Lidia Strigari1, Andrea Attili, Andrea Duggento, Agostino Chiaravalloti, Orazio Schillaci, Maria Giovanna Guerrisi. 1. aLaboratory of Medical Physics and Expert Systems, Regina Elena National Cancer Institute bIstituto Nazionale Fisica Nucleare, Torino cMedical Physics Section, Department of Biomedicine and Prevention, University "Tor Vergata", Rome dDepartment of Biomedicine and Prevention, University of Rome Tor Vergata, Rome eIRCCS Neuromed, Pozzilli, Italy.
Abstract
OBJECTIVES: The qualitative analysis of interim PET has been reported to be useful for predicting the outcome of Hodgkin's lymphoma (HL) after chemotherapy. As the next step, our study aims to present a quantitative analysis on the basis of both a basal (PET/CT0) and an interim (PET/CT2) scan to improve the prognostic value of imaging in HL patients. PATIENTS AND METHODS: A cohort of 68 patients undergoing a basal and an interim scan with F-fluorodeoxyglucose after two cycles of chemotherapy consisting of adriamycin, bleomycin, vinblastine, and dacarbazine were examined. Two subsets of patients with a positive and a negative interim scan were selected. RESULTS: In patients with a negative scan, a total of 108 lymph node lesions showing a good response to chemotherapy were contoured, whereas in the remaining patients with positive scans, six responder and 12 relapsing lymph node lesions were contoured. Standardized uptake value (SUV) and Hounsfield unit (HU) values were included in the volumes contoured on coregistered basal and interim scans and included in a database. A linear regression model was used to identify the predictor of relapse at the lesion level. The support vector machine analysis and bootstrap approach were used to determine the model capability. The predictive models were presented as nomograms on the basis of basal or both basal and interim studies. SUV at the basal/interim study and basal HU values were predictors of a poor prognosis. In particular, the higher points were associated with lower values of SUV and HU at baseline and the higher values of SUV at the interim study. Using the bootstrap and support vector machine approach, the cut-off of the model increased up to 89%. CONCLUSION: The novel tool enables estimation of the risk of tumor relapse after chemotherapy in HL patients on the basis of basal and interim PET/CT scans including SUV and densitometric information.
OBJECTIVES: The qualitative analysis of interim PET has been reported to be useful for predicting the outcome of Hodgkin's lymphoma (HL) after chemotherapy. As the next step, our study aims to present a quantitative analysis on the basis of both a basal (PET/CT0) and an interim (PET/CT2) scan to improve the prognostic value of imaging in HL patients. PATIENTS AND METHODS: A cohort of 68 patients undergoing a basal and an interim scan with F-fluorodeoxyglucose after two cycles of chemotherapy consisting of adriamycin, bleomycin, vinblastine, and dacarbazine were examined. Two subsets of patients with a positive and a negative interim scan were selected. RESULTS: In patients with a negative scan, a total of 108 lymph node lesions showing a good response to chemotherapy were contoured, whereas in the remaining patients with positive scans, six responder and 12 relapsing lymph node lesions were contoured. Standardized uptake value (SUV) and Hounsfield unit (HU) values were included in the volumes contoured on coregistered basal and interim scans and included in a database. A linear regression model was used to identify the predictor of relapse at the lesion level. The support vector machine analysis and bootstrap approach were used to determine the model capability. The predictive models were presented as nomograms on the basis of basal or both basal and interim studies. SUV at the basal/interim study and basal HU values were predictors of a poor prognosis. In particular, the higher points were associated with lower values of SUV and HU at baseline and the higher values of SUV at the interim study. Using the bootstrap and support vector machine approach, the cut-off of the model increased up to 89%. CONCLUSION: The novel tool enables estimation of the risk of tumor relapse after chemotherapy in HL patients on the basis of basal and interim PET/CT scans including SUV and densitometric information.
Authors: Ping Wang; Byunghee Yoo; Sarah Sherman; Pinku Mukherjee; Alana Ross; Pamela Pantazopoulos; Victoria Petkova; Christian Farrar; Zdravka Medarova; Anna Moore Journal: Int J Cancer Date: 2016-04-15 Impact factor: 7.396
Authors: A Pontecorvi; S Nanni; C Possieri; P Locantore; C Salis; L Bacci; A Aiello; G Fadda; C De Crea; M Raffaelli; R Bellantone; C Grassi; L Strigari; A Farsetti Journal: Endocrine Date: 2020-10-08 Impact factor: 3.633