We use cookies necessary to giving you a better online experience. By using our website you consent to all cookies in accordance with our Privacy Policy.
Yes, I accept
Menu
Search Opener
Copenhagen, Denmark
Onsite/Online

ESTRO 2022

Session Item

Physical aspects of quantitative functional and biological imaging
Poster (digital)
Physics
Delineation-related ADC variation between centres in the Elekta Unity MR linac Consortium
Anne Bisgaard, Denmark
PO-1754

Abstract

Delineation-related ADC variation between centres in the Elekta Unity MR linac Consortium
Authors:

Anne Bisgaard1, A.L.H. van Lier2, Catherine Coolens3, Rick Keesman4, Alison Tree5, Andreas Wetscherek6, Paul Romesser7, Neelam Tyagi8, Monica Lo Russo9, Jonas Habrich10, Danny Vespirini11, Angus Lau12, Stella Mook13, Peter Chung3, Linda Kerkmeijer14, Marlies Nowee15, Tine Schytte16, Ebbe Lorenzen1, Carsten Brink1, Uulke Van der Heide17, Faisal Mahmood18

1Odense University Hospital, Laboratory of Radiation Physics, Department of Oncology, Odense, Denmark; 2University Medical Centre Utrecht, Department of Radiotherapy, Utrecht, The Netherlands; 3Princess Margaret Cancer Center, University Health Network, Radiation Medicine Program, Toronto, Canada; 4Radboud University Medical Centre, Department of Radiation Oncology, Nijmegen, The Netherlands; 5Department of Radiotherapy, The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, London, United Kingdom; 6Joint Department of Physics, The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, London, United Kingdom; 7Memorial Sloan Kettering Cancer Center, Department of Radiation Oncology, New York, NY, USA; 8Memorial Sloan Kettering Cancer Center, Department of Medical Physics, New York, NY, USA; 9Unversity Hospital and Medical Faculty, Eberhard Karls University, Department of Radiation Oncology, Tübingen, Germany; 10University Hospital and Medical Faculty, Eberhard Karls University, Section for Biomedical Physics, Department of Radiation Oncology, Tübingen, Germany; 11Sunnybrook Odette Cancer Center, Department of Radiation Oncology, Toronto, Canada; 12Sunnybrook Research Institute, Department of Physical Sciences, Toronto, Canada; 13University Medical Center Utrecht, Department of Radiotherapy, Utrecht, The Netherlands; 14Radboud University Medical Center, Department of Radiation Oncology , Nijmegen, The Netherlands; 15The Netherlands Cancer Institute, Department of Radiotherapy, Amsterdam, The Netherlands; 16Odense University Hospital, Department of Oncology, Odense, Denmark; 17The Netherlands Cancer Institute, Department of Radiation Oncology, Amsterdam, The Netherlands; 18Odense University Hospital, Laboratory of Radiation Physics, Department of Oncology, Odense , Denmark

Show Affiliations
Purpose or Objective

The apparent diffusion coefficient (ADC), derived from Diffusion-Weighted MRI (DWI), is a promising radiotherapy response biomarker. However, lack of consistency in ADC measurements hampers its translation into clinical use. ADC measurements require delineation of a region of interest (ROI), and both choice of ROI type and inter-/intra-reader variation introduces ADC variation. The aim of this study is to investigate the impact of delineation variation on associated ADC measurements between centres in the Elekta Unity MR-linac consortium utilizing the same patients. 

Material and Methods

Nine centres in the Elekta Unity MR linac consortium participated in this study using data acquired on MR linac (Unity, Elekta) by one of the participating centers (OUH). MRI scans from four clinical cases were used: adrenal gland metastasis, oligo metastasis in pelvis, pancreatic cancer relapse and prostate cancer. T2-weighted images and repeated DWI images (test-retest) were acquired using b-values of 150 and 500 s/mm2.

Delineation of GTV and CTV (prostate only) was done for each case by oncologists from the nine centres in a mutually blind manner using ProKnow (Elekta), otherwise adhering to local workflow. A so-called viable tumour volume (VTV), defined as viable tumour (i.e. excluding necrotic and cystic parts), was delineated as well. Deliberately, we did not give detailed delineation guidelines.

For each type of ROI, delineations were compared pair-wise (for each permutation), to calculate the median Dice Similarity Coefficient, maximum Hausdorff Distance, and Mean Surface Distance (MSD).

ADC values were reported as the median values within each ROI using ADC maps generated on MRL Philips MR console (Release 5.3) at OUH. Retest ADC values were calculated using rigid contour propagation of GTV between test and retest scans. Test-retest ADC variation (based on GTV) was estimated as the difference between the 25th and 75th percentile of ADC differences between test and retest scans. 

Results

Delineation variation was different between the four clinical cases and types of ROI (Figure 1). The largest variation was observed for pancreas and prostate (GTV and VTV), possibly due to small volumes of ROI for pancreas, and the fact that GTV and VTV are not normally delineated on prostate. The between-centre ADC variation followed the same trend as the delineation variation (Figure 2). The test-retest ADC variation was estimated as 0.06*10-3 (adrenal gland), 0.01*10-3 (oligo), 0.09*10-3 (pancreas), and 0.23*10-3 mm2/s (prostate), comparable to the between-centre ADC variation, and possibly overestimated since contour propagation was used instead of re-contouring on retest scans.

 


Conclusion

The observed delineation variation is clearly reflected in the variation of ADC measurements. This indicates that consistency of ADC measurements between centres may be improved by reducing delineation variation, e.g. by using delineation guidelines or computer-aided delineation.