No patient incident reported at this moment - this issue is documented as a precaution.One customer (b)(4) was registered and investigated.Bugs identified: (b)(4): (revert of (b)(4)) iv: radiographic scaling factor derived from distance to patient and detector is being applied to non-projection radiography modalities (b)(4): iv (measurement): measurements are not calibrated when dicom is missing mag factor, but has distance to patient and detector plate.Investigation: there is indeed a problem with the measurements after the fix introduced in (b)(4).Iv now uses the ratio of the distance from the source to patient over distance from source to detector to compute a scaling factor to apply to measurements, but that scaling should only be applied when dealing with a projection radiography modality (cr, dx, mg, px, rf, or xa).Unfortunately, it is now being applied to all modalities.There is a guard in the code that prevents comments about the scaling factor to be displayed on modalities that are not projection radiography, but the scaling factor itself is applied in all cases, which is wrong.Root cause of the problem: since the introduction of a fix to the measurements algorithm ((b)(4)), inteleviewer now uses the ratio of the distance from the source to patient over distance from source to detector to compute a scaling factor to apply to measurements when the estimated radiographic magnification factor (dicom (0018,1114)) is not present.That scaling should only be applied when dealing with a projection radiography modality (cr, dx, mg, px, rf, or xa), however, it ended up being applied to other modalities, such as ct, generating erroneous measurements.Reason for possible patient safety: the radiologist could base their diagnosis on incorrect anatomical measurements.Risk analysis: this problem is documented as a known risk and is mitigated by design: consistent use of mapping functions to patient anatomical space and calibrated units before performing coordinate and intensity calculations.For clients using this specific software version, this problem could will occur with approximately 50% of all ct studies.Risk to the patient could occur only if the modality is ct and if the discrepancy of the measurement is small.In order to correct this issue, intelerad will: immediately: notify to clients to stop using the affected version and to revert to an anterior or ulterior software version.Immediately: remove the affected software version from use - make it unavailable on the clients' servers.Immediately: flag the version internally as a no deploy version.Fix the affected version.Run server side scripts on clients' systems to identify studies read with the affected version.Send recall notice to affected clients ( 4 ).Communicate to clients the list of exams that were read using the defective version.Run additional server-side scripts to ensure that no studies were read with the affected version after clients were notified.Estimated date to complete corrective action: end of february 2016.
|