• Home
  • Print this page
  • Email this page
Home About us Editorial board Ahead of print Current issue Search Archives Submit article Instructions Subscribe Contacts Login 

 Table of Contents  
Year : 2016  |  Volume : 2  |  Issue : 4  |  Page : 113-118

Carcinoma of unknown primary: 35 years of a single institution's experience

1 Department of Radiation Oncology, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia; Department of Oncology, Colchester Hospital University NHS Foundation Trust, Colchester, UK
2 Department of Radiation Oncology, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia; Department of Radiation Oncology, Princess Margaret Hospital, Toronto, Canada
3 Research Unit, Oncology Centre, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia
4 Department of Radiology, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia
5 Department of Medical Oncology, King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia

Date of Submission12-Jun-2016
Date of Acceptance20-Jul-2016
Date of Web Publication26-Aug-2016

Correspondence Address:
Rana I Mahmood
Department of Radiation Oncology, King Faisal Specialist Hospital and Research Centre, Riyadh 11211

Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/2395-3977.189304

Rights and Permissions

Aim: To assess the incidence and characteristics of carcinoma of unknown primary (CUP).
Methods: We retrospectively collected the clinical data of CUP cases treated in our medical center between 1975 and 2010, the results of which were statistically analyzed. Further, a comprehensive literature search, using PubMed database, on CUP cases was performed, and the results were discussed.
Results: A total of 1250 cases were identified, with male to female ratio of 1.2:1. Median age at diagnosis was 56 ± 16 years. Liver was the most common site of metastasis, followed by neck and peritoneum. Majority of cases (54.5%) received supportive treatment only, with the overall survival, over 1 year period, being 49%. Univariate analysis revealed pathology, gender, and site of disease as the significant predictors of survival, whereas pathology failed to reach significance on multivariate analysis, with “lymph nodes only” carrying the best prognosis.
Conclusion: Our data confirm the heterogeneity of CUP cases and variable treatment courses. This highlights the importance of establishing a national registry for this subgroup of cancer patients. Moreover, there is a need to develop multidisciplinary specialist teams and protocols to manage this group of patients, and participation in clinical trials should be strongly encouraged.

Keywords: Carcinoma of unknown primary, epidemiology, metastatic cancer, prognosis

How to cite this article:
Mahmood RI, Aldehaim M, Hussain F, Elhassan TA, Khan ZA, Memon MA. Carcinoma of unknown primary: 35 years of a single institution's experience. Cancer Transl Med 2016;2:113-8

How to cite this URL:
Mahmood RI, Aldehaim M, Hussain F, Elhassan TA, Khan ZA, Memon MA. Carcinoma of unknown primary: 35 years of a single institution's experience. Cancer Transl Med [serial online] 2016 [cited 2020 Aug 9];2:113-8. Available from: http://www.cancertm.com/text.asp?2016/2/4/113/189304

  Introduction Top

Carcinoma of unknown primary (CUP) is defined as histologically confirmed metastatic carcinoma in the absence of a detectable primary cancer after a thorough medical history, complete physical examination, and basic laboratory and radiological investigations.[1],[2] CUP is the seventh to eighth most frequently occurring cancer in the world and the fourth most common cause of cancer death in both males and females.[3] Even after a full diagnostic workup, the primary site of a CUP remains unknown in 20–50% of patients. Patients with unknown primary tumor represent 5–10% of all new cancer patients.[4],[5] In the United States, it has an incidence of 7–12 cases per 100,000 people per year, corresponding to 2–5% of all newly diagnosed malignancies,[6] in Australia, 18–19 cases,[7] and in the Netherlands, 5.3–6.7 cases.[8] According to tumor registry at King Faisal Specialist Hospital and Research Center (KFSH and RC), from 1975 to 2010, there were a total of 1250 cases of CUP, accounting for 1.8% of all malignancies. CUP is heterogeneous neoplasm that continues to be a diagnostic and therapeutic challenge for physicians. In most clinical studies, the prognosis of patients with CUP was poor. Median survival, from the time of diagnosis, ranged from 5 to 11 months with < 25% of patients surviving beyond 1 year.[9],[10] CUP is predominantly classified as adenocarcinoma (50–60%) or poorly differentiated adenocarcinoma or carcinoma (30–40%), with 5–8% of the CUP being squamous cell carcinoma and 2–5% undifferentiated malignancy.[3],[4],[11] With a median age of 65–70 years at presentation, CUP is extremely rare in children. A careful and comprehensive pathological examination of biopsied metastatic lesions is crucial for the diagnosis of CUP.[12],[13]

In this article, our objective is to review the incidence and characteristics of CUP in patients, previously diagnosed and managed in our hospital, and to determine the clinical course of the various neoplastic processes as well as the influence of tumor histology and management on patients' outcome.

  Methods Top

Institutional Review Board approval was obtained for the study. We retrospectively identified all cases of CUP who were diagnosed and managed at KFSH and RC, Riyadh, Saudi Arabia, between 1975 and 2010. Data were obtained from Saudi Cancer Registry data at KFSH and RC, including patients' demographic information - age at diagnosis, gender and clinical information - comorbidities, diagnostic evaluation, histological diagnosis, location of the metastasis, surgical procedures, and treatment received (chemotherapy and/or radiation treatment). Survival data were extracted from Saudi death registry through intermediate company that facilitated the process. Survival was calculated from the date of diagnosis, which is the date of the biopsy or hospitalization.

In this series, majority of patients had microscopically confirmed metastatic malignancies, whereas a few patients had a clinical diagnosis only. These later group of patients would normally fall outside the accepted definition of CUP but were included in this analysis to fully encompass CUP patients in actual clinical practice scenario.

We performed a comprehensive literature search using PubMed database to identify publications relevant to CUP, in English language, focusing on review of diagnosis, incidence, pathology, treatments, and outcomes. Search keywords included unknown primary tumor, epidemiology, metastatic cancer, prognosis, and survival.

Statistical analysis

Data were summarized using frequencies for categorical variables and medians for continuous variables. Overall survival (OS) probabilities were summarized using Kaplan–Meier method with variance estimated by Greenwood's formula. OS curves were compared using log-rank test. SPSS (version 20, IBM Co. Ltd, Armonk, USA) was used to analyze the demographic and survival data. Estimates of the survival distribution of patients were constructed using the Kaplan–Meier method. P < 0.05 was considered statistically significant.

Multivariate analysis was performed using Cox proportional hazard model. All variables were tested for affirmation of proportional hazard assumption using time-dependent approach. Factors that violated the proportional assumption were adjusted through stratification. A stepwise approach was then applied to an OS model. Interaction between the model covariates was tested and adjusted.

  Results Top

A total of 1250 cases were identified at KFSH and RC tumor registry, from 1975 to 2010, of which, 687 were males (55%) and 563 were females (45%) with male to female ratio of 1.2:1. Median age at diagnosis was 56 ± 16 years. Out of 1250 cases, 954 cases (76%) had a confirmed histological diagnosis, whereas remaining cases were confirmed radiologically. Metastasis was most frequently found in liver, followed by neck and peritoneum. The patients' demographic and histologic characteristics are listed in [Table 1].
Table 1. Demographic and histological characteristics of patients with unknown primary tumor

Click here to view

Majority of cases (55.8%) were diagnosed elsewhere and received treatment in KFSH and RC, 24.6% of cases were diagnosed and treated in KFSH and RC, and 17.7% of cases were diagnosed, and only part of their management took place in KFSH and RC. Pathological diagnosis was established in 954 (76%) cases, whereas remaining 24% had only clinical or radiological evidence of CUP. Among various established pathological diagnoses [Figure 1], adenocarcinoma not otherwise specified (NOS) and carcinoma NOS were the most common histological diagnoses. Incidence of CUP showed a steady rise over years until 2003 [Figure 2]. However, there has been a decline since 2003 till 2010, whereas total number of cancers seen has been rising over this period.
Figure 1. Types of pathology

Click here to view
Figure 2. Incidence of carcinoma of unknown primary by years

Click here to view

Majority of cases (54.5%) only received supportive treatment. Chemotherapy, radiotherapy, and surgery were provided in 13.9%, 17.7%, and 3.4% of cases, respectively [Figure 3]. Combined modality therapy comprising radiotherapy and chemotherapy, radiotherapy and surgery, surgery and chemotherapy was used in 3.3%, 3.8%, and 1.1% of patients, respectively. A total of 452 patients (36.2%) passed away while they were hospitalized.
Figure 3. Type of treatment

Click here to view

The OS for entire cohort at 1 year was 49% (standard deviation = 0.02) [Figure 4]. Two relevant questions were explored by subset analysis. First, if there was any statistical trend for OS over recent 5–10 years. In answering this, we found superimposable Kaplan–Meier survival curves [Figure 5] for patients diagnosed within last 5 years (2005–2010) vs. the patients diagnosed in a prior 5-year period with 1 year survival being 52% for either cohort. Second, if there was any association between OS and pathological diagnosis, which was confirmed as “yes” through univariate analysis, revealing that different pathological types had an impact on overall patient survival, P = 0.01 [Figure 6]a. Squamous cell carcinoma and neuroendocrine cancers showed best prognosis, with hazard ratio of 0.54 (0.3–0.8) and 0.51 (0.2–0.8), respectively, as compared to adenocarcinoma. It also revealed that the site of cancer and gender, favoring male patients, had significant impact on overall patient survival, P = 0.003 for both covariates. However, the type of treatment showed no significant effect on OS (P = 0.5) [Figure 6]b and [Table 2].
Figure 4. Overall survival

Click here to view
Figure 5. Overall survival in last 5 and 10 years

Click here to view
Figure 6. Overall survival by (a) pathology, (b) type of treatment, (c) site of cancer, and (d) gender

Click here to view
Table 2. Univariate analysis of overall survival

Click here to view

A multivariate model, adjusting for other significant covariates, revealed that the site of cancer [Figure 6]c and gender [Figure 6]d remained the only significant factors associated with overall patient survival, with P = 0.01 and P = 0.02, respectively. Pathological subtypes tended to be statistically insignificant (P = 0.11), and therefore removed from the final model [Table 3]. All covariates satisfied the proportional assumption, and there were no significant interactions between the variables.
Table 3. Multivariate analysis of overall survival using Cox regression model

Click here to view

  Discussion Top

CUP accounts for 5–10% of all new cancers worldwide.[4] In the present analysis, it accounts for only 1.8% of all cancer cases over a period of 35 years. Low recorded incidence at KFSH and RC could represent referral bias reflecting perceived lack of benefit from available therapeutic options. Moreover, patients may have been considered too unwell to be referred to a tertiary national center of excellence from other parts of the kingdom. However, it is apparent from our data that the number of registered cases of CUP at our institution was rising till 2003, followed by a steady decline since then [Figure 2]. Although it is hard to speculate the reason for this, one could attribute it to the increasing availability and utilization of advanced diagnostics and detailed pathology investigations, resulting in a significant improvement in detection of primary cancer. Furthermore, in the past 5 years, there has been an expansion of cancer treating units across the Kingdom of Saudi Arabia, and that could be another reason of recently noticed declining incidence. No epidemiological study exists verifying the geographical variations across the country and their impact on specific illness such as cancer. In our study, we found differences in the relative distribution of cases from various regions which might be due to a possible variation in the referral pattern to our tertiary cancer center.

Pathological confirmation is recommended where feasible as it could assist in proper diagnosis, help guide appropriate therapeutic intervention and better inform of the prognosis. In a recently published Spanish guideline for diagnosis and management of CUP, a serial approach from conventional microscopy, increasingly complex staining to electron microscopy/chromosomal studies are recommended.[14] In cases where ambiguity persists despite pathological diagnosis, clinical profiling could help assign appropriate prognostic group and treatment protocol. For example, many will consider poorly differentiated carcinoma with midline distribution as germ cell tumor and treat accordingly. Nevertheless, in most of the published studies, almost half of the registered patients with CUP did not have pathological confirmation.[15] Our data show that rather higher number of our patients (76%) had pathological diagnosis and only 24% of them had clinical/radiological diagnosis only. In this subgroup of patients, it is unclear of the reason for not performing pathological diagnosis, but it might be possible that patients were unfit for any intervention. In a recent report, 44% of patients, without histologically confirmed CUP, had supportive care alone.[16] It is also essential to spread the importance and the benefit of any pathological information gained through biopsy report. Modern pathological techniques incorporating immunohistochemical analysis and cytogenetic and molecular subtyping have potential to identify the primary diagnosis and provide predictive and prognostic information.[2] Predictive algorithms (using large-scale RNA profiling and oligonucleotide microarray gene expression technology) have been used to identify 78–90% of primary tumors in some series.[15],[17]

Distribution of pathology subtypes in our study is consistent with other published studies.[11] In our study, out of all histology confirmed cases, 82.7% cases were adenocarcinoma or carcinoma NOS, whereas squamous cell carcinoma accounted for 8.4% and neuroendocrine tumors were 3.7%. In a recent report published from Cardiff, UK,[16] adenocarcinoma and carcinoma NOS were noted to account for 83% of all pathology confirmed cases.

In several publications, median survival from the time of diagnosis ranged from 5 to 11 months, and < 25% of patients survived beyond 1 year.[9],[10] In our study, median survival time reached 12 months with half of patients being still alive at 1 year. It is difficult to speculate the reasons although it could be statistically spread. However, different case mix in the local population cannot be ruled out. Our data are unique as there are no published comparative data related to diagnosis and outcome of CUP for population in this region. Female gender and chemotherapy treatment have been reported as favorable prognostic indicators.[16] However, we did not find any association between type of therapy and survival. Moreover, in our analysis, male patients carried better prognosis along with the “disease limited to lymph nodes.”

To the best of our knowledge, this is the first report of CUP pertaining to Arabian population. Our data confirm the heterogeneity of CUP cases and variable treatment courses. This highlights the importance of establishing a national registry for this subgroup of cancer patients. Moreover, there is a need to develop multidisciplinary specialist teams and protocols to manage this group of patients.[14] In future, novel diagnostic techniques may help classify cases into assigned primary sites with specific management protocols. Nevertheless, it is unclear if currently available therapeutic options would alter the natural course of CUP and affect the outcome. Therefore, participation in clinical trials should be strongly encouraged.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

  References Top

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 (8): 2094–103.  Back to cited text no. 1
Varadhachary GR, Abbruzzese JL, Lenzi R. Diagnostic strategies for unknown primary cancer. Cancer 2004; 100 (9): 1776–85.  Back to cited text no. 2
Krementz ET, Cerise EJ, Foster DS, Morgan LR Jr. Metastases of undetermined source. Curr Probl Cancer 1979; 4 (5): 4–37.  Back to cited text no. 3
Greco FA, Hainsworth JD. Cancer of unknown primary site. In: Vita VD Jr., Hellman S, Rosenberg S, Markoe AM, editors. Cancer: Principles and Practice of Oncology. 8th ed. Philadelphia: Lippincott; 2008. p. 2363–87.  Back to cited text no. 4
Lembersky BC, Thomas LC. Metastases of unknown primary site. Med Clin North Am 1996; 80 (1): 153–71.  Back to cited text no. 5
Muir C. Cancer of unknown primary site. Cancer 1995; 75: 353–6.  Back to cited text no. 6
Currow D, Thomson W, Lu H. Cancer in New South Wales: Incidence and Mortality Report 2010. Sydney: Cancer Institute NSW; 2015. p. 32.  Back to cited text no. 7
The Netherlands Cancer Registry; 2015. Available from: . [Last accessed on 2016 Aug 4].  Back to cited text no. 8
Abbruzzese JL, Abbruzzese MC, Hess KR, Raber MN, Lenzi R, Frost P. Unknown primary carcinoma: natural history and prognostic factors in 657 consecutive patients. J Clin Oncol 1994; 12 (6): 1272–80.  Back to cited text no. 9
Altman E, Cadman E. An analysis of 1539 patients with cancer of unknown primary site. Cancer 1986; 57 (1): 120–4.  Back to cited text no. 10
van de Wouw AJ, Janssen-Heijnen ML, Coebergh JW, Hillen HF. Epidemiology of unknown primary tumours; incidence and population-based survival of 1285 patients in Southeast Netherlands, 1984-1992. Eur J Cancer 2002; 38 (3): 409–13.  Back to cited text no. 11
Pavlidis N, Fizazi K. Carcinoma of unknown primary (CUP). Crit Rev Oncol Hematol 2009; 69 (3): 271–8.  Back to cited text no. 12
Bugat R, Bataillard A, Lesimple T, Voigt JJ, Culine S, Lortholary A, Merrouche Y, Ganem G, Kaminsky MC, Negrier S, Perol M, Laforêt C, Bedossa P, Bertrand G, Coindre JM, Fizazi K; FNCLCC. Summary of the standards, options and recommendations for the management of patients with carcinoma of unknown primary site (2002). Br J Cancer 2003; 89 Suppl 1: S59–66.  Back to cited text no. 13
Collado Martín R, García Palomo A, de la Cruz Merino L, Borrega García P, Barón Duarte FJ; Spanish Society for Medical Oncology. Clinical guideline SEOM: cancer of unknown primary site. Clin Transl Oncol 2014; 16 (12): 1091–7.  Back to cited text no. 14
Su AI, Welsh JB, Sapinoso LM, Kern SG, Dimitrov P, Lapp H, Schultz PG, Powell SM, Moskaluk CA, Frierson HF Jr., Hampton GM. Molecular classification of human carcinomas by use of gene expression signatures. Cancer Res 2001; 61 (20): 7388–93.  Back to cited text no. 15
Shaw PH, Adams R, Jordan C, Crosby TD. A clinical review of the investigation and management of carcinoma of unknown primary in a single cancer network. Clin Oncol (R Coll Radiol) 2007; 19 (1): 87–95.  Back to cited text no. 16
Ramaswamy S, Tamayo P, Rifkin R, Mukherjee S, Yeang CH, Angelo M, Ladd C, Reich M, Latulippe E, Mesirov JP, Poggio T, Gerald W, Loda M, Lander ES, Golub TR. Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci U S A 2001; 98 (26): 15149–54.  Back to cited text no. 17


  [Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]

  [Table 1], [Table 2], [Table 3]


Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
Article Figures
Article Tables

 Article Access Statistics
    PDF Downloaded284    
    Comments [Add]    

Recommend this journal