Endoscopic management of patients with familial adenomatous polyposis after prophylactic colectomy or restorative proctocolectomy – systematic review of the literature

Abstract Background Patients with familial adenomatous polyposis (FAP) develop early colorectal adenomas and if left untreated, progression to cancer is an inevitable event. Prophylactic surgery does not prevent further development of cancer in the rectal remnant, rectal cuff in patients with ileal pouch anal anastomosis (IPAA) and even on the ileal mucosa of the pouch body. The aim of this review is to assess long-term rates of cancer and adenoma development in patients with FAP after prophylactic surgery and to summarise current recommendations for endoscopic management and surveillance of these patients. Materials and methods A systematic literature search of studies from January 1946 through to June 2023 was conducted using the PRISMA checklist. The electronic database PubMed was searched. Results Fifty-four papers involving 5010 patients were reviewed. Cancer rate in the rectal remnant was 8.8–16.7% in the western population and 37% in the eastern population. The cumulative risk of cancer 30 years after surgery was 24%. Mortality due to cancer in the rectal remnant is 1.1–11.1% with a 5-year survival rate of 55%. The adenoma rate after primary IPAA was 9.4–85% with a cumulative risk of 85% 20 years after surgery and a cumulative risk of 12% for advanced adenomas 10 years after surgery. Cumulative risk for adenomas after ileorectal anastomosis (IRA) was 85% after 5 and 100% after 10 years. Adenomas developed more frequently after stapled (33.9–57%) compared to hand-sewn (0–33%) anastomosis. We identified reports of 45 cancers in patients after IPAA of which 30 were in the pouch body and 15 in the rectal cuff or at the anastomosis. Conclusions There was a significant incidence of cancer and adenomas in the rectal remnant and ileal pouch of FAP patients during the long-term follow-up. Regular endoscopic surveillance is recommended, not only in IRA patients, but also in pouch patients after proctocolectomy.


Introduction
Familial adenomatous polyposis (FAP) is an autosomal dominant inherited disease caused by pathogenic variants in the adenomatous polyposis coli (APC) gene 1 with reported incidence of one in 8,000 to 12,000 live births. 2 The main hallmark of the disease is the presence of multiple colorectal adenomas, leading to a 100% lifetime risk of developing cancer if the colon remains in situ. 3To prevent the development of cancer, prophylactic colectomy or proctocolectomy is performed when the adenoma burden cannot be managed endoscopically or at the age of 18-25 years old.The following types of surgery are available 4 : total colectomy with ileorectal anastomosis (IRA) or ileosigmoid anastomosis (ISA); proctocolectomy with/ without mucosectomy and stapled ileal pouchanal anastomosis (IPAA) or hand-sewn IPAA; and total proctocolectomy with end ileostomy.Until restorative proctocolectomy with IPAA and pouch reconstruction was described in the 1970s, colectomy with IRA or end ileostomy was the only surgical prophylactic procedure available and was associated to a considerable high CRC incidence and mortality. 5After this, proctocolectomy with pouch reconstruction (IPAA) was the technique of choice in patients with a high adenoma burden and was sought to eliminate the risk of CRC in FAP patients.However, since the first report of pouch cancer in 1994 6 , there has been a substantial increase in published literature reporting rates of adenoma and cancer development after primary IPAA.The development of adenomas along life in remnant rectal mucosa is a natural phenomenon in this population.[9][10] As there are no randomised trials comparing endoscopic surveillance and management strategies for FAP patients with IRA and IPAA, we aimed to systematically evaluate adenoma and cancer development after prophylactic surgery, define potential risk factors and to summarise endoscopic practices from published series.

Materials and methods
Our review is reported according to the PRISMA guidelines. 11

Search strategy
We searched PUBMED from inception to June 2023 to identify studies evaluating long-term adenoma and cancer development in patients with FAP after prophylactic surgery.Deduplication was performed using Zotero software. 12Reference lists of included studies were hand-searched for additional relevant studies.The search was limited to studies, published in English.We used the following keywords: "FAP", "IRA", "IPAA", "familial adenomatous polyposis" and "proctocolectomy".

Inclusion criteria
We included single-or multicentre retrospective cohort studies, prospective cohort studies and retrospective analyses of polyposis registries.Due to the rarity of the events, we only considered case reports for inclusion when summarising reports on cancers after primary IPAA.Only the most recent series from the same institution or polyposis registry were included in the analysis, as some research groups regularly publish retrospective analyses of their cohorts or polyposis registries.Full-text screening and data extraction were performed by a single researcher (AG).Manuscripts of three case reports could not be obtained, data were summarised from the two review articles. 13,14

Studies identified
Of 97 full-text articles screened for eligibility (Figure 1), 46 met our inclusion criteria.A further 8 articles were identified by hand searching the reference lists of the included studies (6 case reports, 1 retrospective cohort, 1 polyposis registry analysis).We included 22 retrospective analyses, 14 case reports (carcinoma development after primary IPAA), 15 retrospective analyses of prospectively maintained polyposis registries and 3 prospective cohort studies.Only 5 studies were multicentre and 1 was bi-centre.The studies were published between 1994 and 2023.The studies included between 1 and 925 patients.A total of 5010 patients were included in the review.Summary characteristics of the included studies are shown in Table 2.

Adenomas
Five studies described the rate of adenoma development in the residual rectum (Supplementary Table 1).In 8 studies that analysed the frequency of secondary proctectomy due to endoscopically unmanageable polyposis, the rate of proctectomy ranged from 3.7% to 35%. 15 Five studies described adenoma evaluated in the neoterminal ileum (Table 3), with a high variance in reported rates from 0% 16 to 47.6% in patients followed-up for median of > 20 years 17 in one study including a paediatric cohort 18 , 2 patients required resection of the terminal ileum and construction of a new IRA, one due to low grade dysplasia (LGD) and one due to high grade dysplasia (HGD) adenoma.

Rectal cancer
The reported rate of cancer in the rectal remnant (Table 4) after primary IRA is 8.8% 18 to 16.7% 19 with a median follow-up from surgery 19 of 91.1 months (3-557 months).However, studies from Japan report higher rates of up to 37% 20 , but this is due to the inclusion of in situ carcinoma in the cancer definition.The same study had the longest median follow-up of 21.1 years (3-35).On the other hand, a small cohort of 21 patients from France reported zero cases of cancer during a median follow-up of 8.4 years.Jenner et al. 21only included patients with a confirmed mutation.Five studies reported a cumulative incidence of rectal cancer ranging from 3% 22 to 17.2% 19 at 5 years, 7.7% 23 to 24.1% 19 at 10 years, 11% 22 to 23% 23 at 20 years, and 24% 22 at 30 years after the primary IRA.In one of the largest studies 24 , which analysed data from 4 national registries and 776 patients, the 10-year cumulative risk of residual rectal cancer was 4.4% (95% CI, 2.6-6.2) for patients who underwent surgery before 1990 and only 2.5% (0-5.5) after the 1990.Only one study reported the time from surgery to cancer diagnosis (median 102 months [1-26 years]) 23 ; other studies reported follow-up time from surgery, but did not clearly define when follow-up started nor the surveillance regime.Five studies reported mortality ranging from 1.6% 23 to 11.1% 20 in which 3 out of 27 patients died from cancer in the rectal remnant.Only one of two studies that examined long-term survival after diagnosis of residual rectal cancer reported a 5-year survival rate of 55%. 22In a study from Japan, 5-year survival was 94% 25 , but the excellent survival was explained by the inclusion of carcinoma in situ despite the exact proportion of these was not given.

Risk factors for progressive phenotype of rectal remnant
Eleven studies reported nine risk factors predictive of the progressive rectal residual phenotype (Supplementary Table 2).Four studies analysed the genotype-phenotype relationship; The presence of a pathogenic variant between codons 1250-1464 was an independent risk factor for subsequent cancer development (HR 4.4 [1.3-15.0] 23and for the secondary proctectomy 26,27 (HR 3.91 [1.45-10.51],P = 0.007).In a small study of 25 patients, all patients (n = 3) with carpeting rectal remnant polyposis had a pathogenic variant in codon 1309, but this was only descriptive data. 28An aggressive colonic phenotype with at least 500 polyps at time for surgery was identified as a risk factor in three studies (Supplementary Table 2).Two studies 15,25 have identified > 20 rectal remnant polyps at the time of surgery or during the endoscopic surveillance 26 as an independent risk factor for secondary proctectomy (HR 30.99 [9.57-100.32]P < 0.001), while in one study a cut-off of > 10 rectal adenomas 28 was associated with a more aggressive phenotype, as these patients developed a mean of 9.

Adenomas
Seventeen studies (Table 5) reported on the development of adenomas after IPAA, of which eight studies differentiated between the pouch body The reported rate of adenoma in the pouch body ranged from 9.4% 29 to 76.9%. 30The proportion of HGD histology among adenomas at the polyp level ranged from 5.9% 17 to 53.2%. 31In one study, the proportion of advanced adenomas on a per-patient basis was 11.2%. 31The cumulative risk of adenoma development after primary IPAA was 12% and 58% at 5 and 20 years after the surgery respectively. 16According to the analysis from Poland 32 , 50% of all patients would develop LGD 15 years after the surgery, while HGD is estimated to be present in half of the patients 17.5 years after the surgery.Six studies analysed the rate of adenoma development in the neo terminal ileum, the proportion of patients with histologically confirmed adenoma varied from 4.2% 33 to 23.1% 30 with at a median follow-up from surgery of 6.5 34 to 23.1 years. 16The cumulative risk of developing an adenoma in the neo terminal ileum was 4.4% at 20 years and increased to 36% at 30 years after the surgery as reported in the same study.The presence of pouch body ad-enomas was the only independent risk factor for the neo terminal ileum adenomas (OR, 2.16, P = 0.007). 35

Cancer
Since the first case report of cancer arising in the ileal pouch of a FAP patient in 1994 6 , we have identified 45 (Table 6) cancers that have developed in FAP patients after primary IPAA.Of these, 30 were located in the pouch body and 15 in the anastomosis/rectal cuff.The time from surgery to cancer diagnosis was reported for 22 patients and ranged from 2.3 36 to 33 years. 37The information about the interval since last follow-up was reported for only 15 patients.The shortest interval between normal endoscopic surveillance and cancer diagnosis was 9 months. 16Of the studies that reported the final outcome, 13 (28.9%)patients were alive at the last follow-up (range 8 months to 6 years) after surgical therapy and 9 patients died of disseminated cancer (1 month to 4 years after diagnosis), most despite an initial R0 resection.

Risk factors for adenoma development after primary IPAA
Nine studies analysed risk factors for adenoma development (Supplementary Table 4).None of the seven studies found a genotype-phenotype association.There was no association between colon adenoma burden at the time of surgery and subsequent development of pouch adenomas in three out of four studies.In the only positive study, none of the patients with < 200 colon polyps developed pouch adenomas, whereas almost half of the patients with > 1000 colon polyps later developed later pouch adenomas.Three studies have identified age of the pouch as a risk factor, while three others found no association between time since surgery and the rate of pouch adenomas.An association between the Spigelman score and the development of pouch adenomas was not confirmed.One study identified the presence of gastric adenomas as an independent risk factor for the development of pouch adenomas.

Discussion
Using a systematic approach, we identified a wide range of reported adenoma and cancer rates in the rectal remnant, pouch body, at IPAA and in the neoterminal ileum.The wide range in adenoma rates is probably partly due to the wide range of included studies in terms of year of publication.
The equipment and quality of optical diagnosis has improved considerably in recent years, allowing better detection of adenomas and more precise examination of the pouch and rectal remnants.In addition, the risk stratification of patients at the time of surgery has also improved, allowing patients with a more aggressive phenotype to un- dergo primary restorative proctocolectomy while primary IRA can still be offered to patients with an attenuated phenotype or low rectal disease burden.Indeed, in the largest study of four European national polyposis registries, the cumulative risk of cancer in the rectal remnant (CRR) was 10% in patients operated in the 'pre-pouch' period and only 2% in those who were operated in the 'pouch period. 24Similar findings have been reported from the USA 38 where 8 patients operated before  Recently published data from two Japanese studies reporting an overall CRR rate of 30% 25 -37% 20 must be interpreted with caution as carcinoma in situ was also included in the definition of cancer in their cohorts.The risk of metachronous cancer after IRA has been recognised early and these patients have been advised to undergo regular surveillance of the rectal remnant.Traditionally, surveillance was recommended every 3 to 12 months.This recommendation has been maintained ever since and can be also found in the recently published international guidelines (Table 1).The French national guidelines published in 2005 39 are the only ones to include the genotype information, as they recommend more frequent surveillance if the pathogenic variant is located between codons 1250-1500.However, they were published in 2005.
The main obstacle to refining recommendations for endoscopic surveillance is the lack of highquality, prospective data.Unfortunately, we have not found a single randomised trial that has compared different surveillance strategies or aimed to identify factors that would allow risk stratifi-cation.Members of the International Society for Gastrointestinal Hereditary Tumors (InSiGHT) 40 proposed a staging system 41 and stage-specific interventions for patients with intact colon and those with IRA, but unfortunately no effort has been made to validate this staging system.Data on endoscopic treatment modalities are even more descriptive.In fact, in five international recommendations (Table 1), only Vasen et al. 7 recommended endoscopic removal of all polyps with dysplasia or those larger than 5 mm.Endoscopic management of these patients has therefore been influenced by expert groups.Unfortunately, preferred methods of endoscopic management were rarely described in the reviewed studies.Maehata et al. 20 recommend removal of all polyps larger than 8 mm.A descriptive study with a small sample size (n = 6) 42 showed that large-scale cold snare polypectomy can effectively reduce the polyp burden in the rectal remnant even in cases of very high polyp numbers.The mean number of polyps removed was 78.5 (30-155).During the follow-up (mean 10.7 months), none of the patients developed rectal cancer and there were no complications related to polypectomy.This is in contrast to another study from the USA 30 , which advocates the use of ablative therapy with argon plasma coagulation.A similar  practice was supported by a study published in France in 2007. 17National French guidelines published in 2005 recommend ablation with APC for small polyps (a few millimetres) and mucosectomy for larger polyps. 39mprovements in endoscopic resection techniques have also been applied to the treatment of large lesions in the rectal remnant.Recently two reports, both from Japan 43,44 , have been published of successful endoscopic submucosal dissection (ESD) of 75 mm Is + IIa adenoma and residual adenoma at the IRA.In our endoscopy unit (Hospital Clinic, Barcelona) we also perform advanced endoscopic resection techniques.Figure 2 (A and  B) shows a recent endoscopic mucosal resection (EMR) of an 18mm laterally spreading tumour granular type (LST-G) in the rectal remnant of a patient with FAP.
There is little data on the use of advanced imaging techniques.The study from St. Mark's hospital in London 45 showed no benefit of dye-based chromoendoscopy to detect additional adenomas in the rectal remnant.The European Society of Gastrointestinal Endoscopy (ESGE) guidelines 46 published in 2014 did not recommend the use of advanced endoscopic imaging in patients with FAP, but did not specifically differentiate between the patients with intact colon and those after surgery.On the other hand, the French Society of Endoscopy 39 recommended the use of dye-based chromoendoscopy with indigo carmine.We believe that use of dye-based chromoendoscopy in these patients does not increase the detection of clinically relevant lesions and it is not routinely performed in our unit.Considering the data on a cumulative risk of 57% for CRR 30 years after surgery 20 and the fact that adenoma development in the rectal remnant is an inevitable event 16 , regular endoscopic surveillance is mandatory.Our recommendations are in line with other guidelines and our patients are recommended annual endoscopic surveillance, despite alarming data from an early study published in 2001 5 from four European registries in which 75% of patients with CRR had a negative rectoscopy within 12 months and 35% within 6 months prior to diagnosis of CRR.There was no information on the endoscopy equipment used for surveillance.We believe that the high rates of negative rectoscopies prior to cancer diagnosis may -to some extent -be influenced by the quality of endoscopy, which has been limited by the technical aspects of the equipment used in the past.This problem needs to be addressed again in the light of developments in endoscopic equipment.
When restorative proctocolectomy with IPAA was first described in 1978 47 , it was believed that this operation would eliminate the risk of colorectal cancer in patients with FAP.However, a few years later, as the first pouches began to age, case reports of cancers arising in the pouch began to appear in the literature. 6Since then, reports have become more frequent and we have identified 45 cases of cancer after primary IPAA, of which 26 arose in the ileal mucosa of the pouch body and 15 at the anastomosis.Furthermore, we now know that cancer can develop even after mucosectomy down to the dentate line 48 , because even after removal of all visible rectal mucosa, some microscopic rectal columnar epithelium remains at the ATZ. 49In the study from the Heidelberg Polyposis

A B
Registry with 100 patients 50 , rectal residual mucosa (defined as visible mucosa or detected by histology from blinded biopsies) was found in 42 (84%) cases after stapled and in 21 (42%) cases after hand-sewn anastomosis.
Researchers from Japan 16 found a 70% incidence of adenomas in the pouch body with one of the longest follow-up periods reported to date (> 20 years).Similarly, in a study from France, 74% of patients had at least one adenoma in the pouch, but with a mean follow-up of only 5.4 years.In contrast, one study found that isolated rectal cuff adenomas were more common than isolated pouch adenomas (49.1% vs. 6.8%), while 18.7% of patients had both pouch and rectal cuff adenomas.Cumulative 5-year, 10-year and 20-year risks for pouch adenomas were 32%, 52% and 68% in the Japanese study 16 , a slightly lower 5-year cumulative risk but a similarly high 10-year risk was observed in a Dutch study 31 ; 16% and 42%, but the authors of this paper did not specifically define the exact location of the adenomas.The authors also reported a 10-year cumulative risk of developing precancerous adenomas of 12.8%.
On the other hand, the adenoma rates -at least in the stapled group -seem to be higher in the studies that only looked at the anastomosis and compared hand-sewn with stapled: 0-33% vs. 33.9-57%.In view of these figures, it is essential that patients with primary IPAA also undergo regular endoscopic surveillance.Particular attention should be paid to the rectal cuff and anastomosis, and the pouch should be examined in both forward and retroflexed position.
International guidelines most commonly recommend annual endoscopy examination, whereas ASCO guidelines 9 advocate 'case-by-case' interval allocation.In 11 of only 12 studies that described a surveillance protocol, an interval of 12 months was recommended except in Brazil where endoscopy of the pouch was recommended every 2 years.
Interestingly, in the Netherlands pouch endoscopy was recommended every 1 to 3 years in the late 1990s but in 2001 the protocol was changed to annual endoscopic surveillance regardless of the anastomotic technique (hand-sewn or stapled).
One of the main concerns is the short interval (< 1 year) between the last normal endoscopy and the cancer diagnosis and the aggressive course of the disease despite an initial R0 resection (Supplementary Table 4).It is not entirely clear whether the adenoma-carcinoma sequence is faster in the ileal mucosa compared with the colon and rectum, or whether "negative" endoscopies prior to cancer diagnosis could be explained by the poor quality of pouch endoscopy.Chromoendoscopy improves the detection of diminutive adenomas 31 and lymphoid hyperplastic nodules 45 , but its use is discouraged 33,35 for the same reasons as in the examination of rectal remnants -increased of detection of clinically irrelevant polyps.Endoscopy should be performed with a gastroscope or paediatric colonoscope, as stricture can occur at the anastomosis, especially after hand suturing.
There are no official recommendations for endoscopic management of FAP patients after IPAA.We have found considerable heterogeneity in local practice.Italian authors recommend resection of

A B
all adenomas > 3 mm. 51On the contrary, ablation with argon plasma coagulation is the preferred resection technique in a French study. 17Ablative techniques were also supported by the study from the Mayo Clinic. 30In a small descriptive cohort of only 5 patients 42 , large-scale cold snare polypectomy with a mean of 110.6 (30-342) resected polyps demonstrated the efficacy of cold snare in controlling large polyp burden (> 30 polyps) with no reported polypectomy related complication.In our unit we do not use nor encourage use of argon plasma coagulation.We recommend resection of all polyps > 3 mm.Advanced resection techniques, when performed in the tertiary centres, may be a viable alternative prior to surgical resection.A case report of successful en bloc ESD of a 15 mm 'non-lifting' HGD adenoma in the ileal pouch has recently been published. 52Figure 3 (A and B) shows an EMR of 25 mm LST in a patient with FAP after IPAA.The polyp was located in the rectal cuff and extended from the anastomosis to the dentate line.
The procedure was performed at our Endoscopy Unit.It should be emphasised that the wall of the ileum is very thin and special care must be taken when resecting larger lesions.
Although there is no randomised trial comparing different endoscopic surveillance intervals, it is unlikely that prospective data will be available in the future.The main reason is ethical issue, as these patients are at increased risk of colorectal cancer.However, with the introduction of high quality colonoscopy and improvements in endoscopy technique, a 'negative' endoscopy before cancer diagnosis should become highly unlikely if not impossible.

FIGURE 1 .
FIGURE 1. Flowchart of the systematic review according to the Preferred Reporting Items for Systematic Reviews (PRISMA) schema.

FIGURE 2 .
FIGURE 2. Surveillance endoscopy in a 48-year old patient with FAP after colectomy with IRA revealed 18 m LST-G (A).After submucosal injection with gelofusine, indigo carmine and adrenaline, piecemeal endoscopic mucosal resection (pEMR) (B) was performed.

FIGURE 3 .
FIGURE 3. Surveillance endoscopy in a 49-year old patient with FAP after proctocolectomy with IPAA revealed 25 mm LST-G mixed type lesion in the rectal cuff.Lesion was spreading from the anastomosis to the dentate line.Patient had undergone surgery five years earlier and did not show up for endoscopy follow-up since then (A).Lesion was removed with pEMR (B).

TABLE 1 .
Summary of recommendations from the international guidelines ACG = American College of Gastroenterology; ASCO = American Society of Clinical Oncology; ASCRS = American Society of Colon and Rectal Surgeons; ASGE = American Society for Gastrointestinal Endoscopy; ESGE = European Society of Gastrointestinal Endoscopy; ESMO = European Society for Medical Oncology; IPAA = ileal pouch anal anastomosis; IRA = ileorectal anastomosis

TABLE 2 . Characteristics of included studies First author and publication date (ref.) No. of patients Country Setting Study design Surgery performed (period) Study population
and the anastomosis, one study only reported the anastomotic adenoma rate, while in the remaining seven studies the authors did not precisely define the anatomical location of the adenomas.The median age of patients at the time of surgery ranged from 15.4 to 34.6 years, with a median follow-up from surgery of 5.4 years to a median of 21.6 years.

Table 3
) compared the rates of adenoma development at the anastomosis between hand-sewn and stapled techniques.The incidence of adenoma was lower for hand-sewn FUP = follow up; IPAA = ileal pouch anal anastomosis; IRA = ileorectal anastomosis

TABLE 3 .
Rate of adenoma development in the neoterminal ileum in patients after ileorectal anastomosis (IRA) and ileal pouch anal anastomosis (IPAA)

TABLE 4 .
Patient characteristics and rate of rectal remnant cancer rate in patients after ileorectal anastomosis (IRA)

TABLE 5 .
Patient characteristics and rate of adenomas in patients after primary ileal pouch anal anastomosis (IPAA)

TABLE 6 .
13ncer rate after primary ileal pouch anal anastomosis (IPAA) The data from these cases has been drawn from reviews by Tajika14and Smith13as full-text of the papers were not accessible.