Carcinoma of an unknown primary (CUP) is a proven tumor metastasis without evidence of a primary site. The management of patients with CUP represents a difficult challenge in oncology, as histopathological analysis frequently does not provide sufficient information to locate the primary site, and not all primary tumors are identified despite a comprehensive diagnostic workup.
About 2–5% of all newly diagnosed cancers are said to be CUP [1,2]. The site of the primary tumor is often detected in only 10–35% of all cases after conventional imaging modalities [3,4]. Even after autopsy, the primary site can remain unidentified . The low rate of detection has been attributed to both a lesion size that is smaller than the spatial and contrast resolution of the technique used, and the involution of the primary mass as a result of limited angiogenic competence . Identification of the primary tumor has been shown to impact patient management and thereby life expectancy in patients with CUP [7–9].
Positron emission tomography/computed tomography (PET/CT) has provided new insights in the diagnosis, staging, and follow-up of oncological patients [10–12]. The contribution of 18F-fluorodeoxyglucose (FDG) PET/CT to the diagnosis of CUP has been investigated by various authors [13–17]. In this era, the technique has been shown to be more helpful in searching for the presence of malignancy than in identifying exactly the site of the primary by Fencl et al. . The same group also showed that FDG PET/CT imaging offers accurate prognostification as additional information.
This study aimed to investigate the value of FDG PET/CT in clarifying the primary site in our patient population with proven tumor metastasis or with a high clinical suspicion of malignancy, and the clinical impact of this technique on the management of the patients of whom the primaries could not be found using all available tests.
Materials and methods
A total of 94 patients (66 male, 34 female; age range: 22–84 years; mean ages: 52.74±13.34, 57.06±13.56 years, respectively) from two nuclear medicine centers (Medical Faculties from Cukurova University and Baskent University Adana Hospital) were included in the study. The CUP attribute was assigned to patients in two categories as 78 patients with histologically proven metastatic disease (group I), 16 patients with clinical suspicion of malignancy (group II). In the second group, medical history and laboratory examinations had suggested a high suspicion for malignancy, which was later proved histologically. The histological types of primaries found in these patients are summarized in Table 1. The patients were referred for FDG PET/CT imaging either before or after conventional imaging methods with no suspected primary site.
The histology and/or follow-up data were used as gold standards. This study was reviewed and approved by the ethics review committee of our institute.
Fluorodeoxyglucose positron emission tomography/computed tomography imaging
Data acquisition was performed using two different PET/CT scanners (Biograph 6 LSO PET/CT scanner by Siemens, Knoxville, Tennessee, USA, and Discovery STE 8 PET/CT scanner by General Electric, Milwaukee, Wisconsin, USA). Patients fasted for at least 6 h before scanning to minimize blood insulin levels and glucose utilization of normal tissue. Blood glucose level was checked in each patient before the FDG injection and all patients presented glycemia blood levels below 200 mg/dl. Water-soluble iodinated contrast material diluted in 1500 ml water was given to each patient orally before scanning. Patients were positioned on the table in a supine head-first position and whole-body scanning from the top, including the skull, to the upper thighs was carried out 60–90 min after intravenous injection of FDG (dose range: 222–370 MBq). The parameters used for CT investigation were as follows: effective current, 54 mAs; 130 kV; slice width, 4.0 mm; collimation, 10 mm; caudocranial table feed, 14 mm/rotation for Biograph 6 LSO PET/CT scanner, and effective current, 80 mAs; 140 kV; slice width, 3.75 mm; collimation, 10 mm; caudocranial table feed, 16.75 mm/rotation for the Discovery STE 8 PET/CT scanner. Immediately after the CT examination, the process continued with a three-dimensional PET acquisition in seven bed positions (3 min/bed position). If necessary, two-dimensional PET acquisition was also applied with the Discovery STE. PET images were reconstructed using CT data for attenuation correction with an iterative approach.
Whole-body PET/CT images were evaluated by two nuclear medicine physicians by consensus. Foci showing increased FDG uptake were evaluated by comparison of the background and blood pool activity and the presence of morphological alterations in CT images. Fused FDG PET/CT images were analyzed in at least three planes (coronary, sagittal, and axial) in the hot body color table for PET, and with appropriate settings of center/width for CT.
The analysis of malignant involvement was based on qualitative visual interpretation of the images. The criterion for malignancy was FDG hypermetabolism at the site of pathological changes on CT or marked focal hypermetabolism at physiological uptake sites such as the liver and the bone marrow despite absence of signs of pathology on CT. To obtain the pretherapeutic baseline metabolic status of the lesions, the maximal standardized uptake value (SUVmax) was calculated for all lesions, and these data were used only for the evaluation of chemotherapy response in seven patients as follows: progressive metabolic disease: greater than 25% increase in SUV or new lesions; stable metabolic disease: between less than 25% SUV increase and less than 15% SUV decrease; complete metabolic response: complete resolution of FDG uptake. To identify the primary site, the distribution of pathological lesions, a prior knowledge of the pattern of spread of different tumors, and the patient's history were taken into consideration. When any lesion as the site of the primary in cases with many foci in different organs was not identified, we interpreted the changes as a sign that the CUP syndrome was generalized, with no known site of primary.
The diagnosis of the primary site of malignancy was classified as true positive (TP) only when it was confirmed histologically. If the finding was not confirmed histologically, the diagnosis was classified as false positive (FP). Evaluation was classified as true negative (TN), despite the fact that FDG PET/CT imaging revealed multiple hypermetabolic lesions for which no primary site could have been differentiated if neither FDG PET/CT nor histological findings or clinical follow-up (including subsequent imaging tests) determined the primary site . The finding was classified as false negative (FN), when the site of the primary was not identified, but was proven histologically. In three patients, the primary malignancies were missed and another falsely diagnosed; for statistical reasons, these lesions were regarded as being both FP and FN findings, and the group was regarded as having 93 patients in total.
During overall survival analysis, one patient with a primary site was excluded because of insufficient follow-up. Furthermore, during survival analysis based on disease extension on PET/CT, patients with normal findings were excluded because no death was observed in this group.
The cases that remained primary unknown and were monitored with FDG PET/CT during follow-up were evaluated one by one, because a sufficient number of patients could not be reached for statistical analysis.
TP/FPs and TN/FNs were defined based on the results of the above-mentioned confirmation methods and the sensitivity and specificity values of FDG PET/CT were calculated. The Kaplan–Meier test was used for survival analysis. The log-rank test was applied for comparison of the survivors.
Among the total study population, 74 (82%) patients had histologically proven metastasis (HPM), 16 (18%) had a high clinical suspicion of malignancy (CSM) (Table 1). In the second group, the primaries were proven histologically in all patients except one. Although the primary could not be located in one patient, FDG PET/CT revealed generalized multiple skeletal lesions, which were later proven to be metastatic lesions histologically.
The histopathological and localization profiles of the metastases in the patients with HPM are shown in Table 2. Of these, 30 of 74 (41%) metastases were localized in lymph nodes, of which most were in the cervical region (21 of 30, 70%). Most of the metastases were of epithelial origin (55 of 74, 74%). The histological subgroup of the malignancy in metastatic lesions could not be differentiated in 22% of the patients (16 of 74).
Among 90 patients, five (6%) were normal on FDG PET/CT. In 85 patients (94%) with pathological findings on FDG PET/CT, 27 patients (32%) had solitary and 58 (68%) patients had multiple organs that were affected. Primary malignancy was discovered in 53 of 90 patients (59%) histologically. FDG PET/CT was able to reach primaries in 12 of 27 (44%) patients with single site and in 39 of 58 (67%) patients with multiple site involvement. In one patient, a second primary tumor was also identified. During the follow-up period, 37 patients (41%) remained primary unknown (RPU).
Accuracy of FDG PET/CT in discovering the primary site regarding the whole patient population (n=90)
In the identification of a primary site, the findings of FDG PET/CT were concordant with those of the gold standard in 70 of 90 patients (78%) with regard to all the study population. Of these, 31 were TN and 39 were TP. In 20 of 90 patients (22%), the findings of FDG PET/CT were discordant with those of the gold standard (Tables 3 and 4). In three patients, the localization of the primaries, which FDG PET/CT suggested, was discordant with the correct sites found by the gold standard. These results were regarded as both FP and FN, and the number of patients was accepted as 93 during the calculation of sensitivity and specificity because of statistical considerations. Thus, FN and FP cases were 14 and 9, respectively. Finally, a sensitivity of 74% and a specificity of 78% were calculated. Positive predictive value (PPV) and negative predictive value (NPV) were 81 and 69%, respectively.
Accuracy of FDG PET/CT in discovering the primary site of the patients with histologically proven tumor metastasis (n = 74)
When the results of the other group (the patients with CSM) were excluded, the final values that were found were as follows: TP=32, TN=30, FP=7, FN=6. Accordingly, the calculated sensitivity, specificity, PPV, and NPV values were 84, 81, 82 and 83%, respectively.
FDG PET/CT performance in patients with clinical suspicion of malignancy (n=16)
Among 16 patients with CSM, FDG PET/CT revealed a solitary organ involvement in seven patients and multiple organ involvement in eight patients. Of these, eight of 14 primaries were correctly identified (TP and TN), whereas six were missing (FN). Primary malignancies were missing in two patients and were falsely diagnosed in another, and these lesions were regarded as being both FP and FN findings. In one patient, there were multiple generalized skeletal system lesions, which were diagnosed as metastatic lesions histologically. The result of FDG PET/CT was regarded as TN for this patient, because the primary could not be found by the gold standard either. There was one normal FDG PET/CT study considered as FN, because this patient was later diagnosed as having chronic lymphocytic leukemia (CLL). Although the number of these patients was not sufficient for individual statistics, calculated sensitivity, specificity, PPV, and NPV values were 47, 33, 78 and 11%, respectively (TP=7, TN=1, FP=2, FN=8).
Survival analysis (n=90)
According to our findings, the identification of the primary site did not change life expectancy in this clinical setting (log rank, Mantel–Cox), because there was no statistical difference between the mean survival times of the patients for whom primaries were reached (PR) and the RPU patients (P = 0.232). In our patient population, the ratios of PR and RUP patients were 59% (53 of 90) and 41% (37 of 90), respectively. The death rate was 32% (17 of 53) for the group of PR and 43% (16 of 37) for the group of RPU. The mean survival times were 17.75±2.21 months for the patients of PR and 16.31±1.95 months for RPU patients When the patients were evaluated on the basis of the extension of disease on FDG PET/CT, their mean survival time with disseminated disease was significantly shorter than that of patients with single or no lesion (13.44±1.61, 20.98±2.0 and 26.67±2.73 months, respectively, P=0.014).
Considering the group including RPU patients, the mean survival time of the patients with multiple system involvement on PET/CT was significantly shorter than that of the patients with single or no lesion (13.95±1.69 and 20.69±2.33, 26.67±2.73, respectively; P=0.041) (Fig. 1). According to the PET/CT findings, 51% (19 of 37) of these patients revealed generalized disease with multiple organ involvement, 40.5% (15 of 37) had single organ involvement, and 8% (three of 37) were normal. Among RPU patients, 43% (16 of 37) of the patients (12 of 19 generalized disease, three of 15 single organ involvement, one of three normal) died with a mean survival time of 17.75±2.21 months (mean±SE). In the surviving group, 33% (7 of 21) of the patients revealed generalized disease with multiple organ involvement, 57% (12 of 21) had single organ involvement, and 10% (2 of 21) were normal. The mean survival time of these patients was 16.31±1.95 months.
Follow-up data of patients with remained primary unknown
Among 21 surviving patients during follow-up, two patients had normal findings on FDG PET/CT, whereas 19 patients showed pathological findings. Two of the normal patients were female and metastatic lesions were detected from the axillary lymph node in one and T12 vertebra in the other (histopathological diagnoses were metastatic adenocarcinoma and malign epithelial tumor, respectively). These two patients have been alive without any symptom for 26 and 19 months, respectively.
Among 19 patients with pathological findings, 12 had single and seven had multiple organ involvement. Four solitary lesions were evaluated as suspicious for malignancy (SUVmax=3–3.5). Of these, one patient has been alive for 26 months without any symptoms and there were no findings from various diagnostic tests. One patient has been alive for 25 months with bone metastasis detected in the last year with magnetic resonance imaging (MRI) from the thoracic vertebra. The other two patients have been relatively newly diagnosed and have been followed for 8 and 3 months.
In eight patients, including two with single and six with multiple organ involvement, FDG PET/CT examination was again performed to monitor the disease (Table 5). One patient underwent FDG PET/CT to clear some symptoms and a bone metastasis was missed, which was reached by MRI later. All the others were studied to search for the response to therapy. A complete response was achieved in three patients, of whom one relapsed 6 months later from the last normal FDG PET/CT examination. In the other four patients, of whom two had progressive disease and two had stable disease, chemotherapy regimes were modified.
In this study, including 90 patients suffering from the CUP syndrome, 18F-FDG PET/CT was able to identify 39 sites among 53 primaries confirmed by the gold standard. Our primary detection rate of 43% seems comparable with the report by Seve et al. , in which the 10 previous reports were reviewed showing detection rates ranging from 24 to 63%. As found in this review, the most common localization site of the primary tumor was the lung, with a ratio of 43%, with disseminated findings on FDG PET/CT in our study.
All but one lung primary were correctly identified with their metastases by FDG PET/CT. There were three patients in whom the FDG PET/CT directed the diagnosis of the primaries to the incorrect organs. All of them had disseminated disease on PET/CT and the incorrect organs that were suggested as the primaries were liver, lung, and colon. The correct localization sites were discovered to be lung, bone marrow (CLL), and ovarium, respectively.
Among the other 11 patients with FN results, the FDG distribution in the whole body was exclusively within normal limits in two patients, one with CLL and the other with myeloproliferative disease. The other types of malignancy missed on FDG PET/CT images were prostatic cancer, neurofibroma, testicular cancer, cutaneous epidermoid cancer, multiple myeloma, low-grade NHL, and papillary thyroid cancer. There were two patients in this group for whom the primaries were diagnosed by the gold standard as prostatic cancer in one and testicular cancer in the other. They had multiple hypermetabolic lymphoid masses in all the lymphoid compartments over the whole body from inguinal to the cervical region without any sign with regard to the other organs. When we retrospectively re-analyzed these organs of the primary sites, we could not find any pathological uptake except physiological FDG distribution.
In a fairly large series by Fencl et al. , overall sensitivity and specificity values in detecting the primary site were reported as 62 and 82%, respectively. When they divided the whole population into two groups as patients with HPM and CSM, the sensitivity and specificity were calculated as 54.5 and 75% for the first group and 76 and 86%, respectively, for the second group. In our study, although most of the patients had been studied by previous conventional imaging methods, some patients were initially studied with FDG PET/CT. For this reason, the overall sensitivity calculated in our study (74%) was rather higher than that found by Fencl et al. The overall specificity (78%) was found to be comparable. In contrast with their results, when the same calculations were carried out with regard to patients with HPM only, the sensitivity and specificity were relatively increased to 84 and 81%, respectively. As the ratio of the tumor types that could be easily missed by FDG PET/CT was rather higher in the group of patients with CSM in our study (Table 1) the value of a negative finding in this group was significantly lower (10%).
In the meta-analysis by Seve et al.  from 10 earlier reports, the sensitivity and specificity were calculated as 92 and 82%, respectively. Although the specificity values are comparable, the sensitivity is higher than that in our study. The reviewers of the meta-analysis commented that patients that were included in the PET studies were selected and did not reflect the overall population of patients with CUP, because 91% of the meta-analysis patients had only a single metastatic site, whereas the rate in the literature ranged from 29 to 43%. In accordance with this observation, the ratio of patients with single-site metastasis among all the groups (35%) was found to be within these limits in our study.
FDG PET/CT has also been shown to be useful for detecting second primary cancers . In our series, a second primary cancer was confirmed in one patient when we were searching for the primary tumor with HPM in the adrenal gland. In this patient, not only did we locate the primary tumor in the lung, but, after FDG PET/CT imaging, we identified a second primary tumor in the colon. The second primary could be proved in only one patient, because an invasive application, such as biopsy, was not accepted by most of the patients in whom a primary site had already been reached and who had generalized disease.
Owing to the differences in patient selection criteria and natural heterogeneity between the groups, it was not possible to make direct comparisons with previous studies. Our experience supports the finding that FDG PET/CT is a useful tool for identifying the primary tumor site in addition to the metastatic sites involved. However, the literature reveals some doubts regarding whether the diagnostic tests, including FDG PET with or without CT, make any sense in the diagnosis of CUP, which looked like a sleuth in the investigation of a sinister crime . From this perspective, we focused on the prognostic value of the test in addition to the value in identifying the primary site in our patient population. The identification of the primary site did not change life expectancy in survival analysis when the mean survival of the patients with PR and RPU was compared. However, the mean survival time of the patients with multiple lesions was found to be significantly shorter than that of patients with single/no lesion on FDG PET/CT with regard to all the study population. When the same analysis was carried out for only RPU patients, the finding did not change and suggested that sensitive staging plays an important role in determining life expectancy even in the patients with RPU. This finding was supported by the previous report by Fencl et al. , who showed a shorter life expectancy in patients with a positive FDG PET finding in comparison with patients with a negative FDG PET finding. The investigators concluded that the presence of a hypermetabolic lesion and the increase in the number of involved organs are sensitive prognostic indicators of a shorter life expectancy in patients with CUP.
In this study, FDG PET/CT was also used to monitor the chemotherapy response in some patients with RPU. Although the number of these patients was not large enough to draw a conclusion in statistical means, the test effectively showed complete remission in three patients. Among them, relapse was diagnosed in one patient during follow-up. In four patients, FDG PET/CT revealed unchanged or increased findings after chemotherapy and their therapies were modified. These findings showed that the test has a potential in therapy monitoring in patients with RPU and effectively identifies the primary sites and gives the opportunity to optimize prognosis.
According to our findings, whole-body FDG PET/CT has to be considered as a useful method in CUP syndrome. In the management of these patients, the method offers several advantages such as (i) identification of the primary tumor in nearly half of the patients, (ii) optimal staging and thereby an opportunity to give a prognosis even when the primary could not be found, and (iii) identification of chemotherapy response. The role of the test seems important, especially in monitoring the chemotherapy response considering the metastatic state in most of these patients. Our preliminary results showed that FDG PET/CT could potentially detect therapeutic efficacy. Pretherapeutic scanning could also be useful in evaluating the response after therapy, in addition to helping in the decision of the therapeutic approach according to the findings with regard to the primary site and metastatic state.
These results suggest that FDG PET/CT can be used reliably in an early phase of the diagnostic workup of the patients with CUP syndrome to optimize their management.
The authors thank all the members of the related departments from two centers for their contributions to the routine workup of nuclear medicine, medical oncology, and pathology divisions.
1. Muir C. Cancer of unknown primary site. Cancer 1995; 75:353–356.
2. Levi F, Te VC, Erler G, Randimbison L, La Vecchia C. Epidemiology of unknown primary tumors. Eur J Cancer 2002; 38:1810–1812.
3. Abbruzzese JL, Abbruzzese MC, Lenzi R, Hess KR, Raber MN. Analysis of a diagnostic strategy for patients with suspected tumors of unknown origin. J Clin Oncol 1995; 13:2094–2103.
4. Chorost MI, Lee MC, Yeoh CB, Molina M, Ghosh BC. Unknown primary. J Surg Oncol 2004; 87:191–203.
5. Le Chevalier T, Cvitkovic E, Caille P, Harvey J, Contesso G, Spielmann M, Rouesse J. Early metastatic cancer of unknown primary origin at presentation. Arch Inten Med 1988; 148:2035–2039.
6. Naresh KN. Do metastatic tumors fom an unknown primary reflect angiogenic incompetence of the tumor at the primary site? A hypothesis. Med Hypothesis 2002; 59:357–360.
7. Haas I, Hoffmann TK, Engers R, Ganzer U. Diagnostic strategies in cervical carcinoma of an unknown primary (CUP). Eur Arch Otorhinolaryngol 2002; 259:325–333.
8. Issing WJ, Taleban B, Tauber S. Diagnosis and management of carcinoma of unknown primary in the head and neck. Eur Arch Otorhinolaryngol 2003; 260:436–443.
9. Pavlidis N, Fizazi K. Cancer of unknown primary (CUP). Crit Rev Oncol Hematol 2005; 54:243–250.
10. Townsend DW. A combined PET/CT scanner: the choices. J Nucl Med 2001; 42:533–534.
11. Hany TF, Steinert HC, Goerres GW, Buck A, Schulthess GK. PET diagnostic accuracy: improvement with in-line PET-CT systems-initial results. Radiology 2002; 225:575–581.
12. Bar-Shalom R, Yefremov N, Guralnik L, Gaitini D, Frenkel A, Kuten A, et al. Clinical performance of PET/CT in evaluation of cancer: additional value for diagnostic imaging and patient management. J Nucl Med 2003; 44:1200–1209.
13. Gutzeit A, Antoch G, Kühl H, Egelhof T, Fischer M, Hauth E, et al. Unknown primary tumors: detection with dual-modality PET/CT-initial experience. Radiology 2005; 234:227–234.
14. Freudenberg LS, Fischer M, Antoch G, Jentzen W, Gutzeit A, Rosenbaum SJ, et al. Dual modality of 18F-fluorodeoxyglucose-positron emission tomography/computed tomography in patients with cervical carcinoma of unknown primary. Med Princ Pract 2005; 14:155–160.
15. Ambrosini V, Nanni C, Rubello D, Moretti A, Battista G, Castellucci P, et al. 18F-FDG PET/CT in the assessment of carcinoma of unknown primary origin. Radiol Med 2006; 111:1146–1155.
16. Pelosi E, Pennone M, Deandreis D, Douroukas A, Mancini M, Bisi G. Role of whole body positron emission tomography/computed tomography scan with 18F-fluorodeoxyglucose in patients with biopsy proven metastases from unknown primary site. Q J Nucl Med Mol Imag 2006; 50:15–22.
17. Paul SAM, Stoeckli SJ, von Schulthess GK, Goerres GW. FDG PET and PET/CT for the detection of the primary tumour in patients with cervical non-squamous cell carcinoma metastasis of an unknown primary. Eur Arch Otorhinolaryngol 2007; 264:189–195.
18. Fencl P, Belohlavek O, Skopalova M, Jaruskova M, Kanturova I, Simonova K. Prognostic and diagnostic accuracy of [18F]FDG-PET/CT in 190 patients with carcinoma of unknown primary. Eur J Nucl Med Mol Imag 2007; 34:1783–1792.
19. Seve P, Billotey C, Broussolle C, Dumontet C, Mackey JR. The role of 2-deoxy-2-(F-18)fluoro-D-glucose positron emission tomography in disseminated carcinoma of unknown primary site. Cancer 2007; 109:292–299.
20. Kim Sy, Roh JL, Yeo NK, Kim JS, Lee JH, Choi SH, Nam SY. Combined 18F-fluorodeoxyglucose-positron emission tomography and computed tomography as a primary sceening method for detecting second primary cancers and distant metastases in patients with head and neck cancer. Ann Oncol 2007; 18:1698–1703.
21. Dowden S, Miller W. New look at the old management of unknown primary. J Surg Onc 2008; 98:71–72.
© 2010 Lippincott Williams & Wilkins, Inc.