Citations

AbstractBackgroundDual‐energy CT with virtual noncalcium (VNCa) images allows the evaluation of focal intramedullary bone marrow involvement in patients with multiple myeloma. However, current commercial VNCa techniques suffer from excessive image noise and artifacts due to material decomposition used in synthesizing VNCa images.ObjectivesIn this work, we aim to improve VNCa image quality for the assessment of focal multiple myeloma, using an Artificial intelligence based Generalizable Algorithm for mulTi‐Energy CT (AGATE) method.Materials and methodsAGATE method used a custom dual‐task convolutional neural network (CNN) that concurrently carries out material classification and quantification. The material classification task provided an auxiliary regularization to the material quantification task. CNN parameters were optimized using custom loss functions that involved cross‐entropy, physics‐informed constraints, structural redundancy in spectral and material images, and texture information in spectral images. For training data, CT phantoms (diameters 30 to 45 cm) with tissue‐mimicking inserts were scanned on a third generation dual‐source CT system. Scans were performed at routine dose and half of the routine dose. Small image patches (i.e., 40 × 40 pixels) of tissue‐mimicking inserts with known basis material densities were extracted for training samples. Numerically simulated insert materials with various shapes increased diversity of training samples. Generalizability of AGATE was evaluated using CT images from phantoms and patients. In phantoms, material decomposition accuracy was estimated using mean‐absolute‐percent‐error (MAPE), using physical inserts that were not used during the training. Noise power spectrum (NPS) and modulation transfer function (MTF) were compared across phantom sizes and radiation dose levels. Five patients with multiple myeloma underwent dual‐energy CT, with VNCa images generated using a commercial method and AGATE. Two fellowship‐trained musculoskeletal radiologists reviewed the VNCa images (commercial and AGATE) side‐by‐side using a dual‐monitor display, blinded to VNCa type, rating the image quality for focal multiple myeloma lesion visualization using a 5‐level Likert comparison scale (−2 = worse visualization and diagnostic confidence, −1 = worse visualization but equivalent diagnostic confidence, 0 = equivalent visualization and diagnostic confidence, 1 = improved visualization but equivalent diagnostic confidence, 2 = improved visualization and diagnostic confidence). A post hoc assignment of comparison ratings was performed to rank AGATE images in comparison to commercial ones.ResultsAGATE demonstrated consistent material quantification accuracy across phantom sizes and radiation dose levels, with MAPE ranging from 0.7% to 4.4% across all testing materials. Compared to commercial VNCa images, the AGATE‐synthesized VNCa images yielded considerably lower image noise (50–77% noise reduction) without compromising noise texture or spatial resolution across different phantom sizes and two radiation doses. AGATE VNCa images had markedly reduced area under NPS curves and maintained NPS peak frequency (0.7 lp/cm to 1.0 lp/cm), with similar MTF curves (50% MTF at 3.0 lp/cm). In patients, AGATE demonstrated reduced image noise and artifacts with improved delineation of focal multiple myeloma lesions (all readers comparison scores indicating improved overall diagnostic image quality [scores 1 or 2]).ConclusionsAGATE demonstrated reduced noise and artifacts in VNCa images and ability to improve visualization of bone marrow lesions for assessing multiple myeloma.

Abstract Objective. To develop a convolutional neural network (CNN) noise reduction technique for ultra-high-resolution photon-counting detector computed tomography (UHR-PCD-CT) that can be efficiently implemented using only clinically available reconstructed images. The developed technique was demonstrated for skeletal survey, lung screening, and head angiography (CTA). Approach . There were 39 participants enrolled in this study, each received a UHR-PCD and an energy integrating detector (EID) CT scan. The developed CNN noise reduction technique uses image-based noise insertion and UHR-PCD-CT images to train a U-Net via supervised learning. For each application, 13 patient scans were reconstructed using filtered back projection (FBP) and iterative reconstruction (IR) and allocated into training, validation, and testing datasets (9:1:3). The subtraction of FBP and IR images resulted in approximately noise-only images. The 5-slice average of IR produced a thick reference image. The CNN training input consisted of thick reference images with reinsertion of spatially decoupled noise-only images. The training target consisted of the corresponding thick reference images without noise insertion. Performance was evaluated based on difference images, line profiles, noise measurements, nonlinear perturbation assessment, and radiologist visual assessment. UHR-PCD-CT images were compared with EID images (clinical standard). Main results. Up to 89% noise reduction was achieved using the proposed CNN. Nonlinear perturbation assessment indicated reasonable retention of 1 mm radius and 1000 HU contrast signals (>80% for skeletal survey and head CTA, >50% for lung screening). A contour plot indicated reduced retention for small-radius and low contrast perturbations. Radiologists preferred CNN over IR for UHR-PCD-CT noise reduction. Additionally, UHR-PCD-CT with CNN was preferred over standard resolution EID-CT images. Significance. CT images reconstructed with very sharp kernels and/or thin sections suffer from increased image noise. Deep learning noise reduction can be used to offset noise level and increase utility of UHR-PCD-CT images.

Abstract Background A common rule of thumb for object detection is the Rose criterion, which states that a signal must be five standard deviations above background to be detectable to a human observer. The validity of the Rose criterion in CT imaging is limited due to the presence of correlated noise. Recent reconstruction and denoising methodologies are also able to restore apparent image quality in very noisy conditions, and the ultimate limits of these methodologies are not yet known. Purpose To establish a lower bound on the minimum achievable signal‐to‐noise ratio (SNR) for object detection, below which detection performance is poor regardless of reconstruction or denoising methodology. Methods We consider a numerical observer that operates on projection data and has perfect knowledge of the background and the objects to be detected, and determine the minimum projection SNR that is necessary to achieve predetermined lesion‐level sensitivity and case‐level specificity targets. We define a set of discrete signal objects that encompasses any lesion of interest and could include lesions of different sizes, shapes, and locations. The task is to determine which object of is present, or to state the null hypothesis that no object is present. We constrain each object in to have equivalent projection SNR and use Monte Carlo methods to calculate the required projection SNR necessary. Because our calculations are performed in projection space, they impose an upper limit on the performance possible from reconstructed images. We chose to be a collection of elliptical or circular low contrast metastases and simulated detection of these objects in a parallel beam system with Gaussian statistics. Unless otherwise stated, we assume a target of 80% lesion‐level sensitivity and 80% case‐level specificity and a search field of view that is 6 cm by 6 cm by 10 slices. Results When contains only a single object, our problem is equivalent to two‐alternative forced choice (2AFC) and the required projection SNR is 1.7. When consists of circular 6‐mm lesions at different locations in space, the required projection SNR is 5.1. When is extended to include ellipses and circles of different sizes, the required projection SNR increases to 5.3. The required SNR increases if the sensitivity target, specificity target, or search field of view increases. Conclusions Even with perfect knowledge of the background and target objects, the ideal observer still requires an SNR of approximately 5. This is a lower bound on the SNR that would be required in real conditions, where the background and target objects are not known perfectly. Algorithms that denoise lesions with less than 5 projection SNR, regardless of the denoising methodology, are expected to show vanishing effects or false positive lesions.

Objective The aim of this study was to evaluate the clinical impact of a higher spatial resolution, full field-of-view investigational photon-counting detector computed tomography (PCD-CT) on radiologist confidence in imaging findings and diagnosis of usual interstitial pneumonia (UIP) compared with conventional energy-integrating detector CT (EID-CT). Materials and Methods Patients suspected of interstitial lung disease were scanned on a PCD-CT system after informed consent and a clinically indicated EID-CT. In 2 sessions, 3 thoracic radiologists blinded to clinical history and scanner type evaluated CT images of the right and left lungs separately on EID- or PCD-CT, reviewing each lung once/session, rating confidence in imaging findings of reticulation, traction bronchiectasis, honeycombing, ground-glass opacities (GGOs), mosaic pattern, and lower lobe predominance (100-point scale: 0–33, likely absent; 34–66, indeterminate; 67–100, likely present). Radiologists also rated confidence for the probability of UIP (0–20, normal; 21–40, inconsistent with UIP; 41–60, indeterminate UIP; 61–81; probable UIP; 81–100, definite UIP) and graded image quality. Because a confidence scale of 50 represented completely equivocal findings, magnitude score (the absolute value of confidence scores from 50) was used for analysis (higher scores were more confident). Image noise was measured for each modality. The magnitude score was compared using linear mixed effects regression. The consistency of findings and diagnosis between 2 scanners were evaluated using McNemar test and weighted κ statistics, respectively. Results A total of 30 patients (mean age, 68.8 ± 11.0 years; M:F = 18:12) underwent conventional EID-CT (median CTDIvol, 7.88 mGy) and research PCD-CT (median CTDIvol, 6.49 mGy). The magnitude scores in PCD-CT were significantly higher than EID-CT for imaging findings of reticulation (40.7 vs 38.3; P = 0.023), GGO (34.4 vs 31.7; P = 0.019), and mosaic pattern (38.6 vs 35.9; P = 0.013), but not for other imaging findings (P ≥ 0.130) or confidence in UIP (34.1 vs 22.2; P < 0.059). Magnitude score of probability of UIP in PCD-CT was significantly higher than EID-CT in one reader (26.0 vs 21.5; P = 0.009). Photon-counting detector CT demonstrated a decreased number of indeterminate GGO (17 vs 26), an increased number of unlikely GGO (74 vs 50), and an increased number of likely reticulations (140 vs 130) relative to EID-CT. Interobserver agreements among 3 readers for imaging findings and probability of UIP were similar between PCD-CT and EID-CT (intraclass coefficient: 0.507–0.818 vs 0.601–0.848). Photon-counting detector CT had higher scores in overall image quality (4.84 ± 0.38) than those in EID-CT (4.02 ± 0.40; P < 0.001) despite increased image noise (mean 85.5 vs 36.1 HU). Conclusions Photon-counting detector CT provided better image quality and improved the reader confidence for presence or absence of imaging findings of reticulation, GGO, and mosaic pattern with idiosyncratic improvement in confidence in UIP presence.

AbstractPurposeThe purpose of this work is to evaluate the scaled computed tomography (CT) number accuracy of an artificial 120 kV reconstruction technique based on phantom experiments in the context of radiation therapy planning.MethodsAn abdomen‐shaped electron density phantom was scanned on a clinical CT scanner capable of artificial 120 kV reconstruction using different tube potentials from 70 to 150 kV. A series of tissue‐equivalent phantom inserts (lung, adipose, breast, solid water, liver, inner bone, 30%/50% CaCO3, cortical bone) were placed inside the phantom. Images were reconstructed using a conventional quantitative reconstruction kernel as well as the artificial 120 kV reconstruction kernel. Scaled CT numbers of inserts were measured from images acquired at different kVs and compared with those acquired at 120 kV, which were deemed as the ground truth. The relative error was quantified as the percentage deviation of scaled CT numbers acquired at different tube potentials from their ground truth values acquired at 120 kV.ResultsScaled CT numbers measured from images reconstructed using the conventional reconstruction demonstrated a strong kV‐dependence. The relative error in scaled CT numbers ranged from 0.6% (liver insert) to 31.1% (cortical bone insert). The artificial 120 kV reconstruction reduced the kV dependence, especially for bone tissues. The relative error in scaled CT number was reduced to 0.4% (liver insert) and 2.6% (30% CaCO3 insert) using this technique. When tube potential selection was limited to the range of 90 to 150 kV, the relative error was further restrained to <1.2% for all tissue types.ConclusionPhantom results demonstrated that using the artificial 120 kV technique, it was feasible to acquire raw projection data at the desired tube potential and then reconstruct images with scaled CT numbers comparable to those obtained directly at 120 kV. In radiotherapy applications, this technique may allow optimization of tube potential without complicating clinical workflow by eliminating the necessity of maintaining multiple sets of CT calibration curves.