| 1. |
- Lempart, Michael, et al.
(författare)
-
A deeply supervised convolutional neural network ensemble for multilabel segmentation of pelvic OARs
- 2021
-
Ingår i: Radiotherapy and Oncology. - 1879-0887. ; 161:Suppl 1, s. 1417-1418
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Accurate delineation of organs at risk (OAR) is a crucial step in radiation therapy (RT) treatment planning but is a manual and time-consuming process. Deep learning-based methods have shown promising results for medical image segmentation and can be used to accelerate this task. Nevertheless, it is rarely applied to complex structures found in the pelvis region, where manual segmentation can be difficult, costly and is not always feasible. The aim of this study was to train and validate a model, based on a modified U-Net architecture, for automated and improved multilabel segmentation of 10 pelvic OAR structures (total bone marrow, lower pelvis bone marrow, iliac bone marrow, lumosacral bone marrow, bowel cavity, bowel, small bowel, large bowel, rectum, and bladder).
|
|
| 2. |
- Lempart, Michael, et al.
(författare)
-
Volumetric modulated arc therapy dose prediction and deliverable treatment plan generation for prostate cancer patients using a densely connected deep learning model
- 2021
-
Ingår i: Physics and imaging in radiation oncology. - : Elsevier BV. - 2405-6316. ; 19, s. 112-119
-
Tidskriftsartikel (refereegranskat)abstract
- Background and purpose: Radiation therapy treatment planning is a manual, time-consuming task that might be accelerated using machine learning algorithms. In this study, we aimed to evaluate if a triplet-based deep learning model can predict volumetric modulated arc therapy (VMAT) dose distributions for prostate cancer patients. Materials and methods: A modified U-Net was trained on triplets, a combination of three consecutive image slices and corresponding segmentations, from 160 patients, and compared to a baseline U-Net. Dose predictions from 17 test patients were transformed into deliverable treatment plans using a novel planning workflow. Results: The model achieved a mean absolute dose error of 1.3%, 1.9%, 1.0% and ≤ 2.6% for clinical target volume (CTV) CTV_D100%, planning target volume (PTV) PTV_D98%, PTV_D95% and organs at risk (OAR) respectively, when compared to the clinical ground truth (GT) dose distributions. All predicted distributions were successfully transformed into deliverable treatment plans and tested on a phantom, resulting in a passing rate of 100% (global gamma, 3%, 2 mm, 15% cutoff). The dose difference between deliverable treatment plans and GT dose distributions was within 4.4%. The difference between the baseline model and our improved model was statistically significant (p < 0.05) for CVT_D100%, PTV_D98% and PTV_D95%. Conclusion: Triplet-based training improved VMAT dose distribution predictions when compared to 2D. Dose predictions were successfully transformed into deliverable treatment plans using our proposed treatment planning procedure. Our method may automate parts of the workflow for external beam prostate radiation therapy and improve the overall treatment speed and plan quality.
|
|
| 3. |
- Rugbjerg, Peter, 1988, et al.
(författare)
-
Short and long-read ultra-deep sequencing profiles emerging heterogeneity across five platform Escherichia coli strains
- 2021
-
Ingår i: Metabolic Engineering. - : Elsevier BV. - 1096-7176 .- 1096-7184. ; 65, s. 197-206
-
Tidskriftsartikel (refereegranskat)abstract
- Reprogramming organisms for large-scale bioproduction counters their evolutionary objectives of fast growth and often leads to mutational collapse of the engineered production pathways during cultivation. Yet, the mutational susceptibility of academic and industrial Escherichia coli bioproduction host strains are poorly understood. In this study, we apply 2nd and 3rd generation deep sequencing to profile simultaneous modes of genetic heterogeneity that decimate engineered biosynthetic production in five popular E. coli hosts BL21(DE3), TOP10, MG1655, W, and W3110 producing 2,3-butanediol and mevalonic acid. Combining short-read and longread sequencing, we detect strain and sequence-specific mutational modes including single nucleotide polymorphism, inversion, and mobile element transposition, as well as complex structural variations that disrupt the integrity of the engineered biosynthetic pathway. Our analysis suggests that organism engineers should avoid chassis strains hosting active insertion sequence (IS) subfamilies such as IS1 and IS10 present in popular E. coli TOP10. We also recommend monitoring for increased mutagenicity in the pathway transcription initiation regions and recombinogenic repeats. Together, short and long sequencing reads identified latent low-frequency mutation events such as a short detrimental inversion within a pathway gene, driven by 8-bp short inverted repeats. This demonstrates the power of combining ultra-deep DNA sequencing technologies to profile genetic heterogeneities of engineered constructs and explore the markedly different mutational landscapes of common E. coli host strains. The observed multitude of evolving variants underlines the usefulness of early mutational profiling for new synthetic pathways designed to sustain in organisms over long cultivation scales.
|
|
