Bedriye Dogan

The exponential growth in the size of databases due to technological advancements has led to challenges in locating and accessing specific components of the data. While deep learning and other machine learning architectures have shown promise in retrieving data components, their efficacy is more pronounced when addressing disease cohorts. Contrarily, this effectiveness diminishes when accessing large datasets. This study focuses on the analysis of brain magnetic resonance imaging (MRI) images and, specifically, to differentiate between benign and malignant lesions associated with Alzheimer’s disease, multiple sclerosis (MS), and intracranial regions, all of which are medically significant with distinct treatment modalities. A hybrid model was first devised to facilitate image retrieval by employing a pre-trained EfficientNet-b0 and local binary pattern (LBP) for feature extraction. These extracted features were then amalgamated to encompass diverse aspects of each image. To improve model performance, redundant features were pruned using the minimum redundancy maximum relevance (mRMR) technique. As a result, the proposed model demonstrated efficacy in analyzing a diverse dataset encompassing three distinct diseases and eight unique classes. Notably, existing machine architectures already published in the literature have struggled to achieve comparable success rates in discerning such closely related yet distinct disease groups. Our study underscores the challenge posed by increasing class diversity on the performance of deep learning architectures and obtained an accuracy of 98.9% in classifying three diseases and eight unique classes. As a result, the same model was used as the base in both the classification and CBIR processes for MRI detection, yielding competitive results when compared with the literature and other models.

Objective: To investigate the incidence, timing, and the factors predictors radionecrosis (RN) development in brain metastases (BMs) undergoing stereotactic radiotherapy (SRT). Methods: The study evaluated 245 BMs who exclusively received SRT between 2010 and 2020. RN was detected pathologically or radiologically. Results: The median of follow-up was 22.6 months. RN was detected in 18.4% of the metastatic lesions, and 3.3% symptomatic, 15.1% asymptomatic. The median time of RN was 22.8 months (2.5-39.5), and the rates at 6, 12, and 24 months were 16.8%, 41.4%, and 66%, respectively. Univariate analysis revealed that Graded Prognostic Assessment (P = .005), Score Index of Radiosurgery (P = .015), Recursive Partitioning Analysis (P = .011), the presence of primary cancer (P = .004), and localization (P = .048) significantly increased the incidence of RN. No significant relationship between RN and brain-gross tumour volume doses, planning target volume, fractionation, dose (P > .05). Multivariate analysis identified SIR > 6 (OR: 1.30, P = .021), primary of breast tumour (OR: 2.33, P = .031) and supratentorial localization (OR: 3.64, P = .025) as risk factors. Conclusions: SRT is used effectively in BMs. The incidence of RN following SRT is undeniably frequent. It was observed that the incidence rate increased as the follow-up period increased. We showed that brain-GTV doses are not predictive of RN development, unlike other publications. In study, a high SIR score and supratentorial localization were identified as factors that increased the risk of RN. Advances in knowledge: RN is still a common complication after SRT. Symptomatic RN is a significant cause of morbidity. The causes of RN are still not clearly identified. In many publications, brain dose and volumes have been found to be effective in RN. But, with this study, we found that brain dose volumes and fractionation did not increase the incidence of RN when brain doses were taken into account. The most important factor in the development of RN was found to be related to long survival after SRT.

Background: To evaluate the factors influencing brain metastases (BM) local control (LC) after stereotactic radiotherapy (SRT).Methods: Between 2010 and 2020, a cohort of 145 patients (246 BM) treated consecutively with robotic radiosurgery was analysed.Results: Median age was 61 years (range, 29–90 years). Median radiological follow‐up of the lesions was 21.7 months (range, 3–115 months). The mean overall survival and LC were 33.0 and 82.7 months, respectively. On univariate analysis, sex, primary cancer site, histological type, use of systemic steroids, maximum diameter, volume, early MRI response, isodose line, number of fractions, BED10 value, and BED10 value proportional to volume and maximum diameter were significant factors for LC. On multivariate analysis, female sex (hazard ratio [HR]: 2.10 P: 0.035), adenocarcinoma histology (HR: 6.54 P: 0.001), no steroid use (HR: 3.60 P: 0.001), maximum diameter (≤1 cm) (HR: 2.64 P: 0.018), complete response of lesion at first follow‐up MRI compared to stable or progressive disease (HR: 4.20, P = 0.024; HR: 19.15, P < 0.001), isodose line (≥90%) (HR: 2.00 P: 0.036), and tumour volume (PTV ≤2 cc) (HR: 5.19 P: 0.001) were independent factors improving LC.Conclusions: SRT is an effective treatment for patients with a limited number of BM with a high LC rate. There are many factors related to the patient, tumour, and radiotherapy plan that have an impact on LC after SRT in brain metastases. These results warrant further investigation in a prospective setting.