Extracting Radiomics from Different Dose Distribution in the same planC

[EvandroMolinari1998]

In my planC I have 3 different dose distributions (IMRT, VMAT and 3D) generated by 3 different RTplans. I wish to select from wich dose distribution matlab will extract the radiomics.

[adityaapte]

The dose can be accessed as follows: global planC % or load from file indexS = planC{end}; doseNum = 1; doseS = planC{indexS.dose}(doseNum);

DVH-based features, for example can be extracted as follows:

[doseV,volV] = getDVH(structNum,doseNum,planC);[doseBinsV,volHistV] = doseHist(doseV,volV,binWidth);dvhFeatS.meanDose = calc_meanDose(doseBinsV, volHistV);dvhFeatS.d10 = calc_Dx(doseBinsV, volHistV, 10, 1);dvhFeatS.v30 = calc_Vx(doseBinsV, volHistV, 30, 0);dvhFeatS.MOH55 = calc_MOHx(doseBinsV, volHistV, 55);

Note that dose can also be converted to scan with the following call (use "testing" branch of CERR from GitHub for this). doseNum = 1; scanType = 'Dose summary'; gridType = 'dose'; planC = dose2scan(doseNum,scanType,gridType,planC);

The above code will store dose as a scan entry in planC. IBSI compliant radiomics can be computed as follows: scanNum = length(planC{indexS.scan});

paramS = getRadiomicsParamTemplate(sampleParamFile);

radiomicsFeatS = calcGlobalRadiomicsFeatures...(scanNum, structNum, paramS, planC);

On Thu, Sep 1, 2022 at 2:56 PM EvandroMolinari1998 @.***> wrote:

In my planC I have 3 different dose distributions (IMRT, VMAT and 3D) generated by 3 different RTplans. I wish to select from wich dose distribution matlab will extract the radiomics.

— Reply to this email directly, view it on GitHub https://github.com/cerr/CERR/issues/755, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNJB354KEKKNXRSK5SOW3V4D355ANCNFSM6AAAAAAQCUD52U . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[EvandroMolinari1998]

Thanks a lot! When I use "planC = dose2scan(doseNum,scanType,gridType,planC)" I get: Error using dose2scan Too many output arguments.

In adition, once I fix this issue i wanna know if it's possible to use this planC from dose2scan to other feature classes.

[adityaapte]

Please use the "testing" branch. It has the correct output argument planC. https://github.com/cerr/CERR/blob/testing/CERR_core/Utilities/dose2scan.m

The pseudo-scan in planC can be used to compute radiomics features.

-Aditya

On Fri, Sep 2, 2022 at 9:31 AM EvandroMolinari1998 @.***> wrote:

Thanks a lot! When I use "planC = dose2scan(doseNum,scanType,gridType,planC)" I get: Error using dose2scan Too many output arguments.

In adition, once I fix this issue i wanna know if it's possible to use this planC from dose2scan to other feature classes.

— Reply to this email directly, view it on GitHub https://github.com/cerr/CERR/issues/755#issuecomment-1235509691, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNJB3OCMU2NAG5LZXFH5DV4H6RFANCNFSM6AAAAAAQCUD52U . You are receiving this because you commented.Message ID: @.***>

[EvandroMolinari1998]

I tested this 'testing' branch and got this pseudo-scan in planC, nonetheless, when I select the three different dose distributions with the 'doseNum' I compute the same values for the First order statistics features.

In adition, when I try de command 'paramS = getRadiomicsParamTemplate(sampleParamFile)' I get this error: Unrecognized function or variable 'sampleParamFile'.

[adityaapte]

sampleParamFile is a path to the .json file containing calculation settings. For example,

sampleParamFile = fullfile(getCERRPath,'PlanMetrics','heterogenity_metrics','sample_radiomics_extraction_settings.json');

Here is the description of the file. https://github.com/cerr/CERR/wiki/Global-radiomics-feature-calculation-parameters

On Fri, Sep 2, 2022 at 11:45 AM EvandroMolinari1998 < @.***> wrote:

I tested this 'testing' branch and got this pseudo-scan in planC, nonetheless, when I select the three different dose distributions with the 'doseNum' I compute the same values for the First order statistics features.

In adition, when I try de command 'paramS = getRadiomicsParamTemplate(sampleParamFile)' I get this error: Unrecognized function or variable 'sampleParamFile'.

— Reply to this email directly, view it on GitHub https://github.com/cerr/CERR/issues/755#issuecomment-1235652732, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNJB2EL55SYYL6GQNYLHTV4IOKDANCNFSM6AAAAAAQCUD52U . You are receiving this because you commented.Message ID: @.***>

[EvandroMolinari1998]

When I select the three different dose distributions with the 'doseNum' I compute the same values for the First order statistics features. And when I compare the planC before and after 'planC = dose2scan(doseNum,scanType,gridType,planC);' I realize that there is no difference. How can I fix this?

[adityaapte]

Are you selecting different scan number as an input for radiomics calculation? For example, after dose2scan, use the following code to obtain scan number corresponding to the converted dose.

scanNumForDose1 = length(planC{indexS.scan})

On Mon, Sep 5, 2022, 3:40 PM EvandroMolinari1998 @.***> wrote:

When I select the three different dose distributions with the 'doseNum' I compute the same values for the First order statistics features. And when I compare the planC before and after 'planC = dose2scan(doseNum,scanType,gridType,planC);' I realize that there is no difference. How can I fix this?

— Reply to this email directly, view it on GitHub https://github.com/cerr/CERR/issues/755#issuecomment-1237415871, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNJB3TDJLFKDNCPHKU7GTV4ZEBRANCNFSM6AAAAAAQCUD52U . You are receiving this because you commented.Message ID: @.***>

[EvandroMolinari1998]

Yes! I can see the scan number value changing, but when I try in this next code, it doesn't change This is the code i'm using:

load('ESTUDO_CERR.mat') indexS = planC{end}; doseNum = 1; doseS = planC{indexS.dose}(doseNum); scanType = 'Dose summary'; gridType = 'dose'; planC = dose2scan(doseNum,scanType,gridType,planC); CERR>> Using zValues to compute voxel thicknesses. scanNum = length(planC{indexS.scan}); sampleParamFile =fullfile(getCERRPath,'PlanMetrics','heterogenity_metrics','sample_radiomics_extraction_settings.json'); structNum = 1; ctOffset = 1000; binWidth = 25; statsFeatureS = radiomics_first_order_stats(planC,structNum,ctOffset,binWidth);

[EvandroMolinari1998]

When I do the whole process and use 'radiomicsFeatS = calcGlobalRadiomicsFeatures(scanNum, structNum, paramS, planC);' to compute the features I get the error:

Unable to perform assignment because the size of the left side is 97-by-162-by-40 and the size of the right side is 283-by-368-by-40. Error in preProcessForRadiomics (line 33) mask3M(:,:,uniqueSlices) = maskUniq3M; Error in calcGlobalRadiomicsFeatures (line 15) [volToEval,maskBoundingBox3M,gridS,paramS] = preProcessForRadiomics(scanNum,...

[adityaapte]

The structure needs to be copied over to the new scan before calling radiomics_first_order_stats. This can be achieved with the following.

planC = copyStrToScan(structNum,scanNum,planC);

Then, use this new structure for feature extraction.

structNum = length(planC{indexS.structures});

On Mon, Sep 5, 2022, 4:27 PM EvandroMolinari1998 @.***> wrote:

Yes! I can see the scan number value changing, but when I try in this next code, it doesn't change This is the code i'm using:

load('ESTUDO_CERR.mat') indexS = planC{end}; doseNum = 1; doseS = planC{indexS.dose}(doseNum); scanType = 'Dose summary'; gridType = 'dose'; planC = dose2scan(doseNum,scanType,gridType,planC); CERR>> Using zValues to compute voxel thicknesses. scanNum = length(planC{indexS.scan}); sampleParamFile =fullfile(getCERRPath,'PlanMetrics','heterogenity_metrics','sample_radiomics_extraction_settings.json'); structNum = 1; ctOffset = 1000; binWidth = 25; statsFeatureS = radiomics_first_order_stats(planC,structNum,ctOffset,binWidth);

— Reply to this email directly, view it on GitHub https://github.com/cerr/CERR/issues/755#issuecomment-1237440048, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAGNJB2AHKNU2ECJNKYFMSLV4ZJSNANCNFSM6AAAAAAQCUD52U . You are receiving this because you commented.Message ID: @.***>