Impact of ¹⁸F-Fluorodeoxyglucose Positron Emission Tomography on Management of Cancer of Unknown Primary: Systematic Review and Meta-Analysis

Abstract

Background:

Cancers of unknown primary (CUP) have traditionally been treated empirically, with a dismal prognosis. Compared with standard diagnostic tests, including CT and MRI, imaging with ¹⁸F-fluorodeoxyglucose (FDG) PET or PET/CT has shown the capacity to better identify the primary tumor site and detect additional sites of metastasis. However, its clinical impact is not well established. We performed a systematic review and meta-analysis of prior studies to assess the impact of FDG-PET or PET/CT on the management of patients with CUP.

Materials and methods:

Pubmed and EMBASE databases were searched up to February 4, 2021. Studies that reported the proportion of patients with CUP who experienced a management change after FDG-PET or PET/CT were included, and the proportions were pooled using the random-effects model. Study quality was assessed using QUADAS-2. Subgroup analysis was conducted to explore heterogeneity.

Results:

Thirty-eight studies (involving 2795 patients) were included. The pooled proportion of patients with management changes was 35% (95% confidence interval 31%–40%). There was substantial heterogeneity among the studies (Q-test, p <0.01; I² = 82%). The specific reason for management change was more commonly detection of the primary site (22% [95% CI 18–28%]) than detection of additional metastatic sites (14% [95% CI 10–19%]). The pooled proportions of patients with management changes were similar among numerous subgroups (range, 32.8% to 38.2%).

Conclusion:

FDG-PET or PET/CT had a meaningful impact on the management of patients with CUP. Approximately a third of patients had their management changed due to FDG-PET or PET/CT results, and this finding was consistent across numerous subgroups.

INTRODUCTION

Cancers of unknown primary (CUP) origin are a group of histologically-proven metastatic cancers for which the primary site cannot be identified after a comprehensive diagnostic work-up (1). It is estimated that approximately 32,880 cases of CUP will be diagnosed in the US in 2021, accounting for approximately 2% of all cancers (2). Patients typically undergo a thorough physical examination, laboratory tests, and diagnostic procedures including chest radiographs, computed tomography (CT) and/or magnetic resonance imaging (MRI), ultrasonography, mammography, and endoscopy with or without biopsies (2). Even after comprehensive sets of tests, the majority (80–85%) of the patients cannot be assigned a presumed primary site of origin and in turn receive empirical systemic treatment, usually a platinum- or taxane-based chemotherapy regimen, with a dismal prognosis (e.g., median survival of 6–9 months) (3). Therefore, many research efforts have been focused on evaluating methods that might improve the search for the putative primary site, including immunohistochemistry (4), genomic profiling (5), and advanced imaging techniques (6).

With respect to imaging, contrast-enhanced CT and MRI (of the breast) have been widely used since their introduction decades ago and have been incorporated as key imaging studies for baseline evaluation of CUP in major guidelines such as those of the National Comprehensive Cancer Network (NCCN) (2, 7). However, interest in using ¹⁸F-fluoro-2-deoxyglucose positron emission tomography (FDG-PET), alone or combined with CT, to evaluate CUP has been growing. In comparison to CT and MRI, FDG-PET or PET/CT has been found capable of better identifying the primary tumor site. Previous meta-analyses showed that it could detect the primary site in more than one third of patients with CUP – for example, in 44% (95% confidence interval [CI] 31%–58%) of patients presenting with cervical nodes (8) and in 41% (95% CI 39%–43%) of those presenting with extracervical manifestations of CUP (9). Additionally, investigators have found that PET can detect previously unknown sites of metastasis in patients with CUP (10). With these strengths, PET has the potential to modify the chemotherapy regimen (so it can be tailored to the suspected primary site) or alter the goal of management (e.g., from curative to palliative or vice versa); however, the frequency and nature of management changes brought about by PET in patients with CUP have not yet been well established, owing to the heterogeneity of patient populations and study designs. Therefore, we performed a systematic review and meta-analysis of prior studies to assess the impact of FDG-PET or FDGPET/CT on the management of patients with CUP.

MATERIALS AND METHODS

This systematic review and meta-analysis was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, and the protocol was registered in PROSPERO (registration no. CRD42021264130) (11). The research question was: “How many of the patients that have cancer of unknown primary have their initial management plan changed after they have undergone FDG-PET (or FDG-PET/CT)?”

Literature search

A systematic search using databases Pubmed and EMBASE was done through February 4, 2021 using the following search query: (FDG OR fluorodeoxyglucose) AND (PET OR “positron emission tomography”) AND (treatment OR management OR therap*) AND (impact OR change OR alter OR modif* OR influence OR affect OR effect) AND (((cancer OR neoplasm OR malignancy OR tumor OR tumour OR metastas*) AND (occult OR “unknown primary” OR “unknown origin”)) OR CUP). The reference lists of initially screened articles were also checked to find additional relevant articles. No language limit was applied.

Study selection

Inclusion Criteria

Studies were included if they met the following criteria for the Patient-Index test-Comparison-Outcome-Study design (PICOS) (12): (1) Patients with CUP; (2) FDG-PET or FDG-PET/CT as the Index test; (3) conventional diagnostic studies of any type, extent, and combination with other clinical investigations as the Comparison; (4) proportion of patients who had their management plan changed as a result of FDG-PET or FDG-PET/CT as the Outcome; and (5) any type of Study design, including prospective or retrospective cohort study, published as either a full paper or a conference abstract.

Exclusion criteria

Exclusion criteria were (1) less than 10 patients; (2) other publication types (e.g., review articles, editorials, guidelines, and case reports); (3) insufficient information for extracting the proportion of patients with management change; and (4) overlapping patients. When overlap was present, we selected either the study that had the larger cohort or the one that provided more comprehensive information relevant to our research question.

The study selection process was conducted by two reviewers (S.W. and A.S.B.), and when disagreement was present, agreement was reached through discussion with a third reviewer (H.A.V.).

Data extraction and quality assessment

The following information was extracted from the included studies using a standardized form: author(s), publication year, patient population (country, institution, enrollment period, age, and gender), study design (prospective or retrospective, multicenter or single center, and whether enrollment was consecutive or not), details of the conventional imaging studies (e.g., CT and/or MRI) to which PET was compared, detection rate of primary site using PET, proportion of additional metastases detected by PET, percentage of patients that had their management changed owing to PET findings (stratified by whether this was due to detection of the primary site or additional metastases), location and histology of metastases, PET acquisition details (PET/CT vs PET, scanner, injected radiotracer dose and injection delay), and whether PET interpretation was done blinded to the conventional studies. The quality of the included studies was evaluated using the revised Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool (13). Data extraction and quality assessment were done in an identical manner by the same three authors mentioned above.

Data synthesis and analysis

The primary outcome was the “impact of FDG-PET or FDG-PET/CT on management,” measured as the proportion of patients that had a change in their management due to imaging findings on FDG-PET or FDG-PET/CT. The secondary outcomes included: (1) subgroup analysis based on various patient-, study-, and PET-related factors; (2) stratification of the “impact of FDG-PET on management” by reason (detection of primary site vs additional metastases); and exploration of what types of management change had occurred (e.g., from radiation treatment to chemotherapy, from surgery to chemotherapy, etc.). Proportions were pooled by inverse variance weighting with a random-effects model using “meta” and “metafor” packages in R software (version 3.4.1; R Foundation for Statistical Computing, Vienna, Austria) (14). Visualization was done using Forest plots and Sankey diagrams. Publication bias was evaluated using the funnel plot and Egger′s test (15). Heterogeneity was evaluated using the Cochran’s Q-test and Higgins I² test.

RESULTS

Literature search

Seven-hundred and sixty-two articles were initially retrieved. After removing 165 duplicates and excluding 529 articles by screening their titles and abstracts, there were 68 potentially eligible articles. Full-text reviews were performed and 32 studies were excluded. Bibliography searches yielded 3 additional relevant articles, and after the exclusion of one article due to overlapping patients, 38 studies comprising 2795 patients were included for the qualitative and quantitative analyses (10, 16–52). The study selection process is shown in Figure 1.

Fig. 1.

Flow diagram showing study selection

Characteristics of included studies

Characteristics of the included studies are summarized in Tables 1 and 2. In brief, study design was prospective in 6, retrospective in 29, and not reported in 3 studies. Approximately half (18/38 [47.4%]) of the studies were published after 2010. The numbers of patients ranged from 10 to 449 with median ages of 50–64 years. CT and/or MRI was performed prior to FDG-PET or FDG-PET/CT in all patients in 11/38 (28.9%) studies and in almost all patients in 7/38 (18.4%) studies. FDG-PET/CT (rather than FDG-PET) was used in all or almost all patients in 21 (55.2%) studies. FDG-PET or FDG-PET/CT identified the primary site in 35% (95% CI 31%–40%) of the patients. Additional sites of metastases were identified by PET in 25% (95% CI 21%–31%) of the patients.

Table 1.

Study characteristics

Author, published year	Country	Institution	Enrollment period	Conventional imaging	Reference standard	No. of patients	Design	Location of metastases	Histology	Primary location
Alberini Jl, 2003	Belgium	University Hospital Sart Tilman	NR	CT all	mixed	41	R	Bone : 14 Intracranial : 10 LN_others : 8 Skin : 2 Pleura : 1	AdenoCa : 20 PoorlyDiffCa : 5 SmallCellCa : 5 SqcCa : 4 Lymphoma : 2 Neuroendocrine : 2 Mesothelioma : 1 Myeloma : 1 Other : 1	Lung : 16 Breast : 3 Colorectal : 3 Hepatobiliary : 2 Head and Neck : 1 Pancreas : 1
Bakhshayeshkaram M, 2016	Iran	Masih Daneshvari Hospital	May 2013-May 2015	CT all	mixed	62	R	Bone : 11 Other : 10 Liver : 8 Lung : 8 Intracranial : 5 LN_cervical : 4 Other (not reported) : 4 Peritoneum : 2	AdenoCa : 29 PoorlyDiffCa : 14 Unknown : 12 SqcCa : 5 Neuroendocrine : 2	Lung : 13 Hepatobiliary : 2 Pancreas : 2 Colorectal : 1 Head and Neck : 1 Muscle : 1
Bruna C, 2007	France	Hopital de Brabois	January 2003-January 2006	CT almost all; MRI for some	mixed	37	R	Multiple : 14 LN_cervical : 10 LN_axilla : 6 Bone : 5 LN_inguinal : 5	AdenoCa : 17 SqcCa : 14 PoorlyDiffCa : 6	Breast : 3 Lung : 3 Colorectal : 2 Brain : 1 Head and Neck : 1 Ovarian : 1 Renal : 1 Thyroid : 1 Uterus : 1
Dakendar MR, 2011	India	Tata Memorial Hospital	June 2005-July 2009	CT or MRI almost all	Pathology & not done	40	R	LN_cervical : 5	NR	Esophageal : 2 Lung : 2 Head and Neck : 1
Deron PB, 2011	Belgium	University Hospital of Ghent	November 2002-November 2007	CT all	Mixed	18	R	LN_cervical : 18	SqcCa: 18	NA
Elboga U, 2014	Turkey	Gaziantep University	October 2009-May 2014	NR	path only	112	R	LN_cervical : 40 LN_others : 33 Bone : 15 Lung : 13 Liver : 8 Other (not reported) : 3	AdenoCa : 61 PoorlyDiffCa : 32 SqcCa : 19	Lung : 18 NA : 7 Pancreas : 5 Breast : 2 Colorectal : 1 Thyroid : 1 Uterus : 1
Fulop M, 2012	Hungary	National Institute of Oncology (Budapest)	January 2006-December 2010	NR	Pathology & unknown	77	R	LN_cervical : 77	NR	NA
Garin E, 2007	France	(2 centers) Centre Eugene Marquis & Centre Rene Gauducheau	March 2002 - January 2005	CT all	Mixed	51	P	LN_cervical : 15 Multiple : 14 LN_others : 11 Intracranial : 3 Bone : 2 Pleura : 2 Cardiac : 1 Pancreas : 1 Skin : 1 Visceral : 1	AdenoCa : 20 SqcCa : 19 PoorlyDiffCa : 11 Sarcoma : 1	Lung : 7 Head and Neck : 3 Esophageal : 1 Peritoneum : 1
Gutte H, 2005	Denmark	Copenhagen University Hospital (no overlap; checked)	January 2000-December 2010	NR	NR	71	R	NR	NR	NA
Hu M, 2011	China	Shandong Cancer Hospital	October 2004 - January 2009	NR	mixed	149	R	LN_cervical : 34 LN_others : 30 Bone : 24 Lung : 14 Other (not reported) : 13 Pleura : 13 Peritoneum : 10 Skin : 6 Liver : 5	PoorlyDiffCa : 49 AdenoCa : 47 SqcCa : 20	NA
Joshi U, 2004	Netherlands	VU University Medical Center (no overlap; checkd)	July 1997-December 2000	NR	mixed	50	R	LN_cervical : 17 Liver : 7 Multiple : 7 Bone : 6 Lung : 6 Pleura : 6 Soft_tissue : 6 Intracranial : 4 Skin : 4	AdenoCa : 48 Unknown : 6 Neuroendocrine : 2 PoorlyDiffCa : 2 SmallCellCa : 2 SqcCa : 2	Lung : 8 Breast : 2 Colorectal : 1 Hepatobiliary : 1 Renal : 1 Stomach : 1 Thyroid : 1 Uterus : 1
Jungehülsing M, 2000	Germany	University of Cologne	May 1994-July 1998	NR	mixed	27	R	LN_cervical : 26 Intracranial : 1	SqcCa: 18 AdenoCa : 4 PoorlyDiffCa : 2 Melanoma : 1 Other : 1 Unknown : 1	Head and Neck : 5 Lung : 2
Karapolat I, 2012	Turkey	Şifa University School Medicine	March 2008-May 2011	CT or MRI almost all	mixed	20	R	LN_cervical : 20	SqcCa: 16 PoorlyDiffCa : 3 AdenoCa : 1	Head and Neck : 5 Esophageal : 1 Lung : 1
Kolesnikov-Gauthier H, 2005	France	Henri Mondor University	April 1999-March 2003	NR	mixed	25	P	LN_others : 6 Intracranial : 5 LN_mediastinum : 5 Multiple : 3 LN_cervical : 2 Pleura : 2 Bone : 1 Peritoneum : 1	AdenoCa : 13 PoorlyDiffCa : 12	Lung : 6 Ovarian : 1
Lassen U, 1999	Denmark	Rigshospitalet, University Hospital, Copenhagen (no overlap; checked)	April 1996-September 1997	NR	mixed	20	R	LN_cervical : 11 Bone : 4 Chest : 1 Intracranial : 1 Pericardial : 1 Pleura : 1 Skin : 1	PoorlyDiffCa : 10 SqcCa : 6 AdenoCa : 4	Lung : 8 Head and Neck : 1
Lonneux M, 2000	Belgium	Cliniques Universitaires Saint-Luc	NR	NR	mixed	24	R	Intracranial : 8 Liver : 3 LN_cervical : 3 Pleura : 3 Bone : 2 LN_axilla : 2 Soft_tissue : 2 Eye : 1	AdenoCa : 18 PoorlyDiffCa : 2 SqcCa : 2 Unknown : 2	Lung : 9 Breast : 1 Colorectal : 1 Head and Neck : 1 Stomach : 1
Mantaka P, 2003	Germany	Johann Wolfgang Goethe University of Frankfurt	NR	CT or MRI all	mixed	25	P	LN_cervical : 14 LN_others : 7 Lung : 3 Liver : 1	AdenoCa : 8 SqcCa : 7 PoorlyDiffCa : 3 SmallC ellCa : 3 Other : 2 Unknown : 2	Lung : 6 Breast : 3 Ovarian : 1 Pancreas : 1 Stomach : 1
Padovani D, 2009	Italy	Padova University Hospital	January 2001-September 2006	CT all; MR in some	NR	13	R	LN_cervical : 13	SqcCa : 5 Lymphoma : 1 Sarcoma : 1 Unknown : NA	Head and Neck : 5 Bone : 1 Lung : 1
Park JS, 2011	Korea	Seoul National University Hospital	January 2003-September 2005	NR	mixed	20	R	LN_cervical : 6 Bone : 4 LN_abdomen : 3 LN_axilla : 2 Peritoneum : 2 Intracranial : 1 Skin : 1 Visceral : 1	PoorlyDiffCa : 11 AdenoCa : 6 SqcCa : 2 Sarcoma : 1	NA
Pelosi E, 2006	Italy	ASQ San Giovanni Battista	June 2004-June 2005	NR	mixed	68	R	LN_cervical : 21 LN_axilla : 11 Pleura : 10 Bone : 7 Liver : 6 LN_inguinal : 4 Multiple : 3 Intracranial : 2 LN_RP : 2 LN_mediastinum : 1 Skin : 1	PoorlyDiffCa : 37 AdenoCa : 18 SqcCa : 8 Melanoma : 4 Other : 1	Lung : 9 Head and Neck : 6 Pancreas : 5 Colorectal : 2 Uterus : 2
Rades D, 2001	Germany	Hannover Medical College	May 1998 - October 2000	CT almost all	mixed	42	R	LN_cervical : 25 LN_axilla : 5 Bone : 4 Liver : 2 LN_inguinal : 2 Intracranial : 1 LN_mediastinum : 1 LN_RP : 1 Pleura : 1	SqcCa : 24 AdenoCa : 11 PoorlyDiffCa : 7	Head and Neck : 7 Lung : 5 Anal : 1 Breast : 1 Hepatobiliary : 1 Ovarian : 1 Urethra : 1 Vagina : 1
Regelink G, 2002	Netherlands	(2 centers) University Hospital Groningen and VU Medical Centre (no overlap; checked)	January 1994-November 2000	CT or MRI all	mixed	50	R	LN_cervical : 50	SqcCa : 30 AdenoCa : 19 Neuroendocrine : 1	Head and Neck : 13 Esophageal : 2 Lung : 1
Reinert CP, 2020	Germany	Tuebingen University Hospital	April 2013 - June 2018	CT or MRI almost all	NR	155	R	Other (not reported) : 45 LN_cervical : 32 Soft_tissue : 17 Liver : 14 Bone : 13 Intracranial : 12 LN_axilla : 11 Lung : 7 Peritoneum : 4	SqcCa : 67 AdenoCa : 26 Neuorendocrine : 19 PoorlyDiffCa : 3 Other : 2	Others : 15 Lung : 9 Breast : 6 Head and Neck : 6 Pancreas : 1
Rossi M, 2015	Italy	Ospedale Maggiore	NR	NR	pathology only	172	R	NR	NR	Lung : 11 Head and Neck : 10 Colorectal : 6 Lymph Nodes : 6 Hepatobiliary : 4 Ovarian : 4 Breast : 3 Pancreas : 3 Muscle : 2 Stomach : 2 Chest : 1 Prostate : 1 Renal : 1 Testicular : 1 Thyroid : 1
Saidha NK, 2013	India	Command Hospital Air Force	May 2008-May 2010	CT or MRI all	Pathology & unknown	50	R	LN_cervical : 16 Peritoneum : 8 LN_axilla : 7 LN_inguinal : 7 LN_mediastinum : 4 Liver : 3 Bone : 2 Intracranial : 2 LN_RP : 1	AdenoCa : 22 PoorlyDiffCa : 15 SqcCa : 13	Lung : 6 Head and Neck : 5 Colorectal : 3 Stomach : 2 Anal : 1 Breast : 1 Esophageal : 1 Pancreas : 1 Uterus : 1
Sarma M, 2015	India	Amrita Institute of Medical Sciences and Research Center	NR	NR	mixed	38	P	LN_cervical : 38	NR	NA
Scott CL, 2005	Australia	Peter MacCallum Cancer Centre	January 1997-May 2000	CT almost all; MRI some	mixed	31	R	Liver : 9 Lung : 9 Abdomen : 4 Other (not reported) : 3 Intracranial : 2 LN_axilla : 2 Bone : 2 Skin : 1 Chest : NA	AdenoCa : 22	Lung : 4 Pancreas : 2 Esophageal : 1 Renal : 1
Stokkel MP, 1999	Netherlands	University Hospital Utrecht	NR (consecutive)	CT all	mixed	10	R	LN_cervical : 10	SqcCa : 9 PoorlyDiffCa : 1	Head and Neck : 5
Su YY, 2016	Taiwan	Chang Gung Memorial Hospital	September 2006-May 2014	NR	pathology only	54	R	LN_cervical : 54 Bone : 4 Multiple : 4 LN_others : 2 Visceral : 2	AdenoCa : 44 PoorlyDiffCa : 8 Unknown : 2	Head and Neck : 7 Esophageal : 4 Lung : 1 Stomach : 1
Subramaniam RM, 2016	USA	Multicenter (National Oncologic PET Registry)	May 2006-April 2009	NR	NR	403	P	Liver : 115 Bone : 78 LN_others : 69 Lung : 46	NR	NR
Talaat O, 2019	Egypt	(multi) National Cancer Institute of Egypt, Zagazig University Hospital	January 2011-October 2015	CT or MRI all	mixed	123	P	Bone : 36 LN_others : 27 Liver : 21 Intracranial : 13 Other (not reported) : 11 Pleura : 10 Lung : 5	PoorlyDiffCa : 52 Unknown : 35 NA : 34 SqcCa : 2	Lung : 16 Breast : 9 Pancreas : 9 Prostate : 6 Colorectal : 5 Ovarian : 5 Lymph Nodes : 4 Head and Neck : 3 Hepatobiliary : 3 Thyroid : 3 Chest : 2 Renal : 2 Uterus : 2 Bladder : 1 Brain : 1
Trampal C, 2000	Sweden	Uppsala University	NR	NR	mixed	12	R	Multiple : 3 LN_axilla : 2 LN_cervical : 2 Adrenal : 1 Cava : 1 Intracranial : 1 LN_mediastinum : 1 Lung : 1	NR	Breast : 2 Lung: 1 Pancreas : 1
Wang G, 2013	China	Sun Yat-sen University	January 2006-June 2010	NR (but described as poor work-up)	mixed	142	R	LN_cervical : 56 Bone : 31 Liver : 25 Pleura_and_peritoneum : 12 Lung : 8 LN_axilla : 4 Pleura : 4 Intracranial : 3 LN_inguinal : 2 LN_RP : 1 Thyroid : 1	Unknown : 64 AdenoCa : 41 SqcCa : 35 Other : 3 Sarcoma : 1 NA : 1	Lung : 38 Head and Neck : 11 Colorectal : 6 Pancreas : 4 Hepatobiliary : 2 Ovarian : 2 Thyroid : 2 Prostate : 1
Wartski M, 2007	France	Cancer Research Centre Rene	October 2002 - March 2005	CT or MRI all	pathology & not done	38	R	LN_cervical : 23	SqcCa : 32 PoorlyDiffCa : 4 Other : 2	Head and Neck : 13
Wong WL, 2003	United Kingdom	Mount Vernon Hospital	January 1999 - September 2000	CT or MRI all	mixed	17	R	LN_cervical : 34	SqcCa : 16 PoorlyDiffCa : 1	Head and Neck : 5
Wu ZJ, 2007	China	Huazhong University of Science and Technology Union Hospital	September 2003-November 2005	CT in more than half	mixed	34	R	LN_cervical : 10 Peritoneum : 4 Bone : 3 LN_axilla : 3 Intracranial : 2 Liver : 2 LN_inguinal : 2 Kidney : 1 Pleura : 1 Pleura_and_peritoneum : 1 Thyroid : 1	AdenoCa : 22 Unknown : 6 SqcCa : 3 Melanoma : 2 PoorlyDiffCa : 1	Lung : 8 Colorectal : 3 Ovarian : 2 Breast : 1 Esophageal : 1 Head and Neck : 1 Pancreas : 1
Yu X, 2016	China	Tianjin Cancer Institute and Hospital	January 2006 - October 2014	NR	mixed	449	R	LN_cervical : 174 LN_others : 169 Bone : 49 Intracranial : 22 Liver : 15 Lung : 8 Pleura : 5 Adrenal : 2 Ovary : 2 Pericardial : 2 Bladder : 1 Eye : 1 Peritoneum : 1	AdenoCa : 211 SqcCa : 121 PoorlyDiffCa : 76 Melanoma : 19 Other : 18 Sarcoma : 4	Lung : 40 Head and Neck : 31 Pancreas : 13 Small intestine : 8 Stomach : 6 Ovarian : 4 Colorectal : 3 Esophageal : 3 Renal : 2 Breast : 1 Hepatobiliary : 1 Penis : 1 Prostate : 1 Uterus : 1
Zhao K, 2012	China	First Affiliated Hospital, College of Medicine, Zhejiang University	July 2007-July 2011	NR	pathology & unknown	25	R	LN_cervical : 25	PoorlyDiffCa : 13 SqcCa : 11 Other : 1	Head and Neck : 6 Lung : 3 Thyroid : 2 Chest : 1

CT = computed tomography; LN = lymph node; MRI = magnetic resonance imaging; NA = not available; NR = not reported; P = prospective; R = retrospective; RP = retroperitoneal; SmallCellCa = small cell carcinoma

Table 2.

¹⁸F-FDG-PET and ¹⁸F-FDG-PET/CT characteristics

Author, publication year	Modality	Scanner (Vendor)	Delay after injection (min)	Injected dose	Blinded interpretation	Primary site by PET	New metastatic site by PET
Alberini Jl, 2003	PET	Penn PET 240H (UGM)	60–90	4 MBq/Kg	Yes	0.63	0.02
Bakhshayeshkaram M, 2016	PET/CT	Discovery 690 (GE)	NR	4.6 MBq/Kg	NR	0.32	0.56
Bruna C, 2007	PET/CT	Biograph (Siemens)	60	5.5 MBq/Kg	NR	0.38	0.27
Dakendar MR, 2011	PET/CT	Discovery ST (GE)	60	370 MBq	NR	0.12	0.22
Deron PB, 2011	PET/CT	Gemini PET-CT Imaging system (Philips)	60	3.7 MBq/kg	yes	0.00	0.00
Elboga U, 2014	PET/CT	Biograph 2 (Siemens)	60	369.3 MBq* (296–444)	NR	0.33	0.29
Fulop M, 2012	PET/CT	NR	NR	NR	NR	0.44	0.17
Garin E, 2007	PET	PET Advence camera or Discovery LS (GE)	61* (50–77)	339 MBq* (197–540)	Yes	0.24	0.41
Gutte H, 2005	PET	GE 4096 or GE Advance (GE)	60	NR	NR	NA	NA
Hu M, 2011	PET/CT	Discovery LS (GE)	60 (52–69)	350–425 MBq	NR	0.25	0.30
Joshi U, 2004	PET	ECAT EXACT HR+ (Siemens)	60	370 MBq	Yes	0.25	NA
Jungehülsing M, 2000	PET	CTI ECAT Exact (Siemens)	60–90	370 MBq	NR	0.26	0.26
Karapolat I, 2012	PET/CT	HI-REZ biograph 6 (Siemens)	60	555 MBq	Yes	0.35	0.35
Kolesnikov-Gauthier H, 2005	PET	ECAT EXACT HR+ (Siemens) or GE Advance (GE)	60	370 MBq*	NR	0.24	NA
Lassen U, 1999	PET	GE Advance (GE)	40	350–400 MBq	No	0.45	0.20
Lonneux M, 2000	PET	ECAT EXACT HR (Siemens)	60	370 MBq	NR	0.54	0.29
Mantaka P, 2003	PET	ECAT Exact 47 (Siemens)	60	NR	No	0.48	NA
Padovani D, 2009	PET	IRIX (Marconi)	60	370 MBq	NR	0.54	0.31
Park JS, 2011	PET	NR	60–90	555–740 MBq	Yes	0.00	0.35
Pelosi E, 2006	PET/CT	Discovery ST (GE) or Gemini Tomography (Philips)	60	222–370 MBq	No	0.35	0.23
Rades D, 2001	PET	ECAT 951/31 or ECAT 922/47 (Siemens)	60–150	370–740 MBq	NR	0.43	0.38
Regelink G, 2002	PET	ECAT 951/31 or Siemens HR+ (Siemens)	60–90	370–490 MBq	Yes	0.32	0.14
Reinert CP, 2020	PET/CT	Biograph mCT (Siemens)	60	300–350 MBq	No	0.23	NA
Rossi M, 2015	PET/CT	NR	60	370 Mbq	NR	0.33	0.28
Saidha NK, 2013	PET/CT	Biograph (Siemens)	NR	NR	No	0.42	0.06
Sarma M, 2015	PET/CT	NR	NR	NR	NR	0.29	0.11
Scott CL, 2005	PET	GE300H Quest (GE)	≥60	74–111 MBq	Yes	0.26	0.48
Stokkel MP, 1999	PET	Vertex MCD (ADAC)	60	185 MBq	NR	0.60	0.50
Su YY, 2016	PET/CT	Discovery ST 16 (GE)	60	370–555 MBq	NR	0.24	NA
Subramaniam RM, 2016	PET/CT	NR	NR	NR	NR	NA	NA
Talaat O, 2019	PET/CT	NR	NR	5.2 MBq/Kg	NR	0.58	0.41
Trampal C, 2000	PET	ECAT EXACT HR1 (Siemens)	60	400 MBq	NR	0.33	0.33
Wang G, 2013	PET/CT	Discovery DST (GE)	60–75	270–370 MBq	No	0.47	0.10
Wartski M, 2007	PET/CT	Discovery LS (GE)	60	4–5 MBq/kg	Yes	0.34	0.08
Wong WL, 2003	PET	ECAT Exact 47 (Siemens)	60	350 MBq	Yes	0.29	0.12
Wu ZJ, 2007	PET/CT	Discovery LS (GE)	NR	5.55 MBq/Kg	NR	0.50	0.50
Yu X, 2016	PET/CT	Discovery ST or Discovery 710 (GE)	60	4.1–4.8 MBq/Kg	No	0.26	NA
Zhao K, 2012	PET/CT	Biograph Sensation 16 (Siemens)	60–90	5.5–7.4 MBq/Kg	Yes	0.44	NA

CT = computed tomography; NR = not reported; PET = positron emission tomography; ¹⁸F-FDG = Fluorine-18 Fluorodeoxyglucose

^*Mean with range in parentheses

Quality assessment

The quality of the majority of the studies was not considered good, with 73.7% (28/38) satisfying less than 5 of the 7 QUADAS-2 domains (Figure 2). In the patient selection domain, three studies were considered at high risk for bias, as two of them specifically stated that work-up before FDG-PET was poor and therefore the patients did not strictly meet the definition of CUP, while the third study was based on a random, rather than consecutive, sample of patients from a national registry. For 16 studies, the risk of bias due to patient selection was unclear because the full details on work-up prior to FDG-PET or FDG-PET/CT was not provided (N = 12) or it was not stated whether the enrollment of patients was consecutive (N = 4). In the index test domain, 21 studies were considered at high risk of bias, 17 of them because FDG-PET was used (as opposed to FDG-PET/CT) and 4 because PET was interpreted without blinding to conventional assessments including CT and MRI. For 13 additional studies, the risk of bias in this domain was uncertain because the status of blinding was not clear. Applicability was a concern for 18 studies: 17 of these used FDG-PET alone (i.e., without the CT component, which is now standard), and 1 used a somatostatin receptor PET/CT for 10.3% of patients; the bulk of patients included in our meta-analysis had undergone FDG PET/CT (which was the index test of our research question). In the reference standard domain, 2 studies were at high risk of bias because some of the patients did not have a reference standard. Seven studies were at unclear risk of bias and concern for applicability because there was no explanation of how the reference standard was established for all or some of the patients. In the flow and timing domain, 28 studies were at high risk of bias because an identical reference standard was not used for all patients within the study. The seven studies for which the reference standard was not reported also had an unclear risk of bias in this domain.

Fig. 2.

Grouped bar charts showing the summary of risk of bias (left) and concern for applicability (right) of the included studies according to the revised QUADAS-2 tool.

Impact of PET on patient management

The proportion of patients who had their management influenced by FDG-PET or PET/CT in each study is shown in Figure 3. The proportions ranged from 0% to 73% among the 38 studies, with a pooled proportion of 35% (95% CI 31%–40%). There was substantial heterogeneity among the studies based on the Q-test (p <0.01) and Higgins I² value (82%). No publication bias was present based on the Egger’s test (p = 0.127) and visual assessment of the funnel plot (Figure 4). The specific reason for management change was more commonly detection of the primary site (in 22% [95% CI 18–28%] of patients) than detection of additional metastatic sites (in 14% [95% CI 10–19%] of patients) (Figure 5). Details of the pre- and post-FDG-PET or FDG-PET/CT management plans and how they were changed are shown in the Sankey plot (Figure 6). Table 3 summarizes the results of multiple subgroup analyses. The pooled proportions of patients that had their management changed by FDG-PET or FDG-PET/CT were similar across various subgroups and ranged from 32.8% to 38.2%.

Fig. 3.

Forest plots showing pooled proportion of patients that had their management changed due to ¹⁸F-FDG-PET or PET/CT.

Fig. 4.

Funnel plot and Egger’s test results suggesting that possibility of significant publication bias is low (p = 0.9755).

Fig. 5.

Stacked bar charts showing the reasons why ¹⁸F-FDG-PET and PET/CT altered management, stratified to (1) detection of primary site (blue) and (2) detection of new sites of metastases (green). Studies in which these reasons were not separable are shown in grey.

Fig. 6.

Sankey diagram shows management decisions before (left) and after (right) ¹⁸F-FDG-PET and PET/CT in patients with cancer of unknown primary. Diagram was based on the 24 studies included in the meta-analysis that provided information on the management plans before and after PET. Boxes on the left indicate pre-PET management plans and those on the right indicate post-PET management plans annotated with each of their respective types of management (e.g., surgery, radiation, chemotherapy, and chemoradiation). *Unknown = Pre-PET management plan not reported in the included paper. ^†Unchanged: Unknown = Details of the post-PET management plan was not reported but was indicated in the paper that there was not change from the pre-PET management decision. ^‡Change: Unknown = The management plan was changed between before and after PET, however the paper did not clarify what the specific management plans were.

Table 3.

Results of subgroup analyses for the impact of ¹⁸F-FDG-PET or ¹⁸F-FDG-PET/CT on management in patients with CUP

Variable	Category	No. of studies	Proportion (95% CI)	I2 (%)
Study design	Prospective	6	38% (23%–55%)	91.1
	Retrospective or unknown	32	35% (30%–40%)	75.1
Publication year	2010 or earlier	20	38% (32%–45%)	65.0
	After 2010	18	34% (26%–40%)	88.4
Study population size	≥50	17	35% (29%–42%)	89.0
	<50	21	36% (29%–44%)	64.1
Imaging modality	PET/CT	17	34% (27%–42%)	65.5
	PET	21	36% (30%–43%)	87.1
Metastatic site(s)	Cervical node only	13	35% (26%–46%)	74.0
	Cervical node >1/3 patients	6	37% (27%–47%	81.5
	Cervical node <1/3 patients	19	35% (28%–42%)	84.5
Conventional imaging workup	CT or MRI in all or most patients	18	37% (28%–47%)	85.2
	CT or MRI not done in all or most patients	20	34% (30%–39%)	74.3

CI = confidence interval; CT = computed tomography; CUP = cancer of unknown primary; FDG = Fluorodeoxyglucose; MRI = magnetic resonance imaging; PET = positron emission tomography

DISCUSSION

Our meta-analysis assessed the impact of FDG-PET or FDG-PET/CT on the management of patients with CUP. We found that FDG-PET altered the management approach in approximately a third of patients (35%; 95% CI 31%–40%). This relatively substantial degree of impact was found to be consistent across numerous subgroups stratified by patient, study, and PET-related factors, for whom the pooled proportions of patients with post-FDG-PET management changes ranged from 32.8% to 38.2%. Until now, most of the literature investigating the role of FDG-PET and FDG-PET/CT in patients with CUP has centered on its ability to detect the primary cancer site; in contrast, our study focuses on the clinical relevance of FDG-PET for these patients and shows that it substantially alters their management. At the moment, guidelines regarding the baseline diagnostic work-up of patients with CUP do not require the use of FDG-PET, even though it has increasingly been applied for various purposes in these patients (2, 7). Our results support the incorporation of FDG-PET into the baseline diagnostic work-up of patients with CUP as a means of improving their management.

When FDG-PET changed the management plan in patients with CUP, it was typically for one of two reasons: detection of the primary site (in 22% [95% CI 18–28%] of patients) or detection of additional sites of metastasis (in 14% [95% CI 10–19%] of patients). Identification of the primary site has the potential to change management in multiple ways. Not only can it change the chemotherapeutic regimen to one tailored to the newly established primary site (as opposed to an empirical platinum- or taxane-based chemotherapy regimen), it can occasionally alter the intent of treatment (e.g., from palliative to curative) even in the presence of metastases. For example, whereas a pancreatic cancer with hepatic metastases would be managed with a palliative intent, colorectal primary cancer with hepatic metastases could be treated with a curative intent). On the other hand, detection of additional sites of metastasis would usually cause a change from a potentially curative approach to a palliative approach.

Interestingly, not all patients where PET identified the primary site or new metastatic sites had their management altered. For example, PET detected the primary site in 35% (95% CI 31%–40%) of patients, but management was changed due to detection of the primary site in only 22% (95% CI 18%–28%) of patients; PET detected additional sites of metastases in 25% (95% CI 21%–31%) of patients, but management was altered due to detection of additional metastases in only 14% (95% CI 10%–19%) of patients. The differences between the detection rates and the frequency of resulting management changes can be attributed to the following reasons: First, the empirical taxane- or platinum-based chemotherapeutic regimens typically used for CUP are standard of care for many types of cancers with metastases. Second, when widespread metastatic disease is already known to be present, finding additional sites of metastasis would not affect the prior decision to administer chemotherapy with a palliative intent. Therefore, it can be hypothesized that the patients with CUP who are most likely to benefit from FDG-PET are those for whom the initial management plan is aggressive and/or of curative intent and could potentially be altered by the discovery of new sites of disease.

Our meta-analysis had certain limitations. First, based on assessment with the revised QUADAS-2 tool, many of the included studies had definite or possible methodological weaknesses that could have led to bias and/or limited applicability. For example, most studies were retrospective, and in a non-negligible portion of the studies, the extent of the baseline conventional diagnostic work-up varied among the patients. It is possible that the impact of FDG-PET or FDG-PET/CT would have been smaller in a more homogeneous population of patients who had all undergone a comprehensive workup. On the other hand, one could argue that using FDG-PET or FDG-PET/CT at an early point, before a more comprehensive conventional work-up had been done, may have reduced the need for other diagnostic tests. Given these uncertainties, better data concerning CUP management is needed to establish an optimal, standardized approach to the work-up of patients with CUP and determine what types of conventional diagnostic tests should be done prior to obtaining FDG-PET or FDG-PET/CT. Second, there was substantial heterogeneity among the included studies (Higgin’s I² =82%). However, we aimed to mitigate the effects of this heterogeneity by performing subgroup analyses, where we found that the impact of FDG-PET or FDG-PET/CT on management was consistent across multiple subgroups. Additionally, we stratified results based on the specific reasons FDG-PET or FDG-PET/CT impacted management. Third, we still do not know whether changing management after PET actually improves outcomes (e.g., prolongs survival). For example, randomized controlled trial showed that tailoring the chemotherapy regimen according to the predicted primary cancer based on gene expression profiling did not improve survival, though the predicted primary site itself had prognostic value (53). Further well-designed prospective studies may be needed to confirm whether management changes based on FDG-PET or FDG-PET/CT findings result in better outcomes. Fourth, the literature regarding imaging in CUP is heavily based on pretreatment evaluation. FDG-PET or FDG-PET/CT can also be used to assess treatment response and may provide additional, clinically relevant information not provided CT or MRI, which generally rely on size measurements. However, this topic was beyond the scope of our meta-analysis and needs to be addressed in future studies. Fifth, we only considered FDG-PET and FDG-PET/CT, as FDG is one of the most widely used tracers for cancer imaging. Somatostatin receptor PET is already actively used for neuroendocrine cancers, and a meta-analysis of 12 studies (383 patients) already showed that it identified the primary site of neuroendocrine tumors of unknown origin in about half of patients (56%; 95% CI 48–63%) and ultimately impacted management in about a fifth (20%; 95% CI: 10–33%) (54). Additionally, novel tracers such as fibroblast activation protein inhibitor (FAPI)-PET, which in early studies showed high tumor uptake in various types of cancers and low background uptake, are promising and need to be investigated separately (55). An early study showed that FAPI-PET/CT was able to better show the primary tumor than FDG PET/CT in 8 patients with head and neck manifestations of CUP (56).

4. Conclusions

FDG-PET or FDG-PET/CT had a meaningful impact on the management of patients with cancer of unknown primary origin. Approximately a third of the patients had their management changed based on FDG-PET or FDG-PET results, and this finding was consistent across numerous subgroups.

Highlights.

1. In 38 studies with cancer of unknown primary, FDG-PET changed management 35%

2. These was related to finding primary site (22%) or additional metastatic sites (14%)

3. Similar pooled proportions were seen across subgroups (32.8–38.2%)

Abbreviations:

CI

confidence interval

CT

computed tomography

CUP

Cancers of unknown primary

FDG

18F-fluoro-2-deoxyglucose

MRI

magnetic resonance imaging

NCCN

National Comprehensive Cancer Network

PET

positron emission tomography

PICOS

Patient-Index test-Comparison-Outcome-Study design

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

QUADAS-2

Quality Assessment of Diagnostic Accuracy Studies-2