Generally applicable window settings of low-keV virtual monoenergetic reconstructions in dual-layer CT-angiography of the head and neck

Background

Increased vessel contrast in low-keV virtual monoenergetic images (VMI) in spectral detector CT angiography of the head and neck requires adaption of window settings. Aim of this study was to define generally applicable window settings of low-keV VMI.

Methods

Two radiologists determined ideal subjective window settings for VMI_{40-70 keV} in 54 patients. To obtain generally applicable window settings, center and width values were modeled against the attenuation of the internal carotid artery (HUICA). This modeling was performed with and without respect to keV. Subsequently, image quality of VMI_{40-70 keV} was assessed using the model-based determined window settings.

Results

With decreasing keV values, HUICA increased significantly in comparison to conventional images (CI) (P<0.05 for 40-60 keV). No significant differences between modelled and individually recorded window settings were found confirming validity of the obtained models (P values: 0.2-1.0). However, modelling with respect to keV was marginally less precise.

Conclusions

Window settings of low-keV VMI can be semi-automatically determined in dependency of the ICA attenuation in spectral detector CTA of the head and neck. The reported models are a promising tool to leverage the improved image quality of these images in clinical routine.

Dual-Energy Computed Tomography for the Diagnosis of Mediastinal Lymph Node Metastasis in Lung Cancer Patients: A Preliminary Study

OBJECTIVE

This study explored the feasibility of dual-energy computed tomography (DECT) for the diagnosis of mediastinal lymph node (LN) metastasis in patients with lung cancer.

METHODS

Forty-two consecutive patients with lung cancer, who underwent DECT, were included in this retrospective study. The attenuation value (Hounsfield unit) in virtual monochromatic images and the iodine concentration in the iodine map were measured at mediastinal LNs. The slope of the spectral attenuation curve (K) and normalized iodine concentration (in thoracic aorta) were calculated. The measurement results were statistically compared using 2 independent samples t test. Receiver operating characteristic curve analysis, net reclassification improvement, and integrated discrimination improvement were used to evaluate the diagnostic performance of DECT for mediastinal LN metastasis.

RESULTS

A total of 74 mediastinal LNs were obtained, including 33 metastatic LNs and 41 nonmetastatic LNs. The attenuation value at the lower energy levels of virtual monochromatic images (40-90 keV), K, and normalized iodine concentration demonstrated a significant difference between metastatic LNs and nonmetastatic LNs. The attenuation value at 40 keV was the most favorable biomarker for the diagnosis of mediastinal LN metastasis (area under curve, 0.91; sensitivity, 0.94; specificity, 0.81), which showed a much better performance than the LN diameter-based evaluation method (area under curve, 0.72; sensitivity, 0.66; specificity, 0.82; net reclassification improvement, 0.359; integrated discrimination improvement, 0.330).

CONCLUSIONS

Dual-energy computed tomography is a promising diagnostic approach for the diagnosis of mediastinal LN metastasis in patients with lung cancer, which may help clinicians implement personalized treatment strategies.

Determination of the optimal range for virtual monoenergetic images in dual-energy CT based on physical quality parameters

PURPOSE

Virtual monoenergetic images (VMI) obtained from Dual-Energy Computed Tomography (DECT) with iodinated contrast are used in radiotherapy of the Head and Neck to improve the delineation of target volumes and organs at-risk (OAR). The energies used to vary from 40 to 70 keV, but noise at low keV and the use of Single Energy CT (SECT) at low kV_p settings may shrink this interval. There is no guide about how to find out the optimal range where VMI has a significant improvement related to SECT images. Our study proposes a procedure to determine this optimal range, based on common image quality parameters, and establishes this range in a Siemens Somatom Confidence and a Head and Neck protocol.

METHODS

We compared the quality of the VMI series at 40-60 keV versus single X-ray tube voltage computed tomography (SECT) at 80 and 120 kV_p . Our reference was 120 kV_p . DECT images were sequentially acquired using the Siemens Somatom Confidence RT Pro CT according to the head and neck protocol in our department. VMI series were constructed using the Syngo Via software Monoenergetic+ algorithm. Quality parameters were: image uniformity, high- and low-contrast resolution, noise, and sensitivity to the iodinated contrast. We used the Catphan 604 phantom for quality control, except when assessing iodine sensitivity. To evaluate high contrast resolution, we calculated the modulation transfer function (MTF) using the point spread function estimation of a point bead and the slanted edge methods. For the low-contrast resolution, we used a statistical method for assessing differences between contrast structures and local noise. To measure the absolute value of noise and compare its texture, we used the standard deviation and the noise power spectrum. We measured iodine sensitivity by dissolving the Optiray Ultraject iodinated contrast in water in concentrations of 0 to 4500 mg/l and then compared the contrast to noise ratio (CNR) and analyzed the linear correlation between concentration and HU.

RESULTS

The entire series met the minimum quality requirements. However, the one at 40 keV presented uniformity at the limits of acceptability. The high- and low-contrast resolutions were similar between series. The noise of the VMI series decreased with increasing energy, while sensitivity to the contrast displayed the opposite behavior. All series showed linearity of HUs from very low iodine concentrations. Images at 60 keV presented lower iodine sensitivity than SECT at 80 kV_p , while those at 55 keV were similar to them.

CONCLUSIONS

Our method of image comparison based on standard quality parameters in phantom gave clear results about the optimal range and can be used as a guide to characterize any other DECT imaging protocols. The optimal range for using VMI images in iodinated contrasts in the Siemens system was 45-55 keV. Lower energies lacked noise and uniformity, while higher ones could be substituted by SECT images at low kilovoltage (80 kV_p ).

Usefulness of Virtual Monochromatic Images and Iodine Maps Derived from Dual-Energy Computed Tomography for Diagnosing Deep Neck Abscesses

OBJECTIVE

We aimed to determine whether dual-energy computed tomography (CT) is useful for evaluating deep neck abscesses.

METHODS

This study included 22 consecutive patients who were clinically suspected of having a deep neck abscess and underwent dual-energy CT. Conventional 120-kVp images, 70- and 40-keV virtual monochromatic images (VMIs), and iodine maps were inspected to calculate the contrast ratio of the abscess rim (AR) to the abscess center (AC) or to the adjacent muscle (M). The diagnostic certainty of abscesses was assessed on these images.

RESULTS

Twenty (91%) of 22 patients had a definitive diagnosis. The contrast ratio for AR/AC and AR/M was significantly higher on 40-keV VMIs and iodine maps than on 120-kVp images and 70-keV VMIs (P < 0.05). On both 40-keV VMIs and iodine maps, the diagnostic certainty of abscess improved in 3 (15%) cases compared with 120-kVp images and 70-keV VMIs.

CONCLUSIONS

Dual-energy CT-based 40-keV VMIs and iodine maps are useful for evaluating deep neck abscesses and may improve diagnostic certainty.

Spectral detector CT applications in advanced liver imaging

OBJECTIVE

Spectral detector CT (SDCT) has many applications in advanced liver imaging. If appropriately utilized, this technology has the potential to improve image quality, provide new diagnostic information, and allow for decreased radiation dose. The purpose of this review is to familiarize radiologists with the uses of SDCT in liver imaging.

CONCLUSION

SDCT has a variety of post-processing techniques, which can be used in advanced liver imaging and can significantly add value in clinical practice.

Appropriate iMAR presets for metal artifact reduction from surgical clips and titanium burr hole covers on postoperative non-contrast brain CT

PURPOSE

To evaluate suitable iterative metal artifact reduction (iMAR) presets for titanium neurosurgical clips and burr hole covers (BHCs) on postoperative non-contrast computed tomography (NCCT).

METHOD

Twenty-two patients who underwent NCCT after intracranial aneurysmal clipping were included. NCCT images were postprocessed using eight currently available iMAR presets. In each image, a circular region of interest (ROI) was placed around clip, BHC, and on parietal lobe as reference. Standard deviation (SD) and attenuation value (HU) were measured in each ROI to obtain artifact index (AI) and contrast-to-noise ratio (CNR). For each iMAR preset, SD, AI, HU, and CNR were compared with those without iMAR for clips and BHCs. Visual assessment around each clip and BHC was performed by two neuroradiologists using three-point visual score (VS) (1 = no apparent, 2 = minor, and 3 = severe artifacts).

RESULTS

Among the presets, the neuro-coils preset (iMAR-NC) showed the lowest SD, AI, and VS for clips (P < 0.001). For BHCs, HU, CNR, and VS with iMAR-NC were significantly higher than those without iMAR (P < 0.001). SD, AI, and VS with the shoulder implants preset (iMAR-ShI) were significantly lower than those without iMAR for clips (P = 0.002, 0.002, and P <  0.001, respectively). For BHCs, VS with iMAR-ShI was lowest among the presets (P = 0.004).

CONCLUSIONS

Although iMAR-NC reduces metal artifacts from clips, it strengthens artifacts from BHCs. For postoperative NCCT, iMAR-ShI most effectively reduces metal artifacts from both clips and BHCs in a single preset.

Deep-learning-based direct synthesis of low-energy virtual monoenergetic images with multi-energy CT

Purpose: We developed a deep learning method to reduce noise and beam-hardening artifact in virtual monoenergetic image (VMI) at low x-ray energy levels. Approach: An encoder-decoder type convolutional neural network was implemented with customized inception modules and in-house-designed training loss (denoted as Incept-net), to directly estimate VMI from multi-energy CT images. Images of an abdomen-sized water phantom with varying insert materials were acquired from a research photon-counting-detector CT. The Incept-net was trained with image patches ( 64 × 64    pixels ) extracted from the phantom data, as well as synthesized, random-shaped numerical insert materials. The whole CT images ( 512 × 512    pixels ) with the remaining real insert materials that were unseen in network training were used for testing. Seven contrast-enhanced abdominal CT exams were used for preliminary evaluation of Incept-net generalizability over anatomical background. Mean absolute percentage error (MAPE) was used to evaluate CT number accuracy. Results: Compared to commercial VMI software, Incept-net largely suppressed beam-hardening artifact and reduced noise (53%) in phantom study. Incept-net presented comparable CT number accuracy at higher-density ( P -value [0.0625, 0.999]) and improved it at lower-density inserts ( P - value = 0.0313 ) with overall MAPE: Incept-net [2.9%, 4.6%]; commercial-VMI [6.7%, 10.9%]. In patient images, Incept-net suppressed beam-hardening artifact and reduced noise (up to 50%, P - value = 0.0156 ). Conclusion: In this preliminary study, Incept-net presented the potential to improve low-energy VMI quality.

Threshold-dependent iodine imaging and spectral separation in a whole-body photon-counting CT system

OBJECTIVE

To evaluate the dual-energy (DE) performance and spectral separation with respect to iodine imaging in a photon-counting CT (PCCT) and compare it to dual-source CT (DSCT) DE imaging.

METHODS

A semi-anthropomorphic phantom extendable with fat rings equipped with iodine vials is measured in an experimental PCCT. The system comprises a PC detector with two energy bins (20 keV, T) and (T, eU) with threshold T and tube voltage U. Measurements using the PCCT are performed at all available tube voltages (80 to 140 kV) and threshold settings (50-90 keV). Further measurements are performed using a conventional energy-integrating DSCT. Spectral separation is quantified as the relative contrast media ratio R between the energy bins and low/high images. Image noise and dose-normalized contrast-to-noise ratio (CNRD) are evaluated in resulting iodine images. All results are validated in a post-mortem angiography study.

RESULTS

R of the PC detector varies between 1.2 and 2.6 and increases with higher thresholds and higher tube voltage. Reference R of the EI DSCT is found as 2.20 on average overall phantoms. Maximum CNRD in iodine images is found for T = 60/65/70/70 keV for 80/100/120/140 kV. The highest CNRD of the PCCT is obtained using 140 kV and is decreasing with decreasing tube voltage. All results could be confirmed in the post-mortem angiography study.

CONCLUSION

Intrinsically acquired DE data are able to provide iodine images similar to conventional DSCT. However, PCCT thresholds should be chosen with respect to tube voltage to maximize image quality in retrospectively derived image sets.

KEY POINTS

• Photon-counting CT allows for the computation of iodine images with similar quality compared to conventional dual-source dual-energy CT. • Thresholds should be chosen as a function of the tube voltage to maximize iodine contrast-to-noise ratio in derived image sets. • Image quality of retrospectively computed image sets can be maximized using optimized threshold settings.

Dual energy computed tomography evaluation of skeletal traumas

Skeletal traumas are among the most common routine challenges faced by Emergency Radiologists, in particular in case of radiographically occult nondisplaced fractures or in case of soft tissue injuries. With the development of Dual Energy Computed Tomography (DECT) technology, new post-processing applications have gained a useful diagnostic role in many fields of musculoskeletal imaging including acute skeletal trauma imaging. In addition to conventional CT images, DECT allows for the generation of virtual calcium-suppressed images subtracting calcium from unenhanced CT images based on the fact that material attenuation varies at different energy levels. In this way, virtual-non-calcium (VNC) images can precisely characterize traumatic bone marrow edema in both axial and appendicular skeleton, facilitating prompt clinical decision, especially when magnetic resonance method is contraindicated or unavailable. Other DECT emerging applications in the trauma setting include metal artifact reduction and collagen mapping for the evaluation of injuries affecting ligament, tendon, and intervertebral disk. This review focuses on the basic principles of DECT and related post-processing algorithms, highlighting the current advantages and limitations of these new imaging advances in the Emergency Department related to skeletal traumas.

Cyclodextrin-Based Contrast Agents for Medical Imaging

Cyclodextrins (CDs) are naturally occurring cyclic oligosaccharides consisting of multiple glucose subunits. CDs are widely used in host–guest chemistry and biochemistry due to their structural advantages, biocompatibility, and ability to form inclusion complexes. Recently, CDs have become of high interest in the field of medical imaging as a potential scaffold for the development of a large variety of the contrast agents suitable for magnetic resonance imaging, ultrasound imaging, photoacoustic imaging, positron emission tomography, single photon emission computed tomography, and computed tomography. The aim of this review is to summarize and highlight the achievements in the field of cyclodextrin-based contrast agents for medical imaging.

Value of spectral detector CT for pretherapeutic, locoregional assessment of esophageal cancer

PURPOSE

To investigate the diagnostic value of spectral detector dual-energy CT-derived low-keV virtual monoenergetic images (VMI) and iodine overlays (IO) for locoregional, pretherapeutic assessment of esophageal cancer.

METHOD

74 patients with biopsy-proven esophageal cancer who underwent pre-therapeutic, portal-venous-phase staging examinations of the chest and abdomen were retrospectively included. Quantitative image analysis was performed ROI-based within the tumor, healthy esophageal wall, peri-esophageal lymph nodes, azygos vein, aorta, liver, diaphragm, and mediastinal fat. Two radiologists evaluated delineation of the primary tumor and locoregional lymph nodes, assessment of the celiac trunk and diagnostic certainty regarding tumor infiltration in conventional images (CI), VMI from 40 to 70 keV and IO. Moreover, presence/absence of advanced tumor infiltration (T3/T4) was determined binary using all available images.

RESULTS

VMI_40-60keV showed significantly higher attenuation and signal-to-noise ratio compared to CI for all assessed ROIs, peaking at VMI_40keV (p < 0.05). Contrast-to-noise ratio of tumor/esophagus (VMI_40keV/CI: 7.7 ± 4.7 vs. 2.3 ± 1.5), tumor/diaphragm (VMI_40keV/CI: 9.0 ± 5.5 vs. 2.2 ± 1.7) and tumor/liver (4.3 ± 5.5 vs. 1.9 ± 2.1) were all significantly higher compared to CI (p < 0.05). Qualitatively, lymph node delineation and diagnostic certainty regarding tumor infiltration received highest ratings both in IO and VMI_40keV, whereas vascular assessment was rated highest in VMI_40keV and primary tumor delineation in IO. Sensitivity/Specificity/Accuracy for detecting advanced tumor infiltration using the combination of CI, VMI_40-70keV and IO was 42.4 %/82.0 %/56.3 %.

CONCLUSIONS

IO and VMI_40-60keV improve qualitative assessment of the primary tumor and depiction of lymph nodes and vessels at pretherapeutic SDCT of esophageal cancer patients yet do not mitigate the limitations of CT in determining tumor infiltration.

Building a dual-energy CT service line in abdominal radiology

As the access of radiology practices to dual-energy CT (DECT) has increased worldwide, seamless integration into clinical workflows and optimized use of this technology are desirable. In this article, we provide basic concepts of commercially available DECT hardware implementations, discuss financial and logistical aspects, provide tips for protocol building and image routing strategies, and review radiation dose considerations to establish a DECT service line in abdominal imaging. KEY POINTS: • Tube-based and detector-based DECT implementations with varying features and strengths are available on the imaging market. • Thorough assessment of financial and logistical aspects is key to successful implementation of a DECT service line. • Optimized protocol building and image routing strategies are of critical importance for effective use and seamless inception of DECT in routine clinical workflows.

Dual-energy CT enterography in evaluation of Crohn's disease: the role of virtual monochromatic images

PURPOSE

To assess the use of virtual monochromatic images (VMI) for discrimination of affected and non-affected bowel walls in patients with Crohn's disease (CD) as well as to compare mural enhancement between patients with and without CD.

MATERIALS AND METHODS

This retrospective study included 61 patients (47 with CD, 14 without CD). Attenuation value (AV), signal-to noise ratio (SNR), and contrast-to-noise ratio (CNR) were obtained at VMI energy levels from 40 to 110 keV in 10 keV increment. Analyses were performed among affected and non-affected bowel walls in CD patients, as well as from bowel walls in patients without CD. Image quality and mural enhancement were evaluated at VMI energy levels at 40, 70, and 110 keV.

RESULTS

At all energy levels of VMI, each quantitative data for AV, SNR, and CNR showed statistically significant difference between diseased and non-diseased bowel walls in CD patients. In the quantitative assessment of patients with and without CD, the optimal AV and SNR were obtained at 40 keV, and the optimal CNR was obtained at 70 keV. For the qualitative assessment, the best image quality and mural enhancement were obtained at 70 keV and 40 keV, respectively.

CONCLUSION

VMI are helpful for the differentiation of affected bowel walls in CD patients, providing high diagnostic accuracy.

Gastrointestinal Applications of Iodine Quantification Using Dual-Energy CT: A Systematic Review

Dual-energy computed tomography (DECT) can estimate tissue vascularity and perfusion via iodine quantification. The aim of this systematic review was to outline current and emerging clinical applications of iodine quantification within the gastrointestinal tract using DECT. The search was conducted with three databases: EMBASE, Pubmed and The Cochrane Library. This identified 449 studies after duplicate removal. From a total of 570 selected studies, 30 studies were enrolled for the systematic review. The studies were categorized into four main topics: gastric tumors (12 studies), colorectal tumors (8 studies), Crohn’s disease (4 studies) and miscellaneous applications (6 studies). Findings included a significant difference in iodine concentration (IC) measurements in perigastric fat between T1–3 vs. T4 stage gastric cancer, poorly and well differentiated gastric and colorectal cancer, responders vs. non-responders following chemo- or chemoradiotherapy treatment among cancer patients, and a positive correlation between IC and Crohn’s disease activity. In conclusion, iodine quantification with DECT may be used preoperatively in cancer imaging as well as for monitoring treatment response. Future studies are warranted to evaluate the capabilities and limitations of DECT in splanchnic flow.

Virtual monoenergetic reconstructions of dynamic DECT acquisitions for calculation of perfusion maps of blood flow: Quantitative comparison to conventional, dynamic 80 kVp CT perfusion

PURPOSE

Investigation of potential improvements in dynamic CT perfusion measurements by exploitation of improved visualization of contrast agent in virtual monoenergetic reconstructions of images acquired with dual-energy computed tomography (DECT).

METHOD

For 17 patients with pancreatic carcinoma, dynamic dual-source DECT acquisitions were performed at 80kVp/Sn140kVp every 1.5 s over 51 s. Virtual monoenergetic images (VMI) were reconstructed for photon energies between 40 keV and 150 keV (5 keV steps). Using the maximum-slope model, perfusion maps of blood flow were calculated from VMIs and 80kVp images and compared quantitatively with regard to blood flow measured in regions of interest in healthy tissue and carcinoma, standard deviation (SD), and absolute-difference-to-standard-deviation ratio (ADSDR) of measurements.

RESULTS

On average, blood flow calculated from VMIs increased with increasing energy levels from 114.3 ± 37.2 mL/100 mL/min (healthy tissue) and 45.6 ± 25.3 mL/100 mL/min (carcinoma) for 40 keV to 128.6 ± 58.9 mL/100 mL/min (healthy tissue) and 75.5 ± 49.8 mL/100 mL/min (carcinoma) for 150 keV, compared to 114.2 ± 37.4 mL/100 mL/min (healthy tissue) and 46.5 ± 26.6 mL/100 mL/min (carcinoma) for polyenergetic 80kVp. Differences in blood flow between tissue types were significant for all energies. Differences between perfusion maps calculated from VMIs and 80kVp images were not significant below 110 keV. SD and ADSDR were significantly better for perfusion maps calculated from VMIs at energies between 40 keV and 55 keV than for those calculated from 80kVp images. Compared to effective dose of dynamic 80kVp acquisitions (4.6 ± 2.2mSv), dose of dynamic DECT/VMI acquisitions (8.0 ± 3.7mSv) was higher.

CONCLUSIONS

Perfusion maps of blood flow based on low-energy VMIs between 40 keV and 55 keV offer improved robustness and quality of quantitative measurements over those calculated from 80kVp image data (reference standard), albeit at increased patient radiation exposure.

Impact of dual energy cardiac CT for metal artefact reduction post aortic valve replacement

PURPOSE

Assess image quality of dual-energy (DE) and single-energy (SE) cardiac multi-detector computed tomographic (MDCT) post aortic valve replacement (AVR) on a dual source MDCT scanner.

METHODS

Eighty patients with cardiac MDCT acquisitions (ECG gated, dual-source) post-surgical and transcatheter AVR were retrospectively identified. Forty DE (cohort 1) and 40 SE acquisitions (cohort 2; 100 or 120 kVp) were reviewed. Metal artefact at valve coaptation (VC) and valve insertion site (VIS), and contrast enhancement were assessed. Valve leaflet edge definition was graded on a 4-point scale by three radiologists.

RESULTS

The mean percentage valve area obscured by metal artifact differed between the cohorts; cohort 1 DE blended, high keV and low keV: 14.8 %, 11.1 % and 17.8 % at VC and 16.4 %, 13 %, 20.4 % at VIS respectively. Cohort 2: 25.8 % and 33.6 % (VC and VIS); each DE reconstruction vs SE: P < 0.0001. Average contrast opacification and coefficient of variance for cohort 1: 562.9 ± 144.7, 281.1 ± 60.3 and 1132.7 ± 300.8 Hounsfield Units (HU) and 9.6 %, 10 % and 8.9 %. For cohort 2: 437.2 ± 119.2 HU and 10.8 % (P < 0.01). Average leaflet edge definition cohort 1: 2.3 ± 0.4, 2.7 ± 0.2 and 2.3 ± 0.2, and cohort 2: 2.9 ± 0.2.

CONCLUSION

DE high keV renderings can result in up to 17.2 % less metal artefact compared to standard SE acquisition for cardiac CT. Contrast opacification and homogeneity is higher for DE blended and low keV renderings compared to SE acquisition with leaflet visibility preferred for low keV and blended DE renderings.

Dual energy CT in clinical routine: how it works and how it adds value

Dual energy computed tomography (DECT), also known as spectral CT, refers to advanced CT technology that separately acquires high and low energy X-ray data to enable material characterization applications for substances that exhibit different energy-dependent x-ray absorption behavior. DECT supports a variety of post-processing applications that add value in routine clinical CT imaging, including material selective and virtual non-contrast images using two- and three-material decomposition algorithms, virtual monoenergetic imaging, and other material characterization techniques. Following a review of acquisition and post-processing techniques, we present a case-based approach to highlight the added value of DECT in common clinical scenarios. These scenarios include improved lesion detection, improved lesion characterization, improved ease of interpretation, improved prognostication, inherently more robust imaging protocols to account for unexpected pathology or suboptimal contrast opacification, length of stay reduction, reduced utilization by avoiding unnecessary follow-up examinations, and radiation dose reduction. A brief discussion of post-processing workflow approaches, challenges, and solutions is also included.

Resectable pancreatic ductal adenocarcinoma: association between preoperative CT texture features and metastatic nodal involvement

Background

To explore the relationship between the lymph node status and preoperative computed tomography images texture features in pancreatic cancer.

Methods

A total of 155 operable pancreatic cancer patients (104 men, 51 women; mean age 63.8 ± 9.6 years), who had undergone contrast-enhanced computed tomography in the arterial and portal venous phases, were enrolled in this retrospective study. There were 73 patients with lymph node metastases and 82 patients without nodal involvement. Four different data sets, with thin (1.25 mm) and thick (5 mm) slices (at arterial phase and portal venous phase) were analysed. Texture analysis was performed by using MaZda software. A combination of feature selection algorithms was used to determine 30 texture features with the optimal discriminative performance for differentiation between lymph node positive and negative groups. The prediction performance of the selected feature was evaluated by receiver operating characteristic (ROC) curve analysis.

Results

There were 10 texture features with significant differences between two groups and significance in ROC analysis were identified. They were WavEnLH_s-2(wavelet energy with rows and columns are filtered with low pass and high pass frequency bands with scale factors 2) from wavelet-based features, 135dr_LngREmph (long run emphasis in 135 direction) and 135dr_Fraction (fraction of image in runs in 135 direction) from run length matrix-based features, and seven variables of sum average from coocurrence matrix-based features (SumAverg). The ideal cutoff value for predicting lymph node metastases was 270 for WavEnLH_s-2 (positive likelihood ratio 2.08). In addition, 135dr_LngREmph and 135dr_Fraction were correlated with the ratio of metastatic to examined lymph nodes.

Conclusions

Preoperative computed tomography high order texture features provide a useful imaging signature for the prediction of nodal involvement in pancreatic cancer.