Introduction
Methods
Search strategy
# | Search string | PubMed/MEDLINE | Embase |
---|---|---|---|
1 | Fluorodeoxyglucose OR 2-fluoro-2-deoxy-D-glucose OR FDG OR positron emission tomography OR positron-emission tomography OR PET | 39,616 | 46,549 |
2 | Computed tomography OR computerized tomography OR computed tomographic OR CT OR CAT | 430,940 | 317,859 |
3 | Unknown primary OR unidentified primary OR occult primary OR unknown origin OR unidentified origin | 32,203 | 25,105 |
4 | #1 AND #2 AND #3 | 230 | 185 |
Study selection
Data abstraction
Study quality
Quality item | Positive score |
---|---|
Was the spectrum of patients representative of the patients who will receive the test in practice? | Patients with histologically confirmed metastatic cancer, in whom medical history, physical examination, full blood count, basic biochemistry battery, urinalysis, stool occult blood testing, immunohistochemistry with specific markers as well as imaging technology with chest X-ray, computed tomography of the chest abdomen and pelvis or mammography and MR imaging in certain cases failed to detect the primary tumor, were included |
Were selection criteria clearly described? | It was clear how patients were selected for inclusion |
Is the reference standard likely to correctly classify the target condition? | Histopathological analysis of tissue obtained by biopsy or surgery, or imaging procedures or clinical follow-up if no histopathological proof could be obtained |
Did the whole sample, or a random selection of the sample, receive verification using a reference standard of diagnosis? | All patients, or a random sample of patients who underwent FDG-PET/CT, also underwent the reference standard |
Was the reference standard independent of the index test (i.e., the index test did not form part of the reference standard)? | FDG-PET/CT did not form part of the reference standard |
Was the execution of the index test described in sufficient detail to permit replication of the test? | All of the following parameters were described: |
-Time of fasting before FDG administration, FDG dose, time interval between FDG administration and scanning | |
-Application of intravenous and/or oral CT contrast | |
-Scanned area | |
-Evaluated images (AC and/or nAC) | |
-Interpreter(s) of FDG-PET/CT mentioned | |
Was the execution of the reference standard described in sufficient detail to permit replication? | Besides histopathological analysis of FDG-PET/CT positive findings, additional diagnostic procedures (e.g., gastroscopy, CT, MR imaging) and duration of follow-up were described, if applicable |
Were the index test results interpreted without knowledge of the results of the reference standard? | FDG-PET/CT was interpreted without knowledge of the findings of the reference standard |
Were the same clinical data available when test results were interpreted as would be available when the test is used in practice? | Interpreter(s) of FDG-PET/CT was/were aware of the histologic nature of the metastases of unknown primary |
Were uninterpretable/intermediate test results reported? | All FDG-PET/CT results, including uninterpretable/ indeterminate/intermediate were reported |
Were withdrawals from the study explained? | It is clear what happened to all patients who entered the study |
Was comparator review bias avoided? | Blinding FDG-PET/CT to the other imaging modality, if more than one imaging modality was applied |
Statistical analysis
Results
Literature search
Study and year | Country of origin | Study design | No. of patients | Age in years (mean, range) | Sex (M/F) |
---|---|---|---|---|---|
Fencl et al. [12], 2007 | Czech Republic | Retrospective | 82 | NR | NR |
Nassenstein et al. [13], 2007 | Germany | NR | 39 | 60, 39–89 | 31/8 |
Fleming et al. [14], 2007 | USA | Retrospective | 22 | NR | NR |
Bruna et al. [15], 2007 | France | Retrospective | 37 | 59, 31–85 | 14/23 |
Wartski et al. [16], 2007 | France | Retrospective | 38 | 57, 36–80 | 31/7 |
Ambrosini et al. [18], 2006 | Italy | NR | 38 | 59, 41–77 | 22/16 |
Fakhry et al. [20] 2006 | France | Retrospective | 22 | 48, 43–71 | 17/5 |
Pelosi et al. [22], 2006 | Italy | Retrospective | 68 | 63, 42–79 | 36/32 |
Nanni et al. [26], 2005 | Italy | NR | 21 | 60, 41–87 | 12/9 |
Freudenberg et al. [27], 2005 | Germany | Retrospective | 21 | 64, 46–94 | 16/5 |
Gutzeit et al. [29], 2005 | Germany | Retrospective | 45 | 57, 29–95 | 26/19 |
Study and year | Location of metastases of unknown primary (N) | Histology of metastases of unknown primary (N) | Diagnostic workup before FDG-PET/CT |
---|---|---|---|
Fencl et al. [12], 2007 | -Cervical (20) | -Anaplastic carcinoma (35) | In all patients detailed medical history, full physical and laboratory examinations and diagnostic imaging methods |
-Extracervical (61) | -Adenocarcinoma (24) | ||
-Squamous-cell carcinoma (5) | |||
-Spinocellular carcinoma (7) | |||
-Mucinous carcinoma (10) | |||
-Small-cell carcinoma (1) | |||
Nassenstein et al. [13], 2007 | Cervical (39) | -Squamous-cell carcinoma (27) | In all patients tumor workup including physical examination, ultrasound, chest X-ray as well as complete endoscopic exploration with multiple blind biopsies of the nasopharynx, tonsils and tongue base |
-Adenocarcinoma (5) | |||
-Undifferentiated carcinoma (2) | |||
-Lymphoepithelioid cancer (1) | |||
-Malignant melanoma (1) | |||
-Neuroendocrine cancer (1) | |||
-Papillary carcinoma (1) | |||
-Undifferentiated carcinoma (1) | |||
Fleming et al. [14], 2007 | Cervical and extracervical (22) | NR | NR |
Bruna et al. [15], 2007 | Cervical and extracervical (37) | -Adenocarcinoma (17) | In all but three patients, CT of the neck and thorax; the three patients without a CT of the neck and thorax had at least a chest X-ray. In all patients CT and/or ultrasound of the abdomen and pelvis. In 13/23 females mammography and ultrasound of the breasts, in five females an additional MR of the breasts. In 25 patients, invasive diagnostic tests (endoscopic or surgical), of which 16 were bronchoscopies, 10 upper airway endoscopies, 9 colonoscopies and 8 gastroscopies |
-Squamous-cell carcinoma (14) | |||
- Undifferentiated carcinoma (6) | |||
Wartski et al. [16], 2007 | Cervical (38) | -Squamous-cell carcinoma (32) | In all patients systematic palpation, fiber-optic laryngoscopy and nasopharyngoscopy, CT and/or MR imaging with sections from the skull base to the mediastinum and rigid panendoscopy with randomized biopsies at the most frequent sites of primary tumor |
-Undifferentiated carcinoma (4) | |||
-Mucoepidermoid carcinoma (2) | |||
Ambrosini et al. [18], 2006 | NR | -Adenocarcinoma (13) | In all patients physical examination and negative laboratory and imaging tests; all patients underwent multislice CT and MR imaging |
-Epithelial carcinoma (8) | |||
-Squamous-cell carcinoma (5) | |||
-Mucoid adenocarcinoma (2) | |||
-Poorly differentiated carcinoma (2) | |||
-Undifferentiated adenocarcinoma (2) | |||
-Flat-cell tumor (1) | |||
-Germ-cell tumor (1) | |||
-Melanoma (1) | |||
-Spindle-cell tumor (1) | |||
-Spinous-cell carcinoma (1) | |||
-Transitional-cell carcinoma (1) | |||
Fakhry et al. [20] 2006 | -Cervical (22) | -Squamous-cell carcinoma (22) | In all patients CT and nasofibroscopy |
Pelosi et al. [22], 2006 | -Cervical (18) | -Undefined carcinoma (32) | In all patients physical examination, laboratory tests and conventional diagnostic procedures, i.e., chest X-ray, abdominal contrast enhancement CT and, on the basis of suspected primary disease, chest contrast enhancement CT, MR imaging, ultrasonography, mammography and endoscopic procedures |
-Extracervical (50) | -Adenocarcinoma (18) | ||
-Squamous-cell carcinoma (8) | |||
-Poorly differentiated carcinoma (5) | |||
-Melanoma (4) | |||
-Urothelial-cell carcinoma (1) | |||
Nanni et al. [26], 2005 | -Cervical (3) | -Adenocarcinoma (8) | In all patients physical examination (digital rectal examination with tests for occult blood in the stool, breast palpation and pelvic examination in women, prostate and testicular examination in men) and traditional diagnostic procedures according to international guidelines (complete blood counts, liver and renal function tests, urine analysis, chest radiography, CT and/or MRI of the abdomen and pelvis plus X-ray mammography in women and prostate-specific antigen test in men) |
-Extracervical (17) | -Squamous-cell carcinoma (7) | ||
-Cervical and extracervical (1) | -Poorly differentiated carcinoma (1) | ||
-Melanoma (1) | |||
-Transitional-cell carcinoma (1) | |||
-Germ-cell tumor (1) | |||
-Spindle-cell carcinoma (1) | |||
-Flat-cell tumor (1) | |||
Freudenberg et al. [27], 2005 | Cervical (21) | -Squamous-cell carcinoma (14) | In all patients clinical, endoscopic, sonographic and planar radiological staging (none of the patients had received a dedicated head and neck CT before) |
-Adenocarcinoma (4) | |||
-Undifferentiated malignancy (3) | |||
Gutzeit et al. [29], 2005 | -Cervical (18) | -Adenocarcinoma (25) | In all patients a complete medical history, thorough physical examination and conventional diagnostic strategies (including comprehensive laboratory analysis, projectional and cross-sectional imaging and endoscopic procedures where indicated) |
-Extracervical (27) | -Squamous-cell carcinoma (15) | ||
-Undifferentiated carcinoma (5) |
Study and year | Time of fasting before FDG administration (h) | FDG dose (MBq) | Time interval between FDG administration and data acquisition (min) | No. of CT detector rows | Reconstructed slice width (mm) | Intravenous CT contrast | Oral CT contrast | Area of body scanned |
---|---|---|---|---|---|---|---|---|
Fencl et al. [12], 2007 | ≥6 | 350–450 | 60–90 | 2 | 4.0 | Yes and no | Yes | From the skull base to below the groin |
Nassenstein et al. [13], 2007 | ≥4 | 350 | 60 | 2 | -3.0 (head and neck) | Yes | Yes | From the head to the upper thigh |
-5.0 (chest and abdomen) | ||||||||
Fleming et al. [14], 2007 | NR | 555 | 75 | 16 | NR | NR | NR | From the top of the head to midthigh |
Bruna et al. [15], 2007 | ≥6 | 5.5/kg (max. 550) | 60 | NR | NR | NR | NR | From the top of the head to midthigh |
Wartski et al. [16], 2007 | ≥6 | 4–5/kg | 60 | NR | 5.0 | No | No | From the skull to the midthigh |
Ambrosini et al. [18], 2006 | 6 | 370 | 60–90 | NR | NR | NR | NR | “Whole-body” |
Fakhry et al. [20] 2006 | ≥6 | 260–330 | 60 | NR | NR | NR | NR | From the skull base to the thighs |
Pelosi et al. [22], 2006 | ≥6 | 222–370 | 60 | NR | NR | NR | NR | From neck to pelvis or from skull to feet |
Nanni et al. [26], 2005 | ≥6 | 370 | 60–90 | NR | NR | NR | NR | “Total body scan” |
Freudenberg et al. [27], 2005 | ≥10 | 360 | 60 | 2 | -3.0 (head and neck) | -No (head and neck | -No (head and neck) | Head, neck, thorax, abdomen and pelvis |
-5.0 (from thorax to pelvis) | -Yes (thorax to pelvis) | -Yes (thorax to pelvis) | ||||||
Gutzeit et al. [29], 2005 | ≥4 | 350 | 60 | 2 | 5.0 | Yes | Yes | Head, neck, thorax, abdomen and pelvis |
Study and year | Evaluated images | Interpreter(s) of FDG-PET/CT | Criteria for positivity | Reference standard |
---|---|---|---|---|
Fencl et al. [12], 2007 | AC and nAC | Seven physicians experienced in both PET and CT reading were randomly involved in routine evaluation of findings; in the event of any uncertainty, a second or even a third opinion was solicited | FDG hypermetabolism at the site of pathological changes on CT or marked focal hypermetabolism at sites suggestive of malignancy (liver parenchyma, bone marrow) despite absence of signs of pathology at those sites on CT | A diagnosis of the primary site of a malignancy was classified as true positive only when it was confirmed histologically. If the finding was not confirmed histologically, the diagnosis was classified as false positive. An evaluation was classified as true negative if neither FDG-PET/CT nor histological findings or clinical follow-up (including subsequent imaging tests) determined the site of the primary. When the site of the primary was not identified, but was proven histologically, the finding was classified as false negative |
Nassenstein et al. [13], 2007 | NR | Two different reading teams, each consisting of a radiologist and a nuclear medicine physician | Contrast-enhancing mass on CT or focally increased FDG uptake on PET | “Full medical history was available for all patients” |
Fleming et al. [14], 2007 | NR | One of three neuroradiologists | An SUV level greater than 2.5 was considered consistent with abnormal, hypermetabolic activity in primary, regional and distant disease | Each site of increased PET metabolic activity was compared with operative histopathology records |
Bruna et al. [15], 2007 | NR | NR | NR | -“Follow-up” |
-Histology (n = 7) | ||||
-Complimentary examination (n = 3), among which were CT of the abdomen and pelvis (n = 2) | ||||
Wartski et al. [16], 2007 | AC and nAC | Two experienced nuclear medicine physicians, independently | Increased FDG focal uptake indicative of a primary tumor in the head and neck and/or chest regions | FDG-PET/CT results were correlated to the patient’s medical record concerning pathological results and treatment. A FDG-PET/CT result was considered as a true positive when an FDG focus matched the primary tumor found during the second panendoscopy, a false positive when the increased FDG focal uptake did not match panendoscopy results and a false negative when the second panendoscopy detected malignant lesions with no corresponding increased FDG focal uptake |
Ambrosini et al. [18], 2006 | NR | Three nuclear medicine physicians in consensus | NR | -PET/CT findings were subsequently confirmed by surgery or biopsy of the primary tumor |
-Gastroscopy and 3-month follow-up in one patient | ||||
Fakhry et al. [20] 2006 | NR | Two nuclear medicine physicians | NR (visual interpretation) | Histology and/or clinical follow-up >6 months in all patients |
Pelosi et al. [22], 2006 | NR | Two nuclear medicine physicians in consensus | NR | The FDG pathological findings, suspected for primaries, were further investigated with other imaging examination, biopsy and/or surgery and clinical follow-up (minimum follow-up of 3 months after the FDG-PET/CT study) |
Nanni et al. [26], 2005 | NR | Three skilled nuclear medicine physicians; in case of discrepancy, the FDG-PET/CT interpretation was reached by consensus | NR | -FDG-PET/CT findings were subsequently confirmed by surgery or biopsy of the primary tumor |
-Gastroscopy in one patient | ||||
Freudenberg et al. [27], 2005 | NR | Two experienced nuclear medicine physicians in consensus (FDG-PET) and two radiologists (CT) | -FDG-PET: regions of focally increased tracer uptake (a maximum SUV of >2.5 was considered to represent malignancy in otherwise equivocal findings) | Histopathology (n = 14) and clinical follow-up ≥9 months (n = 7) with subsequent panendoscopy with biopsy of the most probable tumor sites (n = 7), ultrasound (n = 7), CT (n = 6), MRI (n = 6), diagnostic tonsillectomy (n = 4) and additional biopsies (n = 4) |
-CT: contrast-enhancing masses or asymmetries typical of malignancies | ||||
Gutzeit et al. [29], 2005 | AC and nAC | A nuclear medicine physician and a radiologist, both with 2 years of PET/CT experience | Contrast material-enhanced mass on CT or focally increased glucose metabolism with a SUV exceeding 2.5 on FDG-PET | -All potential sites of the primary tumor depicted by FDG-PET/CT were histologically verified |
-Axillary lymph node dissection in one FDG-PET/CT-negative patient | ||||
-Endoscopy and biopsy of the esophagus in one FDG-PET/CT-negative patient |
Methodological quality assessment
Study and year | Quality items | % of maximum score | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | ||
Fencl et al. [12], 2007 | - | + | + | + | + | + | - | - | - | + | + | + | 67 |
Nassenstein et al. [13], 2007 | - | - | - | + | - | + | - | - | - | + | + | + | 42 |
Fleming et al. [14], 2007 | - | + | - | - | + | - | - | - | - | + | + | + | 42 |
Bruna et al. [15], 2007 | - | + | + | + | - | - | - | - | - | + | + | + | 50 |
Wartski et al. [16], 2007 | - | - | - | + | + | + | + | + | - | + | + | + | 67 |
Ambrosini et al. [18], 2006 | - | - | - | - | - | - | + | - | + | + | + | + | 42 |
Fakhry et al. [20] 2006 | - | + | + | + | - | - | - | - | - | + | + | + | 50 |
Pelosi et al. [22], 2006 | + | - | + | + | - | - | - | - | - |
+
| + | + | 50 |
Nanni et al. [26], 2005 | + | - | + | - | + | - | + | + | + |
+
| + | + | 75 |
Freudenberg et al. [27], 2005 | - | - | + | + | + | + | + | - | - |
+
| + | + | 67 |
Gutzeit et al. [29], 2005 | - | - | - | - | + | + | + | + | - |
+
| + | + | 58 |
Diagnostic performance
Study and year | Primary tumor detection rate (%) | Sensitivity (%) | Specificity (%) | ||
---|---|---|---|---|---|
Value | 95% CI | Value | 95% CI | ||
Fencl et al. [12], 2007 | 22 | 55 | 38–70 | 75 | 62–85 |
Nassenstein et al. [13], 2007 | 28 | 100 | 74–100 | 85 | 69–94 |
Fleming et al. [14], 2007 | 73 | 94 | 73–99 | 100 | 61–100 |
Bruna et al. [15], 2007 | 38 | 93 | 70–99 | 77 | 57–90 |
Wartski et al. [16], 2007 | 34 | 93 | 69–99 | 73 | 48–89 |
Ambrosini et al. [18], 2006 | 53 | 100 | 84–100 | 95 | 76–99 |
Fakhry et al. [20] 2006 | 32 | 70 | 40–89 | 75 | 47–91 |
Pelosi et al. [22], 2006 | 35 | 83 | 66–93 | 87 | 73–94 |
Nanni et al. [26], 2005 | 57 | 100 | 76–100 | 89 | 57–98 |
Freudenberg et al. [27], 2005 | 57 | 86 | 60–96 | 100 | 65–100 |
Gutzeit et al. [29], 2005 | 33 | 88 | 66–97 | 89 | 73–96 |
Pooled estimate | 37 | 84 | 78–88 | 84 | 78–89 |
Parameter | Value | No. of studies | Relative diagnostic odds ratio (1 vs. 2) | ||
---|---|---|---|---|---|
Value | 95% CI | P-value | |||
Completeness of diagnostic workup before FDG-PET/CT | 1. Complete | 2 | 1.93 | 0.22–17.28 | 0.5072 |
2. Incomplete | 9 | ||||
Location of metastases | 1. Cervical | 6 | 0.38 | 0.02–9.55 | 0.4765 |
2. Extracervical | 2 | ||||
Administration of CT contrast agents | 1. Intravenous and oral | 3 | 2.42 | 0.32–18.15 | 0.3347 |
2. Not reported | 7 | ||||
Evaluated FDG-PET/CT images | 1. AC and nAC | 3 | 0.36 | 0.06–2.09 | 0.2187 |
2. AC only or NR | 8 | ||||
Way of FDG-PET/CT review | 1. Reported blinding | 3 | 1.18 | 0.10–13.54 | 0.8766 |
2. No reported blinding | 8 |