0% Complete
صفحه اصلی
/
اولین کنگره بین المللی رویکردهای نوین سبک زندگی، پیشگیری و درمان سرطان
Radiomics: a developing field for better prognosis and diagnosis of lung metastasis patients
نویسندگان :
Yunus Soleymani (دانشگاه علوم پزشکی تهران) , Davood Khezerloo (دانشگاه علوم پزشکی تبریز)
کلمات کلیدی :
Lung metastasis،Radiomics،Machine learning
چکیده :
Background: About 20-54% of metastatic patients experience lung metastasis (LM) as the second most common site for metastasis in cancer patients. The most common extra-thoracic cancers that lead to LM are breast, colorectal, renal, uterine cancer, leiomyosarcoma, and head and neck carcinoma. The most important challenge is to differentiate the primary or benign lung lesions from LM. Common diagnostic methods for the issue are multidetector computed tomography (MDCT) and 18-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT). Despite the multiple advantages of these two imaging modalities, they could confront false positive or false negative results. On the other hand, biopsy as the gold standard to distinguish between primitive lung tumors or LM is highly invasive and not applicable to all cases. Radiomics is an active area of research to extract quantitative data from diagnostic images, which can serve as useful imaging biomarkers for more effective and customized patient care. Our purpose is to review the current applications of radiomics for lesion characterization, treatment planning, and prognostic assessment in patients with LM. Methods: The keywords of “radiomics”, “metastasis OR metastases”, “machine learning”, and “lung OR pulmonary” were entered into scientific databases of Google Scholar, Scopus, PubMed, and Elsevier. Finally, 8 fully relevant papers (publication year: 2018-2022) were extracted and reviewed. Results: The number of patients evaluated in the 8 studies was variable between 51 and 769. The studies could be categorized into 1. Studies distinguishing histological subtypes of the LM (n = 6), 2. Studies evaluating the mutational status of the LM (n = 1), and 3. Studies evaluating the prognostic ability of radiomics (n = 1). In all papers, the imaging modality was CT or 18F-FDG PET/CT. The segmentation process of the metastatic lesions was conducted manually or semi-automatically using artificial intelligence approaches. In all papers, the machine learning methods’ performance in distinguishing histological subtypes or evaluating the mutational status and prognostic ability was variable (poor performance with AUC=0.57 to excellent performance with AUC=0.98). Conclusions: Despite the potential of overcoming the conventional imaging methods for LM patient management, radiomics is not still well adopted clinically. This could be due to several factors such as the standardization of imaging parameters and radiomics features definitions. Legal entity responsible for the study: The authors. Funding: Has not received any funding. Disclosure: The authors have no conflicts of interest.
لیست مقالات
ثمین همایش، سامانه مدیریت کنفرانس ها و جشنواره ها - نگارش 35.4.1
