The value of "liver windows" settings in the detection of small renal cell carcinomas on unenhanced computed tomography

New liver window width in detecting hepatocellular carcinoma on dynamic contrast-enhanced computed tomography with deep learning reconstruction

Changing a window width (WW) alters appearance of noise and contrast of CT images. The aim of this study was to investigate the impact of adjusted WW for deep learning reconstruction (DLR) in detecting hepatocellular carcinomas (HCCs) on CT with DLR. This retrospective study included thirty-five patients who underwent abdominal dynamic contrast-enhanced CT. DLR was used to reconstruct arterial, portal, and delayed phase images. The investigation of the optimal WW involved two blinded readers. Then, five other blinded readers independently read the image sets for detection of HCCs and evaluation of image quality with optimal or conventional liver WW. The optimal WW for detection of HCC was 119 (rounded to 120 in the subsequent analyses) Hounsfield unit (HU), which was the average of adjusted WW in the arterial, portal, and delayed phases. The average figures of merit for the readers for the jackknife alternative free-response receiver operating characteristic analysis to detect HCC were 0.809 (reader 1/2/3/4/5, 0.765/0.798/0.892/0.764/0.827) in the optimal WW (120 HU) and 0.765 (reader 1/2/3/4/5, 0.707/0.769/0.838/0.720/0.791) in the conventional WW (150 HU), and statistically significant difference was observed between them (p < 0.001). Image quality in the optimal WW was superior to those in the conventional WW, and significant difference was seen for some readers (p < 0.041). The optimal WW for detection of HCC was narrower than conventional WW on dynamic contrast-enhanced CT with DLR. Compared with the conventional liver WW, optimal liver WW significantly improved detection performance of HCC.

Radiological Features of T1a Renal Cell Carcinoma on Axial Unenhanced Computed Tomography

CT has become a commonly used diagnostic procedure in clinical practice, particularly in emergency healthcare delivery. Accordingly, the increase in CT usage has increased the likelihood of incidental detections (ID) of renal cell carcinomas (RCCs). This article discusses key points and limitations associated with the diagnosis and characterization of T1a RCC (≤4 cm in diameter) and shows how to improvise on the differentiation of T1a RCC with unenhanced CT (UE-CT). We retrospectively reviewed UE-CT findings of cases associated with the histopathologic diagnosis of T1a RCC and examined the discrimination capacity and radiological characteristics with regard to small RCCs (SRCCs). Detection and characterization of T1a RCC based on UE-CT are not easy in many cases due to limitations in CT findings, but there are notable radiological features to facilitate detection and differentiation. The growth pattern is important for the detection of SRCCs. Internal characteristic features (average attenuation, heterogeneity) are useful for the characterization of the RCC. In addition, CT image visualization techniques may help improve the detectability of RCCs on UE-CT. Radiological features are important in detecting SRCCs and facilitating further examination. In this study, we discuss some cases of T1a RCCs and evaluate the radiological characteristics of the tumors seen on UE-CT.

Know your way around acute unenhanced CT during global iodinated contrast crisis: a refresher to ED radiologists

Due to a contrast shortage crisis resulting from the decreased supply of iodinated contrast agents, the American College of Radiology (ACR) has issued a guidance statement followed by memoranda from various hospitals to preserve and prioritize the limited supply of contrast. The vast majority of iodinated contrast is used by CT, with a minority used by vascular and intervention radiology, fluoroscopy, and other services. A direct consequence is a paradigm shift to large volume unenhanced CT scans being utilized for acute and post traumatic patients in EDs, an uncharted territory for most radiologists and trainees. This article provides radiological diagnostic guidance and a pictorial example through systematic review of common unenhanced CT findings in the acute setting.

Semi-automatic liver segmentation based on probabilistic models and anatomical constraints

Segmenting a liver and its peripherals from abdominal computed tomography is a crucial step toward computer aided diagnosis and therapeutic intervention. Despite the recent advances in computing methods, faithfully segmenting the liver has remained a challenging task, due to indefinite boundary, intensity inhomogeneity, and anatomical variations across subjects. In this paper, a semi-automatic segmentation method based on multivariable normal distribution of liver tissues and graph-cut sub-division is presented. Although it is not fully automated, the method minimally involves human interactions. Specifically, it consists of three main stages. Firstly, a subject specific probabilistic model was built from an interior patch, surrounding a seed point specified by the user. Secondly, an iterative assignment of pixel labels was applied to gradually update the probabilistic map of the tissues based on spatio-contextual information. Finally, the graph-cut model was optimized to extract the 3D liver from the image. During post-processing, overly segmented nodal regions due to fuzzy tissue separation were removed, maintaining its correct anatomy by using robust bottleneck detection with adjacent contour constraint. The proposed system was implemented and validated on the MICCAI SLIVER07 dataset. The experimental results were benchmarked against the state-of-the-art methods, based on major clinically relevant metrics. Both visual and numerical assessments reported herein indicated that the proposed system could improve the accuracy and reliability of asymptomatic liver segmentation.

CT window trainable neural network for improving intracranial hemorrhage detection by combining multiple settings

Window settings to rescale and contrast stretch raw data from radiographic images such as Computed Tomography (CT), X-ray and Magnetic Resonance images is a crucial step as data pre-processing to examine abnormalities and diagnose diseases. We propose a distant-supervised method for determining automatically the best window settings by attaching a window estimator module (WEM) to a deep convolutional neural network (DCNN)-based lesion classifier and training them in conjunction. Aside from predicting a flexible window setting for each raw image, we statistically identify the top four window settings by calculating the mean and standard deviations for the entire dataset. Images are scaled on each of the top settings estimated by WEM and following lesion classifiers are subsequently trained. We study the effects of only using the flexible window, the single fixed window as either a known default window used by radiologists or an estimated mean value, and two different approaches to combine results from the top window settings to improve the detection of intracranial hemorrhage (ICH) from brain CT images. Experimental results showed that using the top four window settings identified from the window estimator module and combining the results had the best performance.

Stochastic tissue window normalization of deep learning on computed tomography

Tissue window filtering has been widely used in deep learning for computed tomography (CT) image analyses to improve training performance (e.g., soft tissue windows for abdominal CT). However, the effectiveness of tissue window normalization is questionable since the generalizability of the trained model might be further harmed, especially when such models are applied to new cohorts with different CT reconstruction kernels, contrast mechanisms, dynamic variations in the acquisition, and physiological changes. We evaluate the effectiveness of both with and without using soft tissue window normalization on multisite CT cohorts. Moreover, we propose a stochastic tissue window normalization (SWN) method to improve the generalizability of tissue window normalization. Different from the random sampling, the SWN method centers the randomization around the soft tissue window to maintain the specificity for abdominal organs. To evaluate the performance of different strategies, 80 training and 453 validation and testing scans from six datasets are employed to perform multiorgan segmentation using standard 2D U-Net. The six datasets cover the scenarios, where the training and testing scans are from (1) same scanner and same population, (2) same CT contrast but different pathology, and (3) different CT contrast and pathology. The traditional soft tissue window and nonwindowed approaches achieved better performance on (1). The proposed SWN achieved general superior performance on (2) and (3) with statistical analyses, which offers better generalizability for a trained model.

T1a renal cell carcinoma on unenhanced CT: analysis of detectability and imaging features

Background

Increasing use of unenhanced computed tomography (CT) has been associated with the increasing incidental detection of renal cell carcinoma (RCC) at an earlier stage.

Purpose

To evaluate the characteristics in detecting and differentiating T1a RCCs on unenhanced CT.

Material and Methods

We retrospectively reviewed 68 patients with 68 T1a RCCs and 39 benign regions. Two radiologists interpreted the images on unenhanced axial CT and performed a blinded and independent review of T1a RCCs. The readers evaluated the presence of RCC and differentiated the detected lesions.

Results

The consensus of two readers detected 53 (78%) RCCs. Of the 53 detected RCCs, 42 (62%) RCCs were correctly diagnosed and 11 (16%) masses were misdiagnosed as benign. Of the 39 benign regions, 29 (74%) cysts were diagnosed correctly, but 10 (26%) cysts were misdiagnosed as malignant. The following values of the radiologists were obtained by consensus: sensitivity = 61.8% (42/68); specificity = 74.4% (29/39); positive predictive value = 80.8% (42/52); negative predictive value = 55.0% (29/55); accuracy = 66.4% (71/107). The receiver operating characteristic curve of consensus was 0.754. Inter-observer correlation was κ = 0.849. There was a significant difference in tumor size (P = 0.019) and the contour type of tumor (P = 0.0207) between correctly diagnosed RCCs and not correctly diagnosed RCCs.

Conclusion

Our findings showed that tumor size and contour type could affect the detection and differentiation of T1a RCC on unenhanced CT. To detect and differentiate T1a RCC on unenhanced CT is difficult. However, the findings from this study may help detection of RCCs on unenhanced CT.

Renal cancer at unenhanced CT: imaging features, detection rates, and outcomes

PURPOSE

To describe and quantify the rate of detection of renal cancer on unenhanced CT.

METHODS

This retrospective, HIPAA-compliant study was approved by the Institutional Review Board. Electronic health records for all patients who underwent unenhanced abdominal CT at our institution between 2000 and 2005 were reviewed to identify patients subsequently diagnosed with renal cancer during a follow-up period of up to 12 years. Images were reviewed to determine if the cancer was visible at index (first) unenhanced CT and their findings recorded. Original radiology reports were reviewed to determine whether the renal cancer was reported; Fisher's Exact Test compared imaging features of detected and missed cancers. Clinical outcomes including time until diagnosis and stage at diagnosis were used to assess the potential impact of missed cancers.

RESULTS

Of 15,695 patients, 82 (0.52%) were diagnosed with renal cancer. Of these, 43/82 (52%) cancers were retrospectively detectable on index unenhanced CT. Among retrospectively detectable cancers, 63% (27/43) were originally detected and reported on index CT and 37% (16/43) were missed. Size was the only feature associated with detection; 83% (20/24) of cancers > 3.0 cm were detected versus 37% (7/19) of cancers ≤ 3.0 cm (p = 0.0036). Although none of the 16 missed renal cancers developed metastases between index CT and time of diagnosis (median 33.5 months), 4 (25%) progressed in stage.

CONCLUSIONS

Renal cancer was rare in patients undergoing unenhanced abdominal CT. Over one-third of potentially detectable cancers were missed prospectively. However, missed cancers did not metastasize and infrequently progressed in stage before being diagnosed.

Using Correlative Properties of Neighboring Pixels to Improve Gray-White Differentiation in Pediatric Head CT Images

BACKGROUND AND PURPOSE

A lower radiation dose can have a detrimental effect on the quality of head CT images. The aim of this study performed in a pediatric population was to test whether an image-processing algorithm (Correlative Image Enhancement) based on the correlation among intensities of neighboring pixels can improve gray-white differentiation in head CTs.

MATERIALS AND METHODS

Sixty baseline head CT images with normal findings obtained from scans of 30 children were processed using Correlative Image Enhancement to produce corresponding enhanced images. Gray-white differentiation in baseline and enhanced images was assessed quantitatively by calculating the contrast-to-noise ratio and conspicuity in equivalent ROIs in gray and white matter. Two masked readers rated the images for visibility of gray-white differentiation on a 5-point Likert scale. Differences in both quantitative and qualitative measures of gray-white differentiation between baseline and enhanced images were tested for statistical significance. P values < .05 were considered significant.

RESULTS

Image processing resulted in improvement in the contrast-to-noise ratio (from 1.86 ± 0.94 to 2.26 ± 1.00, P = .02) as well as conspicuity (from 37.28 ± 11.56 to 46.4 ± 11.5, P < .001). This was accompanied by improved subjective visibility of gray-white differentiation as reported by both readers (P < .01).

CONCLUSIONS

Image processing using Correlative Image Enhancement had a beneficial effect on quantitative measures of gray-white differentiation. This translated into improved perception of gray-white differentiation by readers. Further studies are needed to assess the effect of such image processing on the detection of disease processes using head CTs.

Minimising the impact of errors in the interpretation of CT images for surveillance and evaluation of therapy in cancer

Radiological error is inevitable and usually multifactorial. Error can be secondary to radiologist-specific causes, including cognitive and perceptive errors or ambiguity of report, or system-related causes, including inadequate, misleading, or incorrect clinical information, poor imaging technique, excessive workload, and poor working conditions. In this paper, we discuss a systematic approach to reduce errors in oncological radiology reporting, thus reducing risk to the patient. Rather than attempt to discuss all types of error, we concentrate on the most important and commonly occurring errors that we have encountered over 20 years of practice, based on weekly discrepancy reviews of our practice and independent reviews of clinical and research imaging from other institutions. This review focuses on computed tomography (CT) reporting for staging, surveillance, and response assessment of cancer patients, but the messages apply to all imaging methods.