Visual Abstract
Abstract
Cancer of unknown primary (CUP) is a heterogeneous entity with a limited prognosis. Novel prognostic markers are needed for patient stratification in prospective clinical trials exploring innovative therapies. Methods: In CUP patients treated at the West German Cancer Center Essen, the prognostic value of 18F-FDG PET/CT at the initial diagnostic workup was analyzed by comparing overall survival (OS) in patients who underwent 18F-FDG PET/CT with those who did not. Results: Of 154 patients with a CUP diagnosis, 76 underwent 18F-FDG PET/CT at the initial diagnostic workup. The median overall survival (OS) of the full analysis set was 20.0 mo. Within the PET/CT subgroup, an SUVmax above 20 was associated with significantly superior OS (median OS, not reached vs. 32.0 mo; hazard ratio, 0.261; 95% CI, 0.095–0.713; P = 0.009). Conclusion: Our retrospective work shows that an SUVmax above 20 on 18F-FDG PET/CT at the initial diagnostic workup is a favorable prognostic factor in patients with CUP. This finding deserves further prospective studies for validation.
Cancer of unknown primary (CUP) is a heterogeneous cancer entity in which the primary tumor remains unknown after a comprehensive diagnostic workup (1). CUP affects 2%–4% of all newly diagnosed malignancies. In the literature, 75%–85% of cases are described as diffusely metastatic at initial diagnosis. The median overall survival (OS) of patients with diffusely metastasized CUP remains poor, at only 8–11 mo (2). Unlike the multimodal concepts in solitary metastatic CUP syndrome with a curative approach, patients with diffuse metastasis usually receive only systemic palliative chemotherapy, with generally poor responses (1–3). To provide a basis toward future treatment approaches stratified by clinical and biologic markers, we analyzed a large cohort of CUP patients receiving treatment at a major German comprehensive cancer center.
In this study, particular attention was paid to the potential prognostic value of 18F-FDG PET/CT imaging in CUP patients.
MATERIALS AND METHODS
Patient Selection
The clinical cancer registry of the West German Cancer Center Essen was searched for patients who received an initial diagnosis of CUP between January 2015 and May 2021, with the data being cut off on December 23, 2021. We stringently applied the predefined CUP criteria of the current European Society for Medical Oncology guidelines to confirm the diagnosis (Fig. 1) (1).
CONSORT (Consolidated Standards of Reporting Trials) flow diagram for patient inclusion in data analysis.
Image Acquisition and Analysis
Images were acquired in accordance with the European Association of Nuclear Medicine guidelines for tumor imaging using 18F-FDG PET (4), as follows: after the intravenous administration of a median of 289 MBq (interquartile range, 97 MBq) of 18F-FDG and a median uptake interval of 73 min (interquartile range, 16 min), image acquisition began. The 18F-FDG activity was based on body weight. Images were analyzed by 3 unmasked readers, among whom were at least 1 board-certified radiologist and 1 nuclear medicine physician, as per the clinical standard. If multiple lesions were present, the most representative malignant lesion was chosen as the target lesion for further analyses.
Statistical Analysis
In the initial, pretherapeutic, 18F-FDG PET/CT scan, we determined the SUVmax of the detected tumor manifestations. The 75th percentile of SUVmax was set as the discriminator between high and low SUVmax. OS was defined as the time from initial diagnosis to death from any cause. If the time point of death was not detectable, patients were censored at the time of last follow-up. Survival was analyzed using the Kaplan–Meier method and log-rank testing, and putative prognostic factors were correlated by multivariate analysis with Cox regression.
Statistical analyses were performed using Excel 2016 (version 2304; Microsoft) and SPSS Statistics (version 26; IBM). A P value of 0.05 or less was defined as statistically significant. All analyses were performed on pseudonymized datasets. The institutional review board approved this retrospective study, and the requirement to obtain informed consent was waived; nevertheless, all subjects gave written informed consent to providing their data for scientific purposes.
RESULTS
Patients
In total, 618 patients with a clinical diagnosis of CUP were identified, of whom 154 fulfilled all the prespecified inclusion criteria of the study (full analysis set). Of those, 76 (49%) had undergone 18F-FDG PET/CT at the initial diagnostic workup (PET/CT set), whereas 78 (51%) had no initial 18F-FDG PET/CT (comparator set, Fig. 1). The characteristics of the different subgroups are summarized in Table 1. The median SUVmax of the PET/CT set was 13.9 (range, 3.5–30.8). The 75th percentile was approximately 20. At the time of data cutoff, 96 (62.3%) patients were deceased, 56 (36.4%) were alive, and 2 (1.3%) were lost to follow-up.
Baseline Characteristics
Survival and Prognostic Factors
The OS was 20.0 mo (95% CI, 10.0–30.0 mo) for the full analysis set (Fig. 2A), 36.0 mo (95% CI, 21.3–50.7) for the PET/CT set (Fig. 2B), and 12.0 mo (95% CI, 8.7–15.3) for the comparator set (Fig. 2B).
Median OS of full analysis set and patient subsets. (A) OS of full analysis set (n = 154) was 20.0 mo (95% CI, 10.0–30.0). (B) At 36.0 mo (95% CI, 21.3–50.7), OS of PET/CT set (18F-FDG PET/CT at initial diagnostic workup, n = 76) was significantly higher than OS of comparator set (without 18F-FDG PET/CT at initial diagnostic workup, n = 78), at 12.0 mo (95% CI, 8.7–15.3; P < 0.001).
Using multivariate Cox regression, we correlated OS with established prognostic clinical parameters. We identified a solitary metastasis as a significant favorable prognostic factor (Fig. 3; Supplemental Figs. 1 and 2; supplemental materials are available at http://jnm.snmjournals.org). Significant unfavorable prognostic factors were involvement of 3 or more organ systems (Fig. 3; Supplemental Figs. 1 and 3) and an age over 60 y (Fig. 3; Supplemental Fig. 1). These prognostic parameters were comparably relevant in the full analysis set, the PET/CT set, and the comparator set (Fig. 3; Supplemental Fig. 1), demonstrating overall a similar distribution of CUP biologies among these populations.
Multivariate analysis for prognostic parameters for PET/CT set (n = 76): hazard ratio with 95% CI and P value. *P < 0.05. **P < 0.01. HR = hazard ratio; NEC = neuroendocrine carcinoma).
We next explored the prognostic impact of tumor metabolic activity as determined by SUVmax at pretherapeutic 18F-FDG PET/CT. An SUVmax of 20 clearly discriminated a favorable subgroup from an unfavorable subgroup in the PET/CT set (median OS, not reached vs. 32.0 mo [95% CI, 23.2–40.8]; hazard ratio, 0.261 [95% CI, 0.095–0.713]; P = 0.009) (Figs. 3 and 4).
Median OS in relation to SUVmax in PET/CT set. In subset with initial 18F-FDG PET/CT, patients with SUVmax above 20 (n = 22) had significantly higher OS (not reached) than patients with SUVmax of 20 or less (n = 54) (32.0 mo; 95% CI, 23.2–40.8; P = 0.022).
DISCUSSION
With a poor OS of 11.0 mo, our data for patients with CUP and diffuse metastasis are consistent with data from other studies (2). Further progress in treatment is hampered by the heterogeneity of the CUP population, which may confound the results of studies exploring novel treatments. Current stratification factors such as age, a low number of metastatic sites, or solitary metastasis, which have already been described in the literature (1,3) and could also be confirmed in this work as prognostic factors, are of limited use. They do not reflect the biologic heterogeneity within the respective subgroups.
Using functional imaging, we here describe an independent association of tumor metabolic activity with survival of patients with a stringently defined CUP diagnosis. An SUVmax above 20, as detected at the pretherapeutic workup by 18F-FDG PET/CT, is a strong predictor of a favorable outcome.
In contrast, Yılmaz et al. (5) among patients with stage III non–small cell lung cancer, and Werner et al. (6) among patients with laryngeal carcinoma, found no influence of SUVmax on OS.
Other studies described a high SUVmax as a negative prognostic factor in tumor entities such as hypopharyngeal or pancreatic cancer (6,7). These studies focused on nonmetastatic disease stages with a curative approach, corresponding to Union for International Cancer Control (UICC) stages I–III. An association of high SUVmax with rapid disease recurrence has often been described and usually has not had a direct influence on OS.
By definition, CUP syndrome is already a metastatic, UICC stage IV disease at initial diagnosis.
For other aggressive tumor entities at stage IV and with palliative therapeutic intention, observations comparable to our results are available in the literature. A higher OS was described for patients with small cell lung cancer at UICC stage IV under systemic palliative therapy with a higher SUVmax, whereas at stages I–III a lower SUVmax showed a better outcome (8). A possible interpretation of better OS with high SUVmax at stage IV of tumor disease, and thus also in CUP syndrome, was provided by Liu et al. in a study of ovarian cancer (9). They were able to show a higher chemosensitivity at a high SUVmax. In future studies, the marker SUVmax should be tested prospectively in patients with CUP syndrome.
CONCLUSION
Our retrospective study found that an SUVmax over 20 on 18F-FDG PET/CT is associated with favorable survival in patients with CUP at the initial diagnostic workup. This finding deserves further investigation in prospective studies. If they validate this finding, they could provide information that goes beyond staging to support the treatment of CUP. These results could optimize patient stratification in future studies of systemic therapies in CUP to increase the likelihood that a true therapeutic effect on survival will be detected.
DISCLOSURE
Wolfgang Fendler reports fees from SOFIE Biosciences, Janssen, Calyx, Bayer, Novartis, and Telix. Ken Herrmann reports fees from Bayer, SOFIE Biosciences, SIRTEX, Adacap, Curium, Endocyte, BTG, IPSEN, Siemens Healthineers, GE Healthcare, Amgen, Novartis, ymabs, Aktis Oncology, Theragnostics, Pharma15, Debiopharm, AstraZeneca, and Janssen and nonfinancial support from ABX. Manuel Weber reports fees from Boston Scientific, Terumo, Advanced Accelerator Applications, and Eli Lilly. Martin Metzenmacher reports honoraria from AMGEN, AstraZeneca, Boehringer-Ingelheim, BMS, MSD, Novartis, Roche, and Takeda. Marcel Wiesweg received honoraria from Amgen, Boehringer-Ingelheim, Daiichi Sankyo, GlaxoSmithKline, Novartis, Pfizer, Roche, and Takeda and research funding from BMS and Takeda. Michael Pogorzelski received honoraria from Boehringer-Ingelheim, BMS, Lilly, Merck Healthcare KGaA, Merck Sharp & Dohme, Roche, Sanofi-Aventis, and Servier. Stefan Kasper received honoraria from Merck Serono, MSD, Novartis, BMS, Amgen, Roche, Sanofi-Aventis, Servier, Incyte, and Lilly and research funding from Merck Serono, Lilly, BMS and Roche. Martin Schuler received research funding from AstraZeneca and BMS; received honoraria from Amgen, Boehringer-Ingelheim, BMS, Janssen, and Novartis; and had a compensated consultant role for Amgen, AstraZeneca, BIOCAD, Boehringer-Ingelheim, BMS, GlaxoSmithKline, Janssen, Merck Serono, Novartis, Roche, Sanofi, and Takeda. No other potential conflict of interest relevant to this article was reported.
KEY POINTS
QUESTION: Is a high SUVmax over 20 on pretherapeutic 18F-FDG PET/CT a prognostic factor in patients with CUP?
PERTINENT FINDINGS: We performed a retrospective analysis of 154 patients with CUP, 76 of whom received initial 18F-FDG PET/CT. Although other prognostic parameters were equivalent in the PET/CT and comparator sets, an SUVmax over 20 was associated with significantly superior OS in the PET/CT set.
IMPLICATIONS FOR PATIENT CARE: SUVmax should be used for patient stratification in future studies of CUP patients and can support shared decision-making regarding the initiation and type of therapy in current clinical practice.
Footnotes
Published online Jun. 15, 2023.
- © 2023 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- Received for publication November 8, 2022.
- Revision received April 4, 2023.
