How to determine if single frame DICOM is actually slice of MRI?

Question

I'm writing an application to count medical records. The application will count pages and divide it by 20 to determine the number of films needed to print the records for the client. I have a closet case where when MRI's are scanned shots (especially if they show as SC), it will only be one frame but the film is already a 20 frame MRI.

How can I determine if the a single DICOM is actually a multi frame MRI?

If I try to do a frame count on the file, I get 1. Here is the DICOM Dump for one of the files:

(0005,0005) CS Private Creator 'NovaRad 6.0'
(0005,0006) SS Private Creator '1'
(0005,000b) CS Private Creator 'DOCTOR BOB'
(0005,0059) PN Private Creator 'PATIENT'
(0008,0000) UL Group Length '296'
(0008,0005) CS Specific Character Set 'ISO_IR 100'
(0008,0016) UI SOP Class UID '1.2.840.10008.5.1.4.1.1.7'
(0008,0018) UI SOP Instance UID '1.2.392.12345.030815.2015.8.4.10.38.18.63'
(0008,0020) DA Study Date '20131229'
(0008,0030) TM Study Time '102725'
(0008,0050) SH Accession Number ''
(0008,0060) CS Modality 'SC'
(0008,0064) CS Conversion Type 'DF'
(0008,0070) LO Manufacturer 'Radlink'
(0008,0080) LO Institution Name ''
(0008,0090) PN Referring Physician's Name 'DOCTOR BOB'
(0008,1010) SH Station Name ''
(0008,1030) LO Study Description 'L/SPINE MRI'
(0008,103e) LO Series Description ''
(0008,1090) LO Manufacturer's Model Name 'Radlink LaserPro16'
(0010,0000) UL Group Length '78'
(0010,0010) PN Patient's Name 'PATIENT'
(0010,0020) LO Patient ID ''
(0010,0030) DA Patient's Birth Date ''
(0010,0040) CS Patient's Sex 'M'
(0010,1000) LO Other Patient IDs ''
(0010,21b0) LT Additional Patient History ''
(0018,0000) UL Group Length '38'
(0018,0015) CS Body Part Examined ''
(0018,1012) DA Date of Secondary Capture '20141229'
(0018,1014) TM Time of Secondary Capture '102725'
(0020,0000) UL Group Length '132'
(0020,000d) UI Study Instance UID ''
(0020,000e) UI Series Instance UID ''
(0020,0010) SH Study ID '1'
(0020,0011) IS Series Number '1'
(0020,0013) IS Instance Number '2'
(0028,0000) UL Group Length '152'
(0028,0002) US Samples per Pixel '1'
(0028,0004) CS Photometric Interpretation 'MONOCHROME2'
(0028,0010) US Rows '2500'
(0028,0011) US Columns '2040'
(0028,0030) DS Pixel Spacing '0.1736328125'
(0028,0100) US Bits Allocated '16'
(0028,0101) US Bits Stored '16'
(0028,0102) US High Bit '15'
(0028,0103) US Pixel Representation '0'
(0028,1050) DS Window Center '42343'
(0028,1051) DS Window Width '29336'
(0028,2110) CS Lossy Image Compression '01'
(0028,2112) DS Lossy Image Compression Ratio '7.512827'
(0032,0000) UL Group Length '26'
(0032,000a) CS Study Status ID ''
(0032,000c) CS Study Priority ID '0'
(0032,1030) LO Reason for Study ''
(0040,0000) UL Group Length '8'
(0040,0241) AE Performed Station AE Title ''
(0040,1003) SH Requested Procedure Priority 'ROUTINE'
(0050,0065) CS Unknown 'DOCTOR BOB'
(7fe0,0000) UL Group Length '10200012'
(7fe0,0010) OW Pixel Data ''

Following is the code for number-of-frames, height and width:

DicomImage image = new DicomImage(ctFile.Dataset);
Console.WriteLine(image.NumberOfFrames);
Console.WriteLine(image.Width);
Console.WriteLine(image.Height);

Gives me:

NumFrames: 1

Width: 2040

Height: 2500

Where are all the frames of the study stored? Are they put in a single folder? If the folder contains more than one frame doesn't make sense there are additional frames by the count of files in the folder? — jdweng
@jdweng from what i can see looks like just one frame. I think it's a single jpeg of all the frames scanned from the MRI. I am self admittedly not an expert on DICOM. Any assistance is greatly appreciated. — bumble_bee_tuna
I'm not expert either. I have read the spec once but do not remember all the details. How many jpeg files are in the folder? What is the size of the file? The size would be very large. I have disk I got from my last body scan next to me right now. If I get a chance tomorrow I will look at disk. — jdweng

Amit Joshi Amit Joshi · Accepted Answer · 2020-01-16T09:17:02

I have a closet case where when MRI's are scanned (especially if they show as SC) shots it will only be one frame but the film is already a 20 frame MRI. How can I determine if the a single dicom is actually a multi frame MRI.

I do not think you can do this in this particular case.
Note that the value of [SOP Class UID (0008,0016)] is '1.2.840.10008.5.1.4.1.1.7' and [Modality (0008,0060)] is 'SC'. Also observe that values for many important attributes ([Study Instance UID (0020,000d)], [Series Instance UID (0020,000e)]) are empty. You cannot even say that this instance (image) is part of any particular study for that patient.

Many medical equipments/applications generate this type of dataset by processing original datasets. We can say this is meta data or additional data of original instances. It might be generating 3D or storing/saving DICOM Film layout for further usage or some AI for auto diagnosis/report-generation or else.

Reading your question, it looks that this dataset is saved DICOM Film for further usage. If this is the case, you may neglect this particular dataset from your calculations; but I cannot be sure about this as I do not know domain for your application very well.

Now about the "slice":

I am not sure what exactly you mean by "slice" here. Slice as a part of multiframe? or slice as a part of series? or else?

In any case, as I said above, you cannot determine it in this particular case.

If it is as a part of multiframe, other answer from @kritzel_sw explains it. There is no [Number Of Frames (0028,0008)] attribute in dataset.

If it is as a part of series, there is no Series UID in dataset; you cannot.

I do not think this dataset is "slice" in any way. This more looks standalone dataset. So again, as said above, you may neglect this from your calculations.

How to determine if single frame DICOM is actually slice of MRI?

2 Answers