were applied at end-diastole as 2 parallel black lines on short-axis and 4-chamber sections, and the fractional shortenings were calculated at 7 LV locations. Results: The following results were obtained with significance: a transmural gradient of contractility in the short-axis section; prolonged late-systolic endocardial shortening and epicardial early termination in the free wall; initial delay of shortening in the anterior wall; apical predominance of contractility; predominance of circumferential shortening in the free wall and of meridional shortening in the septum. These findings could be associated with myocardial fiber architecture, presumed wall stress and temporal asynergy of excitation. Conclusion: Cine MR imaging with myocardial tagging proved to be useful in assessing the nonuniformity of LV contraction.

Purpose: The clinical usefulness of an advanced image-processing system called “dynamic range control processing” was investigated, with which selected parts of the dynamic range of digital chest images could be controlled. Material and Methods: A comparative study of 3 different post-processed formats of storage phosphor (SR) images was performed in 35 patients with abnormalities in the chest. The 3 formats were SR images with standard mode (SR-standard), SR images with strong edge-enhancement (SR-enhanced), and dynamic range controlled SR images (SR-controlled). Results: For lung abnormalities, there was no difference among the 3 SR image formats. For normal mediastinal structures and lung abnormalities covered by the heart or diaphragm, SR-controlled and SR-enhanced images were significantly superior to SR-standard images, while no difference was found between SR-controlled and SR-enhanced images except for the trachea and bony structures. Conclusion: Dynamic range control processing appears to be a useful method for displaying SR chest images with an extremely wide dynamic range.

Twenty-seven consecutive patients were examined by T2-(1 800/70 ms) and postcontrast T1-weighted (600/15) spin echo (SE) or dynamic (200/15) SE MR imaging to determine the usefulness of parasagittal MR imaging in assessing cervical invasion of endometrial carcinoma. The images were obtained in a direction parallel to the longitudinal axis of the uterus (parasagittal). The cervical epithelium, being hyperintense on the late phase dynamic and postcontrast T1-weighted SE images, had disappeared partially or totally in all 4 patients with cervical invasion. The enhanced cervical epithelium was completely seen in one patient with the tumor protruding into the cervical canal in a polyp-like form without cervical epithelial invasion. The same was also seen in the 22 patients with the tumor remaining in the corpus cavity. The enhanced parasagittal MR images facilitated the evaluation of the extent of the endometrial carcinoma.

To compare the clinical usefulness of the single-exposure dual-energy subtraction method with an iterative noise-reduction algorithm.Fourteen radiologists read three sets of images from 44 patients: original computed radiographic images only, original computed radiographic images plus conventional bone-subtracted images, and original computed radiographic images plus iterative noise-reduced bone-subtracted images. Twenty-two patients had one or more (maximum, five) pulmonary nodules; 22 had no pulmonary nodules. Observer performance was evaluated by means of calculation of the average area under the alternative free-response receiver operating characteristic curves (A1).Compared with the original computed radiographic image only, detection of nodules was significantly better with both the original computed radiographic image plus iterative bone-subtracted image (A1 = 0.72 +/- 0.02 and 0.66 +/- 0.02, respectively; P = .01) and the original computed radiographic image plus conventional bone-subtracted image (A1 = 0.66 +/- 0.02 and 0.61 +/- 0.01, respectively; P = .03).The iterative noise-reduction algorithm is superior to conventional methods in detection of pulmonary nodules.

MR angiographic examinations were performed in 26 patients. Approximately 30 contiguous fast spoiled GRASS images (8.1/3.1/30°, TR/TE/flip angle) were acquired during single breath-holding for about 35 s, and then new images were reconstructed with maximum intensity projection technique. Spoiled GRASS images (40/12/40) of 2 to 3 slices were taken during breath-holding for about 13 s, and these processes were repeated to obtain about 30 individual images for conventional MR angiography. The new MR angiograms were compared quantitatively and qualitatively with conventional MR angiograms with arterial portography as the gold standard. The new MR angiograms could visualize blood vessels with smooth margins, and provided almost the same anatomic information about the portal vein and collateral vessels as the conventional MR angiograms. Contrast-to-noise ratios between the portal or hepatic vein and liver parenchyma were significantly higher with fast spoiled GRASS images. The new MR angiograms using fast spoiled GRASS images provided useful diagnostic mapping of the collateral venous pathways within a shorter examination time.

Chondrosarcomas are the most frequently occurring primary malignant chest wall tumors. Furthermore, the lungs serve as the most frequent sites for metastases. Pulmonary metastases from sarcomas usually appear as round nodules of varying sizes on roentgenograms. Here, we report an unusual clinical and radiographic presentation of pulmonary metastasis from a costal chondrosarcoma. Bilateral pulmonary metastases developed soon after wide surgical resection. Thoracic computed tomography revealed unusual radiological findings: consolidation accompanied with ground-glass opacity. To confirm the metastasis, we recommend a transbronchial biopsy in cases where unusual pulmonary findings are detected.

MR images of the neck were prospectively studied in 19 patients with hyperparathyroidism. Fast low angle shot (FLASH) sequence was performed in addition to T1- and T2-weighted spin echo (SE) sequences. FLASH images were obtained with 320/12/20° (TR/TE/flip angle) using presaturation technique. TE of 12 ms was chosen to eliminate high signal of fat tissue. In the evaluation of detectability, a combination of T1-weighted SE and FLASH images (T1WI + FLASH) was compared with a combination of T1- and T2-weighted SE images (T1WI + T2WI). MR imaging correctly depicted 20 of 30 abnormal glands on both T1WI + FLASH and T1WI + T2WI. FLASH imaging effectively eliminated high signal of fat tissue. Nineteen abnormal glands demonstrated higher signal than surrounding tissues on FLASH images, whereas 12 glands were high-intense on T2-weighted SE images. We conclude that FLASH imaging provides improved tissue contrast and anatomic delineation and, thus, may replace T2-weighted SE imaging in the neck.

To evaluate the effect of transcatheter arterial chemoembolization (TACE) with iodized oil for hepatocellular carcinoma (HCC), dynamic turbo-fast low angle shot (turbo-FLASH) (TR/TE/flip angle/TI, 8.5/4.6/10/200) MR imaging with gadopentetate dimeglumine was performed in 10 patients with HCC after TACE with iodized oil and before partial hepatectomy. Immediately after 0.05 mmol/kg b.w. of gadopentetate dimeglumine was administered intravenously, 10 images were obtained in the first 20 s (early phase). Then, one image every 30 s from 1 to 3 min (late phase), and images at 5 min and 7 min (delayed phase) were obtained serially. In the early phase, HCC showed no enhancement in 5 patients, partial hyperintense enhancement in 4, and total hyperintense enhancement in one. Viable regions of the tumor, evaluated at histopathology, showed hyperintense enhancement relative to the surrounding liver parenchyma in the early phase, while necrotic regions showed no enhancement. Both viable and necrotic regions showed lower signal intensities than the surrounding liver parenchyma in both late and delayed phases. By using dynamic turbo-FLASH MR imaging, we were able to accurately evaluate the effect of TACE with iodized oil for HCC in 8 of the 10 patients. In 2 patients, in whom small viable cells were seen in the HCC, viable regions could not be detected with our technique. It is concluded that turbo-FLASH dynamic MR imaging was useful for evaluating the effect of TACE for HCC.

It has been stated, but not adequately assessed, that pulmonary tuberculosis in diabetic or immunocompromised patients often has an atypical pattern and distribution. To evaluate the CT features of pulmonary tuberculosis in diabetic or immunocompromised patients compared with patients without underlying disease, we performed this retrospective study.We reviewed conventional CT scans (n = 100) and high-resolution CT scans (n = 16) of the chest in 110 adult patients with active postprimary tuberculosis. Seventy-one patients had no underlying disease, 31 had diabetes mellitus, and eight were immunocompromised.In patients who had no underlying disease, 44 had nodular opacities, 11 had consolidation, and 13 had consolidation with associated loss of volume. Characteristic features of tuberculosis in this group of patients included segmental distribution (97%), satellite lesions (93%), single cavity within any given lesion (95%), and tendency toward architectural distortion and loss of volume. In diabetic and immunocompromised patients, 15 had nodular opacities, seven had consolidation, and 15 had consolidation with associated loss of volume. Diabetic and immunocompromised patients had a high prevalence of nonsegmental distribution (30%) and multiple small cavities within any given lesion (44%). Unusual localization of tuberculosis, including disease confined to the basal segments of the lower lobes, anterior segment of the upper lobes, or right middle lobe, occurred equally in both groups (17% and 18%).We conclude that diabetic and immunocompromised patients have a higher prevalence of multiple cavities within any given lesion (p.01) and of nonsegmental distribution (p.01) than do patients without underlying disease.

Summer-type hypersensitivity pneumonitis is an immunologic disease that occurs only in Japan. It is a form of hypersensitivity pneumonitis in which the clinical symptoms appear in the summer and subside spontaneously in mid autumn. The purpose of our study was to determine the CT findings in this condition, to compare the CT findings with those on chest radiographs, and to assess the variations in the CT findings over time. Accordingly, high-resolution CT scans and chest radiographs of 15 patients with summer-type hypersensitivity pneumonitis were retrospectively studied. Seven patients had sequential CT examinations 18-37 days apart. The CT scans and chest radiographs were reviewed by two observers independently. CT findings included diffuse micronodules (n = 15), slightly elevated lung density (n = 13), and patchy air-space consolidation (n = 13). In one patient, the findings on a chest radiograph were normal, while CT showed parenchymal abnormalities. In two cases, follow-up CT showed micronodular abnormalities after findings on the chest radiograph had returned to normal. Our results show that high-resolution CT findings of summer-type hypersensitivity pneumonitis include pulmonary micronodules, increased lung density, and air-space consolidation. High-resolution CT appears to be more useful than plain chest radiographs in the evaluation of pulmonary parenchymal abnormalities in this condition.

Findings at myelography and computed tomographic (CT) myelography were reviewed in 21 patients (22 limbs) with birth palsy; nerve root shadows were the focus of this study. Myelography demonstrated 51 completely avulsed roots (78%) and 14 incompletely avulsed roots (22%). A traumatic meningocele was detected at 38 roots on myelograms and 51 roots on CT myelograms. Thirteen avulsed roots (eight completely and five incompletely avulsed roots) (20%) were not associated with a meningocele. In nine patients who underwent brachial plexus exploration, myelographic findings were compared with root somatosensory evoked potential (SEP). SEP was not induced at 22 of 25 completely avulsed roots and was induced at all seven incompletely avulsed roots. Myelography and SEP were consistent in 29 of 32 roots (91%). It is concluded that myelography is indispensable for preoperative evaluation of cervical nerve root avulsion of birth palsy, because CT myelography is not sensitive to nerve root avulsion without a traumatic meningocele, and SEP cannot enable one to discriminate incomplete avulsion from intact roots.

MR imaging and CT were performed prospectively in 35 patients with esophageal carcinoma to determine the resectability of the primary tumors, because at our institution patients with resectable tumors have surgery regardless of the presence of distant metastases. Tumors with evidence of aortic or tracheobronchial invasion on MR or CT were considered to be unresectable. Tracheobronchial invasion was diagnosed when the tumor extended into the lumen of the airway, and aortic invasion was diagnosed when the triangular fat space between the esophagus, aorta, and spine adjacent to the primary tumor was obliterated. Two patients were excluded because of suboptimal MR images produced by motion artifacts. Pathologic proof was obtained from either surgery or autopsy in 31 patients. Of these, six patients (19%) had proved unresectable tumors (three aortic invasion and three tracheobronchial invasion). In all six cases, these features were correctly detected with both MR and CT. One patient had false-positive findings on MR and CT. An indeterminate diagnosis was obtained with MR in three patients and with CT in four patients. These incorrect or indeterminate results were all related to the diagnosis of aortic invasion. No patient had a false-negative result. When indeterminate diagnoses were considered false-positive, sensitivity, specificity, and accuracy for resectability were 100%, 84%, and 87%, respectively, for MR and 100%, 80%, and 84%, respectively, for CT. We conclude that MR and CT have nearly the same accuracy in predicting resectability of tumors in patients with esophageal carcinoma.

One hundred and sixty-one patients with pulmonary metastases were studied with CT. Six of them proved to have endobronchial (intraluminal) metastatic lesions by bronchoscopy. Retrospective analysis of the CT studies showed obstruction and/or narrowing of the bronchi in 5 cases while no lesion was observed in one patient. Although CT can not always demonstrate intraluminal lesions, it should be performed when an endobronchial metastasis from extrathoracic malignancy is seen by bronchoscopy because it will show hilar or mediastinal lymphadenopathy, or single or multiple pulmonary metastases other than the endobronchial lesion.

The serial computed tomography findings and prognosis of the acute exacerbation of idiopathic pulmonary fibrosis (IPF) are not yet well defined in a larger number of cases.To evaluate the parenchymal abnormalities and prognosis using high-resolution computed tomography (HRCT) in acute exacerbation of IPF.The study consisted of clinical, laboratory, and HRCT data before and at the time of acute exacerbation in 64 episodes of 58 patients with IPF. A semiquantitative analysis of overall extent of parenchymal abnormalities, extent of alveolar opacity (ground-glass attenuation and consolidation), and extent of fibrotic opacity (reticulation and honeycombing) on CT was performed by two chest radiologists. The newly appeared parenchymal abnormalities were also classified into three patterns: peripheral, multifocal, and diffuse.In all patients, HRCT scans taken at the exacerbation showed typical signs of IPF and newly developing alveolar opacity. They included 34 patients of peripheral pattern, 8 of multifocal pattern, and 16 of diffuse pattern. Twenty-five patients died and 33 survived after the initial exacerbation. Worse survival was associated with patients with diffuse type compared with patients with multifocal and peripheral type. The CT patterns and overall CT extent were associated with an increased hazard of death after adjusting for age, sex, smoking, baseline diffusion capacity for carbon monoxide, baseline FVC, and disease extent on CT. On multivariate analysis, the strongest correlations were observed between CT patterns (combined diffuse and multifocal versus peripheral) and survival (odds ratio, 4.629; 95% confidence interval, 1.900-11.278; P = 0.001).HRCT extent and patterns are predictive of survival in acute exacerbation of IPF.

Nine patients with dural arteriovenous malformations (AVMs) in the region of the cavernous sinus were treated by means of external carotid artery (ECA) embolization using polyvinyl alcolhol. All AVMs received vascular supply from both the ECA and the internal carotid artery. Seven cases were clinically cured after embolization, while 2 cases with cortical venous drainage and high flow through the shunt were not completely cured. Venous thrombosis was observed in 5 cases before and in 9 after embolization. In 6 cases the drainage pattern changed owing to venous thrombosis. Complete thrombosis of the cavernous sinus was found on a follow-up angiography in 2 cases. Formation of venous thrombosis and occlusion of feeding arteries are crucial factors for success of ECA embolization. Dural AVMs with cortical venous drainage and high flow cannot be relieved by ECA embolization alone owing to difficulty in obtaining thrombosis of the veins.

The purpose of this study was to assess the usefulness of color digital summation radiography (CDSR) for detection of nodules on chest radiographs by observers with different levels of experience.A total of 30 radiographs of chest phantoms with abnormalities and 30 normal ones were arranged at random. Set A was conventional radiographs only. Set B consisted of both conventional radiographs and CDSR images, which were colored with magenta. Five chest radiologists and five residents evaluated both image sets on a TFT monitor. The observers were asked to rate each image set using a continuous rating scale. The reading time for each set was also recorded.In set A, the performance of chest radiologists was significantly superior to that of the residents (P0.05). However, in set B, there was no significant difference in the performance of the chest radiologists and the residents. In both observer groups, the mean reading time per case in set B was significantly shorter than that in set A (P0.01).By using CDSR, the detection capability of observers with little experience improves and is comparable to that of experienced observers. Moreover, the reading time becomes much shorter using CDSR.

The aim of this study was to evaluate the image quality of high-resolution CT (HRCT) reconstructed from volumetric data with 16-channel multidetector-row CT (MDCT).Eleven autopsy lungs that were diagnosed histopathologically were scanned by 16-channel MDCT with the step-and-shoot scan mode and three helical scan modes. Each helical mode had each size of focal spot, pitch, and time of gantry rotation. HRCT images were reconstructed from the volumetric data with each helical mode and axial sequence data. Two observers evaluated the image quality and noted the most appropriate diagnosis for each imaging.Visualization of abnormal structures with one helical mode was equal to those with axial mode, whereas those with the other two helical modes were inferior to those with axial mode (Wilcoxon signed rank test; p0.0001). There was no significant difference in diagnostic efficacy between modes.The image quality of HRCT with appropriate helical mode is equal to that with axial mode and diagnostic efficacy is equal among all modes. These results may indicate that sufficient HRCT images can be obtained by only one helical scan without the addition of conventional axial scans.

Objective To determine whether the subtypes of nonspecific interstitial pneumonia (NSIP) could be differentiated from other idiopathic interstitial pneumonias (IIPs) on the basis of findings on high-resolution computed tomography (CT). Methods Two observers evaluated the high-resolution CT findings in 90 patients with IIPs. The patients included 36 with NSIP, 11 with usual interstitial pneumonia (UIP), 8 with cryptogenic organizing pneumonia (COP), 10 with acute interstitial pneumonia (AIP), 14 with desquamative interstitial pneumonia (DIP) or respiratory bronchiolitis-associated interstitial lung disease (RB-ILD), and 11 with lymphoid interstitial pneumonia (LIP). The NSIP cases were subdivided into group 1 NSIP (n = 6), group 2 NSIP (n = 15), and group 3 NSIP (n = 15). Results Observers made a correct diagnosis with a high level of confidence in 65% of NSIP cases, 91% of UIP cases, 44% of COP cases, 40% of AIP cases, 32% of DIP or RB-ILD cases, and 82% of LIP cases. Group 1 NSIP was misdiagnosed as AIP, DIP or RB-ILD, and LIP in 8.3% of patients, respectively. Group 2 NSIP was misdiagnosed as COP in 10% of patients, LIP in 6.7%, AIP in 3.3%, and DIP or RB-ILD in 3.3%. Group 3 NSIP was misdiagnosed as UIP in 6.7% of patients, COP in 6.7%, and DIP or RB-ILD in 3.3%. Conclusions In most patients, NSIP can be distinguished from other IIPs based on the findings on high-resolution CT. Only a small percentage of patients with predominantly fibrotic NSIP (group 3 NSIP) show overlap with the high-resolution CT findings of UIP.

To clarify the appropriate voxel dimensions required for pathologic evaluation of areas with ground-glass opacity on lung high-resolution computed tomography (HRCT).Synchrotron radiation CT (SRCT) images of autopsied lung speci-mens (n=25) that showed ground-glass opacity on HRCT were reconstructed with 12 different voxel dimensions ranging from 0.006 to 0.6 mm. The specimens were micromorphologically categorized into one of three pathologic groups: alveolar, interstitial, and mixed abnormalities. Each SRCT image was independently diagnosed as one of three pathologic groups by six chest radiologists. The diagnostic accuracy required to estimate the appropriate voxel dimensions was compared among different voxel dimensions by means of the Tukey test.Diagnostic accuracy with voxel dimensions less than or equal to 0.06 mm was significantly higher than that with voxel dimensions of 0.18 mm or more (p0.01). There was, however, no significance of difference in diagnostic accuracy with voxel dimensions of less than or equal to 0.06 mm. In addition, no significant difference in diagnostic accuracy was found with voxel dimensions of 0.18 mm or more.The appropriate voxel dimensions are approximately 0.06 mm for pathologic differentiation of areas with ground-glass opacity on HRCT.

Objectives The aim of the present study was to determine how the pattern and extent of sarcoidosis changes over time on serial high-resolution CT (HRCT) scans and to identify CT scan findings that might be helpful in predicting the prognosis of patients with the disease. Methods The initial and last HRCT scan findings of 40 patients with pulmonary sarcoidosis during a mean follow-up period of 7.4 years (range, 3 to 18 years) were evaluated retrospectively. HRCT scan findings then were correlated with the pulmonary function tests. Results Parenchymal abnormalities in most patients with a predominant nodular pattern (18 patients) and multiple large nodular pattern (8 patients) disappeared or decreased in size on long-term follow-up CT scans. A conglomeration pattern (five patients) shrank and evolved into bronchial distortion. The shrinkage of the conglomeration pattern correlated with a decline of FEV1/FVC ratio, despite an improvement in FVC. A ground-glass opacity pattern (five patients) and a consolidation pattern (three patients) evolved into honeycombing. The evolution of the ground-glass opacity and consolidation patterns into honeycombing occurred along with a decline in FVC, although the parenchymal abnormalities became smaller. Conclusion Patients with a predominantly ground-glass opacity pattern and consolidation pattern seen on the initial CT scan had a worse prognosis and were susceptible to developing severe respiratory insufficiency. The predominant patterns seen on the initial HRCT scan may be helpful in predicting the outcomes of patients with sarcoidosis.

The purpose of this study was to evaluate the usefulness of the coronal multiplanar reconstruction (MPR) view in comparison with transverse helical thin-section CT for both the determination of malignant or benign lesions and the differential diagnosis of solitary pulmonary nodules.Sixty-eight cases of pathologically proved solitary pulmonary nodule less than 3 cm in diameter were enrolled in this study. For the routine study, transverse helical thin-section CT (1.25 mm collimation, FOV 20 cm) covering the areas with solitary pulmonary nodules as well as whole lung helical thin-section CT (2.5 mm collimation, 1.25 mm reconstruction interval, FOV 34.5 cm, pitch 6:1, high-spatial frequency algorithm) were scanned with a multidetector-row CT (MDCT) scanner. From the whole lung thin-section CT data, coronal MPR views (2.5 mm slice thickness) were reconstructed on a workstation. ROC analysis was used for an observer performance study, in which three observers indicated their confidence level for the determination of malignant or benign lesion for the nodules by means of transverse thin-section CT and coronal MPR. In addition, the observers recorded appropriate disease entities as the final diagnosis of each case. Accuracies of the final diagnosis based on the two sets of images were compared with McNemer' s test.In terms of the determination of malignant or benign lesion, there was no significant difference between the two sets of images (coronal MPR and transverse thin-section CT; mean Az=0.853 and 0.854, respectively). In addition, accuracy of the final diagnosis based on coronal MPR views (74%) was almost equal to that based on transverse thin-section CT (71%) (p=0.3).The diagnostic efficacy of the coronal MPR view is comparable to that of transverse thin-section CT for the evaluation of solitary pulmonary nodules.

To assess the detection and size of mediastinal and hilar lymph nodes by multiplanar reconstruction (MPR) view from isotropic voxel data sets obtained with multidetector-row computed tomography (MDCT).Thin-section CT of 27 patients with mediastinal or hilar lymph node swelling was obtained with a 25.6-cm FOV, 512 x 512 matrix, and two protocols: A) 0.5-mm collimation, 0.3-mm interval, and B) 2-mm collimation, 1-mm interval. MPR views with a 0.5-mm slice thickness were obtained from these two data sets. Postcontrast axial CT used 5-mm collimation (set C). Two observers evaluated the presence and cranio-caudal length of swollen lymph nodes. Two other board-certified chest radiologists evaluated all three sets and established a gold standard by consensus.The accuracy of detection was 76%, 73%, and 68% for sets A, B, and C, respectively. There was a significant difference between sets A and C (McNemar's test: p0.05) but not between sets A and B or B and C (p0.05). The cranio-caudal length of lymph nodes was significantly correlated with the gold standard only in set A (Pearson's correlation coefficient: r=0.53, p0.05).Non-contrast enhanced coronal MPR views constructed from isotropic voxel data sets may be substituted for axial enhanced CT for the evaluation of mediastinal and hilar lymph nodes.

The objective of this study was to determine whether the various chronic cystic lung diseases can be differentiated on the basis of the pattern and distribution of abnormalities on high-resolution CT.High-resolution CT scans in 92 patients with chronic cystic lung diseases (18 with pulmonary Langerhans cell histiocytosis, 18 with pulmonary lymphangioleiomyomatosis, 17 with usual interstitial pneumonia, 16 with lymphocytic interstitial pneumonia, 15 with emphysema, and eight with desquamative interstitial pneumonia or respiratory bronchiolitis interstitial lung disease) were retrospectively assessed by two independent observers without knowledge of the clinical or pathologic data. The observers recorded the abnormalities, the most likely diagnosis, and the degree of confidence in that diagnosis.The two observers made a correct first-choice diagnosis in 148 (80%) of 184 interpretations. The correct diagnosis was made in 100% of interpretations of usual interstitial pneumonia, 81% of desquamative interstitial pneumonia or respiratory bronchiolitis interstitial lung disease, 81% of lymphocytic interstitial pneumonia, 77% of emphysema, 72% of lymphangioleiomyomatosis, and 72% of Langerhans cell histiocytosis. The two observers made a diagnosis with a high degree of confidence in 105 (57%) of 184 interpretations. The confident diagnosis was correct in 98 (93%) of 105 interpretations.Although various chronic cystic lung diseases often have a characteristic appearance that allows their distinction on high-resolution CT, considerable overlap exists among the CT findings. Therefore, lung biopsy is often required for a definitive diagnosis.

The purpose of this study was to compare the quality of coronal multiplanar reconstructions with the quality of direct coronal thin-section multidetector CT (MDCT) scans.Axial multidetector CT (MDCT) scans were obtained through the entire lung in 10 normal autopsy lung specimens using an MDCT scanner. Four protocols were used: 0.5-mm collimation with a 0.5-mm reconstruction interval; 0.5-mm collimation with a 0.3-mm reconstruction interval; 1-mm collimation with a 0.5-mm reconstruction interval; and 2-mm collimation with a 1-mm reconstruction interval. Multiplanar reconstruction images with 0.5-mm slice thickness were obtained from the four types of data sets. Direct coronal thin-section CT of the same 10 autopsy lung specimens was performed using 0.5-mm scan collimation, a 0.3-mm reconstruction interval, a 25.6-cm field of view, and a 512 x 512 matrix. Two independent observers compared the image quality of each of the four coronal multiplanar reconstruction sets with that of direct coronal thin-section CT scans. The observers analyzed visualization of anatomic features and artifacts.The total image quality of the multiplanar reconstructions obtained from 0.5-mm collimation data with or without 0.3-mm overlapping reconstruction was equal to that of direct coronal thin-section CT scans in all 20 interpretations. The image quality of multiplanar reconstruction images from 0.5-mm collimation data either with or without overlapping reconstruction was superior to multiplanar reconstruction images obtained from 1- or 2-mm collimation scans (p0.01, Fisher's exact test). Stairstep artifacts in multiplanar reconstructions using 0.5-mm collimation without overlapping reconstruction were equal to those with overlapping reconstruction and were fewer than those on 1- or 2-mm collimation (p0.01, Mann-Whitney U test).The image quality of coronal multiplanar reconstructions from isotropic voxel data obtained using 0.5-mm collimation, with or without overlapping reconstruction, is similar to that of direct coronal thin-section CT scans.

Our purpose was to assess the CT features of various subtypes of thymic epithelial neoplasms on the basis of the 1999 World Health Organization classification.Thymic epithelial neoplasms in 53 patients who underwent thymectomy were retrospectively assessed histologically according to the 1999 World Health Organization classification. Type A and B neoplasms correspond to thymomas and type C, to thymic carcinoma. The study included four patients with type A, 14 with type AB, nine with type B1, 14 with type B2, four with type B3, and eight with type C epithelial tumors. Two observers independently assessed the CT scans without knowledge of the histologic findings.Type A tumors were more likely to have smooth contours on CT (4/4, 100%) and round shapes (3.5/4, 88%) than any other type of thymic epithelial tumor (all, p0.05). Type C tumors had a higher prevalence of irregular contours (6/8, 75%) than any other type of thymic epithelial tumor (all, p0.05). Calcification was more frequently seen in type B1 (4/9, 44%), type B2 (8.5/14, 61%), and type B3 (3/4, 75%) tumors than in type AB (2/14, 14%) and type C (0.5/8, 6%) tumors (all, p0.05).Smooth contours and a round shape are most suggestive of type A thymic epithelial tumor, whereas irregular contours are most suggestive of type C tumor. Calcification is suggestive of type B tumors. CT is of limited value, however, in differentiating type AB, B1, B2, and B3 tumors.

To evaluate the usefulness of a commercially available computer-aided diagnosis (CAD) system that incorporates temporal subtraction for the detection of solitary pulmonary nodules on chest radiographs by readers with different levels of experience.Sixty pairs of chest radiographs in 30 patients with newly detected solitary pulmonary nodules and 30 normal cases, all confirmed with serial chest computed tomography (CT), were obtained from screen-film or digital radiographic systems and were digitized (spatial resolution, 0.171 mm/pixel). Temporal subtraction images were produced with an iterative image-warping technique. Five chest radiologists and five residents evaluated both image sets for solitary nodules: set A, current and prior radiographs with temporal subtraction images, and set B, current and prior radiographs only. Assessment was performed with receiver operating characteristic (ROC) analysis of the images on a monitor (pixel size, 1,280 x 1,024) equipped with the system. The reading time needed by each reader was recorded in each case.For the chest radiologists, no statistically significant difference was found between set A (area under the ROC curve [A(z)] = 0.934) and set B (A(z) = 0.964). For the residents, however, observer performance in set A (A(z) = 0.907) was superior to that in set B (A(z) = 0.855) (P.05). For both groups, the mean reading time per case for set A (chest radiologists, 16.7 seconds; residents, 15.7 seconds) was significantly (P.05) shorter than that for set B (chest radiologists, 20.4 seconds; residents, 26.2 seconds).For the detection of solitary pulmonary nodules, the CAD system with temporal subtraction can promote efficiency for established chest radiologists and improvement in accuracy for less experienced readers.

The objective of this study is to evaluate usefulness of newly developed interactive multiplanar reconstruction (MPR) viewer for the evaluation of invasion to surrounding organs of esophageal carcinoma.

Peritoneal pseudocysts are created when fluid arising from the ovary is trapped by peritoneal adhesions. We studied the value of MR imaging, compared with sonography and CT, for detecting these abnormalities.Nine women, 27 to 49 years old, who had surgically proved peritoneal pseudocysts were included in the study. All nine patients were premenopausal, and eight had a history of pelvic surgery. All patients were examined with MR imaging and sonography, and five were also examined with CT. Imaging findings were analyzed and compared with the surgical findings.Sonograms in all nine patients showed a thick-walled cyst that contained focal echoes indicating nodules, making it difficult to differentiate these lesions from cystic ovarian tumors. In eight of the nine patients, the fluid in the pseudocysts had low signal intensity on T1-weighted MR images and high signal intensity on T2-weighted spin-echo MR images, suggesting that the fluid was serous. The pseudocyst had an irregular shape on MR images in seven patients and on CT scans in three patients. However, MR provided more information than CT did about the nature of the fluid within the pseudocyst.Our study suggests that MR imaging is more useful than sonography and CT for determining the characteristic findings of peritoneal pseudocysts.

Angiomyolipoma, a rare benign liver tumor, was observed in a 50-year-old woman examined with US, CT, MR imaging and angiography. Dynamic studies using CT and MR imaging were valuable in differentiating the disease from hepatocellular carcinoma with fat deposits.

The objective of this study was to describe the pulmonary abnormalities on high-resolution computed tomography (CT) in patients with mixed connective tissue disease (MCTD). The study included 41 patients who met the diagnostic criteria for MCTD and showed abnormal findings on high-resolution CT. The presence, extent, and distribution of various high-resolution CT findings were evaluated. The predominant abnormalities included areas of ground-glass attenuation (n = 41), subpleural micronodules (n = 40), and nonseptal linear opacities (n = 32). Other common findings included peripheral predominance (n = 40), lower lobe predominance (n = 39), intralobular reticular opacities (n = 25), architectural distortion (n = 20), and traction bronchiectasis (n = 18). Less common findings included honeycombing, ill-defined centrilobular nodules, airspace consolidation, interlobular septal thickening, thickening of bronchovascular bundles, bronchial wall thickening, bronchiectasis, and emphysema. Pulmonary involvement of MCTD is characterized by the presence of ground-glass attenuation, nonseptal linear opacities, and peripheral and lower lobe predominance. Ill-defined centrilobular opacities were uncommonly seen.

Purpose: The purpose of this work was to compare the thin-section CT findings of acute respiratory distress syndrome (ARDS) with those of acute interstitial pneumonia (AIP). Method: The thin-section CT scans from 25 patients with ARDS and 25 with AIP were independently assessed by two observers without knowledge of clinical and pathologic data. The presence, extent, and distribution of various CT findings were independently analyzed. Results: Honeycombing was seen more frequently in lobes of patients with AIP (26%) than in lobes with ARDS (8%) (p < 0.001). Compared with patients with ARDS, a greater number of patients with AIP had a predominantly lower lung zone distribution (p < 0.05) and a symmetric distribution (p < 0.05) of the parenchymal abnormalities. Conclusion: Patients with AIP have a greater prevalence of honeycombing and are more likely to have a symmetric bilateral distribution and a lower lung zone predominance than patients with ARDS. However, significant overlap exists among the CT findings.

The aim of the present study was to assess the evolution of various computed tomographic (CT) findings of lymphocytic interstitial pneumonia (LIP) with determination of potentially reversible or irreversible features. The study included 14 patients with biopsy-proved LIP who had serial thin-section CT examination 4 to 82 months (median 13 months) apart. Initial and follow-up CT scans were evaluated independently and then directly compared with each other by two observers. The main parenchymal abnormalities on the initial CT scan consisted of ground-glass attenuation (n = 14), thickening of interlobular septa (n = 13), centrilobular nodules (n = 12), cystic airspaces (n = 10), and airspace consolidation (n = 4). On follow-up CT, nine patients improved, one showed no change, and four showed increased extent of disease. With the exception of cysts, the parenchymal opacities were reversible. On follow-up CT, new cysts were seen in three patients; these developed mainly in areas with centrilobular nodules on initial CT. Honeycombing was seen on follow-up CT in four patients; in three patients it developed in areas of airspace consolidation and in one patient it developed in an area with ground-glass attenuation on initial CT. The majority of patients with LIP improved on follow-up. However, airspace consolidation may progress to honeycombing and centrilobular nodules may precede cystic formation.