ROS1 genomic rearrangements are rare actionable drivers in microsatellite stable colorectal cancer.

c-Ros oncogene 1, receptor tyrosine kinase (ROS1) genomic rearrangements have been reported previously in rare cases of colorectal cancer (CRC), yet little is known about the frequency, molecular characteristics, and therapeutic vulnerabilities of ROS1-driven CRC. We analyzed a clinical dataset of 40 589 patients with CRC for ROS1 genomic rearrangements and their associated genomic characteristics (Foundation Medicine, Inc [FMI]). We moreover report the disease course and treatment response of an index patient with ROS1-rearranged metastatic CRC. ROS1 genomic rearrangements were identified in 34 (0.08%) CRC samples. GOPC-ROS1 was the most common ROS1 fusion identified (11 samples), followed by TTC28-ROS1 (3 samples). Four novel 5' gene partners of ROS1 were identified (MCM9, SRPK1, EPHA6, P4HA1). Contrary to previous reports on fusion-positive CRC, ROS1-rearrangements were found exclusively in microsatellite stable (MSS) CRCs. KRAS mutations were significantly less abundant in ROS1-rearranged vs ROS1 wild type cases. The index patient presented with chemotherapy-refractory metastatic right-sided colon cancer harboring GOPC-ROS1. Molecularly targeted treatment with crizotinib induced a rapid and sustained partial response. After 15 months on crizotinib disseminated tumor progression occurred and KRAS Q61H emerged in tissue and liquid biopsies. ROS1 rearrangements define a small, yet therapeutically actionable molecular subgroup of MSS CRC. In summary, the high prevalence of GOPC-ROS1 and noncanonical ROS1 fusions pose diagnostic challenges. We advocate NGS-based comprehensive molecular profiling of MSS CRCs that are wild type for RAS and BRAF and patient enrollment in precision trials.

Spiral breast computed tomography (CT): signal-to-noise and dose optimization using 3D-printed phantoms.

OBJECTIVES: To investigate the dependence of signal-to-noise ratio (SNR) and calculated average dose per volume of spiral breast-CT (B-CT) on breast size and breast density and to provide a guideline for choosing the optimal tube current for each B-CT examination. MATERIALS AND METHODS: Three representative B-CT datasets (small, medium, large breast size) were chosen to create 3D-printed breast phantoms. The phantoms were filled with four different agarose-oil-emulsions mimicking differences in breast densities. Phantoms were scanned in a B-CT system with systematic variation of the tube current (6, 12.5, 25, 32, 40, 50, 64, 80, 100, 125 mA). Evaluation of SNR and the average dose per volume using Monte Carlo simulations were performed for high (HR) and standard (STD) spatial resolution. RESULTS: SNR and average dose per volume increased with increasing tube current. Artifacts had negligible influence on image evaluation. SNR values ≥ 35 (HR) and ≥ 100 (STD) offer sufficient image quality for clinical evaluation with SNR being more dependent on breast density than on breast size. For an average absorbed dose limit of 6.5 mGy for the medium and large phantoms and 7 mGy for the small phantom, optimal tube currents were either 25 or 32 mA. CONCLUSIONS: B-CT offers the possibility to vary the X-ray tube current, allowing image quality optimization based on individual patient's characteristics such as breast size and density. This study describes the optimal B-CT acquisition parameters, which provide diagnostic image quality for various breast sizes and densities, while keeping the average dose at a level similar to digital mammography. KEY POINTS: • Image quality optimization based on breast size and density varying the tube current using spiral B-CT.

INSIDEnet: Interpretable NonexpanSIve Data-Efficient network for denoising in grating interferometry breast CT.

PURPOSE: Breast cancer is the most common malignancy in women. Unfortunately, current breast imaging techniques all suffer from certain limitations: they are either not fully three dimensional, have an insufficient resolution or low soft-tissue contrast. Grating interferometry breast computed tomography (GI-BCT) is a promising X-ray phase contrast modality that could overcome these limitations by offering high soft-tissue contrast and excellent three-dimensional resolution. To enable the transition of this technology to clinical practice, dedicated data-processing algorithms must be developed in order to effectively retrieve the signals of interest from the measured raw data. METHODS: This article proposes a novel denoising algorithm that can cope with the high-noise amplitudes and heteroscedasticity which arise in GI-BCT when operated in a low-dose regime to effectively regularize the ill-conditioned GI-BCT inverse problem. We present a data-driven algorithm called INSIDEnet, which combines different ideas such as multiscale image processing, transform-domain filtering, transform learning, and explicit orthogonality to build an Interpretable NonexpanSIve Data-Efficient network (INSIDEnet). RESULTS: We apply the method to simulated breast phantom datasets and to real data acquired on a GI-BCT prototype and show that the proposed algorithm outperforms traditional state-of-the-art filters and is competitive with deep neural networks. The strong inductive bias given by the proposed model's architecture allows to reliably train the algorithm with very limited data while providing high model interpretability, thus offering a great advantage over classical convolutional neural networks (CNNs). CONCLUSIONS: The proposed INSIDEnet is highly data-efficient, interpretable, and outperforms state-of-the-art CNNs when trained on very limited training data. We expect the proposed method to become an important tool as part of a dedicated plug-and-play GI-BCT reconstruction framework, needed to translate this promising technology to the clinics.

Dual-Energy Low-keV or Single-Energy Low-kV CT for Endoleak Detection?: A 6-Reader Study in an Aortic Aneurysm Phantom.

OBJECTIVES: The aim of this study was to compare image quality, conspicuity, and endoleak detection between single-energy low-kV images (SEIs) and dual-energy low-keV virtual monoenergetic images (VMIs+) in computed tomography angiography of the aorta after endovascular repair. MATERIALS AND METHODS: An abdominal aortic aneurysm phantom simulating 36 endoleaks (2 densities; diameters: 2, 4, and 6 mm) in a medium- and large-sized patient was used. Each size was scanned using single-energy at 80 kVp (A) and 100 kVp (B), and dual-energy at 80/Sn150kVp for the medium (C) and 90/Sn150kVp for the large size (D). VMIs+ at 40 keV and 50 keV were reconstructed from protocols C and D. Radiation dose was 3 mGy for the medium and 6 mGy for the large size. Objective image quality and normalized noise power spectrum were determined. Subjective image quality, conspicuity, and sensitivity for endoleaks were independently assessed by 6 radiologists. Sensitivity was compared using Marascuilo procedure and Fisher exact test. Conspicuities were compared using Wilcoxon-matched pairs test, analysis of variance, and Tukey test. RESULTS: The contrast-to-noise-ratio of the aorta was significantly higher for VMI+ compared with SEI (P < 0.001). Noise power spectrum showed a higher noise magnitude and coarser texture in VMI+. Subjective image quality and overall conspicuity was lower for VMI+ compared with SEI (P < 0.05). Sensitivity for endoleaks was overall higher in the medium phantom for SEI (60.9% for A, 62.2% for B) compared with VMI+ (54.2% for C, 49.3% for D) with significant differences between protocols B and D (P < 0.05). In the large phantom, there was no significant difference in sensitivity among protocols (P = 0.79), with highest rates for protocols B (31.4%) and C (31.7%). CONCLUSIONS: Our study indicates that low-keV VMI+ results in improved contrast-to-noise-ratio of the aorta, whereas noise properties, subjective image quality, conspicuity, and sensitivity for endoleaks were overall superior for SEI.

Metal Artifact Reduction in Pelvic Computed Tomography With Hip Prostheses: Comparison of Virtual Monoenergetic Extrapolations From Dual-Energy Computed Tomography and an Iterative Metal Artifact Reduction Algorithm in a Phantom Study.

OBJECTIVE: The aim of this study was to directly compare metal artifact reduction (MAR) of virtual monoenergetic extrapolations (VMEs) from dual-energy computed tomography (CT) with iterative MAR (iMAR) from single energy in pelvic CT with hip prostheses. MATERIALS AND METHODS: A human pelvis phantom with unilateral or bilateral metal inserts of different material (steel and titanium) was scanned with third-generation dual-source CT using single (120 kVp) and dual-energy (100/150 kVp) at similar radiation dose (CT dose index, 7.15 mGy). Three image series for each phantom configuration were reconstructed: uncorrected, VME, and iMAR. Two independent, blinded radiologists assessed image quality quantitatively (noise and attenuation) and subjectively (5-point Likert scale). Intraclass correlation coefficients (ICCs) and Cohen κ were calculated to evaluate interreader agreements. Repeated measures analysis of variance and Friedman test were used to compare quantitative and qualitative image quality. Post hoc testing was performed using a corrected (Bonferroni) P < 0.017. RESULTS: Agreements between readers were high for noise (all, ICC ≥ 0.975) and attenuation (all, ICC ≥ 0.986); agreements for qualitative assessment were good to perfect (all, κ ≥ 0.678). Compared with uncorrected images, VME showed significant noise reduction in the phantom with titanium only (P < 0.017), and iMAR showed significantly lower noise in all regions and phantom configurations (all, P < 0.017). In all phantom configurations, deviations of attenuation were smallest in images reconstructed with iMAR. For VME, there was a tendency toward higher subjective image quality in phantoms with titanium compared with uncorrected images, however, without reaching statistical significance (P > 0.017). Subjective image quality was rated significantly higher for images reconstructed with iMAR than for uncorrected images in all phantom configurations (all, P < 0.017). CONCLUSIONS: Iterative MAR showed better MAR capabilities than VME in settings with bilateral hip prosthesis or unilateral steel prosthesis. In settings with unilateral hip prosthesis made of titanium, VME and iMAR performed similarly well.