Sarcomas are a heterogenous group of rare tumors that arise predominantly from the embryonic mesoderm. Only 15% of soft-tissue sarcomas originate in the retroperitoneum, and these constitute only 1–2% of all solid malignancies 1. The overall incidence of retroperitoneal sarcoma (RS) is 0.3–0.4%/100 000 of the population. The peak incidence is in the fifth decade of life, although it can occur at any age 2.
Currently, more than 50 histologic types of soft-tissue sarcomas have been identified. The two most common histologic types of RS are liposarcomas and leiomyosarcomas 1.
The onset of clinical signs and symptoms is determined by the anatomic site (depth), grade, and size of the tumor 1.
RS arise in the large potential spaces of the retroperitoneum and can grow very large without producing symptoms 2, which are often nonspecific, such as abdominal pain and fullness, and easily dismissed as being caused by other less serious conditions 3. RS are usually very large at the time of diagnosis 2.
Nearly 50% of RS are larger than 20 cm at the time of diagnosis. Occasionally, patients may initially exhibit obstructive gastrointestinal or neurologic symptoms related to compression of the lumbar or pelvic nerves 1.
Biopsy should be considered in patients with all types of RS to establish a definitive diagnosis 1.
Survival rates at 5 and 10 years are, respectively, 40–50 and 36% 4.
The best chance for long-term survival for patients with RS is achieved with a margin-negative resection. About 75% of patients with RS ultimately die of locally recurrent disease without distant metastasis 1.
However, because these tumors often involve vital structures, surgical resection is not always possible, and even if surgical resection can be performed the margins are often compromised because of anatomic constraints 1.
Because surgery plays such a central role in the treatment for recurrent RS, it is beneficial when recurrences are identified while they are small in size, improving the likelihood of complete resection 5,6.
To date, there is a paucity of information on the imaging of recurrent RS; in particular, protocols for their recognition and distinction from fibrous tissue are not well defined 5.
Similar to primary RS, recurrent tumors can grow quite large without producing any symptoms. Recurrences typically appear within 2 years of initial surgery. According to previous results, 76% of tumors recurred in the first 2 years 5.
Lymph node involvement of soft-tissue sarcomas is rare; less than 5% show nodal spread from only a few histologic subtypes 1.
The dominant pattern of metastasis is hematogenous. Distant metastases occur most often to the lung. Few patients with pulmonary metastases may survive for long periods after surgical resection of metastatic tumors. RS also show a propensity to metastasize to the liver, peritoneum, and less frequently to bones 1.
Contrast enhancement computed tomography (CECT) is the most commonly used morphological modality for identification, localization, staging, and definition of RS and its relationship to surrounding structures, particularly vascular structures 1,7. This technique is also the most used in the early follow-up of patients with RS.
Nuclear medicine functional imaging has been proposed for the whole-body evaluation of RS 8.
In recent years, 18-fluorine-labeled 2-deoxy-2-fluoro-D-glucose PET/computed tomography (18F-FDG-PET/CT) imaging has been used to assess the tumor grade and evaluate patients for tumor recurrence 2.
18F-FDG-PET/CT combines the relatively high spatial resolution with increased lesion contrast and the ability to assay pathological disease throughout the body with coregistration of functional and anatomical information, which improves the localizing ability of PET 9,10.
Today, 18F-FDG-PET/CT can be used in the pretreatment assessment (soft-tissue sarcoma vs. benign soft-tissue mass) and detection of local recurrences (particularly of lesions indeterminable on radiologic tomography) of solid tumors 11.
To date, there are no studies focusing on the role of 18F-FDG-PET/CT in the evaluation of RS recurrence.
The aim of this study was to evaluate the validity of 18F-FDG-PET/CT in comparison with CECT performed within 2 years after treatment, the period with the highest incidence of recurrence (identified as early follow-up), in patients with a previous diagnosis of primary RS.
For both techniques we evaluated the overall sensitivity and specificity and the agreement for each site of relapse.
From January 2010 to March 2011 at our nuclear medicine unit 142 patients underwent 18F-FDG-PET/CT for sarcomas, from whom 57 with retroperitoneal localization of the primitive tumor were selected. Fifteen of them underwent 18F-FDG-PET/CT for initial staging of RS and were excluded, whereas 18 were excluded because of the absence of a previous CECT performed within 30 days.
We retrospectively analyzed 24 patients (19 women and five men; age 23–81 years, mean age 58.4 years) who had undergone 18F-FDG-PET/CT and CECT for early follow-up (within 2 years after therapy) of RS.
The CT scan was performed on average 18 days before the 18F-FDG-PET/CT (range: 1–30 days). Whole-body CT scans were taken for 14 patients; CT scans of the chest and abdomen were taken for six patients, and only abdominal CT scans were taken for four patients (Table 1).
For all patients 18F-FDG-PET/CT was performed at least 45 days (mean 12 months) after surgical resection or during the last cycle of adjuvant chemotherapy (Table 1).
All patients had a primary RS with different histological types: 10 liposarcomas, nine leiomyosarcomas, three unspecified sarcomas, and two schwannomas.
Seventeen patients had previously undergone only surgical therapy; seven underwent surgical resection followed by adjuvant chemotherapy.
18F-FDG-PET/CT and CECT results were compared with the gold standard, established as histological examination, for 10 patients who underwent further surgery and with clinical–instrumental follow-up for at least 12 months (specific and/or nonspecific symptoms, US, CECT, and MRI) for the remaining 14 patients.
All patients had already given their consent for the use of their personal data for clinical research by filling out a form at the nuclear medicine unit.
CT examinations were performed using CTMD up to 16 layers (TSX–101°, Aquilion 16; Toshiba Medical Systems, Tokyo, Japan) using the following acquisition parameters: slice thickness 1 mm, pitch 1.75; increment 0.6 mm, rotation time 0.5 s, and 120 kV/250 mAs.
All examinations included administration of a contrast enhancer.
PET/CT images were acquired with a combined-modality PET/CT (GE Discovery LSA; GE Healthcare, Waukesha, Wisconsin, USA) that integrates a PET scanner (Advance nxI) to a 16-slice CT (Light Speed Plus). Before administration of 18F-FDG all patients were asked to fast for at least 8 h and their capillary blood glucose level was measured (<160 mg/ml). To avoid artifacts caused by the muscle’s physiological uptake, they were instructed to limit any physical activity after radiopharmaceutical administration and before the examination.
Images were acquired 50 min after the intravenous injection of 37 MBq/kg 18F-FDG.
During this time patients were hydrated by giving them 500 ml of water to drink and were allowed to urinate as needed. Muscle relaxant drugs were not administered.
The PET/CT scan was carried out from the external acoustic meatus to the root of the thigh. Patients were positioned on the table in a supine position with their hands above their head.
Five to seven bed positions, according to the height of the patient, were acquired. Patients were instructed to breathe gently during the examination. The CT acquisition parameters were as follows: 340 mA (auto), 120 kV, slice thickness 3.75 mm, tube rotation time 0.8 ms, and collimation field of view 50 cm. The CT images were reconstructed with a filtered back projection. The CT data were used for attenuation correction of PET scanning, which was performed immediately after the acquisition of CT images. The CT scans were obtained without administration of a contrast enhancer. The PET acquisition was obtained in craniocaudal direction, with 3D interactive reconstruction.
Each patient was interviewed by a physician for collecting anamnestic data and obtaining informed consent.
18F-FDG-PET/CT and CECT of the selected patients were evaluated by a nuclear physician and a radiologist, respectively, with at least 4 years of experience.
Both were unaware of the patient’s medical history. In cases of doubt, 18F-FDG-PET/CT images taken by a second nuclear medicine physician were analyzed as well.
CECT was considered positive for recurrence in the retroperitoneum in the case of patients with a description of a retroperitoneal lesion; it was considered positive for lymph node involvement for those with a description of at least one lymph node enlargement in the abdomen and positive for lung and liver involvement for those with a description of at least one lesion in those sites.
The images of 18F-FDG-PET/CT were displayed in three orthogonal planes as PET, CT, and fusion images.
PET/CT images were analyzed for the presence, location, and type of lesion (focal or diffuse uptake).
Abnormal 18F-FDG uptake was defined as a focal area of uptake higher than the background activity in the soft tissue.
18F-FDG-PET/CT was considered positive for recurrence in the retroperitoneum for patients who had a description of a focal area of 18F-FDG uptake in this location; it was considered positive for lymph node involvement in the case of those who had 18F-FDG uptake in at least one abdominal lymph node and positive for lung and liver involvement for those with a description of at least one area of 18F-FDG uptake at those sites.
Regions of interest (ROIs) in PET data were also determined. Circular or elliptic ROIs were placed over the increased uptake areas on transaxial images. Sagittal and coronal image reconstruction was performed to ensure correct ROI placement. The SUVmax for each ROI, expressed in g/ml, was automatically calculated by the workstation software that calculates the ratio between the accumulation of 18F-FDG (MBq/ml) in an area of interest (drawn on the images corrected for attenuation) and the activity administered (MBq) for the weight (kg).
We compared the results of CECT and 18F-FDG-PET/CT for retroperitoneal recurrence with the results of histologic examinations and subsequent clinical–instrumental follow-up, considered as the gold standard, calculating sensitivity and specificity and comparing the performance of the two techniques by the McNemar test.
Negative predictive values (NPVs) and positive predictive values (PPVs) were also calculated.
Moreover, we evaluated the agreement between CECT and 18F-FDG-PET/CT using the positive percentage agreement (PPA) and the negative percentage agreement (NPA) for retroperitoneal recurrence, the two most represented histological types (liposarcomas and leiomyosarcomas), and for metastases to regional lymph nodes, the liver, and lungs.
Ten of 24 patients underwent further surgery for excision of retroperitoneal recurrence. In all these patients, histologic examination on surgical specimens confirmed the presence of sarcoma. In four of 10 patients CECT and 18F-FDG-PET/CT were concordant in detection of RS recurrence (Fig. 1); in two of 10 patients, CECT did not find any lesion that was not found by 18F-FDG-PET/CT; in two of 10 patients, CECT found a retroperitoneal lesion, whereas 18F-FDG-PET/CT did not; in two of 10 patients, both techniques were negative for RS detection.
The remaining 14 of 24 patients did not undergo any subsequent surgery. Two of them showed the onset of recurrent disease: one of them underwent radiotherapy on vertebral recurrence, whereas the other did not undergo any further treatment.
18F-FDG-PET/CT yielded positive results for lymph node involvement in two of 24 patients, both localized on the mesenteric root, one of which was also seen on CECT. CECT identified lymph node involvement in five of 24 patients, which was not detected on 18F-FDG-PET/CT; the lymph node involved was localized on the mesenteric root (2/5 patients), along the gastric lesser curvature (1/5 patients), and along the great retroperitoneal vessels (2/5 patients).
Besides the most common recurrence sites already considered, other sites have been detected with 18F-FDG-PET/CT. It identified recurrences in D6, D8, and S1 vertebrae and in the left wing of the sacrum bone in one patient, concordant with CECT findings, and the patient was started on chemotherapy (Fig. 2).
In another patient whole-body CECT identified skeletal recurrences on the D7 and L2 vertebrae, which did not show 18F-FDG uptake, excluding the necessity to perform subsequent therapy.
The sensitivity of CECT and 18F-FDG-PET/CT is 58.3 and 66.7%, respectively (P=1).
The specificity of CECT and 18F-FDG-PET/CT is 50 and 100%, respectively (P<0.001).
For 18F-FDG-PET/CT, PPV was 100% [95% confidence interval (CI): 67–100%] and NPV was 75% (95% CI: 58–75%); for CECT, PPV was 54% (95% CI: 33–73%) and NPV was 55% (95% CI: 30–78%).
With respect to the four of 24 patients who underwent only abdominal CECT, 18F-FDG-PET/CT was negative in two, it confirmed the lesions identified on CECT in one patient and recognized a pulmonary lesion not seen on CECT in another.
The analysis of PPA and NPA of patients and lesion sites is shown in Table 2.
SUVmax was calculated for the eight patients for whom 18F-FDG-PET/CT was positive for retroperitoneal recurrences and it ranged from 2.6 to 9.5 (average: 6.3); the lesion’s size ranged from 2 to 13 cm.
In the three liposarcoma cases the SUVmax was between 2.6 and 8.7 (average: 4.9), whereas in the four leiomyosarcoma cases the SUVmax was between 5.4 and 9.5 (average: 7.6).
Research has shown that the prognosis of RS patients is related to the possibility of surgeons to completely remove the tumor. Unfortunately, surgery is curative in only a minority of patients, and primary RS has a high rate of recurrence; 43–82% of patients with primary RS developed recurrence after local resection, even when the surgical margins were negative for tumor 5.
It is only when the volume of tumor becomes large enough to be detected by an imaging technique, by direct visualization, or by palpation that a mass lesion becomes clinically evident 8.
Prognosis of RS is determined not only by the histologic type but also by its size and site, particularly if vessels, nerves, and vital structures are involved.
The differential diagnosis of a retroperitoneal tumor and its recurrences includes inflammatory diseases, benign lesions, lymphoma, germ-cell tumors, undifferentiated carcinomas, and metastatic diseases 1.
Most recurrences appear within the first 2 years after therapy, a period that can be defined as ‘early follow-up’ 2,12.
In a review by McGrath 13, up to 75% of first recurrences occurred at the site of the original tumor. Similarly, a study by Van Doorn et al. 14 found local recurrences in 23 of 32 patients with primary RS.
Follow-up imaging is usually performed with CECT with the frequency of follow-up being often chosen on the basis of the completeness of the tumor resection as well as by tumor type and grade 15. One suggested follow-up scheme is to obtain images at regular intervals by CECT or MRI every 3–4 months for 2 years, followed by every 4–6 months for 3–5 years and every 12 months thereafter. Although follow-up beyond 5 years is recommended, most sarcomas (whether of high grade or low grade) recur within 2 years; however, delayed appearance of recurrent disease is not unusual 2.
Radiologists must carefully scrutinize the operative bed and the abdominal cavity in patients from whom an RS has been removed because morphological alterations can be due to postsurgical fibrosis or recurrences 6.
CECT may present pitfalls because of altered fascial planes, fibrosis, scarring, or tissue edema resulting from previous treatment, but if new soft tissue is observed it may also represent recurrent tumor 5,11.
For example, areas of secondary fibrosis may lead to misdiagnosis of a high-grade sarcoma. Because of the risks associated with seeding of the biopsy, the need to repeat diagnostic biopsies may have adverse consequences for patients 16.
Bland et al. 17 found the sensitivity and specificity of CECT for soft-tissue sarcomas to be 73.3 and 87.5%, respectively.
In our study the sensitivity and specificity of CECT in the selected group of patients with recurrence of RS were 58.3% (95% CI: 36–79%) and 50% (95% CI: 27–71%), respectively.
False-positive results are possible even for an expert radiologist and these considerations make 18F-FDG-PET/CT imaging valuable for soft-tissue sarcomas 11.
18F-FDG concentrates in many histotypes and subtypes of sarcoma. One of the most important characteristics of sarcomas is the increased proliferative activity 8.
However, the ability of 18F-FDG-PET/CT to differentiate between benign and malignant lesions has shown varied results across studies 2.
Unlike low-grade soft-tissue sarcomas, benign soft-tissue tumors uniformly showed no 18F-FDG uptake, allowing the differentiation only from higher-grade soft-tissue sarcomas 6. This is consistent with the findings of other authors and may represent a powerful diagnostic tool for the differentiation of primary masses suspicious for soft-tissue sarcoma on CECT 18,19.
A previous study conducted by Swartzbach et al. 11 showed sensitivity of 88% and specificity of 92% for whole-body 18F-FDG-PET/CT in patients with suspected recurrences of sarcomas; however, the study was conducted on 28 patients, and only 10 of them had lesions located in the retroperitoneum.
In our series of 24 patients with RS, the sensitivity and specificity of 18F-FDG-PET/CT were, respectively, 66.7% (95% CI: 44–67%) and 100% (95% CI: 78–100%).
On comparing the performance of the two techniques, sensitivity was found to be higher for 18F-FDG-PET/CT than for CECT (66.7 vs. 58.3%), but the difference was not statistically significant. In contrast, specificity of 18F-FDG-PET/CT was significantly higher than that of CECT (100 vs. 50%).
Following surgery, radiotherapy, and (less often) chemotherapy, scar tissue in the region of the primary tumor can distort normal anatomical structures. Furthermore, scar tissue may coexist with residual sarcoma. A high accumulation of contrast medium in cECT is observed in sarcoma and also in organizing scar tissue. In contrast, low 18F-FDG uptake is seen in chronic scar tissue, whereas most malignant tumors have high cellularity and metabolic activity 9.
In addition, 18F-FDG-PET/CT can detect the most metabolically active portion of a tumor mass and can guide biopsy to a site most likely to contain malignant tissue 9.
18F-FDG-PET/CT resolves the morphological imaging problems after surgery but does not allow differentiation of low-grade RS from phlogosis; for this reason we delay 18F-FDG-PET/CT by at least 45 days from the end of therapy.
As subsequent prognosis in these patients is affected by the ability to completely resect the local recurrences, early detection of tumor recurrence is important. When reresections are performed early, they are successful in up to 90% of patients 4. Detection of early local recurrences can be difficult but their early detection, before the onset of symptoms, is very important to begin a new therapy and increase the survival rate 2.
With respect to retroperitoneal recurrences there is no agreement between 18F-FDG-PET/CT and CECT. The PPA values are 38, 42.8, and 40%, whereas NPA values are 72, 100, and 50%, respectively, for retroperitoneal recurrences (no histotype specified), liposarcomas, and leiomyosarcomas.
In particular, PPA and NPA values are higher for liposarcomas (42.8 and 100%).
In our study, 18F-FDG-PET/CT recognizes lymph node involvement in two patients and TC in six; there was negative agreement between the two techniques (NPA: 94%; PPA: 16.6%).
On CECT, false-negative results at N staging can be due to small positive nodes that were missed, whereas on 18F-FDG-PET/CT false-positive interpretations of the nodes can be the result of reactive lymph nodes. However, findings of combined 18F-FDG-PET/CT and conventional imaging are used to guide sampling of areas of unexpected nodal metastasis 20.
With respect to involvement of the lungs, there is negative agreement between 18F-FDG-PET/CT and CECT (NPA: 100%).
CECT has been shown to be more sensitive than 18F-FDG-PET/CT for the detection of small lung metastases. However, because lung nodules are relatively common in the general population, false-positive results are common on CECT. For lung lesions with no significant 18F-FDG uptake, the likelihood of a metastatic basis is very low, particularly if the primary tumor has or had high 18F-FDG avidity 9.
The number of patients was relatively small, but it should be remembered that the disease is rare, especially in the retroperitoneal site, and prior to our use of 18F-FDG-PET/CT the largest study was performed by Schwarzbach et al. 11 and included 10 of 47 (22%) patients.
Our evaluation has focused on retroperitoneal lesions because of the availability of histological data, but at the same time we considered lymph node, lung, and liver lesions. To overcome the limitation of the lack of histologic sampling for all the lesion sites, our clinical follow-up was extended for another 12 months.
However, despite these limitations, our results are significant and valuable.
Advantages of 18F-FDG-PET/CT consisted of its noninvasive nature and the lack of adverse reactions.
In addition, 18F-FDG-PET/CT allows the evaluation of all body segments using a single dose of radiopharmaceutical, unlike CECT, in which the whole-body examination is only on request with a higher exposure to ionizing radiation at each body segment added.
Our data show higher overall specificity of 18F-FDG-PET/CT compared with CECT in identifying areas of recurrence, demonstrating its validity for early whole-body detection of possible lesions. Integration of 18F-FDG-PET/CT with CECT could improve the follow-up management of RS patients.
Conflicts of interest
There are no conflicts of interest.