search

finnlennartsson / kpum_noddi

repo for processing of DKI/NODDI data for preterm project at KPUM
0 stars 0 forks source link

anonymization not complete on original DCM data #1

Closed finnlennartsson closed 1 year ago

finnlennartsson commented 1 year ago

Original DCM is not properly anonymized, e.g. listed below are PatientName, PatientID and PatientBirthDate (in the below this entries have been exchanged for PATIENT^NAME, PATIENT_ID and DoB)

(base) finn@laptop:/media/finn/HD-PCFSU3-A/1NODDI/Finn/NODDI_DICOM$ mrinfo 8379225_20221012/
mrinfo: [done] scanning DICOM folder "8379225_20221012/"
Select patient (q to abort):
   1 - 001 (PATIENT_ID) DoB
   2 - PATIENT^NAME (PATIENT_ID) DoB

Dr A said that the patient name had been replaced during the anonymization with 001. This seems to have been the case (PATIENT^NAME => 001), but PATIENT_ID and DoB have not. Likewise, there seems to be both the (patient name) anonymized dataset and the original dataset.

Testing whether with rename routine dcm2niix -r y makes any different, as in routine DCMdicomdir => DCMsourcedata. 

dcm2niix -b o -r y -w 1 -o /media/finn/ID1037/Kyoto_collab/preterm_noddi/data/sourcedata -f sub-001/%t_s%2s_%d/%d_%5r.dcm 8379225_20221012/NODDI_DICOM/DICOM

but it does NOT

(base) finn@laptop:/media/finn/HD-PCFSU3-A/1NODDI/Finn/data/sourcedata$ dcmdump sub-001/20221012112601_s09_DKI_2.5mm_iso_DFC_MIX/DKI_2.5mm_iso_DFC_MIX_00001.dcm | grep Patient
(0010,0010) PN [001]                                    #  18, 1 PatientName
(0010,0020) LO [PATIENT_ID]                             #  10, 1 PatientID
(0010,0030) DA [DoB]                               #   8, 1 PatientBirthDate
(0010,0040) CS [F]                                      #   2, 1 PatientSex
(0010,1010) AS [003M]                                   #   4, 1 PatientAge
(0010,1020) DS [0.39]                                   #   4, 1 PatientSize
(0010,1030) DS [2]                                      #   2, 1 PatientWeight
(0010,4000) LT [20221005]                               #   8, 1 PatientComments
(0018,5100) CS [HFS]                                    #   4, 1 PatientPosition
(0020,0032) DS [-764.20561026947\-579.288908303\159.45494567086] #  48, 3 ImagePositionPatient
(0020,0037) DS [0.99555789647773\-0.0900822965147\0.02737982131953\0.0941079408018... # 102, 6 ImageOrientationPatient
finnlennartsson commented 1 year ago

Solution is to adequately anonymize the DICOMs in the DCM data extraction.

Until then we need to keep /dicomdir and /sourcedata folders on external HD