Polymerase Proofreading-Associated Polyposis (POLE/POLD1)

Autosomal dominant syndrome predisposing to colorectal adenomas and carcinomas, endometrial cancer and other tumours, caused by specific missense variants in the exonuclease (3′→5′ proofreading) domains of deoxyribonucleic acid (DNA) polymerases ε and δ; ultramutated tumours, generally microsatellite-stable.

Literature and regulatory information updated: 6 October 2026.

0. Initial Reference Table

1. Introduction

Colorectal adenomatous polyposis is genetically heterogeneous. The Oxford Desk Reference notes that up to 30% of the general population develops one or two adenomas by age 70, but only approximately 1 in 1000 people develops more than two, and that multiple adenomas are empirically associated with a higher risk of CRC; it also records that 20–30% of cases and families with fewer than 100 adenomas remained without a molecular diagnosis after APC and MUTYH testing [31]. Polymerase proofreading-associated polyposis was identified in this group of patients with dozens of adenomas, often before age 60, with or without a family history [32].

The founding paper was published by Claire Palles, Ian Tomlinson and colleagues in Nature Genetics in 2013. Within the WGS500 project (Oxford-Illumina), the authors performed whole-genome sequencing in 15 probands with at least 10 colorectal adenomas diagnosed by age 60, in whom known syndromes had previously been excluded. In family SM2702, combined linkage and sequencing data pointed to a single variant shared by all affected individuals, POLE L424V (c.1270C>G). In a validation phase involving 3805 British patients with CRC enriched for family history, multiple adenomas or early onset and 6721 controls, 12 additional unrelated carriers and no controls were identified (P=2.46×10⁻⁶); the variant was also absent from 10,755 exomes in the NHLBI GO Exome Sequencing Project (National Heart, Lung, and Blood Institute Grand Opportunity). In family SM1645 and a third proband, the same strategy identified POLD1 S478N (c.1433G>A), occupying the position homologous to L424 in polymerase ε. Tumours were microsatellite-stable, with chromosomal instability in carcinomas and an excess of base substitutions; the variant equivalent to S478N in Schizosaccharomyces pombe increased the mutation rate 12-fold [32].

The authors immediately described two features that continue to define the condition: phenotypic variability, with some patients showing a multiple-adenoma phenotype resembling MUTYH-associated polyposis and others having large adenomas or early CRC resembling Lynch syndrome; and endometrial cancer, observed in seven POLD1 carriers and in none of the 13 POLE families in the initial series. They also proposed that, pending further data, carriers should be managed using an approach intermediate between Lynch syndrome and MUTYH-associated polyposis, with regular colonoscopy and polypectomy and consideration of prophylactic surgery [32]. The Oxford Desk Reference incorporated the condition under its own subheading in the adenomatous polyposis chapter, emphasising that the variants map to equivalent sites in the proofreading domains of polymerases ε and δ and are predicted to impair correction of mismatched bases during replication [31].

The following decade brought four changes. The first was expansion of the tumour spectrum to include duodenal, ovarian and breast cancer and brain tumours, documented in Spanish and British series and a review of 132 heterozygotes published in 2022 [7, 33]. The second was quantification of CRC risk, still imprecise, with estimates differing between clinical cohorts and ascertainment-adjusted segregation analyses [7, 12]. The third was demonstration that only missense variants (or in-frame alterations) in the exonuclease domain cause the syndrome, whereas loss-of-function variants and variants outside the domain do not, leading to specific classification rules based on ACMG/AMP (American College of Medical Genetics and Genomics/Association for Molecular Pathology) criteria [2, 13]. The fourth was recognition that some POLE variants produce a phenotype resembling constitutional mismatch repair deficiency (CMMRD, constitutional mismatch repair deficiency) in childhood [10, 34, 35].

Clinical management draws on documents with different scopes and levels of evidence: the 2024 NCCN (National Comprehensive Cancer Network) guidelines, the 2024 joint European Hereditary Tumour Group (EHTG) and European Society of Coloproctology (ESCP) guidelines, the 2019 ESMO (European Society for Medical Oncology) guidelines, the 2020 British BSG/ACPGBI/UKCGG guidelines, ClinGen actionability assessment AC152 and recommendations from the 2022 British cohort [7, 8, 17, 18, 21, 22]. The 2019 European Society of Gastrointestinal Endoscopy (ESGE) guideline on endoscopic management of polyposis explicitly excluded POLE/POLD1 from its scope because of low prevalence [36]. Familial adenomatous polyposis (FAP) and Lynch syndrome feature alongside MUTYH-associated polyposis (MAP) and NTHL1 tumour syndrome as differential diagnoses.

MBD4 tumour predisposition syndrome shares with PPAP adenomatous polyposis in young adults, CRC and tumours hypermutated through failure of a DNA fidelity mechanism, but it has recessive inheritance and an extracolonic spectrum dominated by AML. In 2018, Rodrigues and colleagues identified germline truncating MBD4 variants, with somatic inactivation of the wild-type allele, in a patient with metastatic uveal melanoma and an exceptional response to a PD-1 (programmed cell death protein 1) inhibitor, whose tumour carried a disproportionate burden of CpG>TpG transitions [37]. In the same year, Sanders and colleagues showed, from three early-onset AMLs, that germline loss of MBD4 initiates 5-methylcytosine-dependent hypermutation and predisposes to clonal haematopoiesis and AML [38]. In 2022, Palles and colleagues, through genome or exome sequencing in 309 patients from 198 families with multiple adenomas or familial CRC without a molecular diagnosis and targeted sequencing in a further 1,611 patients, demonstrated that the same deficiency causes colorectal adenomatous polyposis associated with AML, uveal melanoma and other tumours, and proposed the name MANS [28]. OMIM lists the entity as tumour predisposition syndrome 2 (619975), Orphanet as MBD4-related tumor predisposition syndrome (ORPHA:661526), and GeneReviews published its first chapter in August 2026 under the name MBD4-TPDS [26, 27]. The ESMO Handbook includes biallelic MBD4 among the recessive CRC and polyposis syndromes caused by a defect in a base excision repair gene, alongside MUTYH and NTHL1, with AML and uveal melanoma as extracolonic tumours [29].

2. Molecular Genetics and Pathophysiology

2.1 Genes, enzymes and domains

Replication of the eukaryotic genome depends on polymerases α, ε and δ; only the latter two have an active 3′→5′ exonuclease domain, which removes incorrectly incorporated nucleotides from the newly synthesised strand. In humans, polymerases ε (Polε) and δ (Polδ) each have four subunits, and the largest, containing the polymerisation and exonuclease domains, is encoded by POLE and POLD1, respectively [2]. Polε primarily synthesises the leading strand and Polδ the lagging strand; Polε proofreading can correct its own errors and mismatches produced by polymerases lacking proofreading, such as Polα, and Polδ also participates in mismatch repair and base excision repair pathways [32]. According to GeneGuidance, Polδ also contributes to nucleotide excision repair synthesis, break-induced replication and translesion synthesis, and requires POLD2, POLD3 and POLD4 for full activity [3].

POLE and POLD1 belong to the family B polymerases and have their greatest homology precisely in the exonuclease domain (23% identity and 37% similarity between residues 223–517 of POLE and 245–571 of POLD1) [32]. For diagnostic purposes, the 2023 classification recommendations and NCCN guidelines define the exonuclease domain as amino acids 268–471 of POLE (exons 9–14) and 304–533 of POLD1 [2, 8]. Within the domain, the Exo motifs concentrate the catalytic and DNA-binding residues; the catalytic residues are D275 and E277 in POLE and D316 and E318 in POLD1 [2]. For POLD1, a 2025 report defines motifs ExoI (amino acids 312–326), ExoII (393–406), ExoIV (470–478), ExoV (485–493) and ExoIII (506–519) [39].

2.2 Disease mechanism: lost proofreading, preserved polymerisation

The causal variants are exonuclease domain missense variants that leave the polymerase able to synthesise DNA but unable to excise misincorporated bases. The 2022 British cohort frames the problem by predicting that proofreading loss increases point mutations approximately 100-fold, theoretically in all dividing cell types [7]. In the founding paper, POLE L424 maps directly to the exonuclease active site (L479 in yeast polymerase δ, where the yeast L479S substitution, with no described human germline equivalent, produces a mutator phenotype comparable to the pol3-01 allele, with less than 0.5% exonuclease activity); structural modelling placed L424 approximately 5 Å from single-stranded DNA and showed that substitutions at POLE 424 and POLD1 478 alter packing of two helices at the base of the active site [32]. Variants interfering with DNA binding, such as those in the binding cleft, are highly likely to be pathogenic, but the 2020 Spanish series showed that the yeast assay also detects mutator effects in non-pathogenic cleft variants, such as p.(Asp287Glu), limiting its weight as functional evidence [13].

The result is an ultramutated phenotype. In tumours with Polε proofreading deficiency and proficient mismatch repair (MMR, mismatch repair), the mean mutational burden was 144 mutations/Mb (2.6–325), and 255 mutations/Mb (109–531) when MMR deficiency coexisted; only two tumours, both with POLE p.(Leu424Val), had fewer than 25 mutations/Mb. The spectrum is dominated by C>A transversions in a TCT context and C>T transitions in a TCG context, corresponding to COSMIC (Catalogue Of Somatic Mutations In Cancer) signatures SBS10a, SBS10b and SBS28; for Polδ, the specific signatures are SBS10c and SBS10d. When proofreading deficiency coexists with MMR deficiency, SBS14 (Polε) and SBS20 (Polδ) arise. In an analysis of 134 tumours, SBS10a, SBS10b, SBS28 and SBS14 were specific to Polε deficiency, whereas SBS20 also appeared in 15 MMR-deficient tumours without an exonuclease domain variant, leading the authors to advise against its use in variant classification [2]. Three tumours from germline POLD1 carriers (p.(Leu474Pro), p.(Asp316His) and p.(Ser478Asn)) had 59, 114 and 36 mutations/Mb, a 34–68% contribution from SBS10d and copy-number-neutral loss of heterozygosity in the POLD1 region [2].

The involvement of the second allele differs between genes. In POLE, tumour, normal tissue and yeast data indicate that proofreading deficiency is haploinsufficient, meaning that inactivation of the wild-type allele is not required to initiate hypermutation and adenomas; in POLD1, evidence indicates haplosufficiency, requiring loss of heterozygosity or MMR deficiency for hypermutation to arise [9, 41]. In the founding paper, the authors had already observed loss of heterozygosity involving the wild-type allele in some adenomas and carcinomas, but not all, and noted that the L424V and S478N alleles produce stable messenger ribonucleic acid (RNA), that there was no evidence that truncating variants predisposed to tumours and that complete loss of the enzyme is probably lethal; they proposed that pathogenicity depended on coupling functional polymerisation to non-functional proofreading within the same molecule [32]. In mice, murine Pole D272A/E274A variants (experimental substitutions in the mouse protein, with no correspondence to described human variants) near the exonuclease active site cause a mutator phenotype in both heterozygous and homozygous states, with intestinal tumours in homozygotes, whereas the murine Pold1 D400A variant mainly causes skin tumours [32].

The carcinogenic pathway follows, at least partly, the classical adenoma–carcinoma sequence: tumours in L424V and S478N carriers contained mutations in APC, KRAS, BRAF, PIK3CA and FBXW7, all base substitutions, with hotspots uncommon in sporadic tumours (codons 1114 and 1338 of APC and 146 of KRAS) [32]. A proportion of tumours acquire MMR deficiency through biallelic somatic mutations in MMR genes, explaining why some CRCs in patients with PPAP are unstable and may be mistaken for Lynch syndrome [9, 23]. The Oxford Desk Reference summarises the same phenomenon by noting that many tumours with microsatellite instability and abnormal immunohistochemistry, without a constitutional MMR variant, result from somatic MMR mutations caused by constitutional variants in other genes, such as MUTYH and POLD1/POLE [42].

2.3 Pathogenic variant spectrum

Variants causing PPAP are missense variants (and, theoretically, in-frame alterations or in-frame splicing alterations in the exonuclease domain); loss-of-function variants are not associated with PPAP, unlike most cancer predisposition genes [9]. The Spanish series of 2,813 probands identified 12 exonuclease domain missense variants, six loss-of-function variants and 23 missense variants outside the domain predicted to be deleterious; it concluded that loss-of-function variants and variants outside the domain are probably not pathogenic for the syndrome [13]. For a variant to be considered “likely pathogenic”, the 2022 British cohort required cosegregation through more than four meioses, hypermutation in yeast with the equivalent substitution or biochemical evidence of impaired proofreading, without contradictory data [7]. A 2025 report described a variant classified as likely pathogenic outside the exonuclease domain, in the polymerisation domain (POLD1 c.1816C>A p.(Leu606Met)), arising de novo in a 22-year-old woman with polyposis, multiple hypermutated CRCs and glioblastoma, and recommended whole-gene analysis in suspected PPAP; this is an isolated observation [39].

2.4 Genotype–phenotype correlations

In the British cohort, CRC incidence did not clearly differ between heterozygotes for p.(Leu424Val) and other POLE variants, and the authors applied the same recommendations to all patients [7]. The Australian segregation analysis estimated CRC risk by age 70 for p.(Leu424Val) at 97% in men and 92% in women, far above the 28% and 21% estimated for all POLE variants combined, which the authors interpreted as a possible difference in risk between variants [12, 43]. However, some variants are associated with more severe and earlier phenotypes: POLE p.(Glu277Gly), p.(Ser297Phe), p.(Val411Leu), p.(Pro436Arg), p.(Met444Lys), p.(Ala456Pro) and p.(Ser461Thr) have been reported in patients with childhood medulloblastoma, polyposis and CRC in adolescence, café-au-lait macules and pilomatricomas, mimicking CMMRD; coexistence of a pathogenic exonuclease domain variant with a constitutional MMR variant produces similar presentations [9, 41]. The family with POLE p.(Asn363Lys) accounted for five primary brain tumours, four of them glioblastomas between ages 30 and 52 [7]. Between genes, the most consistent differences are the higher endometrial cancer risk with POLD1 and the presence of ovarian and duodenal cancer only with POLE in the British series, although duodenal cancer was subsequently reported in POLD1 carriers [7, 39].

2.5 Allelic disorders

POLE and POLD1 variants with other mechanisms cause disorders without an established predisposition to polyposis. FILS syndrome (OMIM 615139), described in 2012 in 11 members of a large consanguineous French family, combines mild facial dysmorphism, immunodeficiency with recurrent respiratory infections and meningitis from the first year of life, livedo and short stature; all affected individuals were homozygous for a substitution at position +3 of intron 34 (described as g.G4444+3A>G), generating a transcript lacking exon 34 and reducing the wild-type transcript by approximately 90%, and had no cancer susceptibility [14]. IMAGe-I syndrome (OMIM 618336) results from biallelic POLE variants identified in 15 individuals from 12 families, all carrying the splice-altering intronic variant NM_006231.4:c.1686+32C>G in trans with loss-of-function variants; features include intrauterine growth restriction, metaphyseal dysplasia, congenital adrenal hypoplasia, genital anomalies in boys, facial dysmorphism and variable immunodeficiency [5, 15]. MDPL syndrome (mandibular hypoplasia, deafness, progeroid features and lipodystrophy, OMIM 615381) is caused by the recurrent in-frame deletion POLD1 c.1812_1814del p.(Ser605del) in motif A of the polymerisation active site, abolishing polymerisation while reducing exonuclease activity only 1.7–2.5-fold, the reverse of the PPAP mechanism; none of the four patients in the original series had cancer [5, 16]. ClinGen classifies the POLD1–MDPL relationship as definitive (20 Jan 2025) and recessive non-severe combined immunodeficiency due to polymerase δ deficiency as limited (19 Feb 2026) [11].

2.6 Somatic exonuclease domain variants

Somatic exonuclease domain variants in POLE occur in 7–15% of endometrial cancers and 0.5–8% of colorectal tumours, and less often in gliomas, other gastrointestinal cancers and other tumour types; somatic POLD1 variants are extremely rare [2]. In TCGA (The Cancer Genome Atlas) data analysed in 2025, 75% of proofreading-deficient cancers were endometrial, whereas CRC predominated in germline carriers (56% of carriers) [41]. In the original analysis of 224 TCGA CRC exomes, all seven hypermutated, microsatellite-stable or low-instability tumours with a POLE mutation had an exonuclease domain variant [32]. In metastatic CRC, the ESMO Handbook estimates POLE mutation prevalence at less than 1% [44]. This distinction matters: a variant identified in a tumour may be somatic, and only germline testing establishes the syndrome [3, 40].

2.7 MBD4 tumour predisposition syndrome (TPDS2): gene, mechanism and variants

The MBD4 gene (3q21.3; OMIM 603574) encodes methyl-CpG-binding domain protein 4, a DNA glycosylase of the base excision repair (BER) pathway that recognises and removes thymine paired with guanine (G:T mismatch) resulting from spontaneous deamination of 5-methylcytosine at CpG dinucleotides [26, 28]. Without this repair, each uncorrected deamination is converted, at the next replication, into a C>T transition fixed in the CpG context. Biallelic loss of function thus produces a mutator phenotype with an excess of CpG>TpG transitions. In the AMLs of three patients with germline deficiency, the mutational burden was 33 times that of AML in general, and more than 95% of mutations were C>T at CG dinucleotides [38]. In 19 colorectal adenomas from two patients, more than 95% of the excess mutations were CpG>TpG transitions, which accumulated linearly with age, and more than 96% of the mutated sites were methylated in normal colonic mucosa; the median burden of fresh-frozen adenomas from one of the patients was 11.1 mutations/Mb (range 8.5–23.3), compared with 1.8 (1.0–3.1) in nine sporadic adenomas [28]. The 2022 authors described the signature as very close to COSMIC (Catalogue of Somatic Mutations in Cancer) SBS1; GeneReviews and later studies associate MBD4 deficiency with signature SBS96 [26, 28, 45]. The difference from PPAP lies in mechanism and dosage: in PPAP, a dominant missense variant abolishes polymerase proofreading and generates signatures SBS10 and SBS28; in MBD4 syndrome, loss of both alleles is required, and the spectrum is restricted almost entirely to C>T transitions at CpG [2, 28].

The driver genes of the adenomas are those of sporadic adenomas, but with mutations characteristic of the mechanism. All sequenced adenomas had somatic APC variants, with enrichment of the CpG>TpG transition NM_000038.6:c.4348C>T p.(Arg1450Ter); AMER1 mutations were more frequent (12 of 19 adenomas) and KRAS mutations less frequent (3 of 19) than in sporadic adenomas, and 88% of driver mutations were CpG>TpG transitions, compared with 37% in sporadic adenomas [5, 28]. AMLs in patients with the syndrome frequently have biallelic DNMT3A variants and IDH1 or IDH2 hotspot variants, a combination that is rare (<3%) in sporadic AML; serial sequencing showed repeated expansions of progenitors with DNMT3A variants, linking clonal haematopoiesis to leukaemia [26, 38]. The three vestibular schwannomas studied had biallelic somatic inactivation of NF2, the same initiating event as in sporadic schwannomas [26]. In mice, biallelic inactivation of Mbd4 alters the mutational spectrum and modifies tumour predisposition, but does not by itself cause cancer predisposition [26].

The mechanism is loss of function. The variants described are truncating, splice-site or, more rarely, in-frame deletions; no exon deletions or duplications have been described, and sequencing has so far identified all known pathogenic variants [26]. There are no clinically relevant genotype–phenotype correlations. No individual with biallelic variants and no tumours has been described either, but the number of patients is too small to claim complete penetrance [26]. No allelic disorders are known apart from the predisposition of heterozygotes to uveal melanoma [26]. The literature uses two transcripts: the 2018–2022 series number variants on NM_003925.2, whereas the MANE Select transcript is NM_001276270.2, which differs by 18 nucleotides from exon 7 onwards; the correspondences in the following table were confirmed in VariantValidator [5].

Heterozygotes do not have the syndrome, but they are not neutral. In a consecutive series of 1,093 patients with uveal melanoma treated at the Institut Curie between 2013 and 2018, eight had deleterious MBD4 variants, corresponding to an incidence 9.15 times that of the general population (95% confidence interval, 95% CI: 4.24–19.73); all deficient tumours studied had a high mutational burden, with 85.7–96.4% CpG>TpG transitions [48]. The extended series from the same centre (896 patients, 2021–2025) found 23 carriers, with a relative risk of 31.44 (95% CI: 18.18–53.00) compared with non-Finnish Europeans in gnomAD (Genome Aggregation Database) [49]. Among 25 families with monoallelic variants, 18 were identified through uveal melanoma and 7 through breast cancer, with other cancers (breast, kidney, colorectal) in relatives, suggesting a broader spectrum not yet quantified [50]. There is no evidence of increased CRC or polyposis risk in heterozygotes: the excess of carriers observed in the 2022 replication phase (4/1,611 versus 48/64,600 in gnomAD; p=0.0381) was not confirmed in the 100,000 Genomes Project or in UK Biobank, and screening of 728 patients with CRC, polyposis or suggestive phenotypes did not associate heterozygous variants with the disease [26, 28, 51]. In heterozygous carriers of the Israeli founder variant, the proportion of C>T transitions at CpG in blood was higher than in controls (41% versus 31% and 28%), and 3 of 11 carriers, including a 14-year-old girl, had clonal haematopoiesis; the clinical significance of these findings is not established [45].

3. Clinical Spectrum and Manifestations by System

3.1 Sources of clinical data and their biases

Almost everything known about the natural history of PPAP comes from families ascertained in cancer genetics clinics, patients in CRC trials and literature reviews. The 2022 British cohort brought together 132 heterozygotes for exonuclease domain variants considered likely pathogenic (105 POLE, 27 POLD1), from the CORGI study (COloRectal tumour Gene Identification), the VICTOR and QUASAR2 clinical trials (patients with stage II/III CRC treated with curative intent) and a systematic review; cumulative incidence curves were not adjusted for surveillance or surgery, and the authors caution about the small number of families, variable extent of screening and ascertainment bias [7]. The 2023 review accompanying the classification rules totalled 205 heterozygotes for 17 pathogenic or likely pathogenic variants [2], and a 2025 analysis included 70 families and 249 individuals [41]. These denominators measure frequencies in clinical series rather than population risks.

3.2 Colon and rectum

The colorectal phenotype is the most consistent. In the British cohort, 95% (100/105) of POLE heterozygotes and 23 of 27 POLD1 heterozygotes had adenomas, carcinoma or both; polyp burden, known in 59 POLE carriers, ranged from 1 to 100 (median 12), and in 17 POLD1 carriers from 2 to 70 (median 13) [7]. Most polyps were tubular or tubulovillous adenomas, and nine POLE carriers and five POLD1 carriers also had hyperplastic polyps [7]. The term “polyposis” should be interpreted cautiously: 30% (26/86) of CRCs occurred in patients without recorded adenomas, and no clear correlation between polyp burden and cancer risk was demonstrated, so a minority have a Lynch-like phenotype, with early CRC and few polyps [7]. The 2023 review found more than 10 gastrointestinal polyps in 64% (70/108) of carriers with available information; among 35 cancer-free carriers with polyp data, 34 had polyps and, of the 25 with detailed counts, 15 had at least 10 (median age 35 years; 15–53) [2]. The Lynch syndrome GeneReview describes PPAP as having 0 to 100 cumulative adenomas [23].

CRC was diagnosed in 70 of 105 POLE carriers according to the British cohort text (74 in the summary table), with a median age of 43–44.5 years, and in 12 of 27 POLD1 carriers (44%), with a median age of 41 years; no POLD1 carrier with CRC had previously undergone adenoma removal [7]. Fifteen patients developed CRC before age 30 and the earliest diagnosis was at age 14 [7, 10]. Sixteen carriers had multiple CRCs: ten synchronous (2–4 tumours) and the others metachronous, 1–22 years later (median 10 years) [7]. In the 2023 review, 120 of 205 heterozygotes (58%) developed CRC, at a mean age of 41 years (13–80) [2]. Microsatellite instability does not exclude the syndrome: of 28 CRCs in POLE carriers with known status, four (14.3%) were unstable, three with loss of MMR proteins, and no adenoma was unstable [7]. Tumour histology in L424V carriers was described as unremarkable, with no predilection for a particular colonic location [32].

3.3 Duodenum and upper gastrointestinal tract

Duodenal tumours were the most frequent lesions after colorectal lesions in POLE carriers: 10 of 105 (9.5%) developed duodenal cancer (median age 55 years) and 16 (15%) duodenal adenomas (median age 43 years), two before age 18 [7]. More than half of carriers undergoing upper gastrointestinal endoscopy had duodenal adenomas; in an earlier series, 57.1% of 14 POLE carriers undergoing endoscopy had gastroduodenal adenomas [7, 33]. For POLD1, the British cohort recorded duodenal adenomas in 2 of 27 carriers and no duodenal cancer [7]; in 2025, a French report described the first duodenal adenocarcinomas in carriers of likely pathogenic POLD1 variants [39]. In the 2023 review, 9% of heterozygotes had extracolonic gastrointestinal cancers (mean age 45 years; 35–78) [2].

3.4 Endometrium and ovary

Endometrial cancer is the most frequent extraintestinal tumour and the one most clearly differentiating the two genes. It was the most common malignancy in women with POLD1 (9/17, 53%; three before age 50), compared with 5/43 (12%) in POLE carriers (one before age 50; p=0.001); no endometrial cancers occurred before age 45, and risk by age 70 was estimated at approximately 75% with POLD1 and 25% with POLE [7]. In the 2016 series, endometrial cancer affected 1 of 26 POLE carriers and 8 of 14 POLD1 carriers, and the authors acknowledged that inclusion of families meeting Amsterdam or Bethesda criteria might inflate its frequency [33]. Five POLE carriers developed ovarian cancer between ages 33 and 45 (one bilateral at age 40), two simultaneously with endometrial cancer, and no POLD1 carrier did [7]. The 2023 review recorded endometrial cancer in 22% and ovarian cancer in 8% of 95 female carriers [2].

3.5 Breast

Six of 43 POLE carriers (14%; median age 49 years, 38–65) and 4 of 17 POLD1 carriers (24%; median age 62 years, 52–65) developed breast cancer; risk by age 60 was estimated at approximately 20% and 30%, respectively, based on 40 and 17 women, and the authors considered an increase over the general population unproven [7]. The NCCN reports limited evidence of increased breast cancer risk [8].

3.6 Central nervous system

Brain tumours occurred in 9 of 105 POLE carriers (9%): glioblastomas at ages 47 and 61, astrocytoma at age 15 and oligodendroglioma at age 30 in L424V carriers, and four glioblastomas between ages 30 and 52 and one unspecified tumour at age 35 in a family with N363K; one of the first POLD1 S478N carriers developed two astrocytomas at age 26 [7]. The 2023 review recorded brain cancer in 9% of heterozygotes, at a mean age of 28 years (4–66) [2]. Gliomas and medulloblastoma have been described in children and young adults [34, 52].

3.7 Other tumours and cutaneous manifestations

In the British cohort, L424V carriers also developed basal cell carcinomas and ureteral, ampullary, urethral and oesophageal cancers; the N363K family had two pancreatic cancers and one prostate cancer; one POLD1 D316H carrier developed mesothelioma [7]. Only one patient developed prostate cancer, at age 51 [7]. Multiple tumours were frequent: 29% of men and 46.5% of women with POLE, and 2 of 10 men and 10 of 17 women with POLD1 [7]. Café-au-lait macules are frequently observed in PPAP [9], and the 2023 review recorded cancer, preneoplastic lesions or non-tumour manifestations (for example café-au-lait macules) in 88% of heterozygotes [2].

3.8 Paediatric CMMRD-like phenotype

Some POLE variants (p.(Val411Leu), p.(Ala456Pro), p.(Ser297Phe), p.(Pro436Arg), among others) manifest in childhood or adolescence with polyposis, CRC, medulloblastoma, café-au-lait macules and pilomatricomas, a presentation meeting suspicion criteria for CMMRD but without constitutional MMR deficiency [9, 10, 34]. In 2019, a 5-year-old child with classical clinical features of CMMRD and ultra-hypermutated medulloblastoma with preserved MMR expression carried the germline variant POLE p.(Ala456Pro), a known somatic hotspot [34]. A 2022 series excluded CMMRD in three patients with a highly suggestive phenotype and identified in each a constitutional heterozygous POLE variant well known as a somatic mutation in hypermutated tumours [35].

3.9 Penetrance estimates

Gene-specific cumulative incidence curves from the British cohort appear in Figure 2 of the 2022 paper. The estimates are not directly comparable. The 2022 cohort was restricted to variants with strong evidence of pathogenicity, which the authors cite as an explanation for higher values than in earlier series, but included patients whose first cancer developed before any surveillance [7]. The Australian analysis adjusted for ascertainment but classified pathogenicity using in silico tools and had very wide confidence intervals for POLD1 [43]. The first version of that study's abstract had values reversed between genes, corrected in a 2018 erratum [12]. None of these estimates is population-based.

3.10 MBD4 tumour predisposition syndrome

The clinical description is based on 21 patients from 18 families published between 2018 and 2026, identified mainly through early-onset AML or polyposis without a molecular diagnosis; there are no prospective cohorts or cumulative risk estimates, and the following frequencies are proportions of described cases, subject to ascertainment bias [26]. Several patients died of AML, so the data mix natural history and treated disease [28, 46].

Polyposis is the most consistent manifestation. The polyps are mainly tubular adenomas with low- or moderate-grade dysplasia, with rectal predominance in one patient; four patients also had hyperplastic polyps, and one polyp contained a focus of intramucosal adenocarcinoma [26, 28]. The burden ranges from a few adenomas to more than 130, a phenotype close to attenuated FAP and to the polyposes associated with MUTYH, NTHL1, POLE and POLD1 [28]. An ampullary tubulovillous adenoma and a duodenal tubular adenoma have been described, but upper gastrointestinal involvement has not been studied systematically [28, 46]. Ten patients underwent colectomy, at a median age of 38 years (31–53), and in one of them new polyps arose in the remaining colon after hemicolectomy [26].

AML is the main cause of death. In 13 of the 16 patients with AML, polyposis was identified before the leukaemia, which places the gastroenterologist in a position to make the diagnosis before the haematological complication [26]. The leukaemias are hypermutated, with C>T transitions at CpG and DNMT3A, IDH1 or IDH2 variants, and clonal haematopoiesis may precede them by years [26, 38]. The schwannomas include chest wall, cervical and vestibular locations, unilateral or bilateral; one woman with bilateral vestibular schwannomas was first given a clinical diagnosis of NF2-related schwannomatosis [26, 46]. One woman developed two uveal melanomas (38 and 45 years) and liver metastases at age 52 years; another woman had, simultaneously with the uveal melanoma, bilateral diffuse uveal melanocytic proliferation [26, 28].

4. Clinical Recognition and Warning Signs

PPAP has no clinical diagnostic criteria; it is almost always identified through multigene panels requested for polyposis or early-onset CRC [53]. It should be suspected in circumstances overlapping with attenuated FAP, MAP and Lynch syndrome, but with some distinctive clues.

History-taking should record the cumulative number of adenomas and age at detection, upper gastrointestinal endoscopy findings, histology and MMR status of each tumour, and a three-generation pedigree including gynaecological, brain and breast tumours [7, 31]. Physical examination includes looking for café-au-lait macules and nodules suggestive of pilomatricoma, particularly in young patients [10].

5. Diagnosis: Criteria, Tests and Algorithms

5.1 Diagnostic definition and indications for testing

PPAP is diagnosed molecularly: diagnosis is established by identifying a heterozygous pathogenic or likely pathogenic germline variant in the exonuclease domain of POLE or POLD1; no confirmatory biochemical or functional test is available as a clinical service, and an ultramutated tumour phenotype supports but does not confirm diagnosis, because it may result from a somatic variant [2, 3]. There is no consensus on whom to test specifically for PPAP; the genes are included in polyposis and hereditary CRC panels [7, 53].

Diagnostic yield is low. In the CORGI study, 1% of screened patients had POLE L424V or one of three POLD1 variants, and in QUASAR2/VICTOR only 0.1% of patients with CRC had L424V or S478N [7]. In a Catalan series, POLD1 variants explained 1% (4/411) of non-polyposis familial or early-onset CRC with proficient MMR repair and 1.7% (1/60) of Amsterdam-positive families [33]. The British guidelines cite a 1.5% frequency of pathogenic POLE variants in patients with more than 20 synchronous or 40 metachronous adenomas, higher with a family history [21]. The European guidelines estimate that, in patients with 10–20 polyps or older than 60, detection of variants in genes other than APC or MUTYH is below 2–3% [17].

5.2 Molecular tests

A polyposis panel should include APC (with deletions/duplications) and MUTYH, MMR genes (MLH1, MSH2, MSH6, PMS2, MSH3, MLH3), the exonuclease domains of POLE and POLD1, NTHL1, MBD4 and AXIN2, and hamartomatous polyposis genes (STK11, BMPR1A, SMAD4, PTEN, RNF43) because of phenotypic overlap [17]. The 2024 Japanese guideline recommends an equivalent panel for adenomatous polyposis [54]. Historically, screening used targeted L424V and S478N genotyping or Sanger sequencing of exons 9–14 of POLE and 6–12 of POLD1 [10, 32, 33]. Exome or genome sequencing allows the same analysis using virtual panels [17]. In one patient with a CMMRD-like phenotype, the sequence leading to diagnosis was: normal tumour MMR immunohistochemistry and microsatellite instability testing, negative germline instability testing, no variants in MMR or polyposis genes, and only then exonuclease domain sequencing [10].

5.3 Classification of exonuclease domain variants

Classification is the critical step. Unlike most hereditary cancer genes, disease-causing variants are non-disruptive, and generic ACMG/AMP criteria are poorly applicable; in 2023, the Barcelona group published specific recommendations for non-disruptive exonuclease domain variants (POLE NM_006231.4, amino acids 268–471; POLD1 NM_001256849.1 ≈ NM_002691.4, amino acids 304–533), intended to apply until official InSiGHT (International Society for Gastrointestinal Hereditary Tumours)/ClinGen panel specifications become available [2]. That expert panel has POLE and POLD1 specifications under development [9, 40]; in the ClinGen CSpec (Criteria Specification Registry) consulted on 7 October 2026, the panel had published specifications for APC and MMR genes and a pilot specification for MUTYH, without a document for POLE or POLD1 [55].

Applied to 128 variants, these rules classified 7 as pathogenic, 10 as likely pathogenic, 7 as benign and 10 as likely benign; two catalytic residue variants previously considered pathogenic, POLE p.(Asp275Val) and p.(Glu277Gly), became variants of uncertain significance [2]. Before these rules, the British cohort proposed five practical filters: location in the exonuclease domain; frequency <1×10⁻⁵ in non-Finnish Europeans in gnomAD; location within or adjacent to an Exo motif, at a residue conserved between POLE and POLD1; predicted pathogenicity by at least two in silico tools; and prediction of a protein with preserved polymerisation and regulatory functions [7]. Recent structural tools, such as MAVISp (a modular, structure-based framework for the effect of protein variants), applied to more than 60,000 POLE and POLD1 missense variants, assigned PP3 evidence to 364 and BP4 to 323 variants of uncertain significance, but remain research tools not yet incorporated into clinical specifications [56]. Periodic reassessment matters: in a Japanese case published in 2025, a POLE variant of uncertain significance was reclassified as likely pathogenic, establishing the diagnosis [57].

5.4 Contribution of the tumour

Tumour analysis has three functions. First, normal MMR and microsatellite stability are expected in PPAP and do not exclude diagnosis; conversely, some carriers' tumours are unstable because of secondary MMR deficiency [3, 7]. Second, mutational burden and mutational signatures provide functional evidence for germline variant classification, provided that the tumour does not carry another somatic exonuclease domain variant explaining the phenotype [2, 13, 40]. Third, detection of an exonuclease domain variant in a tumour requires asking whether it is somatic or germline, because somatic variants are much more frequent, particularly in the endometrium [40, 41]. High mutational burden alone does not identify proofreading deficiency: the ESMO Handbook notes that microsatellite-stable metastatic CRCs without a POLE mutation respond poorly to immunotherapy even with high mutational burden [44]. As many routine tumour panels do not include POLE, a 2026 study showed that a high number of variants in a small panel (≥6 non-synonymous SNVs) flagged 15 CRCs, of which 5 carried POLE exonuclease domain variants, all hotspots, in MMR-proficient (pMMR) and microsatellite-stable (MSS, microsatellite stable) tumours [58].

5.5 Test limitations and report contents

A panel covering only the exonuclease domain does not detect variants elsewhere; most do not cause PPAP, but at least one likely pathogenic variant has been described in the POLD1 polymerisation domain, leading the authors to suggest whole-gene analysis when suspicion is strong [39]. Yeast functional assays show variability between replicates and may show mutator effects for non-pathogenic DNA-binding cleft variants [2, 13]. A POLE or POLD1 loss-of-function variant should not be reported as a diagnosis of PPAP; GeneGuidance emphasises that the ClinGen haploinsufficiency score for POLD1 is 1 (little evidence for dosage pathogenicity) [3, 13]. De novo cases and a probably mosaic mother (estimated allele fraction 14%) have been described in a family with POLD1 p.(Leu606Met), reinforcing the need to consider allele fraction in the report [7, 39]. A variant of uncertain significance does not become pathogenic because a tumour responds to immunotherapy or because of partial segregation; treatment response depends on the tumour and does not prove the germline mechanism [2, 59].

The report should state the gene, versioned transcript (NM_006231.4 or NM_002691.4), nucleotide and protein HGVS nomenclature, zygosity, location relative to the exonuclease domain and Exo motifs, ACMG/AMP classification with the specific criteria applied, and coverage limitations; where literature uses different numbering or another transcript (for example NM_001256849.1 for POLD1), both should be provided [2, 5]. A negative test does not exclude another hereditary cause when the phenotype is convincing: the European guidelines note that APC mosaicism, undetectable in blood in many cases, explains 20–50% of unexplained polyposis [17].

5.6 Algorithm

5.7 Diagnosis of MBD4 tumour predisposition syndrome

There are no consensus clinical criteria. Suspicion arises with early-onset gastrointestinal polyposis, AML or MDS, schwannoma, CRC or uveal melanoma, particularly in combination, when the family history is compatible with recessive inheritance (affected siblings, consanguinity) or is absent, and when the tumour has an excess of CpG>TpG transitions [26]. The diagnosis is established by identification of biallelic pathogenic or likely pathogenic variants; two variants of uncertain significance, or one pathogenic variant with one of uncertain significance, neither confirm nor exclude the syndrome [26]. Sequencing detects all variants described so far; if only one variant is found, targeted deletion and duplication analysis follows, although none has been described [26]. Apparent homozygosity requires confirmation of the status of both parents, because it may correspond to an undetected deletion on one allele or to uniparental isodisomy [26].

Testing may use a multigene panel or the exome. The authors of the 2022 description proposed including MBD4 in adenomatous polyposis, CRC, early-onset AML and uveal melanoma panels, and a 2026 North American series states that the NCCN now recommends considering MBD4 testing in anyone with more than 10 lifetime adenomas [28, 46]. Patients with suggestive phenotypes and previous negative tests should be reassessed, because the gene was not part of the panels used until a few years ago [46]. In a European genome sequencing programme in 98 patients with strongly suspected undiagnosed hereditary cancer, selected through ERN GENTURIS, pathogenic variants in phenotype-related high-penetrance genes were identified in 6%, including one MBD4 homozygote [60].

In AML, germline testing should not be performed on blood or bone marrow with active disease, where somatic variants confound interpretation; in one of the 2026 patients, the 61-gene germline panel was performed on skin fibroblasts [46]. After allogeneic transplantation, blood DNA is donor-derived, and the germline sample has to be obtained from another tissue [46]. Tumour panels frequently reveal dozens of C>T at CpG variants of uncertain significance; this pattern, a high mutational burden and the combination of biallelic DNMT3A with IDH1/IDH2 in AML should prompt germline MBD4 testing [26, 46]. Testing of tumours from heterozygotes with uveal melanoma shows somatic loss of the wild-type allele and the same mutator phenotype [48]. Once the familial variants have been identified, testing of siblings, prenatal diagnosis and preimplantation genetic testing are possible [26].

6. Differential Diagnosis

Adenomas arising in different adenomatous polyposis syndromes (APC, AXIN2, POLE, POLD1, MMR genes, MUTYH, NTHL1, MBD4, MSH3, MLH3) are morphologically indistinguishable from one another and from sporadic adenomas; they differ in inheritance pattern, extracolonic spectrum and molecular tumour characteristics [9]. The Oxford Desk Reference groups APC, MUTYH, POLE, POLD1 and NTHL1 among genes predisposing to conventional adenomas, contrasting with those predisposing to juvenile polyps (SMAD4, BMPR1A), Peutz–Jeghers hamartomas (STK11) or mixed polyps (GREM1), and with Lynch syndrome, predominantly associated with cancer risk [42]. The 2024 Japanese guideline emphasises that PPAP can mimic both attenuated FAP and Lynch syndrome, and that endometrial cancer makes it a differential diagnosis for the latter [54].

Three situations warrant practical attention. The first is an MMR-deficient tumour without an explanatory germline variant: the Lynch syndrome GeneReview considers the “Lynch-like” category provisional and heterogeneous and lists PPAP among syndromes in which tumour MMR deficiency may be secondary [23]. In a series cited by the British guidelines, 9 of 62 Lynch-like cases (14.5%) had somatic (7) or constitutional (2) variants in the exonuclease domain of POLE/POLD1, with an ultramutated phenotype, and six also carried somatic MMR variants [21]. The second is a child with a CMMRD phenotype without biallelic MMR variants and with preserved tumour MMR expression, in whom a tumour polymerase deficiency signature should prompt POLE testing [10, 34]. The third is a carrier of a POLE or POLD1 variant without a consistent phenotype, in whom the variant may be of uncertain significance, loss-of-function or associated with an allelic disorder; MDPL, for example, shares the POLD1 gene but has no established predisposition to polyposis [3, 16]. Recent reviews also note non-genetic factors, such as previous cancer treatments and colibactin-producing Escherichia coli, as alternative mechanisms in patients with multiple adenomas without a germline variant [9].

In MBD4 syndrome, the differential diagnosis extends to hereditary predispositions to myeloid neoplasms, the schwannomatoses and the uveal melanoma predispositions. One of the 2026 patients successively received clinical diagnoses of FAP, at age 33 years, for more than 70 hyperplastic and adenomatous polyps, and of NF2-related schwannomatosis, for bilateral vestibular schwannomas, before the exome revealed two MBD4 variants in trans [46]. Her brother, with CRC at age 24 years and AML at 30, received a stem cell transplant from her, before any genetic testing [46].

7. Treatments and Risk Management Strategies

7.1 Principles

There is no pharmacological treatment for the predisposition: management relies on endoscopic surveillance with polypectomy, surgery when required by polyp burden or cancer, discussion of risk-reducing gynaecological surgery and family counselling [3, 7]. Evidence quality is low: ClinGen identified no assessment of colonoscopy efficacy in PPAP carriers and extrapolates from benefits demonstrated in Lynch syndrome and FAP [22]. The 2020 British guidelines considered the data insufficient for specific guidance and recommended referral to multidisciplinary specialist centres (low GRADE, weak recommendation, 91% agreement) [21]; the 2024 European guidelines emphasise specialist-centre care and international registries for all rare polyposis syndromes [17].

7.2 Surveillance recommendations: comparative register

Differences reflect the scarcity of data. The NCCN and ClinGen set the starting age between 20 and 30; European sources propose age 18–20, earlier in aggressive phenotypes; and the British cohort, followed by the AACR (American Association for Cancer Research) paediatric group, proposes age 14, because that series included a colorectal adenoma at age 15 and CRC at age 14 [7, 8, 17, 18, 52]. The British cohort justified a biennial rather than annual interval because of uncertainty about progression, but the Australian authors considered that risk in POLE carriers justified management comparable to that of MSH6 carriers in Lynch syndrome [7, 12]. For gynaecological surveillance, ESMO proposes annual transvaginal ultrasound from age 30–35, the Spanish series proposed endometrial screening from age 40 in POLD1 carriers, the British cohort favoured counselling about risk-reducing surgery, citing the absence of robust gynaecological screening data even in Lynch syndrome, and the NCCN considers the evidence insufficient for any strategy [7, 8, 18, 33]. The choice should be made at a specialist centre and explained to the patient as a decision under uncertainty.

7.3 Colorectal management and surgery

Colonoscopy with polypectomy is the central intervention [22]. Prophylactic surgery may be appropriate in patients with a severe polyp phenotype; the British cohort considers colectomy with ileorectal or distal ileosigmoid anastomosis generally the most appropriate option [7]. In patients who have already developed CRC, surgery should be individualised, weighing phenotype, expected functional outcome and metachronous cancer risk, and the authors make no strong recommendation [7]. The Australian authors, acknowledging that metachronous risk is unknown but probably high, suggested considering subtotal rather than segmental colectomy in POLE carriers [43]. In the British cohort, 19 of 24 patients with surgical information underwent surgery after CRC (4 segmental and 15 extended resections) and four underwent prophylactic surgery for multiple adenomas; one patient developed cancer during surveillance and another after discharge from follow-up [7].

7.4 Duodenum and stomach

The frequency of adenomas and duodenal cancer in POLE carriers supports upper gastrointestinal surveillance; lacking specific data, the British cohort proposed following the FAP system from age 25 and considering polypectomy for adenomas from 1 cm, acknowledging that the benefit of duodenal polypectomy is unproven even in FAP [7]. The report of duodenal adenocarcinomas in POLD1 families led its authors to advocate systematic upper gastrointestinal endoscopy for this gene too [39].

7.5 Endometrium and ovary

The options described are gynaecological surveillance (ESMO, Spanish series) and counselling about risk-reducing hysterectomy after completion of childbearing, with bilateral salpingo-oophorectomy to be discussed in POLE carriers because of the five ovarian cancers observed [7, 18, 33]. The British cohort suggests this approach for all women with PPAP, as endometrial cancer risk is appreciable with both genes [7]; a 2025 report notes that data still do not allow accurate estimation of ovarian cancer risk [39]. The NCCN maintains that evidence is insufficient to recommend endometrial cancer risk management strategies [8].

7.6 Breast, brain and other organs

For breast cancer, the British cohort concluded that evidence did not establish sufficient risk to meet British moderate-risk screening criteria, but suggested discussing referral to a high-risk screening service, particularly with familial clustering or very early breast cancer in a first-degree relative, and encouraging breast awareness [7]. Routine brain surveillance is not recommended for PPAP; the AACR paediatric group recommends that families and physicians maintain clinical vigilance for malignant brain tumours [52]. For patients with severe POLE variants and a CMMRD-like presentation, a 2022 series advocated monitoring according to CMMRD protocols, whereas the AACR group considered evidence insufficient to offer this intensive surveillance and EHTG-ESCP proposed only bringing gastrointestinal surveillance forward by 5 years [17, 35, 52]. For other cancers, absolute risk appears small and below the threshold justifying routine surveillance [7].

7.7 Treatment of proofreading-deficient cancers

Proofreading-deficient cancers, sporadic or hereditary, are immunogenic because of their high neoantigen burden and have a favourable prognosis and clinical benefit from immune checkpoint inhibitors (ICIs) [2, 9]. In metastatic CRC, the ESMO Handbook states that tumours with a POLE/POLD1 mutation (more than 100 mutations/Mb) respond well to ICIs, with some evidence of better outcomes than microsatellite-unstable tumours, and classifies a POLE mutation (<1% of cases) as a criterion for a clinical trial with PD-1 (programmed cell death protein 1) inhibitors [44]. In an international retrospective study of 27 patients with proofreading-deficient metastatic CRC treated with anti-PD-(L)1 with or without anti-CTLA-4, response rate was 89%, compared with 54% in 610 patients with dMMR/MSI-H CRC (p=0.01), with better progression-free survival (HR 0.24; 95% CI 0.08–0.74) [62]. In the Korean phase 2 trial of durvalumab 1500 mg every 4 weeks, only 3 of 33 patients had POLE-mutated CRC: the patient with an exonuclease domain variant responded, and the two with variants outside the domain progressed [63]. In children and young adults, ICIs improve survival in aggressive, refractory replication repair-deficient cancers, regardless of germline origin (CMMRD, Lynch or POLE) [52]. These data concern the sequenced tumour and the oncology decision: a germline POLE or POLD1 variant alone does not make an individual eligible for immunotherapy or constitute preventive therapy [3].

The only ICI with an indication applicable to some of these tumours based on a mutational burden biomarker is pembrolizumab (Keytruda®) in the United States: unresectable or metastatic solid tumours with high mutational burden (≥10 mutations/Mb, determined by an FDA-approved test), progressing after prior treatment and without satisfactory alternatives, in adults and paediatric patients; this is accelerated approval based on response rate and duration, and safety and efficacy in paediatric high-burden central nervous system tumours have not been established [19, 64]. Label dosing for this indication is 200 mg intravenously every 3 weeks or 400 mg every 6 weeks in adults, and 2 mg/kg (up to 200 mg) every 3 weeks in paediatric patients, as a 30-minute infusion [19]. Relevant adverse effects are immune-mediated, sometimes severe or fatal (pneumonitis, colitis, hepatitis, endocrinopathies, nephritis, skin reactions, transplant rejection), infusion reactions and embryofoetal toxicity; the label calls for assessment of liver enzymes, creatinine and thyroid function at baseline and periodically, and withholding or discontinuation according to severity [19]. In the European Union, Keytruda® has indications for microsatellite-unstable or dMMR tumours, but not high mutational burden or POLE mutation, so use in a POLE-mutated MSS tumour is off-label [20].

In endometrial cancer, the POLE-mutated molecular class (defined in tumours with somatic variants) is being investigated in adjuvant treatment de-escalation trials, such as RAINBO (Refining Adjuvant treatment IN endometrial cancer Based On molecular features) POLEmut-BLUE and RAINBO BLUE & TAPER (tailored adjuvant therapy in early POLE-mutated or p53 wild-type/NSMP, no specific molecular profile, endometrial cancer); these trials were not designed for germline carriers and their results cannot be extrapolated to PPAP without caution [65, 66].

7.8 Comparative regulatory status

7.9 Trials and studies underpinning decisions

These data do not allow treatments to be ranked. ICI studies predominantly include somatic variants, have small samples, and comparison with dMMR tumours is retrospective and subject to selection; the KEYNOTE-158 TMB-H cohort response rate concerns tumours selected by mutational burden rather than proofreading deficiency, and response was lower in the 10–13 mutations/Mb range [19, 62]. No study assessed surveillance or risk-reducing surgery in PPAP carriers; the recommendations in sections 7.2–7.6 are consensus statements or cohort proposals [7, 22]. Nor are there chemoprevention studies in this syndrome in the sources consulted.

7.10 Extrapolation of evidence and off-label use

Colorectal surveillance in PPAP extrapolates efficacy demonstrated in Lynch syndrome and FAP [22]. Immunotherapy for POLE-mutated MSS cancer is off-label in the European Union; in the United States, it is on-label only when the tumour meets label criteria (TMB ≥10 mut/Mb using an approved test, progressing unresectable or metastatic disease, without a satisfactory alternative) [19, 20]. In children, the intravenous pembrolizumab TMB-H indication includes paediatric patients but does not establish safety and efficacy in central nervous system tumours [19]. No data on pregnancy in PPAP carriers were found in the sources consulted; ICIs have embryofoetal toxicity described in the label [19].

7.11 Developments in the last 12 months (October 2025 to October 2026)

None of these developments changed surveillance recommendations. The search found no authorisation applications, CHMP (Committee for Medicinal Products for Human Use, EMA) opinions or interventional trials in germline PPAP carriers; this absence in the sources consulted does not prove that none exist [20].

7.12 Management of MBD4 tumour predisposition syndrome

There are no published guidelines for MBD4 syndrome. The 2026 GeneReviews recommendations are based on the authors' experience and take up, with few changes, the proposals of the 2022 description, which were modelled on the BER polyposes [26, 28]. At initial assessment, they recommend colonoscopy (from age 18–20 years), full blood count in adults, dilated eye examination in adults and a genetics clinic consultation, and, if the proband has AML, MBD4 testing of any relative considered as a haematopoietic stem cell donor [26]. This last point has direct consequences: a sibling with undiagnosed biallelic variants may have clonal haematopoiesis and go on to develop AML, and a sibling transplant has been described in which the donor, untested beforehand, had the same deficiency and later developed AML at age 35 years [38, 46].

Colorectal management follows the model of the attenuated adenomatous polyposes. Polyps are removed at colonoscopy; subtotal colectomy or proctocolectomy is proposed when polypectomy no longer controls the number and size of polyps, and the choice depends on polyp burden and distribution [26]. After partial resection, the remaining colon continues to produce adenomas [26, 28]. AML, MDS, CRC, schwannomas, meningiomas and thyroid carcinoma are treated according to standard protocols; in the 2026 North American series, the two women treated with intensive induction with cytarabine and daunorubicin ("7 + 3") followed by allogeneic transplantation remained in remission, and the brother of one of them died of relapse after transplantation [26, 46].

Hypermutation has therapeutic implications, so far supported only by retrospective data. In a retrospective French cohort of 300 patients with metastatic uveal melanoma treated with immune checkpoint inhibitors (ICI), the overall objective response rate was 3.3%; of the 131 sequenced tumours, 5 (3.8%) had an MBD4 mutation, and 3 of these 5 responded, compared with 4% of tumours without the mutation (p<0.001), with median progression-free survival of 22.3 versus 4.0 months (HR 0.22) [80]. GeneReviews accepts, on the basis of these data and isolated cases, that patients with MBD4-deficient metastatic uveal melanoma may benefit from ICI, and the 2022 and 2026 authors raise the same hypothesis for the CRCs and AMLs of the syndrome, without clinical data [26, 28, 46]. No medicine is authorised for MBD4 syndrome by the FDA or the EMA. The use of ICI in this indication is extrapolated from the biology of hypermutation. The dosing of pembrolizumab in the United States tumour-agnostic indication is given in section 7.7 [19]. In the ClinicalTrials.gov registry, a search for MBD4 in October 2026 found only studies identifying predisposition genes for uveal melanoma and myeloproliferative neoplasms, with no therapeutic trials [81].

The developments of the last 12 months are observational: the first GeneReviews chapter (August 2026), two case series (four Portuguese families recruited over 20 years and three North American families), the description of a founder variant in Christian Arabs in Israel, the ERN GENTURIS genome sequencing series and the French estimate of the relative risk of uveal melanoma in heterozygotes [26, 45–47, 49, 60].

8. Reproductive Options and Family Counselling

8.1 Transmission, penetrance and de novo cases

PPAP is inherited in an autosomal dominant manner: each child of a carrier has a 50% probability of inheriting the variant at conception, which may be lower if the parent is mosaic [75]. Penetrance is incomplete and age-dependent, and expressivity varies within the same family, with some carriers developing dozens of adenomas, others early CRC without polyposis and others remaining tumour-free at follow-up [2, 7, 32]. Risk estimates in section 3.9 should be presented with their source population and limitations [7, 12]. De novo cases have been described: two de novo heterozygotes for POLE p.(Leu424Val) in the British cohort, one de novo carrier of the same variant with polyposis, CRC and anaplastic oligodendroglioma in the Catalan series, and one de novo carrier of POLD1 p.(Leu606Met) [7, 33, 39]; in the paediatric p.(Val411Leu) case, the variant was absent in the mother and the father was not tested [10]. Absence of a family history therefore does not exclude the diagnosis.

No humans with biallelic pathogenic exonuclease domain variants have been identified; in mice, homozygosity is lethal for Pole P286R and associated with very early cancer for other variants, so the classification-rule authors predict extremely aggressive phenotypes if such cases were observed [2]. A POLE loss-of-function variant identified on a panel does not cause PPAP; its reproductive relevance is carrier status for a recessive Polε disorder, such as IMAGe-I syndrome, in which all reported patients had the intronic variant c.1686+32C>G on one allele and a loss-of-function variant on the other [13, 15].

8.2 Testing relatives

For dominant polyposis syndromes, the European guidelines recommend offering testing to all first-degree relatives followed by cascade testing (low evidence; 97% agreement), explicitly citing PPAP as an example [17]. The age of predictive testing in minors depends on the age at which surveillance starts: the 2026 ERN GENTURIS reproductive counselling guideline states that, when surveillance or preventive measures are indicated in childhood, predictive testing may be considered when that surveillance is recommended [75]. In PPAP, this means coordinating testing with the starting age chosen by the centre (14 in the British and paediatric proposal, 18–20 in European guidelines, 25–30 in the NCCN) and discussing earlier testing in families with variants associated with a CMMRD-like phenotype [7, 8, 17, 52]. Relatives who have not inherited the familial variant return to the risk determined by the remaining family history; published segregation studies showed non-carriers with isolated polyps, consistent with sporadic occurrence [32].

8.3 Reproductive options

The 2026 ERN GENTURIS guideline recommends offering non-directive reproductive counselling to all individuals with a cancer predisposition syndrome and relevant relatives of reproductive age, with comprehensive and timely information (strong recommendations), and access to a multidisciplinary team comprising a geneticist, prenatal diagnosis specialist, reproductive physician and psychologist [75]. Options to discuss include natural conception with acceptance of risk, prenatal diagnosis, in vitro fertilisation with preimplantation genetic testing for monogenic disease (PGT-M), gamete or embryo donation and adoption [75]. Prenatal diagnosis is appropriate only if there is an intention to terminate the pregnancy following a positive result and local law permits this; the associated pregnancy loss risk is estimated below 0.5% [75]. Prenatal diagnosis and PGT-M reveal only the presence of the variant and do not predict clinical severity, a particularly relevant limitation in a syndrome with variable expressivity such as PPAP [7, 75]. Counselling should be repeated at decision points (before testing, at diagnosis, at reproductive age, when planning a pregnancy and following clinical change) and comply with each country's legislation on PGT and pregnancy termination [75].

8.4 Communicating the result

Communication of a PPAP result should explain that the variant is specific to one protein domain, that other variants in the same gene have different implications and that a variant of uncertain significance should not guide testing of relatives [2, 3]. It should also clarify that surveillance recommendations vary between guidelines and that some (gynaecological, breast, brain) are proposals without evidence of benefit [7, 8].

8.5 Counselling in MBD4 tumour predisposition syndrome

MBD4 syndrome is transmitted in an autosomal recessive manner [26]. The parents of an affected individual are presumed to be heterozygous, and their testing is recommended to confirm the phase of the variants and calculate the recurrence risk; if only one parent is a carrier, with confirmed paternity and maternity, one of the variants may be de novo in the patient or in a mosaic parent, and apparent homozygosity may correspond to an undetected deletion or to uniparental isodisomy [26]. When both parents are heterozygous, each sibling has, at conception, a 25% probability of having biallelic variants, 50% of being heterozygous and 25% of inheriting neither variant [26]. The children of an affected individual are obligate heterozygotes, unless the other parent is also a carrier; each sibling of the proband's parents has a 50% probability of being heterozygous [26].

Testing of asymptomatic siblings aims to identify early those who have the syndrome, whose surveillance should begin by age 18–20 years [26]. In general, testing is not recommended before age 18 years, but it should be considered earlier when there are early-onset cancers in the family, according to the 2024 NCCN cited by GeneReviews [26]. In heterozygotes, counselling must distinguish two risks. The relative risk of uveal melanoma is increased (9.15 to 31.44-fold, depending on the series), but, given the low incidence of the tumour, GeneReviews does not recommend systematic ophthalmological surveillance; the risk of polyposis or CRC has not been demonstrated, and colonoscopy follows population screening or family history rules [26, 48, 49]. In endogamous communities, carrier frequency may be relevant to couple counselling: in the Christian Arab community of Israel, 3 of 312 healthy volunteers (about 1%) were heterozygous for the founder variant, which led the authors to propose genetic and reproductive counselling in that population [45]. Prenatal testing and preimplantation genetic testing are possible when both familial variants have been identified, and the discussion should take place before pregnancy [26].

9. Special Considerations: Population Diversity and Epidemiology

9.1 Frequency

There is no validated population estimate of PPAP prevalence. Orphanet assigns the <1/1 000,000 category [1], and ClinGen classifies the prevalence of the disease and pathogenic variants as unknown [22]. Available data are frequencies in selected series, measuring the proportion of diagnoses among tested patients rather than population prevalence:

Somatic variants, by contrast, are frequent in some tumours (7–15% of endometrial and 0.5–8% of colorectal cancers), which should not be confused with syndrome frequency [2]. In non-cancer gnomAD, POLE p.(Leu424Val) is absent from approximately 200,000 individuals, and the most frequent pathogenic variant identified there, POLD1 p.(Asp316His), occurs in 1 of approximately 230,000 alleles [2].

9.2 Recurrent variants and founder effect

POLE p.(Leu424Val) is the most frequent variant but is not a founder variant: in the original series, microsatellite analysis showed no recent common ancestor among the 12 British families, and the 2022 cohort recorded two de novo cases [7, 32]. The 2025 analysis describes L424V and S478N as recurrent non-founder variants [41]. However, the two British families with POLD1 S478N in the original description shared microsatellite alleles consistent with a common ancestor [32].

9.3 Regional data

Data come predominantly from the United Kingdom (studies of people of self-reported European ancestry), Spain, Australia and New Zealand (Australasian registry), France and Germany [10, 13, 32, 39, 43]. The Australian authors cautioned that their risk estimates may not apply to people of non-European ancestry [43]. There are isolated reports from Japan, where PPAP is considered under-recognised, and the 2024 Japanese guideline mentions a Japanese family with a POLD1 variant [54, 57]; a Korean trial included three patients with POLE-mutated CRC [63]. In the United States, the main source is the NCCN guideline, referring to European and Australian cohorts, and a North American paediatric case published in 2019 [8, 34]. In Brazil, a series of 27 patients from Barretos Cancer Hospital with polyposis without APC or MUTYH variants, investigated by exome sequencing, reported no pathogenic POLE or POLD1 variants among the 17 pathogenic or likely pathogenic variants identified [82]. No data from Portugal were found in the sources consulted, which does not prove their absence. Data from Africa, the Middle East, southern Asia and Spanish-speaking Latin America are lacking in the sources analysed.

9.4 Sex and age

Men had a higher cumulative CRC risk than women in the British cohort (p=0.0096), a difference not observed for adenomas or all cancers combined [7]. Children and adolescents are a special population because of variants associated with a CMMRD-like phenotype and early brain tumours, for which international consortia (IRRDC, International Replication Repair Deficiency Consortium, and C4CMMRD, Care for CMMRD) support professionals and families [52].

9.5 Networks and associations

PPAP belongs to thematic group 2 (Lynch syndrome and polyposis) of ERN GENTURIS, the European reference network for genetic tumour risk syndromes, whose CPMS (Clinical Patient Management System) allows complex cases to be discussed with experts from several countries with patient consent [30]. Patient associations participate through the ePAG (European Patient Advocacy Group), and the ERN maintains a disease-specific list of patient associations on its portal [30]. The 2026 ERN GENTURIS reproductive counselling guideline included patient representatives in its working group [75].

9.6 Epidemiology of MBD4 tumour predisposition syndrome

By 2026, 21 patients from 18 families had been described, and the estimated population incidence is below 1 in 1,000,000 [26]. Screening studies confirm the rarity: the 2022 description found two homozygotes among 309 patients with multiple adenomas or familial CRC without a diagnosis sequenced by genome or exome and one among 1,611 patients in the replication phase; a Spanish series of 728 patients with CRC, polyposis or suggestive phenotypes identified no patient with the syndrome [28, 51]. These figures measure the yield in selected populations, not population prevalence.

The patients described come from the United Kingdom, the Netherlands, Australia, the United States, Portugal and Israel [28, 38, 45–47]. In Portugal, an oncology centre in Porto identified five patients from four families with adenomatous polyposis and AML, all homozygous for deleterious MBD4 variants, recruited over about 20 years; one founder variant recurred in three of the families, and another was novel [47]. GeneReviews identifies the variant c.1544-1G>T (MANE; c.1562-1G>T on NM_003925.2) as a suspected ancient founder variant in several populations; the same variant was found in Australian patients and in two French patients with uveal melanoma [26, 28, 47, 48]. In the endogamous Christian Arab community of Israel, the variant c.612_615del has a carrier frequency of about 1% (3/312), compared with an allele frequency of 0.0000399 in gnomAD v4.1 [45]. The sources consulted do not describe Brazilian patients.

In heterozygotes, the proportion of carriers among patients with uveal melanoma ranges from 0% to 2.6% depending on the series: rare in a Finnish series, 8 of 1,093 in a consecutive French series from 2013–2018 and 23 of 896 in the French series from 2021–2025 [26, 48, 49]. ERN GENTURIS is the relevant European Reference Network, and GeneReviews lists as resources the CGA-IGC (Collaborative Group of the Americas on Inherited Gastrointestinal Cancer), the American Cancer Society and the National Cancer Institute [26].

10. Illustrative Clinical Case

10.1 Adolescent with polyposis, rectosigmoid carcinoma and a CMMRD-like phenotype (POLE p.(Val411Leu))

Wimmer and colleagues described in 2017 a 14-year-old boy referred to paediatric haemato-oncology at Hannover Medical School for treatment of a multiloculated mucinous adenocarcinoma of the rectosigmoid junction. He had a tubular duodenal adenoma and multiple tubular and tubulovillous adenomas (more than 10 larger lesions and innumerable smaller ones) in the descending colon, sigmoid colon and rectum, some with high-grade intraepithelial neoplasia; he also had six café-au-lait macules on the trunk and right hip and a firm, pea-sized nodule above the left ear, clinically diagnosed as a pilomatricoma. These findings totalled nine points on the European C4CMMRD consortium scoring system, and CMMRD was initially suspected. His parents were non-consanguineous and reported no Lynch-spectrum tumours in their families [10]. Figure 1 of the paper shows the distribution of café-au-lait macules on a body diagram, a colonoscopic image of multiple polyps in the left hemicolon and photographs of the trunk and hip macules.

Investigation followed a logical sequence. The tumour was microsatellite-stable at all five markers examined and retained expression of all four MMR proteins; germline microsatellite instability testing (gMSI) was negative; MMR gene sequencing, without point variants or copy-number changes, excluded CMMRD; APC, MUTYH, PTEN, STK11, BMPR1A and SMAD4 analysis was also negative. Finally, Sanger sequencing of the exonuclease domains of POLD1 and POLE identified the heterozygous variant NM_006231.4:c.1231G>C p.(Val411Leu) (published on NM_006231.3, with the same numbering), a known recurrent somatic mutation in sporadic ultramutated colorectal and endometrial cancers with demonstrated reduction in proofreading activity in cell-free assays. The variant was absent in his healthy 36-year-old mother; his 43-year-old father was not tested [10]. In the current MANE version, the same variant is described as NM_006231.4:c.1231G>C p.(Val411Leu) [5].

The authors considered him the youngest patient with PPAP and carcinoma reported at the time and suggested that p.(Val411Leu) confers a more severe phenotype than known germline variants, perhaps by acting remotely from the catalytic centre, although they could not exclude modifier variants. They also proposed that café-au-lait macules and pilomatricomas, common in CMMRD, might mark constitutional replication error repair defects in general [10]. This hypothesis has not been validated as a diagnostic criterion. The case has two practical consequences documented in subsequent literature: it was one basis for the British proposal of colonoscopy from age 14 and the European suggestion of earlier surveillance in CMMRD-like phenotypes, and it shows that PPAP should be considered in children with polyposis or CRC and cutaneous signs of CMMRD when tumours retain MMR expression [7, 10, 17].

10.2 Two families with POLD1 and duodenal adenocarcinoma

In 2025, Folletet and colleagues described two French families. In family A, the proband developed colonic adenocarcinoma at ages 43 and 69, endometrial adenocarcinoma at age 54 and bifocal invasive ductal breast carcinoma at age 60, with at least 40 polyps removed since age 41 (low-grade adenomas and serrated polyps in all segments). The family included a half-brother with CRC at age 32, his son with duodenal adenocarcinoma extending to the pancreas (isolated PMS2 loss) and synchronous CRC at age 52, that son's daughter (the proband's great-niece) with duodenal and colorectal adenomas at age 23, and other relatives with CRC, endometrial and rectal cancer; the variant POLD1 NM_002691.4:c.1481T>G p.(Ile494Ser), absent from gnomAD v4, cosegregated through at least six informative meioses, and one polyp from the proband had a mutational burden of 27.1 mutations/Mb, SBS10d signature and proficient MMR repair; it was classified as likely pathogenic using the specific 2023 rules. In family B, a man developed pMMR duodenal adenocarcinoma at age 40, with no other lesions on upper gastrointestinal endoscopy or colonoscopy, and carried POLD1 c.1204G>A p.(Asp402Asn), a variant with one allele in gnomAD v4, in the ExoII motif and previously described in an endometrial cancer with 141 mutations/Mb, SBS10d and loss of heterozygosity in POLD1; his oncologist planned to discuss immunotherapy at a multidisciplinary meeting [39]. Pedigrees for both families appear in Figure 1 of the paper.

The authors present these as the first duodenal adenocarcinomas in POLD1 carriers and advocate systematic upper gastrointestinal endoscopy in PPAP [39]. These are two families, without a risk estimate, and the authors' recommendation is an interpretation of the cases rather than a guideline. The case also illustrates practical application of specific classification criteria (supporting PM2 and PM1, PP3, moderate PP1, moderate or strong PP4 and, for p.(Asp402Asn), supporting PS3), and that PMS2 loss in a carrier's tumour does not exclude PPAP [2, 39].

10.3 Young adult woman with rectal adenoma and hypermutated AML (biallelic MBD4)

A 28-year-old woman presented to an emergency department with 81% circulating blasts. One month earlier, she had undergone transanal excision of a recurrent 3 cm tubulovillous adenoma of the distal rectum, first identified at age 21 years, with slow and painful recovery [46]. The diagnosis was AML. The blasts had a hypermutated profile dominated by CpG>TpG transitions, with somatic variants in BCOR, BCORL1, DNMT3A, IDH2 and WT1; fluorescence in situ hybridisation was negative and the karyotype was 46,XX[20] [46]. She started intensive induction with cytarabine and daunorubicin, complicated by febrile neutropenia, cellulitis and bacteraemia, and two months later the marrow had 1% myeloblasts, with no evidence of disease [46].

A genetics consultation was requested during induction because of her age and the early advanced adenoma. The family history was not suggestive (a distant maternal relative with "leukaemia" at an advanced age) [46]. Her younger brother was the preferred donor, being a full HLA (human leukocyte antigen) match. The 61-gene germline panel for acute leukaemia and bone marrow failure, performed on skin fibroblasts, revealed two truncating MBD4 variants, NM_001276270.2:c.939dup p.(Glu314ArgfsTer13) and NM_001276270.2:c.614_615del p.(Ser205Ter) [5, 46]. Cascade testing gave priority to the matched brother, who carried neither variant. The transplant was successful, and brain magnetic resonance imaging requested later for headaches revealed a vestibular schwannoma. The next two colonoscopies showed two and three tubular adenomas [46]. The pedigree is shown in Figure 1A of the article.

The case brings together the practical points of the syndrome: AML in a young adult with a history of early-onset adenoma should prompt germline testing; the germline sample from a patient with active leukaemia should be from non-haematopoietic tissue; MBD4 testing of the family donor precedes transplantation; and colorectal and neurological surveillance continues after the leukaemia is cured. The management described matches the GeneReviews recommendations [26].

11. Prognosis and Long-Term Follow-Up

Prognosis depends predominantly on early detection of colorectal and duodenal cancers. There are no survival data for germline carriers under surveillance; colorectal and endometrial cancers with somatic exonuclease domain variants have a favourable prognosis, and the British cohort authors considered further study necessary to determine whether the same applies to cancers in germline carriers [2, 7]. Response of metastatic proofreading-deficient tumours to ICIs improves therapeutic prospects but is based on small series [62, 63]. Multiple cancers are frequent, and one patient in the CORGI cohort developed CRC after discharge from follow-up, suggesting surveillance without an end date [7].

Transition of adolescents to adult services is relevant in families starting surveillance at age 14 or with severe-phenotype variants [7, 52]. Follow-up should include updating the family history, because new tumours in relatives may change the colonoscopy starting age according to the NCCN criterion (2–5 years before the earliest family CRC) [8].

In MBD4 tumour predisposition syndrome, prognosis depends mainly on AML. There are no survival data; among the patients described, there are deaths from AML (including one from relapse after transplantation) and patients in prolonged remission after allogeneic transplantation, and one woman had liver metastases from uveal melanoma seven years after the second ocular tumour [28, 46]. Detection of polyposis before AML in 13 of 16 patients makes it possible to organise haematological surveillance and testing of family donors before leukaemia [26]. The following recommendations derive from the authors' experience, not from efficacy studies, and apply to patients with biallelic variants [26, 28].

GeneReviews does not recommend upper endoscopy or surveillance targeted at schwannomas, meningiomas or the thyroid; these organs are assessed according to symptoms, as in the woman in section 10.3 [26, 46].

12. Summary and Final Considerations

Polymerase proofreading-associated polyposis is a rare autosomal dominant syndrome, described in 2013, caused by exonuclease domain missense variants in POLE and POLD1 that eliminate proofreading by polymerases ε and δ without abolishing polymerisation [2, 32]. The biological consequence is an ultramutated tumour phenotype with specific mutational signatures (SBS10a/b and SBS28 in Polε; SBS10c/d in Polδ), generally microsatellite-stable but sometimes with secondary MMR deficiency [2, 7]. The clinical consequence is predisposition to colorectal adenomas (usually 10–100, sometimes none), CRC from adolescence, duodenal tumours (particularly POLE), endometrial cancer (particularly POLD1), ovarian, breast and brain cancer, with frequencies known only from clinical series subject to ascertainment bias [2, 7, 12].

Diagnosis is molecular and depends on domain-specific variant classification: loss-of-function variants and most variants outside the domain do not cause PPAP, and tumour mutational burden and signatures are classification tools [2, 13]. Variants in the same gene also cause different disorders (FILS, IMAGe-I, MDPL), requiring interpretation of the report at protein-domain level [14–16]. Management relies on endoscopic surveillance with a starting age ranging from 14 to 30 according to the source, duodenal surveillance, shared gynaecological decisions and specialist-centre care; none of these strategies has been evaluated in PPAP efficacy studies [7, 8, 17, 22]. Proofreading-deficient cancers respond to immune checkpoint inhibitors, but the only applicable regulatory indication (TMB-H) exists only in the United States [19, 20, 62]. Research priorities identified by the sources include unbiased risk estimates, ClinGen variant classification specifications, calibrated functional assays and international registries [2, 17, 40].

MBD4 tumour predisposition syndrome (TPDS2) is the recessive form of polyposis due to DNA repair deficiency with a hypermutation signature: biallelic loss of the MBD4 glycosylase prevents correction of G:T mismatches generated by deamination of 5-methylcytosine, and tumours accumulate CpG>TpG transitions [26, 28]. The 21 patients described combine adenomatous polyposis in young adults with early-onset AML and, less frequently, schwannomas, CRC, uveal melanoma and meningiomas [26]. Diagnosis requires biallelic variants, and management is based on biennial colonoscopy from age 18–20 years, full blood count and annual eye examination in adults, and testing of family donors before transplantation, without formal guidelines [26]. Heterozygotes have an increased relative risk of uveal melanoma, but no demonstrated risk of CRC, and no specific surveillance is recommended for them [26, 49].

13. References

1. Orphanet. Polymerase proofreading-related polyposis. ORPHA:447877 [Internet]. Disease record. Prevalence class <1/1 000,000; autosomal dominant inheritance; ICD-10 D12.6; ICD-11 2E92.40; OMIM 612591, 615083. Accessed on 7 Oct 2026.
2. Mur P, Viana-Errasti J, García-Mulero S, Magraner-Pardo L, Muñoz IG, Pons T, et al. Recommendations for the classification of germline variants in the exonuclease domain of POLE and POLD1. Genome Med. 2023;15(1):85. PMID:37848928; PMCID:PMC10580551. doi:10.1186/s13073-023-01234-y.
3. GeneGuidance (user database). POLE and POLD1 gene records: identifiers, clinical curation, ClinGen actionability and linked conditions. Export of 26 Sep 2026, accessed on 7 Oct 2026.
4. Cree IA, Arends MJ, Khoury JD, Denton ERE, Gill AJ, Lazar AJ, et al. The WHO Classification of Genetic Tumour Syndromes: Considerations for Genetics. Clin Genet. 2026;110(3):389-401. PMID:42304730; PMCID:PMC13431718. doi:10.1111/cge.70195.
5. VariantValidator. HGVS validation and normalisation (GRCh38; MANE Select transcripts NM_006231.4, NM_002691.4, NM_001276270.2 and NM_000038.6) of variants cited in this article. variantvalidator.org. Accessed on 7 Oct 2026 and 1 Nov 2026.
6. National Center for Biotechnology Information. RefSeq records NM_006231.4/NP_006222.2 (POLE, 2286 amino acids) and NM_002691.4/NP_002682.2 (POLD1, isoform 1, 1107 amino acids), accessed through E-utilities. POLE gene; POLD1 gene. Accessed on 7 Oct 2026.
7. Palles C, Martin L, Domingo E, Chegwidden L, McGuire J, Cuthill V, et al. The clinical features of polymerase proof-reading associated polyposis (PPAP) and recommendations for patient management. Fam Cancer. 2022;21(2):197-209. PMID:33948826; PMCID:PMC8964588. doi:10.1007/s10689-021-00256-y.
8. Hodan R, Gupta S, Weiss JM, Axell L, Burke CA, Chen LM, et al. Genetic/Familial High-Risk Assessment: Colorectal, Endometrial, and Gastric, Version 3.2024, NCCN Clinical Practice Guidelines In Oncology. J Natl Compr Canc Netw. 2024;22(10):695-711. PMID:39689429. doi:10.6004/jnccn.2024.0061. Text and Figure 13 (GENE-11, POLD1/POLE) read; figure visually inspected.
9. Joo JE, Viana-Errasti J, Buchanan DD, Valle L. Genetics, genomics and clinical features of adenomatous polyposis. Fam Cancer. 2025;24(2):38. PMID:40237887; PMCID:PMC12003455. doi:10.1007/s10689-025-00460-0.
10. Wimmer K, Beilken A, Nustede R, Ripperger T, Lamottke B, Ure B, et al. A novel germline POLE mutation causes an early onset cancer prone syndrome mimicking constitutional mismatch repair deficiency. Fam Cancer. 2017;16(1):67-71. PMID:27573199; PMCID:PMC5243902. doi:10.1007/s10689-016-9925-1.
11. Clinical Genome Resource (ClinGen). Gene-Disease Validity: POLE-related polyposis and colorectal cancer syndrome (Definitive, 5 Jul 2023); POLD1-related polyposis and colorectal cancer syndrome (Definitive, 5 Jul 2023); POLD1 mandibular hypoplasia-deafness-progeroid syndrome (Definitive, 20 Jan 2025); POLD1 non-severe combined immunodeficiency due to polymerase delta deficiency (Limited, 19 Feb 2026). Curation file. Downloaded on 7 Oct 2026.
12. Buchanan DD, Stewart JR, Clendenning M, Rosty C, Mahmood K, Pope BJ, et al. Correction: Risk of colorectal cancer for carriers of a germ-line mutation in POLE or POLD1. Genet Med. 2018;20(10):1299. PMID:29388942. doi:10.1038/gim.2017.265.
13. Mur P, García-Mulero S, Del Valle J, Magraner-Pardo L, Vidal A, Pineda M, et al. Role of POLE and POLD1 in familial cancer. Genet Med. 2020;22(12):2089-2100. PMID:32792570; PMCID:PMC7708298. doi:10.1038/s41436-020-0922-2.
14. Pachlopnik Schmid J, Lemoine R, Nehme N, Cormier-Daire V, Revy P, Debeurme F, et al. Polymerase ε1 mutation in a human syndrome with facial dysmorphism, immunodeficiency, livedo, and short stature ("FILS syndrome"). J Exp Med. 2012;209(13):2323-2330. PMID:23230001; PMCID:PMC3526359. doi:10.1084/jem.20121303.
15. Logan CV, Murray JE, Parry DA, Robertson A, Bellelli R, Tarnauskaitė Ž, et al. DNA Polymerase Epsilon Deficiency Causes IMAGe Syndrome with Variable Immunodeficiency. Am J Hum Genet. 2018;103(6):1038-1044. PMID:30503519; PMCID:PMC6288413. doi:10.1016/j.ajhg.2018.10.024.
16. Weedon MN, Ellard S, Prindle MJ, Caswell R, Lango Allen H, Oram R, et al. An in-frame deletion at the polymerase active site of POLD1 causes a multisystem disorder with lipodystrophy. Nat Genet. 2013;45(8):947-950. PMID:23770608; PMCID:PMC3785143. doi:10.1038/ng.2670.
17. Zaffaroni G, Mannucci A, Koskenvuo L, de Lacy B, Maffioli A, Bisseling T, et al. Updated European guidelines for clinical management of familial adenomatous polyposis (FAP), MUTYH-associated polyposis (MAP), gastric adenocarcinoma, proximal polyposis of the stomach (GAPPS) and other rare adenomatous polyposis syndromes: a joint EHTG-ESCP revision. Br J Surg. 2024;111(5):znae070. PMID:38722804; PMCID:PMC11081080. doi:10.1093/bjs/znae070.
18. Stjepanovic N, Moreira L, Carneiro F, Balaguer F, Cervantes A, Balmaña J, et al. Hereditary gastrointestinal cancers: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2019;30(10):1558-1571. PMID:31378807. doi:10.1093/annonc/mdz233.
19. Merck Sharp & Dohme LLC. KEYTRUDA (pembrolizumab) injection, for intravenous use. Full Prescribing Information, version effective 10 Jul 2026 (sections 1, 2, 5, 14.18). DailyMed, set ID 9333c79b-d487-4538-a9f0-71b91a02b287. Accessed through openFDA on 7 Oct 2026.
20. European Medicines Agency. Keytruda (pembrolizumab): EPAR and medicine data (authorisation of 17 Jul 2015; latest record update of 3 Sep 2026; therapeutic indications). EPAR page. Accessed on 7 Oct 2026.
21. Monahan KJ, Bradshaw N, Dolwani S, Desouza B, Dunlop MG, East JE, et al. Guidelines for the management of hereditary colorectal cancer from the British Society of Gastroenterology (BSG)/Association of Coloproctology of Great Britain and Ireland (ACPGBI)/United Kingdom Cancer Genetics Group (UKCGG). Gut. 2020;69(3):411-444. PMID:31780574; PMCID:PMC7034349. doi:10.1136/gutjnl-2019-319915.
22. ClinGen Actionability Working Group. POLE and POLD1 associated susceptibility to CRC. Adult Actionability Summary Report AC152, version 1.1.4 (published on 9 Feb 2022; literature search of 5 Jul 2018); consensus score 9BB (POLE) and 10BB (POLD1). AC152 report (API). Accessed on 7 Oct 2026.
23. Idos G, Hampel H, Valle L. Lynch Syndrome. In: GeneReviews® [Internet]. Seattle: University of Washington; 1993–2026. Published on 5 Feb 2004; updated on 17 Sep 2026 (Table 4 and section on Lynch-like syndrome). NBK1211.
24. Yen T, Stanich PP, Axell L, Patel SG. APC-Associated Polyposis Conditions. In: GeneReviews® [Internet]. Seattle: University of Washington; 1993–2026. Published on 18 Dec 1998; updated on 12 May 2022 (Table 4). NBK1345.
25. de Voer RM, Nielsen M, Gao W, Kuiper RP, Hoogerbrugge N. NTHL1 Tumor Syndrome. In: GeneReviews® [Internet]. Seattle: University of Washington; 1993–2026. Published on 2 Apr 2020; updated on 20 Mar 2025 (Table 2). NBK555473.
26. Bijns C, Palles C, de Voer RM, Nielsen M. MBD4-Related Tumor Predisposition Syndrome. In: GeneReviews® [Internet]. Seattle: University of Washington; 1993–2026. Published on 13 Aug 2026. NBK624097.
27. Orphanet. MBD4-related tumor predisposition syndrome. ORPHA:661526 [Internet]. Disease record. OMIM 619975; gene MBD4. Accessed on 1 Nov 2026.
28. Palles C, West HD, Chew E, Galavotti S, Flensburg C, Grolleman JE, et al. Germline MBD4 deficiency causes a multi-tumor predisposition syndrome. Am J Hum Genet. 2022;109(5):953-960. PMID:35460607; PMCID:PMC9118112. doi:10.1016/j.ajhg.2022.03.018.
29. Harrold, Stadler. Cancer predisposition syndromes: familial adenomatous polyposis. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 27–28 (Table 2).
30. Engels M, Urbanczyk K, Hölzenspies J, Röhl C, Geverink N, Hoogerbrugge N. The European Reference Network on Genetic Tumour Risk Syndromes (ERN GENTURIS): benefits for patients, families, and health care providers. Fam Cancer. 2025;24(2):33. PMID:40159592; PMCID:PMC11955429. doi:10.1007/s10689-025-00457-9.
31. Firth HV, Hurst JA. Familial adenomatous polyposis (FAP) and adenomatous polyposis due to MUTYH, NTHL1, POLE, and POLD1. In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 568–572.
32. Palles C, Cazier JB, Howarth KM, Domingo E, Jones AM, Broderick P, et al. Germline mutations affecting the proofreading domains of POLE and POLD1 predispose to colorectal adenomas and carcinomas. Nat Genet. 2013;45(2):136-144. PMID:23263490; PMCID:PMC3785128. doi:10.1038/ng.2503.
33. Bellido F, Pineda M, Aiza G, Valdés-Mas R, Navarro M, Puente DA, et al. POLE and POLD1 mutations in 529 kindred with familial colorectal cancer and/or polyposis: review of reported cases and recommendations for genetic testing and surveillance. Genet Med. 2016;18(4):325-332. PMID:26133394; PMCID:PMC4823640. doi:10.1038/gim.2015.75.
34. Lindsay H, Scollon S, Reuther J, Voicu H, Rednam SP, Lin FY, et al. Germline POLE mutation in a child with hypermutated medulloblastoma and features of constitutional mismatch repair deficiency. Cold Spring Harb Mol Case Stud. 2019;5(5):a004499. PMID:31624068; PMCID:PMC6824253. doi:10.1101/mcs.a004499.
35. Sehested A, Meade J, Scheie D, Østrup O, Bertelsen B, Misiakou MA, et al. Constitutional POLE variants causing a phenotype reminiscent of constitutional mismatch repair deficiency. Hum Mutat. 2022;43(1):85-96. PMID:34816535. doi:10.1002/humu.24299. Only the abstract was consulted.
36. van Leerdam ME, Roos VH, van Hooft JE, Dekker E, Jover R, Kaminski MF, et al. Endoscopic management of polyposis syndromes: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy. 2019;51(9):877-895. PMID:31342472. doi:10.1055/a-0965-0605.
37. Rodrigues M, Mobuchon L, Houy A, Fiévet A, Gardrat S, Barnhill RL, et al. Outlier response to anti-PD1 in uveal melanoma reveals germline MBD4 mutations in hypermutated tumors. Nat Commun. 2018;9(1):1866. PMID:29760383; PMCID:PMC5951831. doi:10.1038/s41467-018-04322-5.
38. Sanders MA, Chew E, Flensburg C, Zeilemaker A, Miller SE, Al Hinai AS, et al. MBD4 guards against methylation damage and germ line deficiency predisposes to clonal hematopoiesis and early-onset AML. Blood. 2018;132(14):1526-1534. PMID:30049810; PMCID:PMC6172562. doi:10.1182/blood-2018-05-852566.
39. Folletet A, Helyon M, Privat M, Uhrhammer N, Gay-Bellile M, Cavaille M, et al. Case Report: Expansion of the POLD1-related polymerase proofreading-associated polyposis spectrum: first report of duodenal adenocarcinomas and characterization of two likely pathogenic variants. Front Oncol. 2025;15:1727289. PMID:41487566; PMCID:PMC12757262. doi:10.3389/fonc.2025.1727289.
40. de Voer RM, Valle L. Integrating germline and tumor sequencing to improve hereditary cancer diagnosis and care. Eur J Hum Genet. 2026;34(9):1199-1209. PMID:41731180; PMCID:PMC13549967. doi:10.1038/s41431-026-02046-5.
41. Viana-Errasti J, Marín R, García-Mulero S, Pons T, Terradas M, Capellá G, et al. Comparative analysis of somatic and germline polymerase proofreading deficiencies in cancer: molecular and clinical implications. Mod Pathol. 2025;38(12):100843. PMID:40680851. doi:10.1016/j.modpat.2025.100843. Only the abstract was consulted.
42. Firth HV, Hurst JA. Colorectal cancer (CRC). In: Oxford Desk Reference: Clinical Genetics and Genomics. 2nd ed. Oxford: Oxford University Press; 2017. p. 556–558.
43. Buchanan DD, Stewart JR, Clendenning M, Rosty C, Mahmood K, Pope BJ, et al. Risk of colorectal cancer for carriers of a germ-line mutation in POLE or POLD1. Genet Med. 2018;20(8):890-895. PMID:29120461; PMCID:PMC5943186. doi:10.1038/gim.2017.185.
44. Alsina, Dienstmann. Genomics in gastrointestinal cancers: colorectal cancer. In: ESMO Handbook — Cancer Genomics. Lugano: ESMO Press. p. 54–57 (Table 5).
45. Dadi G, Rosen S, Naor H, Hellou E, Chapal-Ilani N, Bacharach T, et al. A founder heterozygous mutation in methyl-CpG binding domain protein 4 (MBD4) prevalent among Israeli Christian Arabs predisposes to increased mutagenesis. Haematologica. 2026;111(3):1034-1038. PMID:40905073; PMCID:PMC12951175. doi:10.3324/haematol.2025.287690.
46. Cooper J, Stewart BL, Bokovitz L, Carraway H, Mukherjee S, Liska D, et al. Biallelic germline MBD4 mutations predispose to colorectal polyposis, hypermutated AML, and schwannomas. Fam Cancer. 2026;25(2):50. PMID:42101541; PMCID:PMC13156161. doi:10.1007/s10689-026-00566-z.
47. Querido I, Pinto C, Arinto P, Brandão A, Santos C, Pinheiro M, et al. A novel deleterious variant and a founder effect in four new families of MBD4-associated neoplasia syndrome recruited over a period of 20 years. Clin Genet. 2026;109(1):50-59. PMID:40605265. doi:10.1111/cge.70014.
48. Derrien AC, Rodrigues M, Eeckhoutte A, Dayot S, Houy A, Mobuchon L, et al. Germline MBD4 mutations and predisposition to uveal melanoma. J Natl Cancer Inst. 2021;113(1):80-87. PMID:32239153; PMCID:PMC7781447. doi:10.1093/jnci/djaa047.
49. Le Ven A, Villy MC, Le Mentec M, Houy A, Dubois d'Enghien C, Matet A, et al. Prevalence of the predisposing gene MBD4 for uveal melanoma. JAMA Ophthalmol. 2026;144(6):536-540. PMID:42060275. doi:10.1001/jamaophthalmol.2026.1071.
50. Villy MC, Le Ven A, Le Mentec M, Masliah-Planchon J, Houy A, Bièche I, et al. Familial uveal melanoma and other tumors in 25 families with monoallelic germline MBD4 variants. J Natl Cancer Inst. 2024;116(4):580-587. PMID:38060262. doi:10.1093/jnci/djad248.
51. Terradas M, Gonzalez-Abuin N, García-Mulero S, Viana-Errasti J, Aiza G, Piulats JM, et al. MBD4-associated neoplasia syndrome: screening of cases with suggestive phenotypes. Eur J Hum Genet. 2023;31(10):1185-1189. PMID:37402954; PMCID:PMC10545785. doi:10.1038/s41431-023-01418-5.
52. Das A, MacFarland SP, Meade J, Hansford JR, Schneider KW, Kuiper RP, et al. Clinical updates and surveillance recommendations for DNA replication repair deficiency syndromes in children and young adults. Clin Cancer Res. 2024;30(16):3378-3387. PMID:38860976. doi:10.1158/1078-0432.CCR-23-3994.
53. Liesegang A, Wu A, Heriot A, Guerra G. The 'other' colonic polyposis syndromes: evolving insights and guidance for endoscopists. Int J Colorectal Dis. 2026;41(1):146. PMID:42348001; PMCID:PMC13553667. doi:10.1007/s00384-026-05186-8.
54. Tanakaya K, Yamaguchi T, Hirata K, Yamada M, Kumamoto K, Akiyama Y, et al. Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2024 for the clinical practice of hereditary colorectal cancer. Int J Clin Oncol. 2026;31(1):1-66. PMID:41214301; PMCID:PMC12770061. doi:10.1007/s10147-025-02892-1.
55. Clinical Genome Resource (ClinGen). Criteria Specification Registry (CSpec): documents of the InSiGHT Hereditary Colorectal Cancer/Polyposis Variant Curation Expert Panel (affiliation 50099; GN089 APC, GN115/GN137/GN138/GN139 MMR genes, GN168 MUTYH in pilot preparation). CSpec (API). Accessed on 7 Oct 2026.
56. Arnaudi M, Krzesińska K, Beltrame L, Besora PS, Tiberti M, Nilbert M, et al. Interpreting the effects of DNA polymerase variants at the structural level. Mol Oncol. 2026;20(8):2054-2084. PMID:42204826; PMCID:PMC13398764. doi:10.1002/1878-0261.70255.
57. Ito H, Chino A, Ueki A, Kaneko K, Saito S. A case of polymerase proofreading-associated polyposis: challenges in genetic diagnosis. JGH Open. 2025;9(8):e70240. PMID:40741356; PMCID:PMC12307963. doi:10.1002/jgh3.70240.
58. Reitsam NG, Schneider KA, Grosser B, Jung A, Kazdal D, Gustav M, et al. Mutation enrichment in targeted panels flags immunotherapy-responsive POLE-driven hypermutated microsatellite-stable colorectal cancers. NPJ Precis Oncol. 2026;10(1):263. PMID:42399417; PMCID:PMC13346926. doi:10.1038/s41698-026-01572-1.
59. Jiang L, Yang Z, Liu X. Immune checkpoint inhibitors in POLE/POLD1 proofreading-deficient CRC: from molecular basis to clinical practice and future directions. Immunotherapy. 2026;18(4):327-337. PMID:42017297. doi:10.1080/1750743X.2026.2662823. Only the abstract was consulted.
60. Martins N, Terradas M, Garcia-Pelaez J, Sommer AK, Demidov G, Matalonga L, et al. Genome sequencing of undiagnosed European patients suspected of hereditary cancer: diagnostic yield and identification of candidate causative variants. JCO Precis Oncol. 2026;10:e2500354. PMID:41990296; PMCID:PMC13098671. doi:10.1200/PO-25-00354.
61. Nielsen M, Infante E, Brand R. MUTYH Polyposis. In: GeneReviews® [Internet]. Seattle: University of Washington; 1993–2026. Published on 4 Oct 2012; updated on 10 Oct 2019; revision of 27 May 2021. NBK107219.
62. Ambrosini M, Rousseau B, Manca P, Artz O, Marabelle A, André T, et al. Immune checkpoint inhibitors for POLE or POLD1 proofreading-deficient metastatic colorectal cancer. Ann Oncol. 2024;35(7):643-655. PMID:38777726. doi:10.1016/j.annonc.2024.03.009. Only the abstract was consulted.
63. Oh CR, Kim JE, Hong YS, Kim SY, Ahn JB, Baek JY, et al. Phase II study of durvalumab monotherapy in patients with previously treated microsatellite instability-high/mismatch repair-deficient or POLE-mutated metastatic or unresectable colorectal cancer. Int J Cancer. 2022;150(12):2038-2045. PMID:35179785. doi:10.1002/ijc.33966. Only the abstract was consulted.
64. U.S. Food and Drug Administration. FDA approves pembrolizumab for adults and children with TMB-H solid tumors. 16 Jun 2020. FDA announcement. Accessed on 7 Oct 2026.
65. ClinicalTrials.gov. NCT05255653 — Refining Adjuvant Treatment IN Endometrial Cancer Based On Molecular Features: the p53abn-RED, MMRd-GREEN, NSMP-ORANGE and POLEmut-BLUE trials (RAINBO). Sponsor: Leiden University Medical Center. Status: recruiting; updated on 15 May 2026. Registry.
66. ClinicalTrials.gov. NCT05640999 — A Phase II Study of Tailored Adjuvant Therapy in POLE-Mutated and p53-Wildtype/NSMP Early Stage Endometrial Cancer (RAINBO BLUE & TAPER). Sponsor: Canadian Cancer Trials Group. Status: recruiting; updated on 24 Sep 2026. Registry.
67. U.S. Food and Drug Administration. Drugs@FDA: BLA 125514 (KEYTRUDA), original approval of 4 Sep 2014 and supplements 071/090 approved on 16 Jun 2020. 2014 approval letter; 2020 approval letter. Accessed through openFDA on 7 Oct 2026.
68. Merck Sharp & Dohme LLC. KEYTRUDA QLEX (pembrolizumab and berahyaluronidase alfa-pmph) injection, for subcutaneous use. Prescribing Information, version effective 31 Jul 2026. DailyMed, set ID 097d166f-b73b-41d3-9b37-7653cd2a0c41. Accessed through openFDA on 7 Oct 2026.
69. AstraZeneca. IMFINZI (durvalumab) injection — prescribing information, BLA 761069. Revision of 31 Aug 2026. Silver Spring (MD): FDA; 2026 [accessed 26 Oct 2026]. Available at: https://www.accessdata.fda.gov/scripts/cder/daf/index.cfm?event=overview.process&ApplNo=761069
70. European Medicines Agency. Imfinzi (durvalumab): EPAR — product information (summary of product characteristics). Amsterdam: EMA [accessed 26 Oct 2026]. Available at: https://www.ema.europa.eu/en/documents/product-information/imfinzi-epar-product-information_en.pdf
71. ClinicalTrials.gov. NCT03435107 — A Phase II Study of Durvalumab in Patients With Mismatch Repair Deficient or POLE Mutated Metastatic Colorectal Cancer. Sponsor: Asan Medical Center. Status: completed; results posted in the registry; updated on 15 Oct 2024. Registry.
72. ClinicalTrials.gov. NCT03827044 — POLEM: Avelumab Plus 5-FU Based Chemotherapy as Adjuvant Treatment for Stage 3 MSI-High or POLE Exonuclease Domain Mutant Colon Cancer. Sponsor: Royal Marsden NHS Foundation Trust. Status: terminated (insufficient recruitment); updated on 7 Oct 2022. Registry.
73. ClinicalTrials.gov. NCT02899793 — A Phase II Evaluation of Pembrolizumab in the Treatment of Persistent or Recurrent Hypermutated/Ultramutated Endometrial Cancer. Sponsor: Yale University. Status: completed; results in the registry; updated on 11 Jun 2025. Registry.
74. ClinicalTrials.gov. NCT05103969 — Multicenter Prospective Cohort of Tumors With POLE/D1 Mutation. Sponsor: Fédération Francophone de Cancérologie Digestive. Status: unknown; updated on 8 Jul 2024. Registry.
75. Farschtschi SC, Kumps C, Milagre TH, Makrythanasis P, Van Tongerloo A, Denayer E, et al. ERN GENTURIS guideline on counselling on reproductive options for individuals with a cancer predisposition syndrome (including genturis). Eur J Hum Genet. 2026;34(3):307-313. PMID:41530368; PMCID:PMC12963376. doi:10.1038/s41431-025-02007-4.
76. Mahmood U, Bhan C, Imber C, Murphy J, Menon K, Barwick TD, et al. ctDNA guided management of POLE mutant GI malignancies promotes exceptional responses and prolonged survival to immunotherapy. Front Immunol. 2026;17:1738295. PMID:41782871; PMCID:PMC12953476. doi:10.3389/fimmu.2026.1738295. Only the abstract was consulted.
77. ClinicalTrials.gov. NCT07642323 — Safety and Efficacy of Nonoperative Management and Organ Preservation First Strategy Versus Radical Operation for dMMR/MSI-H or POLE-Mutated Gastrointestinal Cancers: a single-center, bidirectional registry study. Sponsor: Peking University Cancer Hospital & Institute. Status: not yet recruiting; first posted in the registry on 11 Jun 2026. Registry.
78. ClinicalTrials.gov. NCT07656740 — Non-Operative Management and Following Immunotherapy for Colorectal Cancer and Other GI Cancers (NOMIC). Sponsor: Peking University Cancer Hospital & Institute. Status: not yet recruiting; first posted in the registry on 18 Jun 2026. Registry.
79. Kim EJ, Jin DX, Chibly AM, Dorighi K, Sivakumar S, Kamm J, et al. Deficiency in POLE exonuclease causes synthetic lethality in highly aneuploid cancer cells. Cancer Res. 2026;86(17):4352-4364. PMID:42258764; PMCID:PMC13535316. doi:10.1158/0008-5472.CAN-25-3977. Only the abstract was consulted.
80. Saint-Ghislain M, Derrien AC, Geoffrois L, Gastaud L, Lesimple T, Negrier S, et al. MBD4 deficiency is predictive of response to immune checkpoint inhibitors in metastatic uveal melanoma patients. Eur J Cancer. 2022;173:105-112. PMID:35863105. doi:10.1016/j.ejca.2022.06.033.
81. ClinicalTrials.gov. Search for the term «MBD4»: NCT06550674 (identification of new uveal melanoma predisposition genes) and NCT06923670 (prevalence of germline variants in myeloproliferative neoplasms with a family history). clinicaltrials.gov. Accessed on 1 Nov 2026.
82. Dos Santos W, Pereira AS, Laureano T, de Andrade ES, Reis MT, Garcia FAO, et al. Whole-exome sequencing identifies new pathogenic germline variants in patients with colorectal polyposis. World J Gastroenterol. 2025;31(29):104830. PMID:40809927; PMCID:PMC12344370. doi:10.3748/wjg.v31.i29.104830.