Hepatic imaging with multidetector CT

Three-dimensional conditional generative adversarial network-based virtual thin-slice technique for the morphological evaluation of the spine

Virtual thin-slice (VTS) technique is a generative adversarial network-based algorithm that can generate virtual 1-mm-thick CT images from images of 3–10-mm thickness. We evaluated the performance of VTS technique for assessment of the spine. VTS was applied to 4-mm-thick CT images of 73 patients, and the visibility of intervertebral spaces was evaluated on the 4-mm-thick and VTS images. The heights of vertebrae measured on sagittal images reconstructed from the 4-mm-thick images and VTS images were compared with those measured on images reconstructed from 1-mm-thick images. Diagnostic performance for the detection of compression fractures was also compared. The intervertebral spaces were significantly more visible on the VTS images than on the 4-mm-thick images (P < 0.001). The absolute value of the measured difference in mean vertebral height between the VTS and 1-mm-thick images was smaller than that between the 4-mm-thick and 1-mm-thick images (P < 0.01–0.54). The diagnostic performance of the VTS images for detecting compression fracture was significantly lower than that of the 4-mm-thick images for one reader (P = 0.02). VTS technique enabled the identification of each vertebral body, and enabled accurate measurement of vertebral height. However, this technique is not suitable for diagnosing compression fractures.

Machine learning model development for quantitative analysis of CT heterogeneity in canine hepatic masses may predict histologic malignancy

Tumor heterogeneity is a well-established marker of biologically aggressive neoplastic processes and is associated with local recurrence and distant metastasis. Quantitative analysis of CT textural features is an indirect measure of tumor heterogeneity and therefore may help predict malignant disease. The purpose of this retrospective, secondary analysis study was to quantitatively evaluate CT heterogeneity in dogs with histologically confirmed liver masses to build a predictive model for malignancy. Forty dogs with liver tumors and corresponding histopathologic evaluation from a previous prospective study were included. Triphasic image acquisition was standardized across dogs and whole liver and liver mass were contoured on each precontrast and delayed postcontrast dataset. First-order and second-order indices were extracted from contoured regions. Univariate analysis identified potentially significant indices that were subsequently used for top-down model construction. Multiple quadratic discriminatory models were constructed and tested, including individual models using both postcontrast and precontrast whole liver or liver mass volumes. The best performing model utilized the CT features voxel volume and uniformity from postcontrast mass contours; this model had an accuracy of 0.90, sensitivity of 0.67, specificity of 1.0, positive predictive value of 1.0, negative predictive value of 0.88, and precision of 1.0. Heterogeneity indices extracted from delayed postcontrast CT hepatic mass contours were more informative about tumor type compared to indices from whole liver contours, or from precontrast hepatic mass and whole liver contours. Results demonstrate that CT radiomic feature analysis may hold clinical utility as a noninvasive method of predicting hepatic malignancy and may influence diagnostic or therapeutic approaches.

Computed tomographic features for differentiating benign from malignant liver lesions in dogs

Thus far, there are few computed tomography (CT) characteristics that can distinguish benign and malignant etiologies. The criteria are complex, subjective, and difficult to use in clinical applications due to the high level of experience needed. This study aimed to identify practical CT variables and their clinical relevance for broadly classifying histopathological diagnoses as benign or malignant. In this prospective study, all dogs with liver nodules or masses that underwent CT examination and subsequent histopathological diagnosis were included. Signalments, CT findings and histopathological diagnoses were recorded. Seventy liver nodules or masses in 57 dogs were diagnosed, comprising 18 benign and 52 malignant lesions. Twenty-three qualitative and quantitative CT variables were evaluated using univariate and stepwise multivariate analyses, respectively. Two variables, namely, the postcontrast enhancement pattern of the lesion in the delayed phase (heterogeneous; odds ratio (OR): 14.7, 95% confidence interval (CI): 0.82–262.03, P=0.0429) and the maximal transverse diameter of the lesion (>4.5 cm; OR: 33.3, 95% CI: 2.29–484.18, P=0.0006), were significantly related to the differentiation of benign from malignant liver lesions, with an area under the curve of 0.8910, representing an accuracy of 88.6%. These findings indicate that features from triple-phase CT can provide information for distinguishing pathological varieties of focal liver lesions and for clinical decision making. Evaluations of the maximal transverse diameter and postcontrast enhancement pattern of the lesion included simple CT features for predicting liver malignancy with high accuracy in clinical settings.

ASSOCIATIONS BETWEEN DUAL-PHASE COMPUTED TOMOGRAPHY FEATURES AND HISTOPATHOLOGIC DIAGNOSES IN 52 DOGS WITH HEPATIC OR SPLENIC MASSES

Ability to noninvasively differentiate malignant from nonmalignant abdominal masses would aid clinical decision making. The aim of this retrospective, cross-sectional study was to identify features in dual-phase computed tomographic (CT) studies that could be used to distinguish malignant from nonmalignant hepatic and splenic masses in dogs. Medical records were searched for dogs that had an abdominal dual-phase CT examination, a hepatic or splenic mass, and subsequent histopathologic diagnosis. Computed tomographic images for all included dogs were acquired prior to and <30 s (early phase) and >60 s (delayed phase) after intravenous contrast administration. Fifty-two dogs with 55 masses were studied: 24 hepatic, including 14 (58%) malignant and 10 (42%) non-malignant; 31 splenic, including 18 (58%) malignant and 13 (42%) nonmalignant. There was substantial overlap in the pre- and postcontrast CT features of malignant and nonmalignant hepatic and splenic masses. Regardless of histologic diagnosis, hepatic masses most frequently showed marked, generalized enhancement in early phase images that persisted in the delayed phase. Splenic hemangiosarcoma and nodular hyperplastic lesions most frequently showed marked, generalized enhancement in early phase images that persisted in delayed images whereas most splenic hematomas had slight enhancement in early phase images. All splenic hematomas and 77% of the hemangiosarcomas had contrast accumulation compatible with active hemorrhage. There were no other significant differences in quantitative or categorical CT data between malignant and nonmalignant hepatic or splenic masses. Dual-phase CT of dogs with hepatic or splenic masses provides limited specific diagnostic information.

Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions

Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2 -D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85% is achieved for three out of the five lesion pathologies and for 98% of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.

Content-based image retrieval of multiphase CT images for focal liver lesion characterization

PURPOSE

Characterization of focal liver lesions with various imaging modalities can be very challenging in the clinical practice and is experience-dependent. The authors' aim is to develop an automatic method to facilitate the characterization of focal liver lesions (FLLs) using multiphase computed tomography (CT) images by radiologists.

METHODS

A multiphase-image retrieval system is proposed to retrieve a preconstructed database of FLLs with confirmed diagnoses, which can assist radiologists' decision-making in FLL characterization. It first localizes the FLL on multiphase CT scans using a hybrid generative-discriminative FLL detection method and a nonrigid B-spline registration method. Then, it extracts the multiphase density and texture features to numerically represent the FLL. Next, it compares the query FLL with the model FLLs in the database in terms of the feature and measures their similarities using the L1-norm based similarity scores. The model FLLs are ranked by similarities and the top results are finally provided to the users for their evidence studies.

RESULTS

The system was tested on a database of 69 four-phase contrast-enhanced CT scans, consisting of six classes of liver lesions, and evaluated in terms of the precision-recall curve and the Bull's Eye Percentage Score (BEP). It obtained a BEP score of 78%. Compared with any single-phase based representation, the multiphase-based representation increased the BEP scores of the system, from 63%-65% to 78%. In a pilot study, two radiologists performed characterization of FLLs without and with the knowledge of the top five retrieved results. The results were evaluated in terms of the diagnostic accuracy, the receiver operating characteristic (ROC) curve and the mean diagnostic confidence. One radiologist's accuracy improved from 75% to 92%, the area under ROC curves (AUC) from 0.85 to 0.95 (p = 0.081), and the mean diagnostic confidence from 4.6 to 7.3 (p = 0.039). The second radiologist's accuracy did not change, at 75%, with AUC increasing from 0.72 to 0.75 (p = 0.709), and the mean confidence from 4.5 to 4.9 (p = 0.607).

CONCLUSIONS

Multiphase CT images can be used in content-based image retrieval for FLL's categorization and result in good performance in comparison with single-phase CT images. The proposed method has the potential to improve the radiologists' diagnostic accuracy and confidence by providing visually similar lesions with confirmed diagnoses for their interpretation of clinical studies.

[Radiological diagnosis of primary hepatic malignancy]

Modern radiology offers countless opportunities both in the detection but also in the characterization of primary liver malignancies. Ultrasound remains usually the first exploratory overview study whereat using ultrasound contrast agent for a further characterization of liver lesions improves this technique considerably. Advanced cross-sectional imaging methods can, in most cases, already provide an exact diagnosis. Thus, the CT is already considered a standard technique for liver imaging and magnetic resonance imaging has gained in recent years due to liver-specific contrast agents and faster sequences a central role in liver imaging. The following article provides an overview of these various radiological procedures and describes the different primary liver malignancies and their imaging characteristics.

Small (≤ 20 mm) pancreatic adenocarcinomas: analysis of enhancement patterns and secondary signs with multiphasic multidetector CT

PURPOSE

To evaluate the enhancement patterns, prevalence of secondary signs, and histopathologic features of 20-mm-diameter or smaller pancreatic cancers seen on multiphasic multidetector computed tomographic (CT) images.

MATERIALS AND METHODS

This retrospective study was approved by the institutional review board; the requirement for informed consent was waived. From January 2002 through September 2009, the authors reviewed the clinical and imaging data of 130 consecutive patients (76 men, 54 women; mean age, 64.1 years; age range, 28-82 years) who had surgically proven 30-mm-diameter or smaller pancreatic cancers and underwent preoperative multidetector CT and 33 consecutive patients (17 men, 16 women; mean age, 65.1 years; age range, 48-84 years) who had histopathologically proven pancreatic cancer and underwent incidental multidetector CT before the diagnosis was rendered. Only pancreatic phase CT was performed in two patients, and only hepatic venous phase CT was performed in nine patients. Two radiologists in consensus classified the tumor attenuation as hyper-, iso-, or hypoattenuation during the pancreatic and hepatic venous phases. Accompanying secondary signs, temporal changes in tumor attenuation, and histopathologic findings also were analyzed. The Fisher exact test, χ(2) test, generalized estimating equation, and Student t test were used to compare the variables.

RESULTS

Seventy tumors were 20 mm or smaller, and 93 were 21-30 mm. Isoattenuating pancreatic cancers were more commonly observed among the 20-mm or smaller tumors (16 of 59, 27%) than among the 21-30-mm tumors (12 of 93, 13%) (P = .033). They were also more common among well-differentiated tumors (seven of 12, 58%) than among moderately differentiated (20 of 124, 16%) and poorly differentiated (one of 10, 10%) tumors (P = .001). The prevalence of secondary signs differed significantly according to tumor size (53 [76%] of 70 ≤20-mm tumors vs 92 [99%] of 93 21-30-mm tumors) (P < .001). The prevalence of secondary signs was high among isoattenuating pancreatic cancers (14 [88%] of 16 ≤20-mm tumors vs all 12 [100%] 21-30-mm tumors). Most of the isoattenuating tumors seen at prediagnostic CT were hypoattenuating after 6 months (100% [four of four] during pancreatic phase, 71% [five of seven] during hepatic venous phase).

CONCLUSION

The prevalence of isoattenuating pancreatic cancers differed significantly according to tumor size and cellular differentiation. Most small isoattenuating pancreatic cancers showed secondary signs.

SUPPLEMENTAL MATERIAL

http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.11101133/-/DC1.

Liver Metastases on Serial Contrast-enhanced Multidetector Computed Tomography Examinations

OBJECTIVE

To verify the earliest detectability of liver metastases in patients who underwent serial multidetector computed tomography (MDCT) examinations.

METHODS

We selected 12 patients with known primary cancer who underwent 4 or more contrast-enhanced, 4-detector MDCTs. When metastases had been reported, an evaluation of the preceding MDCT was done to define whether the lesion was detectable, detectable only by minimal signs, undetectable, or detected but misdiagnosed as a benign lesion (MBL).

RESULTS

Eighty-eight lesions were analyzed. Evaluating the preceding examination, we defined detectable (n=8), detectable only by minimal signs (n=5), undetectable (n=74), and MBL (n=1). The group with minimal signs was composed of 4 small hypodense foci and 1 calcification. The MBL was a non-Hodgkin lesion first misdiagnosed as a hemangioma.

CONCLUSION

Approximately 15% of liver metastases were prospectively missed, 9% of them being retrospectively detectable, 6% being retrospectively visible as minimal signs, whereas only 1% of liver metastases were misdiagnosed as a benign lesion.

Liver and Biliary System: Evaluation by Multidetector CT

CT commonly is indicated for the evaluation of suspected hepatic and biliary pathology. The recent introduction of multidetector CT (MDCT) provides unique capabilities that are valuable especially in hepatic volume acquisitions, combining short scan times, narrow collimation, and the ability to obtain multiphase data. These features result in improved lesion detection and characterization. Concomitant advances in computer software programs have made three-dimensional applications practical for a range of hepatic image analyses and displays. This article discusses the specific areas of hepatic and biliary pathology where MDCT has a significant diagnostic impact.

Multidetector CT: contributions in liver imaging

Multidetector CT has had great impact in the evaluation of the liver. Increased speed combined with thin slice collimation improved the spatial and temporal resolution, which in turn provided a higher sensitivity in the detection of focal lesions, particularly for the hepatocellular cancer (HCC) in patients with cirrhosis. There is no consensus as to whether the so-called "double arterial phase scanning" improves the detection rate of hepatocellular cancers, and the majority of the more recent studies suggest acquiring only the late arterial phase images along with the portal venous phase images. However, if the patient is a surgical candidate, the early arterial phase images must also be obtained, because they provide better pre-surgical mapping of the hepatic arteries. In primary and metastatic liver tumors, arterial and portal venous phase images should be acquired with thin collimation and overlapping reconstruction. Maximum intensity projection and volume rendering methods demonstrate the three-dimensional anatomy of the hepatic arteries, portal and hepatic veins successfully, which provide useful information before hepatic resection or intraarterial chemotherapy. In living donor candidates, early arterial and portal venous phase images obtained with thin collimation and overlapping reconstruction are used to reconstruct three-dimensional images with maximum intensity and volume rendering methods. These have a high sensitivity to detect hepatic artery, portal and hepatic vein variations, which could render the surgery difficult or even impossible. Portal venous phase images can also be used to measure total and lobar liver volumes. A virtual hepatectomy plane can be simulated on the three-dimensional model of the liver and hepatic veins, and when necessary, its location can be modified according to the metabolic needs of the recipient and donor. Thin collimation and better temporal resolution may also have beneficial effects in the characterization of liver lesions.

Multidetector CT of hepatocellular carcinoma

Hepatocellular carcinoma is one of the most common causes of cancer death worldwide. Numerous surgical and nonsurgical treatment options are available for the management of patients with HCC. Successful long-term outcome is dependent on early detection of HCC, as well as accurate delineation of the number and location of tumor nodules. We present the different manifestations of hepatocellular carcinoma as depicted by multidetector-row CT with advanced image processing.

Imaging of neuroendocrine tumors: accuracy of helical CT versus SRS

BACKGROUND

We retrospectively compared the accuracy of somatostatin receptor scintigraphy (SRS) with that of helical computed tomography (CT) in the detection and localization of primary and metastatic neuroendocrine tumors.

METHODS

A medical record search identified 27 patients with known or clinically suspected neuroendocrine tumors who underwent helical CT and SRS within 3 months of one another at our institution. CT images were evaluated retrospectively by two blinded radiologists who used consensus reading. Images were evaluated for the presence or absence of primary tumor and hepatic and extrahepatic metastases. CT results were compared with the SRS report as interpreted by the nuclear medicine physicians. The results of the surgical, clinical follow-up, and pathologic findings were considered as the gold standard. Sensitivity, specificity, and accuracy were calculated for both imaging techniques. In addition, McNemar analysis was performed to determine statistically significant differences between CT and SRS.

RESULTS

Helical CT was more sensitive than SRS in the detection of extrahepatic metastases, and the difference between the two imaging modalities was statistically significant (p = 0.0312) as determined by the McNemar chi-square test. However, the difference between CT and SRS in detecting primary neuroendocrine tumors, hepatic metastasis, and combined hepatic and extrahepatic metastasis was not statistically significant (p = 0.625, 1.000, and 1.000, respectively).

CONCLUSION

Helical CT and SRS have similar sensitivity, specificity, and accuracy in detecting primary neuroendocrine tumor and hepatic metastasis. However, helical CT appears to be more sensitive in detecting extrahepatic metastasis from primary neuroendocrine tumors.

Incremental value of advanced image processing of multislice computed tomography data in the evaluation of hypervascular liver lesions

Multiplanar volume rendering and maximum intensity projection of multislice computed tomography data allow for robust evaluation of the hepatic parenchyma, with early detection of neovascularity and tumor stain, and enable better assessment of the tumor burden compared with routine axial images. Lesion characterization can also potentially improve. This is in addition to accurate delineation of the hepatic vascular anatomy, which allows for segmental localization and assists in surgical and nonsurgical planning.

High-resolution multidetector CT in the preoperative evaluation of patients with renal cell carcinoma

OBJECTIVE

The purpose of our study was to evaluate the accuracy of multidetector CT (MDCT) using a high-resolution protocol in the preoperative assessment of patients with renal cell carcinoma who are possible candidates for nephron-sparing surgery.

MATERIALS AND METHODS

Forty patients with suspected renal cell carcinoma underwent MDCT. Contrast-enhanced acquisitions were obtained during arterial, nephrographic, and urographic phases using a thin-slice protocol. One-millimeter-thick source images were evaluated by two observers on a dedicated workstation for the identification and characterization of the tumor, presence of a pseudocapsule or invasion of perirenal fat, involvement of adrenal glands or surrounding tissues, presence of satellite lesions within Gerota's fascia, infiltration of renal vein and inferior vena cava, involvement of lymph nodes, and presence of distant metastases. Imaging findings were compared with surgical specimens using criteria from the Robson and TNM classification systems.

RESULTS

The presence and size of all lesions were correctly shown in all patients. In evaluating Robson stage I of renal cell carcinoma, we were able to diagnose fat infiltration on 1-mm scans with 96% sensitivity, 93% specificity, and 95% accuracy; the positive and negative predictive values were, respectively, 100% and 93%. One hundred percent accuracy was achieved in staging high-grade lesions.

CONCLUSION

High-resolution MDCT is accurate in the preoperative evaluation of patients with renal cell carcinoma.

Detection of hepatocellular carcinoma: comparison of dynamic three-phase computed tomography images and four-phase computed tomography images using multidetector row helical computed tomography

PURPOSE

The purpose of our study was to assess the value of additional early arterial phase computed tomography (CT) imaging in the detection of hepatocellular carcinoma (HCC) by comparing three-phase and four-phase imaging by using multidetector row helical CT.

METHODS

Twenty-five patients with 33 HCCs underwent four-phase helical CT imaging. The diagnosis was established by pathologic examination after surgical resection in 19 patients and by biopsy in six. Four-phase CT imaging comprises early arterial, late arterial, portal venous, and delayed phase imaging obtained 25 seconds, 45 seconds, 75 seconds, and 180 seconds after the start of contrast material injection using multidetector row helical CT. Three-phase CT images (late arterial, portal venous, and delayed phase) and four-phase CT images (early arterial, late arterial, portal venous, and delayed phase) were interpreted independently for the detection of HCC by three blinded observers on a segment-by-segment basis. Sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (Az) for three-phase CT images and four-phase CT images were calculated. The enhancement pattern of HCC was analyzed on early arterial and late arterial phase imaging.

RESULTS

The mean sensitivity of three- and four-phase CT images was 94% and 93%, respectively. The differences between sensitivities were not statistically significant (all p > 0.05). The mean specificities of three- and four-phase CT images were 99% and 98%, respectively. The differences between the specificities were not statistically significantly (all p > 0.05). Neither were the mean areas under the ROC curve for four-phase CT images (Az = 0.976) and three-phase CT images (Az = 0.971) statistically significant (p > 0.05). On early arterial phase imaging, 16 HCCs were hyperattenuating and 17 HCCs were isoattenuating. On late arterial phase imaging, 24 HCCs were hyperattenuating and nine HCCs were isoattenuating.

CONCLUSIONS

Additional early arterial phase imaging did not improve the detection of HCC compared with three-phase CT images, including late arterial, portal venous, and delayed phase imaging.