Aims and objectives
Background
Assessment of treatment response in brain tumours is one of the most important issues in oncology.
After chemo-radiotherapy (CRT),
glioblastoma,
the most common aggressive primary brain tumour,
increases in size in one third of patients.
It appears that treatment is not working and in about half of these cases,
the growth is transient and due to treatment itself,
known as ‘pseudoprogression’,
rather than true progression.1In clinical practice,
it is impossible to differentiate between true progression and pseudoprogression.
Conventional MRI scans also cannot distinguish between...
Methods and materials
We retrospectively analysed 20 MRI studies of patients with biopsy-proven glioblastoma who had standard chemo-radiotherapy treatment and early progressive enhancing disease.
Studies were labelled as true progression (n=11) if there was progression or death within six months or pseudoprogression (n=9) if there was no further progression within six months.
The T1-weighted post-contrast and T2-weighted sequences were co-registered to allow segmentation of tumour components.
Enhancing disease and perilesional oedema were segmented from the two sequences respectively to create volumes of interest (Fig. 2) using ITK-SNAP open-source...
Results
Results showed several features demonstrating significant difference between the true progression and pseudoprogression groups.
For enhancing disease on T1W imaging,
the significant GLCM features were contrast and homogeneity,
and significant GLRLM features were grey level non-uniformity and run length non-uniformity.
For perilesional oedema,
the significant GLRLM features were grey level non-uniformity and run length non-uniformity.
There were also significant differences in the volume of enhancing disease and perilesional oedema between both groups.
Results are summarised in Table 1(Fig. 3).
The results suggest that computer vision...
Conclusion
This pilot study has shown that radiomic texture features can differentiate between early true progression and pseudoprogression in glioblastoma.
The most significantradiomic features distinguishing pseudoprogression from true progression were contrast,
homogeneity,
grey level non-uniformity and run length non-uniformity.
The volumes of enhancing disease and perilesional oedema were also significantly different between both groups.
Big data incorporating machine learning is required to produce strong prediction models for earlier prediction of treatment response.
References
Brandes AA,
Franceschi E,
Tosoni A,
et al.
MGMT promoter methylation status can predict the incidence and outcome of pseudoprogression after concomitant radiochemotherapy in newly diagnosed glioblastoma patients.
J Clin Oncol.
2008;26(13):2192-7.
Gillies RJ,
Kinahan PE,
Hricak H.
Radiomics: Images Are More than Pictures,
They Are Data.
Radiology.
2016;278(2):563-77.
Erickson BJ,
Korfiatis P,
Akkus Z,
Kline TL.
Machine Learning for Medical Imaging.
Radiographics.
2017;37(2):505-515.
Paul A.
Yushkevich,
Joseph Piven,
Heather Cody Hazlett,
Rachel Gimpel Smith,
Sean Ho,
James C.
Gee,
and Guido Gerig.
User-guided 3D...