Quick Answers to Important Questions

ClearRead CT can potentially process all axial CT chest studies that meet the following:

• Consistent slice thickness between 0.5 mm and 3.0 mm
• Consistent slice spacing with no gaps

ClearRead CT CAC basic requirements:

• Whole heart visible in series
• Without contrast

ClearRead CT Nodule basic requirements:

• Input scan is expected to contain both lungs, the field of view, and should not clip the lung
• Full chest wall visible in the series

Common study types processed include:

• Lung cancer screening
• Chest without contrast
• Chest with contrast
• Low dose CT
• Chest/Abdomen/Pelvis exams
• Additional study types can be processed but are best considered in later stages of implementation

Caveats:

• Partial lung exams will have a greater chance of false positives near the break-point of the lung
• Incomplete field of view exams are not recommended for processing

The answer is “yes” for pulmonary nodule detection and “no” for CAC scoring.

Only in the most basic sense – ClearRead CT and CAC require true axial series that are between 0.5 mm and 3.0 mm slice thickness with consistent gapless spacing. Overly sharp enhanced reconstruction kernels can cause false negatives, but most body and lung kernels process without issue.

The only requirement ClearRead CT poses to PACS vendors is being able to use frame of reference sequencing. We use this for synchronizing the source, Vessel Suppress, and Index series and being able to scroll them together.

ClearRead CT is aimed at nodules between 4 mm and 30 mm. Some slightly larger nodules will sometimes be marked, up to approximately 35 mm.

ClearRead CT Vessel Suppress will show smaller and larger nodules, even though Detect only marks between 4 and 30 mm.

ClearRead CT can detect solid, part-solid, and ground glass nodules and provide characterizations of them automatically.

Yes, Vessel Suppress is the basis of our Clear Visual Intelligence focus, providing a much less obstructed view of the lungs. Vessel Suppress finds and removes everything it is certain is normal, leaving behind residuals such as nodules and sometimes other structures.

Detect performs a second pass on the Vessel Suppress output checking all the residuals and comparing them to ClearRead’s internal nodule models, providing characterizations when a residual flags as a nodule.

Compare performs Vessel Suppress and Detect on both a current and a singular prior, then attempts to match the nodules found in the current with appropriate findings in the prior. If it is able to do so, it provides change measurements like average diameter change, volume change, and volume doubling time automatically.

When ClearRead CT receives a new series for analysis, it goes through several steps before producing the Vessel Suppress series:

• Normalization
  • Detector-dependent image features are extracted
  • De-noising algorithms are applied
  • Reformatting to an approximate 1 mm voxel spacing is performed

• Analysis of the volume is performed, mapping out the pulmonary-vascular and quasi-linear structures of the lungs
• Everything ClearRead the system is certain is normal, vessel structure is removed from the derived output series.
• Anything that ClearRead perceives as less than certain is retained as a visual residual
• Image features extracted during normalization are added back in to make the series match the source.
• Vessel Suppress is complete.

For ClearRead, image normalization is the process by which we remove the detector differences, improve the clarity of the series by de-noising, and preparing for analysis. Riverain’s normalization process is strong enough that the only differences to the ClearRead system for analysis are resolution differences.

Different reconstruction kernels can affect this by skewing the baseline beyond what ClearRead can account for with certainty. Most common reconstruction kernels do not pose concern regarding this.

The Compare portion of ClearRead CT is specifically designed to provide current-to-prior comparisons and change calculations. ClearRead CT always re-analyzes the prior to make sure it is measured with the same algorithm as the current study, so ClearRead CT does not need to build a prior base before it can start comparisons.

ClearRead CT Compare provides the following characterizations when processing, diameters and volume are for both current series and prior series:

• Lobar location
• Average diameter (equivalent to Fleischner Mean Diameter)
• Minimum diameter (equivalent to Fleischner Short Diameter)
• Maximum diameter (equivalent to Fleischner Long Diameter)
• Volume (measured in mm³, determined by counting the voxels within the 3-D segmentation)
• Volume change between current and prior, expressed as a percentage (negatives denote nodule shrinking)
• Volume doubling time calculated with using the modified Schwartz formula: VDT = [ln2 × ∆T] / [ln(X2/X1)]