Outcomes of patients with confirmed cancer of unknown primary: a 10-year analysis


Session type:

Aoife Gatenby1, Mark Stares1, Karin Purshouse2, Gillian Knowles1, Rachel Haigh1, Jenny Irvine1, Rebekah Patton1, Jo McGinty1, Alan Christie1, Marjory MacLennan1, Colin Barrie1, Sally Clive1
1Edinburgh Cancer Centre, 2University of Edinburgh



Patients with CUP have pathological evidence of malignancy but no confirmation of a primary site, despite a comprehensive diagnostic work-up. Accounting for only 2-3% of cancers, CUP is the 5th greatest cause of cancer death in the UK. Established ESMO and NICE guidelines have standardised investigation and management of CUP but “real-world” outcomes are not well defined.


Prospective data collection was undertaken for all patients with a diagnosis of confirmed CUP (cCUP) referred to the Edinburgh Cancer Centre CUP Service between 01/09/2010 and 31/08/2020. Patient demographics, clinical, radiological, pathological and treatment data were recorded.


Data were available for 301 patients. Median survival was 4.0 (IQR 2.0-10.0) months. 75% (n=227) had ‘poor prognosis cCUP’ (PP-cCUP). ‘Best fit’ tissue of origin, after comprehensive investigations including immunohistochemistry analyses, showed that 47% (n=142) had features consistent with hepatobiliary or pancreatic cancer and 22% (n=65) were poorly/undifferentiated cancers. 25% (n=74) had features that classified them as ‘favourable prognosis cCUP’ (FP-cCUP). Median survival of patients with FP-cCUP was 5.5 (IQR 3.2-16.4) months compared to 3.7 (IQR 1.8-8.5) months in those with PP-cCUP (p<0.001). 23% (n=69) received palliative radiotherapy. 43% (n=32) of patients with FP-cCUP and 24% (n=55) with PP-cCUP received systemic anticancer therapy (SACT). 84% (n=73) received 1st-line platinum-based doublet chemotherapy. 31% (n=27) received more than one line of SACT. Survival was improved in all patient groups who received SACT. Amongst patients treated with SACT there was no difference in survival between those with FP-cCUP and PP-cCUP (12.6 (IQR 5.3-16.6) v 10.1 (IQR 5.8-14.9) respectively (p=0.101)).


Outcomes for patients with cCUP in this real-world cohort are worse than previously described. We confirm that patients with FP-cCUP have better survival than those with PP-cCUP. Only a minority of patients with cCUP receive SACT. A higher proportion of patients with FP-cCUP received SACT, perhaps reflecting better treatment options for these patients. However, there was no difference in survival between those patients with FP-cCUP and PP-cCUP who received SACT.

Impact statement

This data highlights the importance of careful selection of cCUP patients who are suitable for treatment rather than being guided only by clinico-pathological subgroup.