Somatic POLE proofreading domain mutation, immune response, and prognosis in colorectal cancer: a retrospective, pooled biomarker study

Enric Domingo, Luke Freeman-Mills, Emily Rayner, Mark Glaire, Sarah Briggs, Louis Vermeulen, Evelyn Fressler, Jan Paul Medema, Arnoud Boot, Hans Morreau, Tom van Wezel, Gerrit-Jan Liefers, Ragnhild A. Lothe, Stine A. Danielsen, Anita Sveen, Arild Nesbakken, Inti Zlobec, Alessandro Lugli, Viktor H. Koelzer, Martin D. BergerSergi Castellvi-Bel, Jenifer Muñoz, The EPICOLON Consortium, Marco de Bruyn, Hans W. Nijman, Marco Novelli, Kay Lawson, Dahme Oukrif, Eleni Frangou, Peter Dutton, Sabine Tejpar, Mauro Delorenzi, Rachel Kerr, Ian Tomlinson, David Church, David James Kerr

Research output: Contribution to journalArticlepeer-review

124 Citations (Scopus)


Background: Precision cancer medicine depends on defining distinct tumour subgroups using biomarkers that may occur at very modest frequencies. One such subgroup comprises patients with exceptionally mutated (ultramutated) cancers caused by mutations that impair DNA polymerase epsilon (POLE) proofreading.
Methods: We examined the association of POLE proofreading domain mutation with clinicopathological variables and immune response in colorectal cancers from clinical trials (VICTOR, QUASAR2, and PETACC-3) and colorectal cancer cohorts (Leiden University Medical Centre 1 and 2, Oslo 1 and 2, Bern, AMC-AJCC-II, and Epicolon-1). We subsequently investigated its association with prognosis in stage II/III colorectal cancer by Cox regression of pooled individual patient data from more than 4500 cases from these studies.
Findings: Pathogenic somatic POLE mutations were detected in 66 (1·0%) of 6517 colorectal cancers, and were mutually exclusive with mismatch repair deficiency (MMR-D) in the 6277 cases for whom both markers were determined (none of 66 vs 833 [13·4%] of 6211; p<0·0001). Compared with cases with wild-type POLE, cases with POLE mutations were younger at diagnosis (median 54·5 years vs 67·2 years; p<0·0001), were more frequently male (50 [75·8%] of 66 vs 3577 [55·5%] of 6445; p=0·0010), more frequently had right-sided tumour location (44 [68·8%] of 64 vs 2463 [39·8%] of 6193; p<0·0001), and were diagnosed at an earlier disease stage (p=0·006, χ2 test for trend). Compared with mismatch repair proficient (MMR-P) POLE wild-type tumours, POLE-mutant colorectal cancers displayed increased CD8+ lymphocyte infiltration and expression of cytotoxic T-cell markers and effector cytokines, similar in extent to that observed in immunogenic MMR-D cancers. Both POLE mutation and MMR-D were associated with significantly reduced risk of recurrence compared with MMR-P colorectal cancers in multivariable analysis (HR 0·34 [95% CI 0·11–0·76]; p=0·0060 and 0·72 [0·60–0·87]; p=0·00035), although the difference between the groups was not significant.
Interpretation: POLE proofreading domain mutations identify a subset of immunogenic colorectal cancers with excellent prognosis. This association underscores the importance of rare biomarkers in precision cancer medicine, but also raises important questions about how to identify and implement them in practice.
Funding: Cancer Research UK, Academy of Medical Sciences, Health Foundation, EU, ERC, NIHR, Wellcome Trust, Dutch Cancer Society, Dutch Digestive Foundation.
Original languageEnglish
Pages (from-to)207-216
JournalThe Lancet Gastroenterology & Hepatology
Issue number3
Early online date3 Aug 2016
Publication statusPublished - Nov 2016


Dive into the research topics of 'Somatic POLE proofreading domain mutation, immune response, and prognosis in colorectal cancer: a retrospective, pooled biomarker study'. Together they form a unique fingerprint.

Cite this