Takeo Ishigaki

Preoperative staging of thyroid papillary carcinoma with ultrasonography

OBJECTIVEnTo evaluate the usefulness of ultrasonography including Doppler flow imaging for the preoperative staging of thyroid papillary carcinoma.nnnMATERIALS AND METHODSnIn 77 patients with thyroid papillary carcinoma who underwent total thyroidectomy, the accuracy of ultrasonography in preoperative clinical staging was assessed with use of pathologic examination on the basis of TNM classification by the International Union Against Cancer (UICC).nnnRESULTSnIn 63 (81.8%) cases, T categories were estimated accurately. The sensitivity in depicting tumor extension into the prethyroidal muscle and/or the sternocleidomastoid muscle was 77.8%, whereas the sensitivity for invasion into the trachea and the esophagus was 42.9 and 28.6%, respectively. In 37 (48.1%) cases, N categories were underestimated, and the sensitivity in the detection of regional lymph node metastasis was 36.7%. Doppler flow imaging was performed in 36 patients, and no correlation was found between flow patterns and the presence of local invasion or regional lymph node metastasis.nnnCONCLUSIONnUltrasonography was useful for preoperative investigation of thyroid papillary carcinoma, but several limitations existed, especially in evaluating extracapsular invasion to deep locations and regional lymph node metastasis.

Role of ultrasonography in the detection of intraductal spread of breast cancer: correlation with pathologic findings, mammography and MR imaging

Abstract. The purpose of this study was to assess the role of US in the detection of intraductal spread of breast cancer in comparison with mammography (MMG) and MRI. In 46 patients with breast cancer, US features of the intraductal component were classified as ductal type or distorted type. Histopathologically, 29 of 46 (63 %) cases had intraductal components, and the sensitivity, specificity, and accuracy rates in detection of intraductal spread were 89, 76, and 85 %, respectively. Each US pattern demonstrated good correspondence to the histologic components, and the distorted type correlated well with comedo-type carcinoma. Mammography was performed in all cases, and the sensitivity, specificity, and accuracy rates in detection of intraductal spread were 55, 100, and 72 %, respectively. In comedo type, MMG could diagnose the extent of intraductal spread more accurately compared with US examination. Magnetic resonance imaging comparison was available in 25 cases. Magnetic resonance imaging depicted intraductal extension as an enhanced area during the early phase of a contrast enhancement study with a sensitivity of 93 %. Ultrasound and MRI were closely related in terms of morphologic characteristics: the ductal type of US image correlated well with linear enhancement on MRI, whereas the distorted type correlated with regional or segmental enhancement. Current US examination is useful in depicting the intraductal spread of breast cancer; however, US has a tendency to underestimate intraductal component of comedo type compared with MMG and MRI.

Assessment of the pancreatic and intrapancreatic bile ducts using 0.5-mm collimation and multiplanar reformatted images in multislice CT

Abstract. Our objective was to evaluate the ability of multiplanar reformatted (MPR) images combined with 0.5-mm axial images to depict the pancreatic and intrapancreatic bile ducts and compare the results with those of 0.5-mm axial, 2-mm axial, and 6-mm axial images alone. Seventy-seven patients without obstruction of the main pancreatic ducts (MPD) underwent dual-phase helical scanning of the pancreas using multislice computed tomography (MSCT). The MPR images were generated from 0.5-mm-thick images. Visualization of the pancreatic and intrapancreatic bile ducts and their confluence was graded on a four-point scale by a consensus of two radiologists. The results for 0.5-mm axial images in early-phase CT, 2-mm axial images in early-phase CT, MPR images combined with 0.5-mm axial images in early-phase CT, and 6-mm axial images in late-phase CT were then compared. The relationships of the focal pancreatic lesions with the pancreatic ducts were analyzed. The MPR images combined with 0.5-mm axial images were significantly superior to the other three types of images for the visualization of the pancreatic and intrapancreatic bile ducts and their confluence (p<0.01). The depiction rate of the MPD using MPR images combined with 0.5-mm axial images was 94, 94, 95, and 75%, respectively in the head, neck, body, and tail of the pancreas. Accessory pancreatic ducts, intrapancreatic bile ducts, and duct confluence were depicted in 48, 99, and 92%, respectively. In comparison with evaluation based on axial images alone, the use of MPR images more clearly demonstrated the relationship between the lesions and the pancreatic ducts in 14 of 19 lesions. The MPR images combined with 0.5-mm axial images improve the CT depiction of the pancreatic and intrapancreatic bile ducts in comparison with 0.5-mm axial, 2-mm axial, and 6-mm axial images alone.

Further reduction of radiation dose in helical CT for lung cancer screening using small tube current and a newly designed filter

A new aluminum filter, 5.8 mm thick at the center, was designed. The effective energy, exposure dose, absorbed dose, and noise were measured by using low-dose technique, very low-dose technique with a conventional filter, and very low-dose technique with a new filter on a chest phantom. Accuracy of very low-dose computed tomography (CT) with a new filter was compared against standard helical CT in 40 patients and against chest radiography in 35 patients. Effective energies were 42.6 keV and 51.6 keV at a conventional filter and the new filter, respectively. Compared against 20mA with a conventional filter, exposure dose was reduced by 17%, and absorbed dose was equivalent, at 30 mA with the new filter. Noise was improved by 9%. Compared with standard helical CT, the sensitivity, specificity, and accuracy of very-low-dose helical CT were 100%, 88%, and 95%, respectively. Very-low-dose helical CT was found to be significantly superior to chest radiography in the detection of lung cancers. Using a smaller tube current and an appropriate filter allows a further reduction in radiation dose in helical CT for lung cancer screening.

Relationship between Brier score and area under the binormal ROC curve

If we consider the Brier score (B) in the context of the signal detection theory and assume that it makes sense to consider the existence of B as a parameter for the population (let B be this B), and if we assume that the calibration in the observers probability estimate is perfect, we find that there is a theoretical relationship between B and the area under the binormal receiver operating characteristic (ROC) curve, A(Z). We have derived this theoretical functional relationship between B and A(Z), by using the parameter a and b in the binormal ROC model and the prior probability of signal events (alpha); here, the two underlying normal distributions are N and N; and, a= and b=. We empirically found that, if parameters b and alpha are constant, B values in relation to given A(Z) values monotonically decrease as A(Z) values increase, and these relationship curves have monotonically decreasing slopes.

CRT diagnosis of pulmonary disease: influence of monitor brightness and room illuminance on observer performance

Using a 21-in. cathode ray tube (CRT) monitor (2048 x 2560 x 8bits), six radiologists interpreted 12 images with interstitial lung disease under six conditions of CRT luminance (50 and 400 cd/m(2)) and room illuminance (20, 120 and 480lx), and 10 radiologists interpreted 25 images with pulmonary nodules under nine conditions of CRT luminance (50, 200 and 500 cd/m(2)) and room illuminance (20, 120 and 480lx). Observers performance for interstitial disease was relatively better at 120lx. Four hundred and eighty lux illuminance with 50 cd/m(2) CRT luminance, which degraded the detectability of pulmonary nodule significantly (p<0.05), should be avoided for clinical use.

Detection of subtle pulmonary disease on CR chest images: monochromatic CRT monitor vs color CRT monitor

Abstract To clarify the diagnostic efficacy of color soft-copy computed radiographic (CR) images of the chest in the detection of subtle pulmonary abnormalities. Twenty observers compared 87 soft-copy CR images on four types of CRT monitor (nonmagnified monochromatic CRT, magnified monochromatic CRT, nonmagnified color CRT, and magnified color CRT). Of 87 test images, 45 (including two identical sets of 12 images to test intraobserver variability) were abnormal and 42 (including two identical sets of 12 images) were normal. Of the 45 abnormal images, 15 showed subtle abnormalities, 15 showed mild abnormalities, and 15 showed obvious abnormalities. In the receiver operating characteristic (ROC) analyses, there were no statistically significant differences among the four types of CRT display formats in the detection of subtle abnormalities. Color CRT monitors can replace monochromatic CRT monitors without any loss in the ability to detect subtle interstitial lung disease.

Interobserver agreement in sonographic diagnosis of breast tumors

OBJECTIVEnTo evaluate interobserver agreement in the interpretation of breast ultrasonography.nnnMETHODSn55 breast masses (30 benign, 24 malignant) were interpreted by seven radiologists using a CRT viewing station. US criteria for differentiating between benign and malignant lesions included shape, border, boundary echoes, internal echoes, posterior echoes, and bilateral shadows. Each criterion and the observers final impression was scored using the 5-point rating scales. For analyzing interobserver agreement, the kappa (kappa) values were employed.nnnRESULTSnThe kappa values of shape and posterior echoes were significantly higher than those of the other four criteria (P<0.05). Agreement was intermediate in border and internal echoes, and was low in boundary echoes and bilateral shadows. Agreement in the senior group (four observers) was relatively higher than that in the junior group (three observers) for all criteria but for internal echoes. Easily-diagnosed cases showed significantly higher kappa values compared with more ambiguous cases (P<0.05).nnnCONCLUSIONnInterobserver agreement in shape and posterior echoes was significantly higher than those of the other four criteria. Agreement was significantly dependent on case difficulty.

Interobserver agreement and performance score comparison in quality control using a breast phantom: screen-film mammography vs computed radiography

Abstract. Our objective was to evaluate interobserver agreement and to compare the performance score in quality control of screen-film mammography and computed radiography (CR) using a breast phantom. Eleven radiologists interpreted a breast phantom image (CIRS model X) by four viewing methods: (a) original screen-film; (b) soft-copy reading of the digitized film image; (c) hard-copy reading of CR using an imaging plate; and (d) soft-copy reading of CR. For the soft-copy reading, a 17-in. CRT monitor (1024×1536×8xa0bits) was used. The phantom image was evaluated using a scoring system outlined in the instruction manual, and observers judged each object using a three-point rating scale: (a) clearly seen; (b) barely seen; and (c) not seen. For statistical analysis, the kappa statistic was employed. For mass depiction, interobserver agreement using CR was significantly lower than when using screen-film (p<0.05). There was no significant difference in the kappa value for detecting microcalcification; however, the performance score of microcalcification on CR hard-copy was significantly lower than on the other three viewing methods (p<0.05). Viewing methods (film or CR, soft-copy or hard-copy) could affect how the phantom image is judged. Paying special attention to viewing conditions is recommended for quality control of CR mammograms.

Display method can affect interobserver agreement: comparison of 'zoom-and-pan' and 'browse-and-paste' for primary CT interpretation.

Eight radiologists interpreted body CT images of 30 cases using a viewing station (six 17-in. monitors, 1024x1280). Using two different display methods, zoom-and-pan and browse-and-paste, the readers described the presence or absence of liver tumors using a five-point rating scale and temporal changes between the current and previous studies using a seven-point rating scale. There was no significant difference in kappa values for tumor detection between the two display modes. However, in describing temporal changes, the kappa value of the browse-and-paste was significantly lower than that of zoom-and-pan (p<0.01). Browse-and-paste may have the disadvantage of greater interobserver variation.