Fuhrman Grade Classification of Clear-cell Renal Cell Carcinoma Using Computed Tomography Image Analysis
Meng, Fei (2017-08-15)
Fuhrman Grade Classification of Clear-cell Renal Cell Carcinoma Using Computed Tomography Image Analysis
Meng, Fei
(15.08.2017)
Tätä artikkelia/julkaisua ei ole tallennettu UTUPubiin. Julkaisun tiedoissa voi kuitenkin olla linkki toisaalle tallennettuun artikkeliin / julkaisuun.
Turun yliopisto
Kuvaus
Siirretty Doriasta
Tiivistelmä
The renal cancer is a kind of malignant tumors which is originated in the renal parenchymal urinary epithelial system, accounting for about 2% to 3% of adult malignancies, for 80% to 90% of adult renal malignancies. The clear cell renal cell carcinoma (CCRCC) is the most common type of the renal cancer, about 85% renal cell carcinoma is the CCRCC. The classification of the CCRCC is based on the Fuhrman classification system. Clinically, several studies have shown that the original moderate intra- and inter-observer agreement among pathologists is improved to the substantial agreement when the Fuhrman grading system is collapsed to two categories where Fuhrman grades I and II are considered together as the low grade and Fuhrman grades III and IV are considered together as the high grade. In this work, the classification is also divided into the low grade and high grade.
A set of methods for the CCRCC classification are proposed in this thesis. The system includes the image segmentation, the feature extraction and selection, and the classification. The system is applied to the Fuhrman classification for CT images of 90 renal clear cell carcinoma (24 tumors of the Fuhrman grade I, 26 tumors of the Fuhrman grade II, 27 tumors of the Fuhrman grades III and 13 tumors of the Fuhrman grade IV). Fuhrman grades I and II are the class one which is regarded as the low grade, Fuhrman grades III and IV are the class two which is regarded as the high grade. The classification accuracy reaches 70%, the sensitivity reaches 72%, and the specificity is 75%. The result of the classification is compared to the diagnosis of doctors, which proves that the system improves the accuracy rate. To some degree, it achieves a non-invasive way to achieve the renal clear cell Fuhrman classification.
A set of methods for the CCRCC classification are proposed in this thesis. The system includes the image segmentation, the feature extraction and selection, and the classification. The system is applied to the Fuhrman classification for CT images of 90 renal clear cell carcinoma (24 tumors of the Fuhrman grade I, 26 tumors of the Fuhrman grade II, 27 tumors of the Fuhrman grades III and 13 tumors of the Fuhrman grade IV). Fuhrman grades I and II are the class one which is regarded as the low grade, Fuhrman grades III and IV are the class two which is regarded as the high grade. The classification accuracy reaches 70%, the sensitivity reaches 72%, and the specificity is 75%. The result of the classification is compared to the diagnosis of doctors, which proves that the system improves the accuracy rate. To some degree, it achieves a non-invasive way to achieve the renal clear cell Fuhrman classification.