Alternative Free-response Receiver Operating Characteristic Curve Research Articles

Bone Suppression on Chest Radiographs for Pulmonary Nodule Detection: Comparison between a Generative Adversarial Network and Dual-Energy Subtraction

ObjectiveTo compare the effects of bone suppression imaging using deep learning (BSp-DL) based on a generative adversarial network (GAN) and bone subtraction imaging using a dual energy technique (BSt-DE) on radiologists’ performance for pulmonary nodule detection on chest radiographs (CXRs).Materials and MethodsA total of 111 adults, including 49 patients with 83 pulmonary nodules, who underwent both CXR using the dual energy technique and chest CT, were enrolled. Using CT as a reference, two independent radiologists evaluated CXR images for the presence or absence of pulmonary nodules in three reading sessions (standard CXR, BSt-DE CXR, and BSp-DL CXR). Person-wise and nodule-wise performances were assessed using receiver-operating characteristic (ROC) and alternative free-response ROC (AFROC) curve analyses, respectively. Subgroup analyses based on nodule size, location, and the presence of overlapping bones were performed.ResultsBSt-DE with an area under the AFROC curve (AUAFROC) of 0.996 and 0.976 for readers 1 and 2, respectively, and BSp-DL with AUAFROC of 0.981 and 0.958, respectively, showed better nodule-wise performance than standard CXR (AUAFROC of 0.907 and 0.808, respectively; p ≤ 0.005). In the person-wise analysis, BSp-DL with an area under the ROC curve (AUROC) of 0.984 and 0.931 for readers 1 and 2, respectively, showed better performance than standard CXR (AUROC of 0.915 and 0.798, respectively; p ≤ 0.011) and comparable performance to BSt-DE (AUROC of 0.988 and 0.974; p ≥ 0.064). BSt-DE and BSp-DL were superior to standard CXR for detecting nodules overlapping with bones (p < 0.017) or in the upper/middle lung zone (p < 0.017). BSt-DE was superior (p < 0.017) to BSp-DL in detecting peripheral and sub-centimeter nodules.ConclusionBSp-DL (GAN-based bone suppression) showed comparable performance to BSt-DE and can improve radiologists’ performance in detecting pulmonary nodules on CXRs. Nevertheless, for better delineation of small and peripheral nodules, further technical improvements are required.

Evaluation of Weighted Diffusion Subtraction for Detection of Clinically Significant Prostate Cancer.

Diffusion-weighted imaging (DWI) is an important method for clinically significant prostate cancer (csPCa) diagnosis; however, the Prostate Imaging-Reporting and Data System (PI-RADS) requires the subjective assessment of "markedly hypointense or not" on apparent diffusion coefficient (ADC) map. We hypothesize that weighted diffusion subtraction (WDS) images, created by weighted subtraction of high and low b-value DWIs, might better show areas of ADC values below a set threshold, thus decreasing the subjectivity of the assessment. To evaluate the diagnostic ability of WDS for csPCa by comparing scores based on WDS images (DWI/WDS) with those based on PI-RADS DWI (DWI/ADC). Retrospective. Eighty-six PCa patients. 3.0 T; DWI. Four readers assessed the probability of csPCa in lesions (overall, in the peripheral zone [PZ] and transitional zone [TZ]) using 5-point DWI/ADC and DWI/WDS scores. Prostatectomy specimens were the reference standard. ADC values and contrast between csPCa and normal prostate tissue on ADC maps and WDS images were calculated with reference to the pathological map. Diagnostic ability was evaluated by Jackknife alternative free-response receiver-operating characteristic curve. Figure of merit (FOM), sensitivity, and positive predictive value (PPV) between the DWI/ADC and DWI/WDS scores were compared using paired t-test. Inter-reader agreement was analyzed using κ statistics, and the significance probability was calculated using the Z statistic. Wilcoxon signed-rank test was used to compare contrast between csPCa and normal prostate tissue on ADC maps and WDS images. A P value <0.05 was considered statistically significant. FOM and sensitivity of the DWI/WDS scores were significantly better than those of the DWI/ADC scores overall, in the PZ and TZ (FOM: overall, 0.715 vs. 0.783; PZ, 0.756 vs. 0.815; TZ, 0.653 vs. 0.738. Sensitivity: overall, 0.512 vs. 0.607; PZ, 0.485 vs. 0.573; TZ, 0.636 vs. 0.761). For PPV, a statistically significant difference was observed overall (0.727 vs. 0.777). The κ value of DWI/WDS score was significantly higher than that of DWI/ADC score overall and in the PZ (overall, 0.614 vs. 0.792; PZ, 0.609 vs. 0.797). Contrast was significantly higher overall in the PZ and TZ in WDS images (median, 1.26, 1.19, and 1.61) than in ADC maps (0.46, 0.47, and 0.41). WDS images performed better than ADC maps in the diagnosis of csPCa and in inter-reader agreement of the diagnosis. 4 Technical Efficacy Stage: 2.

Development and Validation of a Deep Learning-Based Automated Detection Algorithm for Major Thoracic Diseases on Chest Radiographs

Background: Interpretation of chest radiographs (CRs) is a challenging task prone to errors, requiring expert readers. An automated system that can accurately classify CRs may help streamline the clinical workflow. Methods: We developed a deep-learning-based automatic detection (DLAD) algorithm for classification of CRs with major thoracic diseases including pulmonary malignancy, active pulmonary tuberculosis, pneumonia, and pneumothorax, using 54,221 normal CRs and 47,917 abnormal CRs, which were labeled and annotated by board certified radiologists. Performance of DLAD was validated using five independent datasets from different institutions. To compare the performances of DLAD with physicians, an observer performance test was conducted by 15 physicians including non-radiology physicians, board-certified radiologists, and thoracic radiologists. Image-wise classification performances and lesion-wise localization performances were measured by area under the receiver operating characteristic (ROC) curves, and area under the alternative free-response ROC curves, respectively. Findings: DLAD demonstrated classification performance of 0*973-1*000 and localization performance of 0*923-1*000 in independent datasets. DLAD demonstrated significantly higher performance than all three physician groups in both image-wise classification (0*983 vs. 0*814-0*932) and lesion-wise localization (0*985 vs. 0*781-0*907). Significant improvements in both image-wise classification (performance increment, 0*026-0*090) and lesion-wise localization (performance increment, 0*031-0*092) were observed in all three physician groups with assistance of the algorithm. Interpretation: Our DLAD demonstrated excellent and consistent performance in the detection of major thoracic diseases on CR, outperforming even thoracic radiologists. Funding Statement: The present study was supported by the Seoul National University Hospital Research fund (grant number: 04- 2016-3000), Lunit Inc. and the Seoul Research & Buisiness Development Program (grant number: FI170002). Declaration of Interests: Chang Min Park reports grants from Seoul National University Hospital, Seoul Metropolitan Government, and Lunit Inc., during the conduct of the study; Sunggyun Park and Jaehong Aum are employees of Lunit Inc. Ethics Approval Statement: The present study was approved by the institutional review boards of all participating institutions. Patients’ informed consents were waived.

Development and Validation of a Deep Learning-based Automatic Detection Algorithm for Active Pulmonary Tuberculosis on Chest Radiographs.

BackgroundDetection of active pulmonary tuberculosis on chest radiographs (CRs) is critical for the diagnosis and screening of tuberculosis. An automated system may help streamline the tuberculosis screening process and improve diagnostic performance.MethodsWe developed a deep learning–based automatic detection (DLAD) algorithm using 54c221 normal CRs and 6768 CRs with active pulmonary tuberculosis that were labeled and annotated by 13 board-certified radiologists. The performance of DLAD was validated using 6 external multicenter, multinational datasets. To compare the performances of DLAD with physicians, an observer performance test was conducted by 15 physicians including nonradiology physicians, board-certified radiologists, and thoracic radiologists. Image-wise classification and lesion-wise localization performances were measured using area under the receiver operating characteristic (ROC) curves and area under the alternative free-response ROC curves, respectively. Sensitivities and specificities of DLAD were calculated using 2 cutoffs (high sensitivity [98%] and high specificity [98%]) obtained through in-house validation.ResultsDLAD demonstrated classification performance of 0.977–1.000 and localization performance of 0.973–1.000. Sensitivities and specificities for classification were 94.3%–100% and 91.1%–100% using the high-sensitivity cutoff and 84.1%–99.0% and 99.1%–100% using the high-specificity cutoff. DLAD showed significantly higher performance in both classification (0.993 vs 0.746–0.971) and localization (0.993 vs 0.664–0.925) compared to all groups of physicians.ConclusionsOur DLAD demonstrated excellent and consistent performance in the detection of active pulmonary tuberculosis on CR, outperforming physicians, including thoracic radiologists.

Utility as a rationale for choosing observer performance assessment paradigms for detection tasks in medical imaging

Studies of lesion detectability are often carried out to evaluate medical imaging technology. For such studies, several approaches have been proposed to measure observer performance, such as the receiver operating characteristic (ROC), the localization ROC (LROC), the free-response ROC (FROC), the alternative free-response ROC (AFROC), and the exponentially transformed FROC (EFROC) paradigms. Therefore, an experimenter seeking to carry out such a study is confronted with an array of choices. Traditionally, arguments for different approaches have been made on the basis of practical considerations (statistical power, etc.) or the gross level of analysis (case-level or lesion-level). This article contends that a careful consideration of utility should form the rationale for matching the assessment paradigm to the clinical task of interest. In utility theory, task performance is commonly evaluated with total expected utility, which integrates the various event utilities against the probability of each event. To formalize the relationship between expected utility and the summary curve associated with each assessment paradigm, the concept of a "natural" utility structure is proposed. A natural utility structure is defined for a summary curve when the variables associated with the summary curve axes are sufficient for computing total expected utility, assuming that the disease prevalence is known. Natural utility structures for ROC, LROC, FROC, AFROC, and EFROC curves are introduced, clarifying how the utilities of correct and incorrect decisions are aggregated by summary curves. Further, conditions are given under which general utility structures for localization-based methodologies reduce to case-based assessment. Overall, the findings reveal how summary curves correspond to natural utility structures of diagnostic tasks, suggesting utility as a motivating principle for choosing an assessment paradigm.

Correlation of free‐response and receiver‐operating‐characteristic area‐under‐the‐curve estimates: Results from independently conducted FROC/ROC studies in mammography

From independently conducted free-response receiver operating characteristic (FROC) and receiver operating characteristic (ROC) experiments, to study fixed-reader associations between three estimators: the area under the alternative FROC (AFROC) curve computed from FROC data, the area under the ROC curve computed from FROC highest rating data, and the area under the ROC curve computed from confidence-of-disease ratings. Two hundred mammograms, 100 of which were abnormal, were processed by two image-processing algorithms and interpreted by four radiologists under the FROC paradigm. From the FROC data, inferred-ROC data were derived, using the highest rating assumption. Eighteen months afterwards, the images were interpreted by the same radiologists under the conventional ROC paradigm; conventional-ROC data (in contrast to inferred-ROC data) were obtained. FROC and ROC (inferred, conventional) data were analyzed using the nonparametric area-under-the-curve (AUC), (AFROC and ROC curve, respectively). Pearson correlation was used to quantify the degree of association between the modality-specific AUC indices and standard errors were computed using the bootstrap-after-bootstrap method. The magnitude of the correlations was assessed by comparison with computed Obuchowski-Rockette fixed reader correlations. Average Pearson correlations (with 95% confidence intervals in square brackets) were: Corr(FROC, inferred ROC) = 0.76[0.64, 0.84] > Corr(inferred ROC, conventional ROC) = 0.40[0.18, 0.58] > Corr (FROC, conventional ROC) = 0.32[0.16, 0.46]. Correlation between FROC and inferred-ROC data AUC estimates was high. Correlation between inferred- and conventional-ROC AUC was similar to the correlation between two modalities for a single reader using one estimation method, suggesting that the highest rating assumption might be questionable.

Quantifying the clinical relevance of a laboratory observer performance paradigm

Laboratory observer performance measurements, receiver operating characteristic (ROC) and free-response ROC (FROC) differ from actual clinical interpretations in several respects, which could compromise their clinical relevance. The objective of this study was to develop a method for quantifying the clinical relevance of a laboratory paradigm and apply it to compare the ROC and FROC paradigms in a nodule detection task. The original prospective interpretations of 80 digital chest radiographs were classified by the truth panel as correct (C=1) or incorrect (C=0), depending on correlation with additional imaging, and the average of C was interpreted as the clinical figure of merit. FROC data were acquired for 21 radiologists and ROC data were inferred using the highest ratings. The areas under the ROC and alternative FROC curves were used as laboratory figures of merit. Bootstrap analysis was conducted to estimate conventional agreement measures between laboratory and clinical figures of merit. Also computed was a pseudovalue-based image-level correctness measure of the laboratory interpretations, whose association with C as measured by the area (rAUC) under an appropriately defined relevance ROC curve, is as a measure of the clinical relevance of a laboratory paradigm. Low correlations (e.g. κ=0.244) and near chance level rAUC values (e.g. 0.598), attributable to differences between the clinical and laboratory paradigms, were observed. The absolute width of the confidence interval was 0.38 for the interparadigm differences of the conventional measures and 0.14 for the difference of the rAUCs. The rAUC measure was consistent with the traditional measures but was more sensitive to the differences in clinical relevance. A new relevance ROC method for quantifying the clinical relevance of a laboratory paradigm is proposed.

Diagnostic accuracy of C-arm CT during selective transcatheter angiography for hepatocellular carcinoma: comparison with intravenous contrast-enhanced, biphasic, dynamic MDCT

This study was aimed to compare the accuracy, sensitivity, and positive predictive value of C-arm CT (CACT) during selective transcatheter angiography with those of multidetector CT (MDCT) in the detection of hepatocellular carcinoma (HCC). In this prospective study, 30 patients (mean age, 73 years) with unresectable HCC were examined with CACT before chemoembolisation. Images of a combination of CACT during arterial portography (CACTAP) and dual-phase CACT during hepatic arteriography (CACTHA) was obtained and images of intravenous contrast-enhanced, biphasic, dynamic, MDCT was also obtained beforehand. Three blinded observers independently reviewed CACT and MDCT. Diagnostic accuracy was evaluated by the alternative free-response receiver operating characteristic (AFROC) method. Sensitivities and positive predictive values (PPV) were analyzed with the paired t-test. In the mean area under the AFROC curve (Az), there was no significant difference between MDCT and CACT (MDCT, mean Az value, 0.83; CACT, 0.85, respectively) (P = 0.32). There was also no significant difference between the two techniques in sensitivity (MDCT, mean 0.65; CACT, 0.60) and PPV (MDCT, mean 0.98; CACT, 0.97) (P = 0.40, P = 0.68, respectively). The diagnostic accuracy of CACT was equivalent to that of biphasic CT in the diagnosis of HCC. C-arm CT helps detection of hepatocellular carcinoma (HCC) during interventional (TACE) treatment. C-arm CT for HCC seemed just as accurate as biphasic CT. TACE can be performed with greater confidence using C-arm CT.

Ideal AFROC and FROC Observers

Detection of multiple lesions in images is a medically important task and free-response receiver operating characteristic (FROC) analyses and its variants, such as alternative FROC (AFROC) analyses, are commonly used to quantify performance in such tasks. However, ideal observers that optimize FROC or AFROC performance metrics have not yet been formulated in the general case. If available, such ideal observers may turn out to be valuable for imaging system optimization and in the design of computer aided diagnosis techniques for lesion detection in medical images. In this paper, we derive ideal AFROC and FROC observers. They are ideal in that they maximize, amongst all decision strategies, the area, or any partial area, under the associated AFROC or FROC curve. Calculation of observer performance for these ideal observers is computationally quite complex. We can reduce this complexity by considering forms of these observers that use false positive reports derived from signal-absent images only. We also consider a Bayes risk analysis for the multiple-signal detection task with an appropriate definition of costs. A general decision strategy that minimizes Bayes risk is derived. With particular cost constraints, this general decision strategy reduces to the decision strategy associated with the ideal AFROC or FROC observer.

Computed Radiography-based Mammography with 50-μm Pixel Size: Intra-individual Comparison with Film-screen Mammography for Diagnosis of Breast Cancers

The aim of this study was to evaluate the breast cancer diagnostic capability of "dual-side readout" computed radiography-based mammography (DRCRM) with a 50-microm pixel size compared to that of conventional film-screen mammography (FSM). Thirty patients who were scheduled for surgical treatment for breast cancer and 10 normal volunteers were enrolled. All 30 patients underwent surgical treatment, and breast cancer was proved histopathologically. Twenty-eight patients had 35 invasive carcinomas, and the remaining two had ductal carcinomas in situ. Each of the 40 women underwent both DRCRM and FSM (with double exposure and the same view, without removing compression). Three observers retrospectively interpreted the mammograms independently and evaluated and rated masses and class categories. The accuracy of the detection of masses was evaluated with alternative free-response receiver-operating characteristic analysis. Sensitivity for the detection of masses and of cancers was also evaluated. The mean areas under the alternative free-response receiver-operating characteristic curves in the detection of the masses were 0.88 for DRCRM and 0.91 for FSM (P = .08). The corresponding values for mean sensitivity for the detection of masses were 0.74 and 0.77 (P = .48) and those for the detection of cancers 0.79 and 0.84 (P = .20). No significant differences were observed between DRCRM and FSM for diagnosis of breast cancers.

Diagnosis of Unruptured Intracranial Aneurysms: 3T MR Angiography versus 64-channel Multi-detector Row CT Angiography

We compared 3-dimensional time-of-flight magnetic resonance angiography (3D TOF MRA) using a 3-tesla (T) MR unit with 64-channel multi-detector row computed tomographic angiography (64-MDCTA) for detection and characterization of angiographically proven unruptured intracranial aneurysms. Thirty-eight patients with 47 aneurysms and 8 patients without aneurysms underwent 3T, 3D TOF MRA; 64-MDCTA; and intra-arterial angiography. As a first study, 3 radiologists blinded to pertinent clinical information independently reviewed MRA and CTA images. We evaluated diagnostic accuracy using an alternative free-response receiver operating characteristic (AFROC) analysis and evaluated the sensitivity and specificity of each technique. Next, 2 radiologists used volume-rendering images generated from MRA or CTA data to evaluate the morphology of the 47 aneurysms detected, and MRA and CTA results were compared. Three-dimensional digital angiography (DA) images were used as the standard of reference. On the AFROC analysis, the value of the mean area under the AFROC curve (A(1)) was 0.91 for both modalities. Mean sensitivity of 89% and specificity of 76% for MRA were not significantly different from sensitivity of 87% and specificity of 79% for CTA. Therefore, when used to evaluate aneurysmal morphology, both modalities appear satisfactory for determining these vascular anomalies. Three-tesla, 3D TOF MRA and 64-MDCTA are excellent modalities with high diagnostic accuracy for evaluating unruptured intracranial aneurysms and no significant difference between them in diagnostic performance.

Detection of Hepatocellular Carcinoma with Ferucarbotran (Resovist)-enhanced Breath-hold MR Imaging: Feasibility of 10 minute-delayed Images

We evaluated the optimal timing for breath-hold MR imaging with bolus-injectable superparamagnetic iron oxide (SPIO) for detecting hepatocellular carcinoma (HCC). Twenty patients with 62 HCCs (52 hypervascular, 10 non-hypervascular) underwent MR imaging that included unenhanced and SPIO-enhanced T1-weighted gradient echo (GRE) and T2-weighted fast spin echo (FSE) sequences, perfusion study, and SPIO-enhanced T2*-weighted GRE sequences. We obtained SPIO-enhanced T2*-weighted sequences 10 and 30 min after injecting SPIO and made 2 image sets, comprising 10- or 30-min delayed T2*-weighted images. Three observers performed alternative free response receiver operating characteristic (AFROC) analysis, and quantitative evaluation was performed. Only Observers 2 and 3 recognized a significant difference in the area under the AFROC curve (Az) value in the 10-min delayed images; no significant difference was observed in the 30-min delayed images. There was no significant difference in the sensitivity of individual observers between 10- and 30-min delayed images. The contrast-to-noise (C/N) ratio of the 30-min delayed images was significantly higher than that of the 10-min delayed images. The C/N ratio of hypervascular HCCs in the 30-min delayed images was significantly higher than in the 10-min delayed images, but that of non-hypervascular HCCs showed no significant difference. In most cases, 10-min delayed SPIO-enhanced T2*-weighted images are sufficient to detect HCCs.

Hepatic Metastases: Detection with Multi–Detector Row CT, SPIO-enhanced MR Imaging, and Both Techniques Combined

To retrospectively compare the accuracy in detection of hepatic metastases among contrast material-enhanced multi-detector row computed tomography (CT) alone, superparamagnetic iron oxide (SPIO)-enhanced magnetic resonance (MR) imaging alone, and a combination of contrast-enhanced CT and SPIO-enhanced MR imaging. The ethics committee did not require its approval or informed consent for this retrospective study, which was compliant with Declaration of Helsinki principles. Data in 38 patients (22 men, 16 women; mean age, 64.5 years; range, 35-78 years) suspected of having hepatic metastases who underwent both contrast-enhanced CT and SPIO-enhanced MR imaging were retrospectively analyzed. Twenty-one of the 38 patients had 61 metastases. Seventeen of the 61 metastases were confirmed histologically; the remaining 44 metastases were defined with imaging follow-up. At MR imaging, SPIO-enhanced heavily T1-weighted images, T2*-weighted gradient echo images, and T2-weighted fast spin-echo images were evaluated. Contrast-enhanced multi-detector row CT images obtained in the portal phase were evaluated. Four blinded observers independently reviewed CT images, MR images, and the combination of CT and MR images. Diagnostic accuracy was evaluated by using the alternative free-response receiver operating characteristic (AFROC) method. Sensitivities and positive predictive values were also analyzed with the Fisher protected least significant difference test and generalized estimating equations. The mean area under the AFROC curve for the combined approach (0.70) was significantly higher than that for SPIO-enhanced MR imaging alone (0.58, P < .05, Fisher protected least significant difference test), and there was no significant difference between each of them and that for contrast-enhanced CT alone (0.66). For all lesions, the mean sensitivity of combined imaging (0.59) was significantly higher than that of CT (0.48) or MR imaging (0.43) alone (P < .05, Fisher protected least significant difference test and generalized estimating equations). For all lesions, the mean positive predictive values were 0.82, 0.89, and 0.81, for combined MR and CT, CT alone, and MR alone, respectively. The addition of SPIO-enhanced MR imaging to contrast-enhanced multi-detector row CT (ie, combined analysis of SPIO-enhanced MR images and contrast-enhanced CT images) can improve sensitivity in the detection of hepatic metastases, although this improvement in sensitivity was not significant at AFROC analysis.

Assessment of Two 3D MDCT Colonography Protocols for Observation of Colorectal Polyps

The objective of our study was to assess the value of two-way interpretation (i.e., from rectum to cecum and vice versa) compared with one-way interpretation (i.e., from rectum to cecum only) in terms of polyp detection and interpretation time on MDCT colonography. Fifty consecutive patients underwent both CT colonography and conventional colonoscopy. Three radiologists independently analyzed the CT colonographic examinations of each patient using a primary 3D method. All examinations were analyzed using two techniques: navigation from rectum to cecum only (one-way) and navigation from rectum to cecum and vice versa (two-way). Sensitivity and positive predictive value were calculated on both a per-polyp basis and a per-patient basis. Alternative free-response receiver operating characteristic (ROC) curve analysis was estimated, and image interpretation time was documented. One hundred fifty-five polyps were depicted in 45 patients by colonoscopy. The mean sensitivity of CT colonography for polyp detection with two-way (88.4%) was significantly superior to that with one-way (78.1%) (p < 0.01). The mean positive predictive value of each observer with one-way was 66.7%, whereas that with two-way was 65.8%. The mean area under the alternative free-response ROC curve (A(z) value) with two-way (0.827) was higher than that with one-way (0.816), but there was not a statistically significant difference. The average interpretation time of each observer with two-way (39 min) was statistically significantly longer than that with one-way (25 min) (p < 0.01). When using a primary 3D interpretation technique at CT colonography, complete 3D navigation from rectum to cecum and from cecum to rectum is mandatory to maximize polyp detection. The image interpretation time for two-way interpretation is statistically significantly longer than that with one-way interpretation.

Superparamagnetic iron oxide-enhanced MR imaging for focal hepatic lesions: a comparison with CT during arterioportography plus CT during hepatic arteriography

We aimed to evaluate the efficacy of a breath-hold superparamagnetic iron oxide (SPIO)-enhanced magnetic resonance (MR) imaging protocol for the detection of focal hepatic lesions, in comparison with a non-breath-hold SPIO-enhanced imaging protocol and computed tomography during arterioportography (CTAP) plus CT during hepatic arteriography (CTHA). Findings of SPIO-enhanced MR imaging and CTAP/CTHA for 24 hepatic metastases in 17 patients and 29 HCCs in 21 patients were analyzed. All patients underwent breath-hold SPIO-enhanced MR imaging (1.5 tesla), breath-hold plus non-breath-hold SPIO-enhanced MR imaging, and CTAP plus CTHA prior to partial hepatectomy or laparoscopic ablation therapy. Histopathology for lesion characterization and intraoperative ultrasound for lesion detection were available for all patient. Breath-hold SPIO-enhanced MR imaging consisted of T2-weighted single-short fast spin echo (FSE), T2-weighted (T2W) FSE, T2*-weighted gradient echo (GRE), and T1-weighted GRE. For the non-breath-hold imaging protocol, respiratory-triggered, fat-suppressed T2W-FSE was added to the breath-hold MR imaging protocol. Double phase CTAP plus CTHA was performed on an angio-CT system. To compare the three imaging protocols, three radiologists performed blind film reading, and all data, on a hepatic segment-to-segment basis, were entered for alternative free-response receiver-operating characteristic (AFROC) analysis. ROC analysis showed that there was no significant difference in the area under the AFROC curve (A1) value for metastases and HCCs among the three protocols; the breath-hold SPIO-enhanced MR imaging protocol, non-breath-hold MR imaging protocol, and CTAP plus CTHA. The breath-hold SPIO-enhanced MR imaging protocol showed a sensitivity, specificity, and accuracy equivalent to the non-breath-hold MR imaging protocol and CTAP plus CTHA. As a preoperative test, SPIO-enhanced MR imaging could have the potential to replace CTAP plus CTHA in a certain clinical setting.

Triple Arterial Phase Dynamic MRI with Sensitivity Encoding for Hypervascular Hepatocellular Carcinoma: Comparison of the Diagnostic Accuracy Among the Early, Middle, Late, and Whole Triple Arterial Phase Imaging

We assessed and compared the diagnostic accuracy of the early, middle, late, and whole triple arterial phase MRI with sensitivity encoding (SENSE) for the detection of hypervascular hepatocellular carcinoma (HCC). Thirty-one patients with 102 HCCs underwent dynamic MRI with SENSE. The findings of CT examinations, combined with those of visceral angiography or histopathologic examination, were used as the gold standard. After acquisition of T1- and T2-weighted images, gadolinium-enhanced triple arterial, portal, and delayed phase images were obtained. Acquisition of the triple arterial phase imaging was started at the timing of peak aortic enhancement and completed within a single breath-hold. Acquisition time for each phase was 8.4 sec. Four image sets including the early, middle, late, and whole triple arterial phase imaging were interpreted separately by four observers. The mean values of area under alternative-free-response receiver operating characteristic (AFROC) curve and of sensitivity were compared among the four image sets. The mean values of area under AFROC curve were 0.52, 0.66, 0.53, and 0.68 and of sensitivity were 45%, 64%, 48%, and 65% for the image sets with the early, middle, late, and whole triple arterial phase imaging, respectively. Both mean values were significantly higher for the image sets with the middle and whole triple arterial phase imaging than for those with the early and late arterial phase imaging. The middle arterial phase imaging with k-space centered at 12.6 sec after the peak aortic enhancement was optimal for detecting HCC and showed diagnostic accuracy equivalent to that of the whole triple arterial phase imaging.

Pulmonary Nodules on Multi–Detector Row CT Scans: Performance Comparison of Radiologists and Computer-aided Detection

To compare the performance of radiologists and of a computer-aided detection (CAD) algorithm for pulmonary nodule detection on thin-section thoracic computed tomographic (CT) scans. The study was approved by the institutional review board. The requirement of informed consent was waived. Twenty outpatients (age range, 15-91 years; mean, 64 years) were examined with chest CT (multi-detector row scanner, four detector rows, 1.25-mm section thickness, and 0.6-mm interval) for pulmonary nodules. Three radiologists independently analyzed CT scans, recorded the locus of each nodule candidate, and assigned each a confidence score. A CAD algorithm with parameters chosen by using cross validation was applied to the 20 scans. The reference standard was established by two experienced thoracic radiologists in consensus, with blind review of all nodule candidates and free search for additional nodules at a dedicated workstation for three-dimensional image analysis. True-positive (TP) and false-positive (FP) results and confidence levels were used to generate free-response receiver operating characteristic (ROC) plots. Double-reading performance was determined on the basis of TP detections by either reader. The 20 scans showed 195 noncalcified nodules with a diameter of 3 mm or more (reference reading). Area under the alternative free-response ROC curve was 0.54, 0.48, 0.55, and 0.36 for CAD and readers 1-3, respectively. Differences between reader 3 and CAD and between readers 2 and 3 were significant (P < .05); those between CAD and readers 1 and 2 were not significant. Mean sensitivity for individual readings was 50% (range, 41%-60%); double reading resulted in increase to 63% (range, 56%-67%). With CAD used at a threshold allowing only three FP detections per CT scan, mean sensitivity was increased to 76% (range, 73%-78%). CAD complemented individual readers by detecting additional nodules more effectively than did a second reader; CAD-reader weighted kappa values were significantly lower than reader-reader weighted kappa values (Wilcoxon rank sum test, P < .05). With CAD used at a level allowing only three FP detections per CT scan, sensitivity was substantially higher than with conventional double reading.

Dual Phase Helical CT Versus Portal Venous Phase CT For the Detection of Colorectal Liver Metastases: Correlation with Intra-operative Sonography, Surgical and Pathological Findings

AIM: To assess whether dual phase helical computed tomography (DPCT) of the liver improves the detection of colorectal liver metastases compared with portal venous phase (PVP) imaging alone.MATERIALS AND METHODS: DPCT was performed in 33 consecutive patients before laparotomy for resection of colorectal liver metastases. CT comprised 8-mm slice collimation with a pitch of 1 to 1.25; imaging was commenced 20–25 and 65–70 s after the start of injection of 150ml of contrast medium at 5ml/s to coincide with hepatic arterial phase (HAP) and PVP contrast enhancement, respectively. Four blinded observers independently reviewed the HAP, PVP and DPCT images recording the site and size of all lesions. Alternative-free response receiver operating characteristic (AFROC) methodology was used to analyse the results, which were correlated with surgery, intra-operative ultrasound and histology.RESULTS: The mean observer sensitivities for malignant lesion detection were 75.3% for DPCT, 69.7% for PVP imaging and 66.7% for HAP imaging alone. There was a statistically significant improvement in malignant lesion detection using DPCT when compared with PVP imaging alone (P<0.05). The mean areas under the AFROC curves were 0.84 for DPCT and 0.82 for PVP (P<0.03) imaging alone.CONCLUSION: The detection of colorectal liver metastases was marginally better with DPCT than with PVP imaging alone, but the discovery of additional lesions did not affect the management of any of the patients in this study. Scott, D.J.et al.(2001). Clinical Radiology56, 235–242.