Search

We correlated high-resolution CT (HR-CT) images of 30 surgically resected epidermoid carcinomas in peripheral lung fields with pathologic findings. We classified peripheral epidermoid carcinomas into three types based on HR-CT findings. All tumors showed a notch along their margin. Type 1 shows scanty speculations without the convergence of surrounding lung structures. Type 2, which has two subtypes, shows a spiculated border with the convergence of peripheral lung structures. Type 2a has fine irregularities in its border, while type 2b has few fine irregularities. The convergence of peripheral lung structures seen in type 2 carcinoma is caused by the presence of scar tissue within tumors. Speculations corresponded to tumor extension or lymphocytic infiltration, sometimes along pulmonary vessels or interlobular septa, and fine irregularities seen in type 2a corresponded to summation of fibrous thickening of alveolar septa. Nine tumors showed varying-sized cavities, which were caused by the circulatory disturbance and/or central necrosis seen in tumor nests. In conclusion, HR-CT images were well correlated with pathologic findings of the resected specimens; however, some type 2a tumors may mimic adenocarcinoma on HR-CT findings.

We examined the usefulness of a new medical facsimile (MFAX) system in recording and transmitting various kinds of medical images, including X-ray images and colored histopathologic images. The system consists of an image scanner, a magnetic disk for image storage, a transmission circuit and a thermal image printer. Transmission time for a FCR (Fuji computed radiography) image by super-fine mode was 6 minutes. We used ROC (receiver operating characteristics) curves to evaluate the ability of eight radiologists to detect the small simulated nodules placed on an anthropomorphic chest phantom and shown on MFAX images. The radiologists observed both the FCR films and MFAX copies and determined the presence or absence of simulated nodules using five confidence levels. The results obtained for FCR films and MFAX images showed no statistically significant difference.

The detectability of simulated nodules placed on a chest phantom and shown on FCR images has been studied. Different tube voltages (130 kVp, 100 kVp), presence or absence of compensation filter, various sites and different sizes (5 mm/10 mm) and configurations (spherical or broad based) of the nodule were examined. 112 FCR chest phantom images were evaluated by ten radiology specialists utilizing ROC analysis. The detection rate for nodules was influenced in general by the site. The detection rate was as follows in decreasing order: intercostal, overlaid with one rib, two ribs, diaphragm and heart. Detection of nodules of 10 mm in height was mainly influenced by the density of the lung where the nodule was located, whereas that of nodules of 5 mm in height was more dependent on the site of the nodule, probably greatly influenced by the amount of structure mottle. Spherical nodules of 5 mm in height were more frequently recognized than the broad-based ones. High kVp (130 kVp) resulted in a higher detection rate for both sizes of nodules than low kVp (100 kVp). Detection of spherical nodules overlying the heart or diaphragm was improved with the use of a compensation filter.

We compared the ability of standard film-screen chest radiographs to detect simulated pulmonary nodules. The radiographs were taken with different X-ray filters and tube voltages. Detection capability was analyzed utilizing receiver operating characteristic (ROC) curves formed from the interpretation of six readers. Nodule detection was slightly better in images obtained with a tube voltage of 135 kVp than with one of 100 kVp (not statistically significant). Approximately equal detection performance was shown by a conventional film-screen system and FCR with three X-ray filter sets: a copper filter backed by aluminium, a tungsten filter backed by yttrium and aluminium, and a lead filter backed by yttrium and aluminium. Results showed less X-ray exposure with the two latter filter sets. Unenhanced FCR images taken with approximately the same X-ray exposure as conventional radiographs showed nearly equal ability to detect the simulated nodules.

To determine a suitable combination of X-ray filter and tube kilovoltage for Fuji Computed Radiography (FCR), to provide better detection of pulmonary nodules and reduce patient exposure, we compared observer performance with different X-ray filters and tube voltages. Radiographs were obtained with a copper filter backed by aluminium, with a tungsten filter backed by yttrium and aluminium, and with a lead filter backed by yttrium and aluminium, at both 100 kVp and 135 kVp. Observer performance in detecting simulated lung nodules, which were placed on the posterior aspect of a chest phantom, was compared using receiver operating characteristic (ROC) techniques for each combination of the X-ray filter and tube voltage. The results of the study indicated that 1) nodule detection was superior for the images obtained with 135 kVp as compared with 100 kVp; 2) approximately equal detection rates were obtained for the three X-ray filter sets, although results with the tungsten filter were slightly inferior to the other two filters; and 3) the absorbed dose may be reduced by 30% with the use of a lead filter backed by yttrium and aluminium compared with a copper filter backed by aluminium. We conclude that chest radiography with FCR should preferably be conducted with a higher kilovoltage, e.g., 135 kVp rather than 100 kVp, to ensure a higher detection rate of pulmonary nodules and in conjunction with a lead filter backed by yttrium and aluminium to reduce X-ray exposure to the patient.

The appropriate radiation dose and field for the treatment of intracranial germinoma were investigated in 33 patients. Recurrences were observed in 4 cases treated with local field irradiation only, and all of them were recognized at the margin of the radiation field or under the dose (less than 25 Gy) area. This suggests that whole cranial irradiation (dose of 25 to 30 Gy) should be added even if the tumor is solitary. The effective dose for cerebrospinal dissemination appears to be 25-35 Gy, but prophylactic CNS irradiation seems unnecessary for patients who have not undergone surgical procedures. Changes in mental status were seen in 5 patients (26.3%). Doses of over 59 Gy may be related to this complication.