elialombardo commited on
Commit
15b5091
·
verified ·
1 Parent(s): 1a460fd

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +36 -13
README.md CHANGED
@@ -14,9 +14,15 @@ size_categories:
14
  challenge_homepage: https://trackrad2025.grand-challenge.org/trackrad2025/
15
  challenge_repository: https://github.com/LMUK-RADONC-PHYS-RES/trackrad2025/
16
  ---
17
- ## **The dataset 🗃️**
18
- Overall, the TrackRAD2025 challenge will have over <u>2.5 million</u> unlabelled sagittal cine-MRI frames from 500 individual patients, and over 10,000 labelled sagittal cine-MRI frames (+8000 from frames with multiple observers) from 110 individual patients. Precisely,
19
- a cohort of <u>500</u> unlabeled and <u>110</u> manually labeled patients has been prepared for participants. For each patient, 2D sagittal cine MRI data (time-resolved sequence of 2D images) has been acquired during the course of radiotherapy treatments at 0.35 T (ViewRay MRIdian) or 1.5 T (Elekta Unity) MRI-linacs from six international centers. Tracking targets (typically tumors) in the thorax, abdomen and pelvis were included as these can be affected by motion and reflect the most often treated anatomies on MRI-linacs. The <u>training set</u>, which comprises the <u>500</u> unlabeled cases plus <u>50</u> labeled cases was publicly released. Participants can further subdivide this dataset locally into training and validation. The remaining <u>60</u> labeled cases building the <u>preliminary and final testing set</u> is only accessible for evaluation via submission to the challenge.
 
 
 
 
 
 
20
 
21
  Detailed information about the dataset are provided in the following paper "<u>link data paper (expected early 2025)</u>".
22
 
@@ -75,7 +81,7 @@ dataset/
75
  | |-- D_001_frames2.mha -> second scan
76
  | `-- D_001_frames3.mha -> third scan from the same patient
77
  |-- F_003/
78
- `-- .../
79
  ```
80
 
81
  Please note that the dataset folder structure does not match the interfaces that submissions need to implement one-to-one. For details regarding submission requirements, please read the corresponding page.
@@ -95,7 +101,8 @@ This challenge provides 2D+t sagittal cine MRI data collected at six internation
95
  - Sichuan Cancer Center, Chengdu
96
  - University Medical Center Utrecht, Utrecht
97
 
98
- For anonymization purposes, the provenance of the data is not provided, and each center is indicated with letters from A to F. One of the centers also provided cine data with an updated MRI sequence for gating at a 1.5 T MRI-linac, this data is indicated with the letter X.
 
99
 
100
  ##### *Training set*
101
 
@@ -106,7 +113,7 @@ For anonymization purposes, the provenance of the data is not provided, and each
106
 
107
  | 1.5 T MRI-linac | B | C | F | Total |
108
  |------------------|---------|--------|--------|--------|
109
- | Unlabeled | 63 | 60 | 110? | 233? |
110
  | Labeled | 15 | 10 | - | 25 |
111
 
112
  For training, centers A, B and C provided both unlabeled and manually labeled data while centers E and F provided solely unlabeled data.
@@ -137,27 +144,42 @@ For preliminary testing and final testing centers A, B, C, D and X provided manu
137
 
138
  ###### *Unlabelled data protocol*
139
 
140
- For the unlabeled data in the training set, cine MRI from one or multiple radiotherapy MRI-linac treatment fractions were included. Cine MRI acquired during treatment for the 0.35 T MRI-linac also include frames with degraded image quality due to gantry rotations, which participants are free to exclude using a method of their choice.
141
 
142
- All frames from the 1.5 T MRI-linac were acquired during treatments only and do not present degradation due to gantry rotations due to machine design. In this set some patients present temporal jumps in the cine MRI due to paused treatment.
 
 
 
143
 
144
  ###### *Labelled data protocol*
145
 
146
  To avoid degraded images during evaluation, labeled frames for the 0.35 T MRI-linac were either chosen from simulation cine MRIs prior to treatment start or, when taken from treatments, a manual selection was performed to avoid periods of gantry rotation.
147
 
148
- All frames from the 1.5 T MRI-linac were acquired during treatments only and do not present degradation due to gantry rotations due to machine design.
149
 
150
- Human observers have generated the reference labels both for the training and testing sets. For dataset A, two observers (a medical student and a dentistry student) labeled the cine MRI frames using an in-house labeling tool developed for the challenge. For dataset B, a medical physics researcher (assistant professor) with more than 10 years experience in radiotherapy used the in-house labeling tool to delineate the frames. For dataset C, two radiation oncologists independently labeled the cine MRI frames using itk-snap. For dataset D, 4 radiation oncologists and one medical physicist have independently labeled the cine MRI frames using software provided by the 0.35 T MRI-linac vendor.
 
 
 
 
151
 
152
  For all labeled data, a medical physics doctoral student with 4 years experience in tumor tracking then reviewed and if necessary corrected all labels used in this challenge using the in-house tool.
153
 
154
  ##### *Data acquisition and pre-processing*
155
 
156
- All images were acquired using the clinically adopted imaging protocols of the respective centers for each anatomical site and reflect typical images found in daily clinical routine. The cine-MRI sequences used at the 0.35 T and 1.5 T MRI-linacs are standardized, which ensures uniformity of the data for a given field strength.
 
157
 
158
- At the centers using the 0.35 T MRI-linac, the 2D cine-MRIs were acquired in sagittal orientation with the patient in treatment position in the MRI-linac bore. During treatment simulation or delivery, the patients performed breath holds to increase the duty cycle of the gated radiation delivery. The breath-holds are followed by periods of regular breathing. The sequence was a 2D-balanced steady-state free precession (bSSFP) at 4 Hz or 8 Hz with a slice thickness of 5, 7 or 10 mm and pixel spacing of 2.4x2.4 or 3.5x3.5 mm2.
 
 
 
159
 
160
- At the centers using the 1.5 T MRI-linac, the 2D cine-MRIs were acquired in interleaved sagittal and coronal orientation with the patient in treatment position in the MRI-linac. For the challenge, only the sagittal plane has been considered. During treatment simulation or delivery, some patients performed breath holds to increase the duty cycle of the gated radiation delivery, while others breathed freely. The breath-holds are followed by periods of regular breathing. The sequence was a balanced fast field echo (bFFE) sequence at 1 Hz to 5 Hz with a slice thickness of 5, 7 or 8 mm and pixel spacing of 1.0x1.0 to 1.7x1.7 mm2.
 
 
 
 
161
 
162
  The following pre-processing steps were performed on the data:
163
 
@@ -165,3 +187,4 @@ The following pre-processing steps were performed on the data:
165
  - Anonymization
166
  - Reshaping and orientation correction
167
  - Resampling to 1x1 mm2 in-plane pixel spacing
 
 
14
  challenge_homepage: https://trackrad2025.grand-challenge.org/trackrad2025/
15
  challenge_repository: https://github.com/LMUK-RADONC-PHYS-RES/trackrad2025/
16
  ---
17
+ ### **The dataset 🗃️**
18
+ Overall, the TrackRAD2025 challenge provides over **2.8 million unlabeled** sagittal cine-MRI frames from 477 individual patients, and over **10,000 labeled** sagittal cine-MRI frames
19
+ (+8000 from frames with multiple observers) from 108 individual patients. Precisely,
20
+ a cohort of **477** unlabeled and **108** manually labeled patients has been prepared for participants. For each patient, 2D sagittal cine MRI data (time-resolved sequence of 2D images)
21
+ has been acquired during the course of radiotherapy treatments at 0.35 T (ViewRay MRIdian) or 1.5 T (Elekta Unity) MRI-linacs from six international centers.
22
+ Tracking targets (typically tumors) in the thorax, abdomen and pelvis were included as these can be affected by motion and reflect the most often treated anatomies on MRI-linacs.
23
+ The **training set**, which comprises the 477 unlabeled cases plus 50 labeled cases was publicly released.
24
+ Participants can further subdivide this dataset locally into training and validation. The remaining 58 labeled cases building the **preliminary and final testing set** is only
25
+ accessible for evaluation via submission to the challenge. A couple of years after the challenge is closed, the testing set data is also going to be uploaded to the same location as the training set.
26
 
27
  Detailed information about the dataset are provided in the following paper "<u>link data paper (expected early 2025)</u>".
28
 
 
81
  | |-- D_001_frames2.mha -> second scan
82
  | `-- D_001_frames3.mha -> third scan from the same patient
83
  |-- F_003/
84
+ `-- ...
85
  ```
86
 
87
  Please note that the dataset folder structure does not match the interfaces that submissions need to implement one-to-one. For details regarding submission requirements, please read the corresponding page.
 
101
  - Sichuan Cancer Center, Chengdu
102
  - University Medical Center Utrecht, Utrecht
103
 
104
+ For anonymization purposes, the provenance of the data is not provided, and each center is indicated with letters from A to F. One of the centers also provided cine data
105
+ with an updated MRI sequence for gating at a 1.5 T MRI-linac, this data is indicated with the letter X.
106
 
107
  ##### *Training set*
108
 
 
113
 
114
  | 1.5 T MRI-linac | B | C | F | Total |
115
  |------------------|---------|--------|--------|--------|
116
+ | Unlabeled | 63 | 60 | 101 | 224 |
117
  | Labeled | 15 | 10 | - | 25 |
118
 
119
  For training, centers A, B and C provided both unlabeled and manually labeled data while centers E and F provided solely unlabeled data.
 
144
 
145
  ###### *Unlabelled data protocol*
146
 
147
+ For the unlabeled training set data, sagittal cine MRI from one or multiple radiotherapy MRI-linac treatment fractions and from MRI-linac simulations were included.
148
 
149
+ Due to machine design, when the gantry of the 0.35 T moves, the image quality is degraded.
150
+ The 0.35 T unlabeled data therefore includes frames with degraded image quality due to gantry rotations, which participants are free to exclude using a method of their choice.
151
+ All frames from the 1.5 T MRI-linac were acquired during treatments only and do not present degradation due to gantry rotations by design.
152
+ The 1.5 T unlabeled data however can present temporal jumps and large changes in contrast within one cine MRI due to treatment interruptions, which during export are combined in a single scan.
153
 
154
  ###### *Labelled data protocol*
155
 
156
  To avoid degraded images during evaluation, labeled frames for the 0.35 T MRI-linac were either chosen from simulation cine MRIs prior to treatment start or, when taken from treatments, a manual selection was performed to avoid periods of gantry rotation.
157
 
158
+ Labeled frames from the 1.5 T MRI-linac were visually inspected and selected to avoid temporal jumps due to treatment interruptions.
159
 
160
+ Human observers have generated the reference labels both for the training and testing sets.
161
+ For dataset A, two observers (a medical student and a dentistry student) labeled the cine MRI frames using a [labeling tool](https://github.com/LMUK-RADONC-PHYS-RES/contouring-tool) developed specifically for the challenge.
162
+ For dataset B, a medical physics researcher (assistant professor) with more than 10 years experience in radiotherapy used the same in-house labeling tool to delineate the frames.
163
+ For dataset C, two radiation oncologists independently labeled the cine MRI frames using itk-snap. For dataset D, 4 radiation oncologists and one medical physicist have independently
164
+ labeled the cine MRI frames using software provided by the 0.35 T MRI-linac vendor.
165
 
166
  For all labeled data, a medical physics doctoral student with 4 years experience in tumor tracking then reviewed and if necessary corrected all labels used in this challenge using the in-house tool.
167
 
168
  ##### *Data acquisition and pre-processing*
169
 
170
+ All images were acquired using the clinically adopted imaging protocols of the respective centers for each anatomical site and reflect typical images found in daily clinical routine.
171
+ The cine-MRI sequences used at the 0.35 T and 1.5 T MRI-linacs are standardized, which ensures uniformity of the data for a given field strength.
172
 
173
+ At the centers using the 0.35 T MRI-linac, the 2D cine-MRIs were acquired in sagittal orientation with the patient in treatment position in the MRI-linac bore.
174
+ During treatment simulation or delivery, the patients performed breath holds to increase the duty cycle of the gated radiation delivery.
175
+ The breath-holds are followed by periods of regular breathing. The sequence was a 2D-balanced steady-state free precession (bSSFP) at 4 Hz or 8 Hz with a slice thickness of 5, 7 or 10 mm and pixel
176
+ spacing of 2.4x2.4 or 3.5x3.5 mm2.
177
 
178
+ At the centers using the 1.5 T MRI-linac, the 2D cine-MRIs were acquired in either interleaved sagittal and coronal or interleaved sagittal, coronal and axial orientations
179
+ with the patient in treatment position in the MRI-linac.
180
+ For the challenge, only the sagittal plane has been considered. During treatment simulation or delivery, some patients performed breath holds to increase the duty cycle of
181
+ the gated radiation delivery, while others breathed freely. The breath-holds are followed by periods of regular breathing. The sequence was a balanced fast field echo (bFFE) sequence
182
+ at 1.3 Hz to 3.5 Hz in the sagittal orientation with a slice thickness of 5, 7 or 8 mm and pixel spacing of 1.0x1.0 to 1.7x1.7 mm2.
183
 
184
  The following pre-processing steps were performed on the data:
185
 
 
187
  - Anonymization
188
  - Reshaping and orientation correction
189
  - Resampling to 1x1 mm2 in-plane pixel spacing
190
+ - Conversion to 16-bit unsigned integer