Authors' objectives:

The objective of this report is to assess, in light of the existing scientific evidence, the relative contribution of PET, PET-CT (positron emission tomography-computed tomography) and CT alone to clinical management of patients with pancreatic, ovarian, esophageal, gastric, breast and testicular cancer. This report aims to answer if this technology is able to provide a higher diagnostic accuracy compared to other available technologies.

Authors' results and conclusions: 509 articles were retrieved, but only 68 were finally selected and 64 included in the meta-analysis (MA) (3,691 patients, 1,780 cases and 1,911 controls). These 64 articles were classified according to their methodological quality and recommendation in classes B (28 articles) and C (36 articles). The pooled sensitivity (Se) of PET was 0.69 (CI 95%, 0.67- 0.71), pooled specificity (Sp) 0.88 (0.86-0.89), and 0.62 (0.59-0.65) and 0.87 (0.85-0.89), respectively, in the case of CT. The area under the SROC of PET was 0.88 (EE = 0.02) and the Inouye-Sox (Q*) point was 0.81 (EE = 0.02), and for CT, 0.93 (0.07) & 0.87 (0.09), respectively. Very close results were found for PET and CT when the MA for the different types of tumor were performed, except for PET in the diagnostic of ovarian cancer relapse: the pooled Se was 0.82 (0.77-0.85) and the pooled Sp was 0.86 (0.80-0.91) and for CT these figures were 0.59 (0.52-0.66) and 0.78 (0.67-0.86), respectively. Twenty nine patients with breast cancer and ten with testicular tumors have been included. For the first, the Se of PET was 0.89 (0.67-0.97) and the Sp was 0.91 (0.62-0.98), and in the second case, 0.92 (0.54-0.99) and 0.75 (0.30-0.95), respectively.

Authors' recommendations: PET is a useful diagnostic technique for malignancy detection, although the acceptable methodological quality in the published works on this technology in pancreatic, ovarian, esophageal and gastric tumours is not is always correct (56% of C4 class). The results of this MA agreed only with two MA from the medical literature. One concerning esophageal patients, with a pooled Se and Sp for PET in the setting of distant metastases close to ours. The second study about PET in ovarian cancer resulted in different scores respecting to the present report, perhaps it was due to the fact that the articles included in this work were of C class (but two of B class, 11.76%), while in our report from 18 articles included, 6 were of B class (33.33%). Notwithstanding, in this study PET doesn't appear as the first discriminative option, because of its low whole Se for these four tumours. Though the PET Sp was good enough (close to that obtained by CT), it can't be considered the first option for confirming cases. The PET OR was superior to the CT OR but the SROC curve was better for CT. Anyway, in ovarian relapse, PET Se was clearly superior to CT Se, but as the majority of articles were of C4 EBM classification, it would be necessary to perform more works to confirm these preliminary results. In the case of PET in breast and testicular cancer, where a low number of patients have been taken into account, the results are favourable for PET, but it must be assessed with more cases and set the clinical indications where PET may have a decisive role.

Authors' methods: Systematic review

Project Status: Completed

URL for project: http://www.isciii.es/aets

Year Published: 2006

English language abstract: An English language summary is available

Publication Type: Not Assigned

Country: Spain

Organisation Name: Agencia de Evaluacion de Tecnologias Sanitarias

Contact Address: Instituto de Salud "Carlos III", Calle Sinesio Delgado 6, Pabellon 4, 28029 Madrid, Spain. Tel: +34 9 1 822 2005; Fax: +34 9 1 387 7841;

Contact Name: Luis M. Sánchez Gómez

Contact Email: luism.sanchez@isciii.es

Copyright: Agencia de Evaluacion de Tecnologias Sanitarias (AETS)