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

Familial adenomatous polyposis (FAP) is an autosomal dominant inherited disease caused by pathogenic variants in the adenomatous polyposis coli (APC) gene¹ 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.³ To 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 available⁴: total colectomy with ileorectal anastomosis (IRA) or ileosigmoid anastomosis (ISA); proctocolectomy with/without mucosectomy and stapled ileal pouch-anal 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.⁵ After 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 1994⁶, 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. Long live periodical surveillance with rectoscopies is widely recommended in international guidelines as shown in Table 1.^4,7,8,9,10 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.

Our review is reported according to the PRISMA guidelines.¹¹

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 undergo 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.²⁴ Similar findings have been reported from the USA³⁸ where 8 patients operated before 1983 (12.9%) were diagnosed with CRR compared to none of those operated after 1983 when pouch surgery was introduced at the Cleveland Clinic. 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 2005³⁹ 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 high-quality, 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 stratification. Members of the International Society for Gastrointestinal Hereditary Tumors (InSiGHT)⁴⁰ proposed a staging system⁴¹ 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 USA³⁰, which advocates the use of ablative therapy with argon plasma coagulation. A similar practice was supported by a study published in France in 2007.¹⁷ National French guidelines published in 2005 recommend ablation with APC for small polyps (a few millimetres) and mucosectomy for larger polyps.³⁹

Improvements 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.

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.

There is little data on the use of advanced imaging techniques. The study from St. Mark’s hospital in London⁴⁵ showed no benefit of dye-based chromoendoscopy to detect additional adenomas in the rectal remnant. The European Society of Gastrointestinal Endoscopy (ESGE) guidelines⁴⁶ 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 Endoscopy³⁹ 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 surgery²⁰ and the fact that adenoma development in the rectal remnant is an inevitable event¹⁶, 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 2001⁵ 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 1978⁴⁷, 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.⁶ Since 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 line⁴⁸, because even after removal of all visible rectal mucosa, some microscopic rectal columnar epithelium remains at the ATZ.⁴⁹ In the study from the Heidelberg Polyposis Registry with 100 patients⁵⁰, 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 Japan¹⁶ 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 study¹⁶, a slightly lower 5-year cumulative risk but a similarly high 10-year risk was observed in a Dutch study³¹; 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 guidelines⁹ 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 adenomas³¹ and lymphoid hyperplastic nodules⁴⁵, 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 all adenomas > 3 mm.⁵¹ On the contrary, ablation with argon plasma coagulation is the preferred resection technique in a French study.¹⁷ Ablative techniques were also supported by the study from the Mayo Clinic.³⁰ In a small descriptive cohort of only 5 patients⁴², 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.⁵² Figure 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.

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).

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.