Diagnostic efficiency of PET/CT in patients with cancer of unknown primary with brain metastasis as initial manifestation and its impact on overall survival

Carcinoma of unknown primary (CUP) is a diagnostic and therapeutic challenge; correct diagnostic approach results in better outcomes in patients with brain metastasis. Whole body 18 F-FDG PET/CT is a helpful modality to detect the origin of primary tumor and assess metastasis in the rest of the body. We assessed the value of PET/CT in detection of unknown primary in 39 patients with brain metastases as initial presentation and detected its impact on their overall survival. PET/CT could correctly detected the primary tumor in 49% of patients which occurred mainly in the lung, 63% of patients had additional extracerebral metastatic sites mostly in lymph nodes. Specificity, sensitivity, and accuracy of PET/CT in CUP patients were 95%, 79%, and 87%, respectively. No significant difference was noticed in estimated overall survival time between patients with an identified primary tumor and patients with unidentified primary tumor. PET–CT was a helpful non-invasive imaging modality in detecting primary in CUP patients with metastatic brain lesion. It helped in the diagnosis of unexpected extracerebral metastatic lesions. However, the identification of the primary tumor by PET/CT has no significant prognostic effect on patient overall survival.


Background
Cancer of unknown primary tumors (CUP) is determined as known metastatic involvement of an organ without definite primary site in spite of detailed clinical investigation. CUP is considered a diagnostic and therapeutic challenge and true diagnostic approach results in favorable outcomes in these patients [1].
Brain metastases (BM) are frequent and devastating complications that occur in 10-30% of patients with advanced cancer. Most patients experience brain metastasis late during their clinical course [2].
Many reports stated that the detection rate of primary tumor in patients with extracranial metastases as first presentation was 40% and the availability of such information for patients with BM-CUPS is still unclear [3].
Therefore, diagnostic work up in CUP patients gained importance for therapeutic management. PET/CT has been proven to be more accurate than contrastenhanced CT or MRI alone in the identification of the primary tumor site, tumor extent, local-regional, and distant metastases. This helps in selection of a more suitable and site-specific therapy and follow-up associated with a better overall survival and improved treatment outcome [4].
Chest/abdomen CT has been considred as the most valuable diagnostic procedure in patients with BM-CUPS and helps in detecting the primary malignancy in more than 80% of patients [5].
The advantage of PET/CT over the other imaging techniques is its ability to reveal abnormal metabolic activity in normal body structures. Also, primary tumor in CUP patients may be located anywhere through the body. Besides, CUP is often associated with unpredictable metastatic lesion so PET/CT is considered the less sophisticated imaging modality in whole body imaging in comparison to other imaging tools [6].
The main disadvantage of PET/CT in brain scanning is the normal high gray matter uptake of 18F-FDG due to high tissue utilization of glucose that may exceed the expected uptake of many malignant tumors and as a result, the diagnosis of brain malignancy will be difficult. Several techniques have been suggested to avoid this problem, like dual time point imaging or glucose loading. However, these new techniques are time-consuming because they require specific dedicated brain protocols which are difficult to utilize in routine PET scans [7].
The aim of this study was to assess the diagnostic efficiency of PET/CT in detection of the unknown primary tumors in patients with confirmed brain metastases as first manifestation and detect its impact on their overall survive (OS).

Materials
This prospective study included 39 patients with brain metastasis confirmed by histopathological results and/or other imaging modalities. There was no information regarding the primary tumors. The study lasted during the period from March 2017 to June 2020.
Patients were selected in our study after fulfilling the following criteria: age greater than 18 years with normal cardiac functions, adequate liver, and kidney functions without comorbidity.
Exclusion criteria were the following: (1) previously diagnosed primary tumor, (2) previous history of chemotherapy or radiotherapy prior to initial PET/CT scan, (3) patients with incomplete medical records or follow-up, and (4) patients did not undergo surgical management for definite confirmation of site of primary tumor.
All patients gave written informed consent according to ethics committee guidelines.

PET/CT protocol
All patients were scanned with a dedicated PET-CT scanner (Philips, Gemini TF). All patients were fasted for at least 6 h before FDG injection. Fasting blood glucose level of less than 150 mg/dl was a requirement in all patients. The scan started 60 min after intravenous administration of 2.516 MBq (0.068 mCi/kg) FDG and the patient was instructed to rest without talking. CT was performed from the skull base to pelvis by performing a scout view followed by a spiral CT with 80 mA and 140 kVp. No oral contrast was given and water only was used to delineate bowel. Intravenous 130 ml of iohexol (omnipaque 300 mg iodine/ml) was administered. 2D PET emission data on completion of CT (4 min per bed position covering an axial FOV of 15.7 cm with a 3-slice overlap) were obtained. The total scanning time varied between 25 and 30 min for every patient. The CT, PET, and PET/CT images were reconstructed in trans-axial, coronal, and sagittal planes.

Image interpretation of PET/CT scan
Any focal active lesion was detected in brain by PET/CT scan that exceeded the activity of normal surrounding gray matter or has a corresponding lesion in CT scan indicating metastasis. Any abnormal FDG uptake in any tissue or organ rather than the uptake in physiologic sites (e.g. bowel, myocardium) without a specific standardized uptake value (SUV) cut-off was suspected as the primary tumor site. The diagnosis of other extracerebral metastasis was done when intensity of uptake paralleled that in the site of primary tumor [7].

Standard of reference
The site of the primary tumor was confirmed by histolopathological analysis of a surgical biopsy or resection specimen of the tumor (a lesion suspicious of primary tumor that matching histology of BM) and by clinical follow-up, at least for 3 months including repeated physical examination and other imaging methods. A new appearance of a lesion or lesion enlargement on subsequent radiological imaging procedures was another point of confirmation.

Statistical analysis
Fisher's exact test was used to assess the significant differences between the pathological and clinical characteristics of included patients. The overall survival (OS) was measured from the time of resection until death or last follow-up. The survival rates were calculated using the Kaplan-Meier method. Using surgery and histopathological results as gold standard, the sensitivity, specificity, and accuracy of the PET/CT in the diagnosis of primary tumor were assessed. The analyses were performed using SAS version 20 (SAS Inc., Cary, NC).

Demographic and clinical characteristics of included patients
Thirty-nine patients (25 men and 14 women; mean age, 54.45 ± 12.16 years) with diagnostic metastatic brain lesions were included in this study. There was no significant difference between patients with PET/CT lesions regarding gender and age (P = 0.362, P = 0.075). Clinical data and pathologic characteristics of patients are illustrated in Table 1.
The results of PET/CT were truly positive in 19 patients confirmed by histopathological examination; these 19 patients were selected for statistical analysis (Fig. 1).
Twelve patients (63%) had different additional extracerebral metastatic sites mostly lymph nodes metastasis  In this study, we found that PET/CT showed specificity, sensitivity, and accuracy PET/CT in detection of primary tumor as 95%, 79%, and 87%, respectively.

Correlation of PET-CT and overall survival status
During the follow-up period which ranged between 6 and 25 months with a median of 12 months, median OS in patients with identified primary tumor were 12 months versus 13 months in patients with no identified primary tumor with no significant difference (P = 0.217) (Fig. 2).

Discussion
Early determination of primary tumor site, extent of metastatic lesions, and best location for biopsy are critical issues for patients with BM-CUPS [8]. There is only one retrospective study of 16 patients that showed a high capability of PET/CT to localize the primary lesion in patients presented with brain metastasis [9].
This study included 39 patients with brain metastasis confirmed by histopathological analysis and/or imaging modalities with unknown primary tumor. The aim of this study was to assess the diagnostic efficiency of PET/ CT in detection of the unknown primary tumors in patients presented with brain metastases as first manifestation and its impact on survival.
There was no significant difference between patients and PET/CT lesions regarding gender and age (P = 0.362, P = 0.075); the same result was seen in the result of Bakhshayeshkaram et al. who stated that mean age did not show significant different between male (61.10 ± 10.91 years) and female (57.84 ± 13.10) (P = 0.36) [10]. Some reports have stated that in spite of complete diagnostic work up, PET/CT may not add any significant benefits in either detection of occult primary tumor or help in patient's management and the additional value of PE/CT in CUP patients may be overestimated [11]. However, identification of primary tumor is not the only problem in CUP patients. Determination of the lesion extent is not helpful in recognition of the small population of CUP patients with favorable outcome but may also help in optimizing treatment planning including field of radiation therapy and also evaluation of response to treatment [12].
Wolpert et al. investigated the diagnostic value of PET/CT in 64 patients with BM of unknown primary and detected additional lesions suspicious of extracerebral metastases in 27 of 64 patients (42%) mainly in lymph nodes and concluded that PET/CT improves the accuracy of tumor staging by detecting more metastases that resulted in adjustment of the therapeutic strategy [13].
In another retrospective study by Koc et al. which included 31 patients with brain metastasis according to histopathology and/or MRI, they showed that PET/CT detected additionally 13 patients with additional metastatic sites mainly in the lung. They concluded that although the lung was the most frequent primary tumor in patients with BM, there may be unexpected metastatic tumors all over the body [14].
Our results agreed with Wolpert and Bakhshayeshkaram et al. as PET/CT detected additional extracerebral metastatic sites in 12/19 patients (63%) mainly in lymph nodes.
In the same study of Bakhshayeshkaram et al., they found that the lung was the most frequent primary source of brain metastasis in 4.93% patients [15]. The same result was detected by Kung et al. in their study which included 40 patients with BM; they found that  Roh et al. showed the sensitivity of PET/CT (87.5%) was significantly higher than that of CT (43.7%) in detection of the primary tumor in patients with brain metastases [18].
Han et al. in their study which included 162 CUP patients presented mainly with brain metastases concluded that sensitivity, specificity, and accuracy of 18 F-FDG PET/CT was 91.5%, 85.2%, and 88.3%, respectively [19].
In another study by Riaz et al. which included 100 patients, the sensitivity, specificity, and accuracy of 18 F-FDG PET/CT in detection of primary tumor were reported as 80%, 74%, and 78% [20].
In our study, the primary tumor could not be detected in (41%) 16/39 patients. Fifteen patients showed true-negative results and one patient was a false-negative result; the tumor was detected by histopathological analysis to be cancer pancreas and showed low uptake of FDG. Four patients showed false-positive results, two of them were diagnosed falsely by PET/CT to be pulmonary carcinoma but one of them was pulmonary hamartoma and the other one was pulmonary inflammatory pseudotumor. The remaining 2 patients were diagnosed by PET/CT to be cancer colon but by colonoscopic biopsy, one patient was diagnosed as polyp and other one was sigmoid diverticulosis. As a result, we found that PET/CT had specificity, sensitivity, and accuracy in detection of primary tumor as 95%, 79%, and 87%, respectively.
During the follow-up period which ranged between 6 and 25 months with a median of 12 months, median OS in patients with identified primary tumor were (12 months) versus (13 months) in patients with no identified primary tumor with no significant difference (P = 0.217) In a study cohort by Reinert   management in CUP patients, he concluded that there was no significant differences in estimated overall survival time could be noticed between patients with an identified primary tumor and patients with unidentified primary tumor [2]. This result is in concordance with other studies; their explanation was that for many patients with identified primary tumor, no management can be provided that improves their overall survival [21].

Conclusion
From our study, we found that PET/CT was a helpful imaging modality in detecting the localization of primary lesion in majority of the CUP patients with metastatic brain lesion as first manifestation. Most of the patients with unknown primary site was from lung cancer but there were unexpected other extra cerebral metastatic lesions mostly lymph nodes metastasis. However, the identification of the primary tumor by PET/CT has no significant prognostic effect on patient overall survival.