لخّصلي

qCT: AI-powered Lung Nodule Analysis

Qure.ai's qCTv1.0 is a deep learning software for analyzing chest CT scans. It detects, labels, visualizes, and tracks lung nodules, aiding in their characterization and progression monitoring. It can process both contrast and non-contrast scans, serving as a second or concurrent reader.

qCT calculates the Brock score, considering factors like nodule size, type, location, count, and spiculation. It operates on-premises and in the cloud, interacting with PACS systems to receive and return results.

qCT highlights segmented lesions on the scan, providing volumetric measurements. These measurements help assess lesion severity and disease progression, enabling volume doubling time (VDT) analysis.

qCT's deep learning algorithms work independently, providing probability scores that are then thresholded to determine the presence or absence of abnormalities. Pre- and post-processing features integrate the software into existing radiology workflows.

The user receives image previews with highlighted findings and quantifications, along with a text report outlining the findings and volumetric measurements.

1 GENERAL INFORMATION
1.1 About qCT
Qure.ai’s computed tomography (CT) scan software, qCTv1.0, is a deep-learning-based
interpretation, detection, labeling, visualization, characterization, and monitoring progression
of lung nodules in Chest CT scans. The software can be used as second or concurrent reader,
which can process both contrast and non-contract CT (NCCT) scans. qCT provides an option
to calculate Brock score, for every nodule by estimating nodule size, nodule type, lobular
localization, nodule count and presence of spiculation (Section 1.4).
qCT consists of an on-premises module and a cloud module. The software interacts with
DICOM’s modality or the user’s picture archiving and communication system (PACS) to
receive de-identified Chest CT scans and returns the results to the same destination. In
addition, lesions segmented and quantified by qCT are presented to users as an additional
overlay, highlighting the location and labeling the volume of the structures and lesions
segmented. qCT accepts contrast and non-contrast adult Chest CT scan DICOM files as
input and provides an overlay highlighting the location of lung nodules in the scan. The
overlays generated by qCT can be presented in two ways to the clinicians:
a. PACS-based mode: As a new series (secondary capture) which are sent to the originating
PACS system.
b. Web-based mode: On Qure’s web application , the overlays are burnt on top of the original
scan.
The volumetric segmentations are most relevant in the assessment of patients with a
potential lung nodule. Clinicians may use these quantitative measurements to assist with
determining the severity of the lesion, or underlying disease, or to assist with the comparison
of multiple CT scans from the same patient acquired on the same CT scanner, enabling
measurement of their progress and changes over time (Volume Doubling Time [VDT]).
The abnormality detection deep learning analysis module underlying qCT consists of a set of
independent algorithms, each providing a probability score. A pre-specified threshold is then
applied to the score to generate a ‘present’ or ‘absent’ status for each abnormality. The set
of deep learning algorithms that have been trained to outline and quantify the target
structures on qCT scans are coupled with pre- and post-processing functionality that allows
the device to work directly with the radiology workflow.
The user is presented with a preview of images highlighting the abnormal findings, a preview
of images highlighting and labeling the quantification based segmented structures, a text
report indicating the type and location of findings, along with the volumetric quantification