We performed a retrospective analysis to evaluate the ability of whole-body 18 F-fluorodeoxyglucose positron emission tomography (FDG PET) to identify local recurrence and pulmonary metastases in patients with soft-tissue tumours after treatment. We compared the results of FDG PET with those of MRI for the detection of local recurrence, and with CT of the chest for pulmonary metastases.We assessed 62 patients of mean age 51 years, who had 15 types of soft-tissue sarcoma, after a mean follow-up of 3 years 2 months. For the detection of local disease, 71 comparisons showed that the sensitivity and specificity of FDG PET were 73.7% and 94.3%, respectively; there were 14 true-positive and five false-negative results. MRI had a sensitivity and specificity of 88.2% and 96.0% respectively. For the identification of lung metastases, 70 comparisons showed that the sensitivity and specificity of FDG PET were 86.7% and 100%, with 13 true-positive results and two false-negative results. CT of the chest had a sensitivity and specificity of 100% and 96.4%. Thirteen other sites of metastases were identified by FDG PET.FDG PET can identify both local and distant recurrence of tumour as a one-step procedure and will detect other metastases. It seems that all three methods of imaging are needed to define accurately the extent of disease, both at initial staging and during follow-up. and 23% of patients will have metastases, the lung being the most common site with one-third of secondary tumours. Deposits in bone, liver, and brain comprise about 40%; the others are found in the regional lymph nodes, retroperitoneum and soft tissues.
3,4After treatment of the primary tumour between 30% and 35% of patients will develop recurrence either locally or at a distant site. 5,6 Of these patients, between 15% and 47% will develop a local recurrence, depending on the margins achieved at surgical resection. Isolated metastases are seen in the lung in 38% to 64%, and from 28% to 34% will develop metastases at multiple sites.
5-7The optimal management of these tumours depends on the site, size and grade of the local growth and accurate staging of the disease when first seen. The site and size are best determined by MRI, [8][9][10] but the accurate diagnosis of distant metastases is confounded by the wide distribution of potential sites. The detection of local recurrence is hampered by difficulty in differentiating between recurrence, the changes after operation, and the effect of any radiotherapy. Local recurrence is best assessed using MRI with gadolinium contrast enhancement; regular clinical examination alone is insensitive. MRI has a sensitivity of 80% to 85% for the detection of local recurrence, better than CT which has a sensitivity of 57% to 70%. [11][12][13][14] Ultrasound has a similar sensitivity to MRI for the detection of local recurrence, but depends very much on the proficiency of the operator. 10,12 A variety of radionucleide methods has been tried with varying success.14,15Lung metastases can be identified using CT, but other