Search

The rapid development of artificial intelligence (AI) has gained importance, with many tools already entering our daily lives. The medical field of radiation oncology is also subject to this development, with AI entering all steps of the patient journey. In this review article, we summarize contemporary AI techniques and explore the clinical applications of AI-based automated segmentation models in radiotherapy planning, focusing on delineation of organs at risk (OARs), the gross tumor volume (GTV), and the clinical target volume (CTV). Emphasizing the need for precise and individualized plans, we review various commercial and freeware segmentation tools and also state-of-the-art approaches. Through our own findings and based on the literature, we demonstrate improved efficiency and consistency as well as time savings in different clinical scenarios. Despite challenges in clinical implementation such as domain shifts, the potential benefits for personalized treatment planning are substantial. The integration of mathematical tumor growth models and AI-based tumor detection further enhances the possibilities for refining target volumes. As advancements continue, the prospect of one-stop-shop segmentation and radiotherapy planning represents an exciting frontier in radiotherapy, potentially enabling fast treatment with enhanced precision and individualization., Competing Interests: Declarations. Conflict of interest: The authors declare no conflicts of interest. Editorial Policies: Springer journals and proceedings: https://www.springer.com/gp/editorial-policies, (© 2024. The Author(s).)

Background and Introduction: Increasing evidence suggests that a subgroup of patients with oligometastatic cancer might achieve a prolonged disease-free survival through local therapy for all active cancer lesions. Our aims are to investigate the impact of brain metastases on the classification, treatment, and outcome in these patients., Materials and Methods: We analyzed a total of 7,000 oncological positron emission tomography scans to identify patients with extracranial oligometastatic disease (defined as ≤ 5 intra- or extra-cranial metastases). Concurrent magnetic resonance imaging brain was assessed to quantify intracranial tumor burden. We investigated the impact of brain metastases on oligometastatic disease state, therapeutic approaches, and outcome. Predictors for transitioning from oligo- to polymetastatic states were evaluated using regression analysis., Results: A total of 106 patients with extracranial oligometastases and simultaneous brain metastases were identified, primarily originating from skin or lung/pleura cancers (90%, n = 96). Brain metastases caused a transition from an extracranial oligometastatic to a whole-body polymetastatic state in 45% (n = 48) of patients. While oligometastatic patients received systemic therapy (55% vs. 35%) more frequently and radiotherapy for brain metastases was more often prescribed to polymetastatic patients (44% vs. 26%), the therapeutic approach did not differ systematically between both sub-groups. The oligometastatic sub-group had a median overall survival of 28 months compared to 10 months in the polymetastatic sub-group (p < 0.01)., Conclusion: In patients with brain metastases, a low total tumor burden with an oligometastatic disease state remained a significant prognostic factor for overall survival. Presence of brain metastases should therefore not serve as exclusion criterion for clinical trials in the field of oligometastatic disease. Moreover, it underscores the importance of considering a multimodality treatment strategy in oligometastatic cancer patients., (© 2024. The Author(s).)

Background: Surgical resection is the standard of care for patients with large or symptomatic brain metastases (BMs). Despite improved local control after adjuvant stereotactic radiotherapy, the risk of local failure (LF) persists. Therefore, we aimed to develop and externally validate a pre-therapeutic radiomics-based prediction tool to identify patients at high LF risk., Methods: Data were collected from A Multicenter Analysis of Stereotactic Radiotherapy to the Resection Cavity of BMs (AURORA) retrospective study (training cohort: 253 patients from 2 centers; external test cohort: 99 patients from 5 centers). Radiomic features were extracted from the contrast-enhancing BM (T1-CE MRI sequence) and the surrounding edema (T2-FLAIR sequence). Different combinations of radiomic and clinical features were compared. The final models were trained on the entire training cohort with the best parameter set previously determined by internal 5-fold cross-validation and tested on the external test set., Results: The best performance in the external test was achieved by an elastic net regression model trained with a combination of radiomic and clinical features with a concordance index (CI) of 0.77, outperforming any clinical model (best CI: 0.70). The model effectively stratified patients by LF risk in a Kaplan-Meier analysis (P < .001) and demonstrated an incremental net clinical benefit. At 24 months, we found LF in 9% and 74% of the low and high-risk groups, respectively., Conclusions: A combination of clinical and radiomic features predicted freedom from LF better than any clinical feature set alone. Patients at high risk for LF may benefit from stricter follow-up routines or intensified therapy., (© The Author(s) 2024. Published by Oxford University Press on behalf of the Society for Neuro-Oncology.)

Background: The diffuse growth pattern of glioblastoma is one of the main challenges for accurate treatment. Computational tumor growth modeling has emerged as a promising tool to guide personalized therapy. Here, we performed clinical and biological validation of a novel growth model, aiming to close the gap between the experimental state and clinical implementation., Methods: One hundred and twenty-four patients from The Cancer Genome Archive (TCGA) and 397 patients from the UCSF Glioma Dataset were assessed for significant correlations between clinical data, genetic pathway activation maps (generated with PARADIGM; TCGA only), and infiltration ( D w ) as well as proliferation (ρ) parameters stemming from a Fisher-Kolmogorov growth model. To further evaluate clinical potential, we performed the same growth modeling on preoperative magnetic resonance imaging data from 30 patients of our institution and compared model-derived tumor volume and recurrence coverage with standard radiotherapy plans., Results: The parameter ratio D w /ρ ( P  < .05 in TCGA) as well as the simulated tumor volume ( P  < .05 in TCGA/UCSF) were significantly inversely correlated with overall survival. Interestingly, we found a significant correlation between 11 proliferation pathways and the estimated proliferation parameter. Depending on the cutoff value for tumor cell density, we observed a significant improvement in recurrence coverage without significantly increased radiation volume utilizing model-derived target volumes instead of standard radiation plans., Conclusions: Identifying a significant correlation between computed growth parameters and clinical and biological data, we highlight the potential of tumor growth modeling for individualized therapy of glioblastoma. This might improve the accuracy of radiation planning in the near future., Competing Interests: M.M. and F.S. serve as part-time consultants for Novocure GmbH., (© The Author(s) 2023. Published by Oxford University Press, the Society for Neuro-Oncology and the European Association of Neuro-Oncology.)

Background: Many automatic approaches to brain tumor segmentation employ multiple magnetic resonance imaging (MRI) sequences. The goal of this project was to compare different combinations of input sequences to determine which MRI sequences are needed for effective automated brain metastasis (BM) segmentation., Methods: We analyzed preoperative imaging (T1-weighted sequence ± contrast-enhancement (T1/T1-CE), T2-weighted sequence (T2), and T2 fluid-attenuated inversion recovery (T2-FLAIR) sequence) from 339 patients with BMs from seven centers. A baseline 3D U-Net with all four sequences and six U-Nets with plausible sequence combinations (T1-CE, T1, T2-FLAIR, T1-CE + T2-FLAIR, T1-CE + T1 + T2-FLAIR, T1-CE + T1) were trained on 239 patients from two centers and subsequently tested on an external cohort of 100 patients from five centers., Results: The model based on T1-CE alone achieved the best segmentation performance for BM segmentation with a median Dice similarity coefficient (DSC) of 0.96. Models trained without T1-CE performed worse (T1-only: DSC = 0.70 and T2-FLAIR-only: DSC = 0.73). For edema segmentation, models that included both T1-CE and T2-FLAIR performed best (DSC = 0.93), while the remaining four models without simultaneous inclusion of these both sequences reached a median DSC of 0.81-0.89., Conclusions: A T1-CE-only protocol suffices for the segmentation of BMs. The combination of T1-CE and T2-FLAIR is important for edema segmentation. Missing either T1-CE or T2-FLAIR decreases performance. These findings may improve imaging routines by omitting unnecessary sequences, thus allowing for faster procedures in daily clinical practice while enabling optimal neural network-based target definitions., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.)

Background: Stereotactic radiotherapy is a standard treatment option for patients with brain metastases. The planning target volume is based on gross tumor volume (GTV) segmentation. The aim of this work is to develop and validate a neural network for automatic GTV segmentation to accelerate clinical daily routine practice and minimize interobserver variability., Methods: We analyzed MRIs (T1-weighted sequence ± contrast-enhancement, T2-weighted sequence, and FLAIR sequence) from 348 patients with at least one brain metastasis from different cancer primaries treated in six centers. To generate reference segmentations, all GTVs and the FLAIR hyperintense edematous regions were segmented manually. A 3D-U-Net was trained on a cohort of 260 patients from two centers to segment the GTV and the surrounding FLAIR hyperintense region. During training varying degrees of data augmentation were applied. Model validation was performed using an independent international multicenter test cohort (n = 88) including four centers., Results: Our proposed U-Net reached a mean overall Dice similarity coefficient (DSC) of 0.92 ± 0.08 and a mean individual metastasis-wise DSC of 0.89 ± 0.11 in the external test cohort for GTV segmentation. Data augmentation improved the segmentation performance significantly. Detection of brain metastases was effective with a mean F1-Score of 0.93 ± 0.16. The model performance was stable independent of the center (p = 0.3). There was no correlation between metastasis volume and DSC (Pearson correlation coefficient 0.07)., Conclusion: Reliable automated segmentation of brain metastases with neural networks is possible and may support radiotherapy planning by providing more objective GTV definitions., (Copyright © 2023 The Authors. Published by Elsevier B.V. All rights reserved.)

Purpose: This study aims to evaluate the association of the maximum standardized uptake value (SUVmax) in positron-emission tomography targeting prostate-specific membrane antigen (PSMA-PET) prior to salvage radiotherapy (sRT) on biochemical recurrence free survival (BRFS) in a large multicenter cohort., Methods: Patients who underwent 68  Ga-PSMA11-PET prior to sRT were enrolled in four high-volume centers in this retrospective multicenter study. Only patients with PET-positive local recurrence (LR) and/or nodal recurrence (NR) within the pelvis were included. Patients were treated with intensity-modulated-sRT to the prostatic fossa and elective lymphatics in case of nodal disease. Dose escalation was delivered to PET-positive LR and NR. Androgen deprivation therapy was administered at the discretion of the treating physician. LR and NR were manually delineated and SUVmax was extracted for LR and NR. Cox-regression was performed to analyze the impact of clinical parameters and the SUVmax-derived values on BRFS., Results: Two hundred thirty-five patients with a median follow-up (FU) of 24 months were included in the final cohort. Two-year and 4-year BRFS for all patients were 68% and 56%. The presence of LR was associated with favorable BRFS (p = 0.016). Presence of NR was associated with unfavorable BRFS (p = 0.007). While there was a trend for SUVmax values ≥ median (p = 0.071), SUVmax values ≥ 75% quartile in LR were significantly associated with unfavorable BRFS (p = 0.022, HR: 2.1, 95%CI 1.1-4.6). SUVmax value in NR was not significantly associated with BRFS. SUVmax in LR stayed significant in multivariate analysis (p = 0.030). Sensitivity analysis with patients for who had a FU of > 12 months (n = 197) confirmed these results., Conclusion: The non-invasive biomarker SUVmax can prognosticate outcome in patients undergoing sRT and recurrence confined to the prostatic fossa in PSMA-PET. Its addition might contribute to improve risk stratification of patients with recurrent PCa and to guide personalized treatment decisions in terms of treatment intensification or de-intensification. This article is part of the Topical Collection on Oncology-Genitourinary., (© 2022. The Author(s).)

Purpose: Ultrasound (US) is a burgeoning diagnostic tool and is often the only available imaging modality in low- and middle-income countries (LMICs). However, bedside providers often lack training to acquire or interpret US images. We conducted a study to determine if a remote tele-intensivist could mentor geographically removed LMIC providers to obtain quality and clinically useful US images., Materials and Methods: Nine Haitian non-physician health care workers received a 20-minute training on basic US techniques. A volunteer was connected to an intensivist located in the USA via FaceTime. The intensivist remotely instructed the non-physicians to ultrasound five anatomic sites. The tele-intensivist evaluated the image quality and clinical utility of performing tele-ultrasound in a LMIC., Results: The intensivist agreed (defined as "agree" or "strongly agree" on a five-point Likert scale) that 90% (57/63) of the FaceTime images were high quality. The intensivist felt comfortable making clinical decisions using FaceTime images 89% (56/63) of the time., Conclusions: Non-physicians can feasibly obtain high-quality and clinically relevant US images using video chat software in LMICs. Commercially available software can connect providers in institutions in LMICs to geographically removed intensivists at a relatively low cost and without the need for extensive training of local providers., (Copyright © 2017 Elsevier Inc. All rights reserved.)

Purpose: Remote telementored ultrasound (RTMUS) systems can deliver ultrasound (US) expertise to regions lacking highly trained bedside ultrasonographers and US interpreters. To date, no studies have evaluated the quality and clinical utility of US images transmitted using commercially available RTMUS systems., Methods: This prospective pilot evaluated the quality of US images (right internal jugular vein, lung apices and bases, cardiac subxiphoid view, bladder) obtained using a commercially available iPad operating FaceTime software. A bedside non-physician obtained images and a tele-intensivist interpreted them. All US screen images were simultaneously saved on the US machine and captured via a FaceTime screen shot. The tele-intensivist and an independent US expert rated image quality and utility in guiding clinical decisions., Results: The tele-intensivist rated FaceTime images as high quality (90% [69/77]) and could comfortably make clinical decisions using these images (96% [74/77]). Image quality did not differ between FaceTime and US images (97% (75/77). Strong inter-rater reliability existed between tele-intensivist and US expert evaluations (Spearman's rho 0.43; P<.001)., Conclusion: An RTMUS system using commercially available two-way audiovisual technology can transmit US images without quality degradation. For most anatomic sites assessed, US images acquired using FaceTime are not inferior to those obtained directly with the US machine., (Copyright © 2016 Elsevier Inc. All rights reserved.)