Search

Rationale and Objectives: Traditionally, multireader receiver operating characteristic (ROC) studies have used a "paired-case, paired-reader" design. The statistical power of such a design for inferences about the relative accuracies of the tests was assessed and compared with alternative designs., Methods: The noncentrality parameter of an F statistic was used to compute power as a function of the reader and patient sample sizes and the variability and correlation between readings., Results: For a fixed-power and Type I error rate, the traditional design reduces the number of verified cases required. A hybrid design, in which each reader interprets a different sample of patients, reduces the number of readers, total readings, and reading required per reader. The drawback is a substantial increase in the number of verified cases., Conclusion: The ultimate choice of study design depends on the nature of the tests being compared, limiting resources, a priori knowledge of the magnitude of the correlations and variability and logistic complexity.

Although the technology exists for accurate noninvasive screening for intracranial aneurysms, the efficacy of screening depends on several key parameters of the natural history of aneurysms. Recent studies suggest that the prevalence of intracranial aneurysms may reach 20% in the subpopulation of patients with a family history of these lesions; other key parameters are less certain. The authors investigated factors that impact the efficacy of screening to establish interim guidelines. Three plausible models for the natural history of aneurysms were constructed. For each model the monetary cost of screening and the average gain in life expectancy were computed for a range of screening ages and prevalence rates. It is shown that the efficacy of screening depends on the pattern of aneurysm rupture. If aneurysms develop and rupture rapidly, then screening has no benefit. On the other hand, if aneurysms remain at risk for some time after formation, then screening may improve average life expectancy depending on when it occurs. The authors recommend that patients with a positive family history of aneurysms who are 30 years of age or younger be screened. This recommendation is based on the belief that the gains attributable to screening, assuming a constant rupture rate, outweigh the losses attributable to screening using a decreasing rupture rate model.

Objective: The purpose of this study was to determine the relative effects of delay time, contrast agent concentration, and contrast agent volume on hepatic enhancement and on aortic and portal vein attenuation for contrast-enhanced spiral CT of the abdomen. The effect of delay time was evaluated by comparing two different delay times for each of three IV contrast media. Three different media were tested to identify whether the effects were different with different iodine concentrations and doses or with different volumes of contrast medium., Subjects and Methods: A total of 169 patients who underwent contrast-enhanced CT of the abdomen with spiral CT imaging were evaluated for enhancement of the liver parenchyma and for postenhancement attenuation of the aorta and portal vein. Total iodine doses of 30 g as 100 ml of iohexol 300 or 125 ml of ioversol 240 and of 40 g as 125 ml of ioversol 320 were evaluated at delay times of 45 and 60 sec and with a low-flow monophasic injection at a rate of 2 ml/sec., Results: For each contrast medium tested, a 60-sec delay provided greater hepatic enhancement than did a 45-sec delay (43 vs 32 H for iohexol 300, 37 vs 28 H for ioversol 240, and 50 vs 39 H for ioversol 320; all comparisons, p < or = .0001). The aortic attenuation achieved in the abdomen after contrast medium administration was significantly greater with a 45-sec delay than with a 60-sec delay for iohexol 300 (p = .0290) but not for the other contrast media. The portal vein attenuation achieved after contrast medium administration was significantly greater with a 60-sec delay than with a 45-sec delay for all three contrast media (p = .0051)., Conclusion: When IV contrast material was injected with a low-flow-rate monophasic bolus, a delay time of 60 sec, longer than that typically used for dynamic incremental scanning, provided greater hepatic enhancement in spiral CT. This result was significant for the three contrast media tested. Aortic attenuation was significantly greater at shorter delay times when a smaller volume of contrast material was injected. Portal vein attenuation was greater at longer delay times, suggesting that the optimal delay time may be longer than 60 sec. Further work to determine the optimal scanning window for hepatic enhancement in spiral CT by use of time-density curves is indicated.

Objective: The purpose of this study was to compare the cost-effectiveness of sonography and CT for the evaluation of renal masses discovered at excretory urography., Materials and Methods: The records of 225 patients with a renal mass shown by urography who then had either sonography or CT within 3 months were reviewed retrospectively. The number, location, and size of lesions; initial and subsequent imaging tests; and final diagnoses were determined. Using the current Medicare reimbursements for sonography and CT, we calculated the economic implications of using sonography or CT as the initial examination. Any effect of the location or size of the lesion on the most cost-effective examination was also determined., Results: Twenty-one percent of patients had both initial sonography and follow-up CT because of indeterminate findings or detection of a solid mass that required further staging. When CT was done first, CT findings were equivocal in 12%, necessitating follow-up sonography. At the prevailing charges, CT would have to be needed in 70% of patients initially imaged with sonography to justify the use of CT as the initial examination. The location and size of the lesion did not affect the need for CT at a rate (greater than 70%) that would economically justify use of CT as the first imaging test. Eighty-six percent of patients with a mass detected by urography had either a simple cyst or no evidence of a mass on sonography or CT., Conclusion: Sonography is the most cost-effective imaging method for the workup of a renal mass detected at urography. The number of sonographic examinations in which findings are indeterminate or positive (for a solid mass) is not sufficiently high to warrant replacement of sonography by CT, regardless of the size and location of the lesion. CT should be reserved for a limited number of specific indications.

Objective: Use of MRA for thoracic aortic disease (TAD) evaluation has been limited. This report describes an initial experience with TAD evaluation using a single MRA volume series., Materials and Methods: A single volume series, based on sequential 2D TOF MRA, was acquired in 30 cases (28 with suspected TAD and 2 normals). Each series was processed using multiplanar reconstruction (MPR) and maximum intensity projection (MIP); resulting tomographic (one base and two MPR) and MIP sets were blindly interpreted by four reviewers to detect TAD and, if present, to diagnose its specific form. For cases incorrectly interpreted, the standard MR images were subsequently interpreted., Results: The TAD categories included aneurysm (n = 13), dissection (n = 9), and arch anomalies (n = 5). Sensitivities were high for TAD overall (89-100%) and TAD in ascending and descending portions; sensitivities were lower for TAD of the arch (two of four reviewers > or = 90% for TAD overall and descending TAD). Specificities for TAD overall had a wider range (67-100%), but were high for ascending, arch, and descending portions (three to four of four reviewers > or = 90% for each). Sensitivities for aneurysms (69-92%) and dissections overall (67-100%) were comparable, as they were in ascending and arch portions; descending dissection was better detected than descending aneurysm (two of four reviewers > or = 90% for ascending or arch aneurysm and for descending dissection); overall specificities (88-100 vs. 81-95%) and specificities in ascending, arch, and descending portions were also comparable (three to four of four reviewers > or = 90% for both in each portion; two of four reviewers > or = 90% for dissection overall). Each reviewer achieved > or = 70% diagnostic accuracy for TAD (one of four reviewers = 85%); accuracies for each category were comparable. Interpretation of standard MR images corrected all detection and most diagnostic (> or = 63%) errors., Conclusion: This initial experience with conventional TOF MRA for TAD evaluation is encouraging, but it indicates the potential for advancements in data acquisition and/or postprocessing.

Purpose: To evaluate the clinical efficacy, cost-effectiveness, and safety of presurgical devascularization of meningiomas., Methods: Matched samples of embolized and nonembolized groups of meningiomas were compared. The study variables for clinical efficacy were estimated blood loss, number of transfusions, surgical resection time, and length of hospitalization. The cost-effectiveness was evaluated by adjusting all hospital costs to 1991 dollar amounts, and adding additional embolization costs and fees to the hospital cost totals for the embolized group. A qualitative comparison of complications was made., Results: All dependent variables evaluating the clinical efficacy of the procedure (estimate blood loss, 533 cc versus 836 cc; number of transfusions, 0.39 units versus 1.56 units; surgical resection time, 305.8 minutes versus 337.5 minutes; and length of hospitalization, 10.6 days versus 15.0 days) displayed trends of higher means in the nonembolized group; however, only the estimated blood loss and number of transfusions variables were significant. The cost-effectiveness of the procedure was not statistically significant. The mean cost was $29,605 for the embolized group and $38,449 for the nonembolized group. There were three major and nine minor complications in the nonembolized group and zero major and six minor complications in the embolized group. There were four additional minor complications caused by the embolization procedure., Conclusion: Endovascular devascularization of meningiomas is beneficial for large meningiomas because it diminishes the necessity of intraoperative transfusions and decreases blood loss. The additional day of hospitalization, emolization costs, and costs of complications do not conversely increase treatment costs. There were no major complications or adverse long-term effects caused by the embolization procedure.

Purpose: To determine whether fast spin-echo sequences could replace conventional spin-echo methods in the evaluation of head and neck neoplasms and associated adenopathy and to evaluate differences in tissue contrast characteristics between conventional spin-echo and fast spin-echo examinations of head and neck disease., Methods: Twenty-seven patients with squamous cell carcinoma were imaged on a 1.5-T imager with both conventional spin-echo and fast spin-echo sequences with identical section thickness and position. Twenty-one of the 27 fast spin-echo studies were performed with frequency-selective fat suppression. Three radiologists independently evaluated the images using a five-point scale to compare primary lesion margin definition and conspicuity, lymph node margin definition and conspicuity, gross motion artifact, and flow artifact. Quantitative percent contrast and contrast-to-noise ratios were calculated and compared in 7 cases with fat-suppressed fast spin-echo., Results: Fast spin-echo was preferred by all three readers for lesion margin conspicuity and lymph node conspicuity. Gross motion and flow artifact demonstrated trends toward reader preference for fast spin-echo. Quantitative contrast values for fast spin-echo were significantly greater than those for conventional spin-echo., Conclusions: Fast spin-echo with fat suppression can replace conventional spin-echo at a time savings of more than 50% and improves tissue contrast and the conspicuity and definition of margins for primary lesions and lymph nodes. Fat-suppression heterogeneity remains the major limitation of this technique. Thus, careful attention to fat-suppression failure and unwanted water saturation is essential.

Background: The relation between abnormalities in the lumbar spine and low back pain is controversial. We examined the prevalence of abnormal findings on magnetic resonance imaging (MRI) scans of the lumbar spine in people without back pain., Methods: We performed MRI examinations on 98 asymptomatic people. The scans were read independently by two neuroradiologists who did not know the clinical status of the subjects. To reduce the possibility of bias in interpreting the studies, abnormal MRI scans from 27 people with back pain were mixed randomly with the scans from the asymptomatic people. We used the following standardized terms to classify the five intervertebral disks in the lumbosacral spine: normal, bulge (circumferential symmetric extension of the disk beyond the interspace), protrusion (focal or asymmetric extension of the disk beyond the interspace), and extrusion (more extreme extension of the disk beyond the interspace). Nonintervertebral disk abnormalities, such as facet arthropathy, were also documented., Results: Thirty-six percent of the 98 asymptomatic subjects had normal disks at all levels. With the results of the two readings averaged, 52 percent of the subjects had a bulge at at least one level, 27 percent had a protrusion, and 1 percent had an extrusion. Thirty-eight percent had an abnormality of more than one intervertebral disk. The prevalence of bulges, but not of protrusions, increased with age. The most common nonintervertebral disk abnormalities were Schmorl's nodes (herniation of the disk into the vertebral-body end plate), found in 19 percent of the subjects; annular defects (disruption of the outer fibrous ring of the disk), in 14 percent; and facet arthropathy (degenerative disease of the posterior articular processes of the vertebrae), in 8 percent. The findings were similar in men and women., Conclusions: On MRI examination of the lumbar spine, many people without back pain have disk bulges or protrusions but not extrusions. Given the high prevalence of these findings and of back pain, the discovery by MRI of bulges or protrusions in people with low back pain may frequently be coincidental.

Rationale and Objectives: Hanley and McNeil (1982) proposed a nonparametric method for computing the standard error of the area under the receiver operating characteristic (ROC) curve. The method has been important in planning the minimum sample size for ROC studies. However, the validity of this method for rating data with various standard deviation ratios has not been investigated., Methods: A simulation study was conducted to compare the empirical standard error of the area under the curve with Hanley and McNeil's estimate over a range of parameters. An alternative method of computing the standard error based on a binormal distribution is proposed., Results: The method of Hanley and McNeil can lead to underestimation of the minimum sample size. The proposed method provides more appropriate estimates of sample size., Conclusions: When determining sample size for a study of the area under the ROC curve where rating data are used, the standard error estimator based on the binormal distribution should be used.

Purpose: To compare conventional T2-weighted spin-echo imaging with a rapid acquisition relaxation enhanced (RARE) technique in the routine evaluation of lumbar degenerative disk disease., Methods: Thirty consecutive patients referred for evaluation of the lumbar spine for suspected degenerative disk disease were evaluated with sagittal and axial T1-weighted spin-echo, conventional T2-weighted spin-echo, and T2-weighted RARE "turbo spin-echo" sequences (4000/93/2 (repetition time/echo time/excitations), 192 x 256, echo train length of 8). Conventional T2-weighted and RARE images were evaluated independently by two neuroradiologists for image quality, presence of artifacts, cerebrospinal fluid signal intensity, extradural interface conspicuity, intradural nerve root conspicuity, soft-tissue detail, and signal intensity of normal and degenerated intervertebral disks., Results: Both readers rated the cerebrospinal fluid signal higher, the extradural interface conspicuity higher, and the nerve root detail greater on the turbo spin-echo than on conventional spin-echo images. Neither reader had a significant difference in ranking "normal" or "degenerated" disk signal on the two sequences. Both readers rated soft-tissue detail higher for conventional than for turbo spin-echo., Conclusion: RARE sequences can replace conventional T2-weighted spin-echo sagittal studies for degenerative lumbar disk disease.

The authors undertook this study to compare bright and dark CSF three-dimensional (3D) gradient-echo (GE) MR techniques to answer the following questions: Could a single Gd-DTPA enhanced T1-weighted GE volume sequence (with multiplanar reformats) be diagnostically equivalent for degenerative cervical disk disease to a standard sequence consisting of sagittal T1-weighted spin echo and axial low flip angle volume GE images (with reformatted images)? Does performing oblique coronal reformats perpendicular to the course of exiting cervical nerve roots improve diagnostic confidence over axial images alone? Thirty-one consecutive patients received a "routine" MR examination consisting of a sagittal T1-weighted spin echo and axial low flip angle volume sequence (FISP) [(35/7/5), 64 slices, 2 mm slice thickness, 192 x 256 matrix, 7.2 min]. Each patient was then given 0.1 mmol/kg Gd-DTPA intravenously, and reimaged with a T1-weighted volume GE sequence [(13/6/12), acquired as 128-1.2 mm coronal partitions, 192 x 256 matrix, 5.5 min]. Sequences were reconstructed on the standard diagnostic console in 1 mm increments. Sets of examinations (routine vs T1-weighted volume) were independently interpreted by three neuroradiologists for location, type, and severity of extradural degenerative disease. There was no strong or consistent trend for increased detection of disease by one imaging sequence over the other. For lateral disk disease, only 3% of the observations were in discordance. For disk disease, there was close agreement in the severity scores. All readers indicated that additional information was provided by the reformatted images more frequently with TurboFLASH (fast low angle shot) than with FISP. All readers indicated that increased confidence was provided by the reformatted images more frequently with TurboFLASH than with FISP. A single 3D contrast-enhanced TurboFLASH sequence is diagnostically equivalent to a set of two-dimensional T1-weighted sagittal spin echo and 3D axial low flip angle sequences for assessing the location and degree of cervical extradural degenerative disease. A screening examination of the cervical spine could be performed with a single contrast-enhanced 5.2 min study, and then relying on computer postprocessing to provide additional imaging planes.

Three-dimensional (3D) volume turbo fast low angle shot (FLASH) techniques have become available which produce heavily T1-weighted images, similar to inversion recovery scans, utilizing the appropriate flip angle and inversion time. The purpose of this study was to compare the sensitivity of a rapid volume gradient echo technique [3D magnetization prepared rapid acquisition gradient echo (MP RAGE)] in identifying multiple sclerosis (MS) plaques with a conventional T2-weighted spin echo (SE) sequence. Ten patients with clinical MS were evaluated. Patients underwent a routine examination consisting of an axial T2-weighted SE sequence (2,500,22/90) and a coronal 3D MP RAGE, 10/4/10, acquired as 128 two mm partitions. In six patients, area measurements of 22 plaques were determined on both the axial T2-weighted SE examinations and the axial reformatted MP RAGE examinations. The overall number of plaques utilizing each technique was approximately the same. One hundred twenty-two plaques were visualized for the 3D MP RAGE sequence, and 128 plaques for the T2-weighted SE sequence. There were differences in detection of plaques in different regions, with plaques in gray matter better demonstrated utilizing the conventional T2-weighted SE sequence. Plaques in the corpus callosum, pons, and brachia pontis were better demonstrated utilizing 3D MP RAGE. No significant difference was found between the areas measured on the MP RAGE sequence and on the T2-weighted SE sequence. Three-dimensional MP RAGE provides a sensitive and complementary method to conventional T2-weighted SE sequences in the evaluation of patients with MS.

The effect of pulmonary artery (PA) catheter insertion on cardiac output measurements and the effect of single-lung ventilation on data derived from catheters placed in both the right and left pulmonary arteries were examined in 17 anesthetized mongrel dogs. Serial tricarboncyanine dye measurements of cardiac output were taken at baseline and after insertion of each of two PA catheters, one catheter in each lung, to detect possible changes in cardiac output due to instrumentation. In the second part of the study, single-lung ventilation was achieved through a Kottmeier double-lumen tube. Right and left paired values of thermodilution cardiac output, PA systolic pressure, and PA PO2 were compared during ventilation of both lungs, during right lung ventilation, during return to ventilation of both lungs, and during left lung ventilation. Mean dye-dilution cardiac outputs were not significantly different after one or two catheters were inserted. Paired right and left values were similar for thermodilution cardiac output, PA PO2, and pulmonary systolic pressure, regardless of ventilation conditions. Even during single-lung ventilation, data obtained from PA catheters in the ipsilateral and contralateral lung were equivalent in supine dogs.

Maximum likelihood estimation, assuming a binormal model, has been used extensively to estimate the area under receiver operating characteristic (ROC) curves and thus provide an index of diagnostic accuracy for the comparison of two imaging systems. However, in some instances, a degenerate data set results in a fitted ROC curve of inappropriate shape. In those instances, a nonparametric Wilcoxon statistic may be used to compare the areas under two ROC curves. Simulation of this approach indicates that the procedure has relatively high statistical power and its use in specific degenerate data sets shows its applications potential.

Receiver-operating characteristic (ROC) analysis has been used in many medical imaging applications during the past decade. In order to ensure that reader-confidence ratings are analyzable (well distributed to meet convergence requirements of curve-fitting algorithms) and meaningful (limit extrapolation of the data), many investigators train readers specifically for this purpose. No experimental data are available concerning the possible effects of such training on the results of ROC studies. We performed a multi-observer, multi-disease study in which 300 chest images were rated by four radiologists before and after they were trained to provide well-distributed confidence ratings. The results indicate that for our data set, reader and disease-specific accuracy was not significantly affected by the training process for interstitial disease and pneumothoraces. However, the accuracy of two readers was significantly affected for the detection of nodules (P less than 0.05), and the overall accuracy of one reader was significantly affected for the classification of normal versus abnormal images (P less than 0.01). Thus, in spite of the difficulties associated with the performance of ROC studies in a free-reading environment, one should carefully consider the possible effects of any intervention on the results prior to conducting ROC studies.

ROC methodology has been expanded in recent years to include multi-disease experiments. To accommodate these changes, different rating formats, general or disease specific, can be used. No experimental data are available concerning the possible effects of the rating format on the results of these studies. We performed a multi-observer, multi-disease study in which 196 chest images were rated using a format where each disease was evaluated individually and one in which the cases were evaluated without scoring a specific disease. The results indicate that for our data set, the overall assessment of accuracy was not significantly affected by the study format used. Thus, in spite of the difficulties in selecting appropriate controls and the necessity of reassessing sample size considerations, the disease-specific format appears to produce an assessment of accuracy equivalent to that produced by the general format. This equivalence permits the use of the disease-specific approach since it more closely simulates the readers' true environment and is more appropriate for comparing imaging systems that may have a relative accuracy that is disease specific.