NLST CT Images -- General Details
Numbers of Images
- LSS: 48,546 CT exams; 12,230,145 images from 17,043 participants.
- ACRIN: 24.572 CT exams; 8,852,357 images from 9,211 participants. A few recently-discovered studies will soon be added.
- Total: 73,118 CT exams; 21,082,502 images from 26,254 participants.
Dates (Study Dates)
The Study Dates are fictitious. While appearing real, they are coded to represent the 3 screening years:
- 19990102 -- Screening Year 0
- 20000102 -- Screening Year 1
- 20010102 -- Screening Year 2
The Patient Identifiers (PIDs), in DICOM tag (0010, 0020) Patient ID, are 6 digit random numbers assigned by IMS. They are called IMS-Global-IDs. PIDs beginning with '1' are LSS images; those beginning with '2' are ACRIN images; examples:
- 101337 is an LSS PID
- 208344 is an ACRIN PID
Prior ACRIN and LSS PIDs may be found in 2 DICOM tags: (0010, 1000) Other Patient IDs and (0013, 0040) Clinical Trial Subject ID. These prior-PIDs are random 4- or 5-digit (ACRIN) or 6-digit (LSS) numbers that contain no screening site information; the LSS prior-PIDs may have leading zeroes that may or may not appear unless specified that they do so; examples:
- 2542 is an ACRIN prior-PID
- 18519 is an ACRIN prior-PID
- 022921 is an LSS prior-PID that may also be rendered as 22921
It is possible, if an LSS prior-PID has been rendered without leading zeroes, that the LSS prior-PID may appear to match an ACRIN prior-PID; but, they are not related in any way and represent two distinct NLST participants.
Patient Names appearing in DICOM tag (0010, 0010) are a construct from the participant's prior PID concatenated with a caret and the string 'ACRIN' or 'LSS'; examples:
Private DICOM Tags
DICOM private tags have been removed from all NLST images because protected health information (PHI) was discovered hidden in some private tags.
The original Series Descriptions, if present, were often vague and/or, at best, inconsistent.
New Series Descriptions consist of 11 coded comma-separated values, listed below. Code value descriptions may be found in the attachment Series-Description-Details.xlsx.
- (1) Screening Year
- (2) Image Type
- (3) Manufacturer
- (4) Scanner Model
- (5) Convolution Kernel
- (6) Reconstruction Diameter
- (7) Slice Thickness
- (8) kVp
- (9) mAs*
- (10) Effective mAs*
- (11) Pitch*
* In order to calculate 'effective mAs' (=mAs/pitch), one needs to first determine mAs and pitch. Understanding both of these require extensive reading of manufacturer DICOM conformance statements and explanations from the manufacturers. Values neede to determine mAs and pitch are sometimes stored in private attributes. Discussions with manufacturers' engineers allowed us to interpret those private attributes and extract the information necessary to to determine effective mAs, mAs, and pitch. Extractions were made prioir to the removal of DICOM private tags because some contained protected health information (PHI). Where pitch could not be determined, a 'null' is reported.
DICOM Standard -- CT Image Module Attributes: References
The portion of the DICOM Standard describing CT Image Module Attributes may be found in the attachment DICOM-CT-Module-11_03pu.pdf. The 3rd column of the attachment specifies the 'Type' for each attribute; DICOM Data Element Types are described in the attachment DICOM-Data-Element-Types-11_05pu.pdf. Addional details regarding DICOM tags found in the DICOM headers of CT images not found in the DICOM CT Module may be found in the attachments DICOM-General-Equipment-Module-11_03pu.pdf and DICOM-Image-Modules-11_03pu.pdf.