| 11. |
- Zubatkina, Irina, et al.
(författare)
-
Clinically Driven Alpha/Beta Ratios for Melanoma Brain Metastases and Investigation of Biologically Effective Dose as a Predictor for Local Control After Radiosurgery : A Proof of Concept in a Retrospective Longitudinal Series of 274 Consecutive Lesions
- 2024
-
Ingår i: Neurosurgery. - : Wolters Kluwer. - 0148-396X .- 1524-4040. ; 94:2, s. 423-430
-
Tidskriftsartikel (refereegranskat)abstract
- BACKGROUND AND OBJECTIVES: Brain metastases (BM) develop in nearly half of the patients with advanced melanoma. The aim of this retrospective historical cohort study was to analyze radiological response of melanoma BM to single-fraction Gamma Knife radiosurgery (GKRS), in relation to biologically effective dose (BED) for various alpha/beta ratios.METHODS: Included in the study were 274 lesions. Primary outcome was local control (LC). Mean marginal dose was 21.6 Gy (median 22, range 15-25). Biologically effective dose was calculated for an alpha/beta ratio of 3 (Gy3), 5 (Gy10), 10 (Gy10), and 15 (Gy15).RESULTS: Receiver operating characteristic value for LC and BED was 85% (most statistically significant odds ratio 1.14 for BED Gy15, P = .006), while for LC and physical dose was 79% (P = .02). When comparing equality of 2 receiver operating characteristic areas, this was statistically significant (P = .02 and .03). Fractional polynomial regression revealed BED (Gy10 and Gy15) as statistically significant (P = .05) with BED of more than 63 Gy10 or 49 Gy15 as relevant, also for higher probability of quick decrease in volume first month after GKRS and lower probability of radiation necrosis. Shorter irradiation time was associated with better LC (P = .001), particularly less than 40 minutes (LC below 90%, P = .05).CONCLUSION: BED Gy10 and particularly Gy15 were more statistically significant than physical dose for LC after GKRS for radioresistant melanoma BM. Irradiation time (per lesion) longer than 40 minutes was predictive for lower rates of LC. Such results need to be validated in larger cohorts.
|
|
| 12. |
- Ödén, Jakob, et al.
(författare)
-
Spatial correlation of linear energy transfer and relative biological effectiveness with treatment related toxicities following proton therapy for intracranial tumors
- 2020
-
Ingår i: Medical physics (Lancaster). - : Wiley. - 0094-2405 .- 2473-4209. ; 47:2, s. 342-351
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: The enhanced relative biological effectiveness (RBE) at the end of the proton range might increase the risk of radiation-induced toxicities. This is of special concern for intracranial treatments where several critical organs at risk (OARs) surround the tumor. In the light of this, a retrospective analysis of dose-averaged linear energy transfer (LETd) and RBE-weighted dose (DRBE) distributions was conducted for three clinical cases with suspected treatment related toxicities following intracranial proton therapy. Alternative treatment strategies aiming to reduce toxicity risks are also presented.Methods: The clinical single-field optimized (SFO) plans were recalculated for 81 error scenarios with a Monte Carlo dose engine. The fractionation DRBE was 1.8 Gy (RBE) in 28 or 30 fractions assuming a constant RBE of 1.1. Two LETd- and α/β-dependent variable RBE models were used for evaluation, including a sensitivity analysis of the α/β parameter. Resulting distributions of DRBE and LETd were analyzed together with normal tissue complication probabilities (NTCPs). Subsequently, four multi-field optimized (MFO) plans, with an additional beam and/or objectives penalizing protons stopping in OARs, were created to investigate the potential reduction of LETd, DRBE and NTCP.Results: The two variable RBE models agreed well and predicted average RBE values around 1.3 in the toxicity volumes, resulting in increased near-maximum DRBE of 7-11 Gy (RBE) compared to RBE=1.1 in the nominal scenario. The corresponding NTCP estimates increased from 0.8%, 0.0% and 3.7% (RBE=1.1) to 15.5%, 1.8% and 45.7% (Wedenberg RBE model) for the three patients, respectively. The MFO plans generally allowed for LETd, DRBE and NTCP reductions in OARs, without compromising the target dose. Compared to the clinical SFO plans, the maximum reduction of the near-maximum LETd was 56%, 63% and 72% in the OAR exhibiting the toxicity for the three patients, respectively.Conclusions: Although a direct causality between RBE and toxicity cannot be established here, high LETd and DRBE correlated spatially with the observed toxicities, whereas setup and range uncertainties had a minor impact. Individual factors, which might affect the patient-specific radiosensitivity, were however not included in these calculations. The MFO plans using both an additional beam and proton track-end objectives allowed the largest reductions in LETd, DRBE and NTCP, and might be future tools for similar cases.
|
|
| 13. |
- Akuwudike, Pamela, et al.
(författare)
-
Cell Type-Specific Patterns in the Accumulation of DNA Damage Following Multifractional Radiation Exposure
- 2022
-
Ingår i: International Journal of Molecular Sciences. - : MDPI AG. - 1661-6596 .- 1422-0067. ; 23:21
-
Tidskriftsartikel (refereegranskat)abstract
- Predicting the risk of second malignant neoplasms is complicated by uncertainties regarding the shape of the dose–response relationship at high doses. Limited understanding of the competitive relationship between cell killing and the accumulation of DNA lesions at high doses, as well as the effects of other modulatory factors unique to radiation exposure during radiotherapy, such as dose heterogeneity across normal tissue and dose fractionation, contribute to these uncertainties. The aim of this study was to analyze the impact of fractionated irradiations on two cell systems, focusing on the endpoints relevant for cancer induction. To simulate the heterogeneous dose distribution across normal tissue during radiotherapy, exponentially growing VH10 fibroblasts and AHH-1 lymphoblasts were irradiated with 9 and 12 fractions (VH10) and 10 fractions (AHH-1) at 0.25, 0.5, 1, or 2 Gy per fraction. The effects on cell growth, cell survival, radiosensitivity and the accumulation of residual DNA damage lesions were analyzed as functions of dose per fraction and the total absorbed dose. Residual γH2AX foci and other DNA damage markers (micronuclei, nuclear buds, and giant nuclei) were accumulated at high doses in both cell types, but in a cell type-dependent manner. The competitive relationship between cell killing and the accumulation of carcinogenic DNA damage following multifractional radiation exposure is cell type-specific.
|
|
| 14. |
- Astaraki, Mehdi, PhD Student, 1984-, et al.
(författare)
-
A Comparative Study of Radiomics and Deep-Learning Based Methods for Pulmonary Nodule Malignancy Prediction in Low Dose CT Images
- 2021
-
Ingår i: Frontiers in Oncology. - : Frontiers Media SA. - 2234-943X. ; 11
-
Tidskriftsartikel (refereegranskat)abstract
- Objectives: Both radiomics and deep learning methods have shown great promise in predicting lesion malignancy in various image-based oncology studies. However, it is still unclear which method to choose for a specific clinical problem given the access to the same amount of training data. In this study, we try to compare the performance of a series of carefully selected conventional radiomics methods, end-to-end deep learning models, and deep-feature based radiomics pipelines for pulmonary nodule malignancy prediction on an open database that consists of 1297 manually delineated lung nodules.Methods: Conventional radiomics analysis was conducted by extracting standard handcrafted features from target nodule images. Several end-to-end deep classifier networks, including VGG, ResNet, DenseNet, and EfficientNet were employed to identify lung nodule malignancy as well. In addition to the baseline implementations, we also investigated the importance of feature selection and class balancing, as well as separating the features learned in the nodule target region and the background/context region. By pooling the radiomics and deep features together in a hybrid feature set, we investigated the compatibility of these two sets with respect to malignancy prediction.Results: The best baseline conventional radiomics model, deep learning model, and deep-feature based radiomics model achieved AUROC values (mean ± standard deviations) of 0.792 ± 0.025, 0.801 ± 0.018, and 0.817 ± 0.032, respectively through 5-fold cross-validation analyses. However, after trying out several optimization techniques, such as feature selection and data balancing, as well as adding context features, the corresponding best radiomics, end-to-end deep learning, and deep-feature based models achieved AUROC values of 0.921 ± 0.010, 0.824 ± 0.021, and 0.936 ± 0.011, respectively. We achieved the best prediction accuracy from the hybrid feature set (AUROC: 0.938 ± 0.010).Conclusion: The end-to-end deep-learning model outperforms conventional radiomics out of the box without much fine-tuning. On the other hand, fine-tuning the models lead to significant improvements in the prediction performance where the conventional and deep-feature based radiomics models achieved comparable results. The hybrid radiomics method seems to be the most promising model for lung nodule malignancy prediction in this comparative study.
|
|
| 15. |
- Astaraki, Mehdi, PhD Student, 1984-
(författare)
-
Advanced Machine Learning Methods for Oncological Image Analysis
- 2022
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Cancer is a major public health problem, accounting for an estimated 10 million deaths worldwide in 2020 alone. Rapid advances in the field of image acquisition and hardware development over the past three decades have resulted in the development of modern medical imaging modalities that can capture high-resolution anatomical, physiological, functional, and metabolic quantitative information from cancerous organs. Therefore, the applications of medical imaging have become increasingly crucial in the clinical routines of oncology, providing screening, diagnosis, treatment monitoring, and non/minimally-invasive evaluation of disease prognosis. The essential need for medical images, however, has resulted in the acquisition of a tremendous number of imaging scans. Considering the growing role of medical imaging data on one side and the challenges of manually examining such an abundance of data on the other side, the development of computerized tools to automatically or semi-automatically examine the image data has attracted considerable interest. Hence, a variety of machine learning tools have been developed for oncological image analysis, aiming to assist clinicians with repetitive tasks in their workflow.This thesis aims to contribute to the field of oncological image analysis by proposing new ways of quantifying tumor characteristics from medical image data. Specifically, this thesis consists of six studies, the first two of which focus on introducing novel methods for tumor segmentation. The last four studies aim to develop quantitative imaging biomarkers for cancer diagnosis and prognosis.The main objective of Study I is to develop a deep learning pipeline capable of capturing the appearance of lung pathologies, including lung tumors, and integrating this pipeline into the segmentation networks to leverage the segmentation accuracy. The proposed pipeline was tested on several comprehensive datasets, and the numerical quantifications show the superiority of the proposed prior-aware DL framework compared to the state of the art. Study II aims to address a crucial challenge faced by supervised segmentation models: dependency on the large-scale labeled dataset. In this study, an unsupervised segmentation approach is proposed based on the concept of image inpainting to segment lung and head-neck tumors in images from single and multiple modalities. The proposed autoinpainting pipeline shows great potential in synthesizing high-quality tumor-free images and outperforms a family of well-established unsupervised models in terms of segmentation accuracy.Studies III and IV aim to automatically discriminate the benign from the malignant pulmonary nodules by analyzing the low-dose computed tomography (LDCT) scans. In Study III, a dual-pathway deep classification framework is proposed to simultaneously take into account the local intra-nodule heterogeneities and the global contextual information. Study IV seeks to compare the discriminative power of a series of carefully selected conventional radiomics methods, end-to-end Deep Learning (DL) models, and deep features-based radiomics analysis on the same dataset. The numerical analyses show the potential of fusing the learned deep features into radiomic features for boosting the classification power.Study V focuses on the early assessment of lung tumor response to the applied treatments by proposing a novel feature set that can be interpreted physiologically. This feature set was employed to quantify the changes in the tumor characteristics from longitudinal PET-CT scans in order to predict the overall survival status of the patients two years after the last session of treatments. The discriminative power of the introduced imaging biomarkers was compared against the conventional radiomics, and the quantitative evaluations verified the superiority of the proposed feature set. Whereas Study V focuses on a binary survival prediction task, Study VI addresses the prediction of survival rate in patients diagnosed with lung and head-neck cancer by investigating the potential of spherical convolutional neural networks and comparing their performance against other types of features, including radiomics. While comparable results were achieved in intra-dataset analyses, the proposed spherical-based features show more predictive power in inter-dataset analyses.In summary, the six studies incorporate different imaging modalities and a wide range of image processing and machine-learning techniques in the methods developed for the quantitative assessment of tumor characteristics and contribute to the essential procedures of cancer diagnosis and prognosis.
|
|
| 16. |
- Astaraki, Mehdi, PhD Student, 1984-, et al.
(författare)
-
Benign-malignant pulmonary nodule classification in low-dose CT with convolutional features
- 2021
-
Ingår i: Physica medica (Testo stampato). - : Elsevier BV. - 1120-1797 .- 1724-191X. ; 83, s. 146-153
-
Tidskriftsartikel (refereegranskat)abstract
- Purpose: Low-Dose Computed Tomography (LDCT) is the most common imaging modality for lung cancer diagnosis. The presence of nodules in the scans does not necessarily portend lung cancer, as there is an intricate relationship between nodule characteristics and lung cancer. Therefore, benign-malignant pulmonary nodule classification at early detection is a crucial step to improve diagnosis and prolong patient survival. The aim of this study is to propose a method for predicting nodule malignancy based on deep abstract features.Methods: To efficiently capture both intra-nodule heterogeneities and contextual information of the pulmonary nodules, a dual pathway model was developed to integrate the intra-nodule characteristics with contextual attributes. The proposed approach was implemented with both supervised and unsupervised learning schemes. A random forest model was added as a second component on top of the networks to generate the classification results. The discrimination power of the model was evaluated by calculating the Area Under the Receiver Operating Characteristic Curve (AUROC) metric. Results: Experiments on 1297 manually segmented nodules show that the integration of context and target supervised deep features have a great potential for accurate prediction, resulting in a discrimination power of 0.936 in terms of AUROC, which outperformed the classification performance of the Kaggle 2017 challenge winner.Conclusion: Empirical results demonstrate that integrating nodule target and context images into a unified network improves the discrimination power, outperforming the conventional single pathway convolutional neural networks.
|
|
| 17. |
|
|
| 18. |
- Ghaderi Aram, Morteza, 1988, et al.
(författare)
-
Radiobiological evaluation of combined gamma knife radiosurgery and hyperthermia for pediatric neuro-oncology
- 2021
-
Ingår i: Cancers. - : MDPI AG. - 2072-6694. ; 13:13
-
Tidskriftsartikel (refereegranskat)abstract
- Combining radiotherapy (RT) with hyperthermia (HT) has been proven effective in the treatment of a wide range of tumours, but the combination of externally delivered, focused heat and stereotactic radiosurgery has never been investigated. We explore the potential of such treatment enhancement via radiobiological modelling, specifically via the linear-quadratic (LQ) model adapted to thermoradiotherapy through modulating the radiosensitivity of temperature-dependent parame-ters. We extend this well-established model by incorporating oxygenation effects. To illustrate the methodology, we present a clinically relevant application in pediatric oncology, which is novel in two ways. First, it deals with medulloblastoma, the most common malignant brain tumour in children, a type of brain tumour not previously reported in the literature of thermoradiotherapy studies. Second, it makes use of the Gamma Knife for the radiotherapy part, thereby being the first of its kind in this context. Quantitative metrics like the biologically effective dose (BED) and the tumour control probability (TCP) are used to assess the efficacy of the combined plan.
|
|
| 19. |
- Ghaderi Aram, Morteza, 1988, et al.
(författare)
-
Radiobiological modeling of hyperthermia combined with Gamma-Knife radiosurgery in pediatric brain cancer
- 2021
-
Konferensbidrag (övrigt vetenskapligt/konstnärligt)abstract
- Assessment of the synergistic effect of radiotherapy (RT) and hyperthermia (HT) in clinical settings is crucial for further expansion of hyperthermia. The radiobiological modeling using an extended version of the LQ model with temperature-dependent radiosensitivity parameters has been suggested in combination with external beam radiotherapy in previous studies. This study investigates the radiobiological effect of intracranial hyperthermia combined with stereotactic radiosurgery (SRS) in pediatric brain cancers. The hyperthermia treatment plan was achieved with an elliptical applicator consisting of 16 ORWG antennas working at 400 MHz and a hybrid Specific Absorption Rate (SAR) optimization procedure based on Time-Reversal and PSO. The radiotherapy plan was created by the treatment planning software of Leksell Gamma Knife® Icon™.
|
|
| 20. |
- Häger, Wille, 1990-
(författare)
-
A Novel Approach for Radiotherapy and Radiosurgery Treatment Planning Accounting for High-Grade Glioma Invasiveness into Normal Tissue
- 2023
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- High-grade gliomas (HGGs) are a type of malignant brain cancer, which include glioblastomas (GBMs). In adults, GBM is the most common malignant primary brain cancer. Attempts to treat patients with GBMs have been conducted for over a century, but the prognosis has only marginally improved. Current standard treatment involves surgical resection of the gross tumor volume (GTV), followed by radiotherapy and chemotherapy. Despite the efforts, the median survival for patients diagnosed with GBMs is less than 15 months. The inability to accurately determine the full extent of the tumor invaded regions in the brain is assumed to be the reason for the incurability of GBMs. In radiotherapy, the microscopic infiltration of normal tissue by tumor cells in the vicinity of the GTV is accounted for by extending the target into a clinical target volume (CTV). Current recommended margin widths for GBMs range from 15 to 30 mm. Despite a generous margin, the persistent recurrence of GBMs following treatment indicates that the CTV delineations currently used might fail to encompass the entirety of the tumor cell distribution, leaving clonogenic tumor cells untreated. To improve the CTV delineation and possibly treatment of GBMs, novel approaches in determining the tumor infiltrated regions have been suggested in the form of mathematical modeling. The aim of this project is to develop a mathematical model for the infiltration of glioma cells into normal brain tissue and implement it into a framework for predicting the full extent of tumor-invaded tissue for HGGs. This thesis is comprised of papers I–II, an overview of the methodology, results, and discussion of the work. The work herein is presented in order of: 1) model development; 2) model verification. Paper I explores the robustness and results of a mathematical model for tumor spread in terms of its input parameters. By applying the model to a large dataset, the behavior of the model could be investigated statistically, and optimal input parameters determined. The results of the tumor invasion simulations were compared in terms of volumes to the conventionally delineated CTVs, which were found not to adhere to the pathways of the simulated spread. Paper II used the resulting simulated invasions from paper I to predict the overall survival (OS) of the same cohort of cases. OS prediction was better predicted by the simulated volumes of the tumor spread than the size of the GTV. The results showed the potential of improving OS prediction and furthermore demonstrated a new methodology for indirect model verification that does not rely on histopathological data. Planned future work will revolve around dose prescription and plan optimization based on the simulated tumor spread, model investigation using artificial intelligence methods, and finally, practical implementation of the model into research versions of treatment planning systems.
|
|
