There is now strong evidence for an alternative pathway of colorectal carcinogenesis implicating hyperplastic polyps and serrated adenomas. This article briefly reviews the evidence for this serrated pathway, provides diagnostic criteria for clinically significant hyperplastic polyps and allied serrated polyps, and suggests how this information may be translated into safe, effective guidelines for colonoscopy-based colon cancer prevention. Consideration also is given to the definition and management of hyperplastic polyposis syndrome. The currently proposed management plan for serrated polyps is tentative because of incomplete knowledge of the nature and behavior of these polyps. This article highlights key areas warranting further research.
The traditional understanding of the evolution of colorectal cancer was based on four straightforward propositions. First, most colorectal cancers arise in precancerous polyps. Second, there are two main types of colorectal polyps: hyperplastic polyps and adenomas. Third, the adenoma is fundamentally an epithelial neoplasm, with malignant potential, whereas the hyperplastic polyp is fundamentally a benign epithelial lesion with no malignant potential. Fourth, adenomas progress to carcinomas through a single linear sequence of genetic alterations involving particular tumor suppressor genes (eg, APC and p53 ) and oncogenes (eg, KRAS ). These propositions were easy to comprehend and to translate into clinical management guidelines and were supported by considerable circumstantial evidence. Therefore, although decades of study identified exceptions to these propositions, the contrary data received no attention. In the last few years, however, these exceptions have been organized to demonstrate an alternative pathway of colorectal carcinogenesis: the serrated pathway involving hyperplastic polyps or related lesions. This article briefly presents the evidence for the serrated pathway, provides diagnostic criteria for clinically significant serrated polyps, and suggests how these data can be translated into safe and effective guidelines for cancer prevention. Because of the incomplete understanding of the behavior of serrated polyps, any clinical management plan must be tentative. The authors therefore highlight key areas warranting further research.
Colorectal Cancer: More Than One Disease
The serrated pathway concept evolved based on (1) identification of a subtype of colorectal cancer that could not be conceived to develop within a pre-existing adenoma and (2) demonstration that hyperplastic polyps and related lesions could fill the gap left by the adenoma. Study of rare forms of hereditary colorectal cancer provided the initial evidence of different molecular mechanisms for the evolution of colorectal cancer. Investigation of familial adenomatous polyposis led to the discovery of the tumor suppressor gene APC and illustrated how mutation and loss of both APC alleles can initiate the development of the sporadic adenoma . The APC gene guards against chromosomal instability, a key step in the progression of adenoma to carcinoma . Investigation of Lynch syndrome or hereditary nonpolyposis colorectal cancer led to the discovery of the DNA mismatch repair gene . Disruption of DNA mismatch repair causes a different type of genetic instability, called “microsatellite instability” (MSI). Genetic instability is fundamental to carcinogenesis; it leads to rapid accumulation of genetic alterations without apoptosis or programmed cell death.
High-level MSI (MSI-H) is not restricted to Lynch syndrome but occurs in 15% of sporadic colorectal cancer . Initially it was assumed that sporadic MSI-H cancers are initiated by APC inactivation. The advent of MSI-H then would accelerate the mutation rate of key cancer genes, including KRAS and p53 . MSI-H, however, was shown to be a rare event in adenomas, and most such examples occurred in patients who had Lynch syndrome . It therefore was assumed that MSI-H occurs at a late stage, perhaps during the transition from adenoma to carcinoma . The earlier genetic steps of mutation and loss of the tumor suppressor gene APC and mutation of the oncogene KRAS therefore should have been demonstrable within sporadic MSI-H colorectal cancer. It took many years to accept that this loss and mutation did not occur, because cancer cell lines were primarily used to screen for somatic mutations, and most MSI-H cell lines were derived from patients who had Lynch syndrome . Mutations of APC and KRAS often occur in Lynch syndrome cancers but are uncommon in sporadic MSI-H colorectal cancer . That sporadic MSI-H colorectal cancers might not be counterparts to the Lynch syndrome was suggested strongly by the identification of molecular alterations in sporadic MSI-H colorectal cancers that were not present in Lynch syndrome cancers, including mutation of the oncogene BRAF and extensive DNA methylation . DNA methylation can silence tumor suppressor genes, including the DNA mismatch repair gene MLH1 .
The characteristic changes of adenomas are not seen in sporadic MSI-H colorectal cancer, and the genetic changes found in MSI-H sporadic cancers ( BRAF mutation, extensive DNA methylation, and MSI) are rarely seen in conventional adenomas (DNA methylation is more evident in villous adenomas). It is impossible for sporadic MSI-H colorectal cancers to arise from conventional adenomas because cancers carry a genetic record of their early origins. This subtype of colorectal cancer must arise de novo, or it must arise in a different type of polyp that is characterized by somatic mutation of BRAF , extensive DNA methylation, and MSI-H. The requisite molecular changes were found to occur in a variant form of hyperplastic polyp .
Hyperplastic Polyps, Serrated Adenomas, and the Serrated Pathway to Colorectal Cancer
For half a century colorectal epithelial polyps were assumed to comprise two main and nonoverlapping groups: hyperplastic polyps and adenomas. Reports of mixed or intermediate types of polyp had occasionally appeared in the pathologic literature, but it was not until 1990 that Longacre and Fenoglio-Preiser explored the concept of an intermediate type of polyp in a detailed survey of more than 18,000 colorectal polyps. Among these polyps, 110 (0.6%) were diagnosed as serrated adenomas. In the original pathologic reports most of these polyps had not been identified as having unusual histology: about one third had been classified as hyperplastic polyps, about one third as typical adenomas, and about one third (0.2% of all polyps) as intermediate lesions. This variability in classification indicates the difficulty in the histologic diagnosis of these rare intermediate lesions. Of note, 11% of the serrated adenomas showed severe dysplasia or intramucosal carcinoma, indicating that these lesions are precancerous and may be particularly prone to malignant transformation. Moreover, these serrated adenomas occurred more often in the proximal colon (35.4%) than did either conventional adenomas (23.3%) or hyperplastic polyps (7.5%). This pathologic study concluded that serrated adenomas were fundamentally adenomas that resembled hyperplastic polyps because the crypt epithelium had adopted a serrated or saw-tooth contour similar to that of hyperplastic polyps. This study also highlighted the existence of mixed hyperplastic and adenomatous polyps. Such lesions were interpreted as chance collisions between the two common types of colorectal polyp.
Although seminal, this study did not conclude that hyperplastic polyps could transform into serrated adenomas even though residual foci of a hyperplastic polyp were sometimes identified within serrated adenomas . The concept of a serrated pathway to colorectal cancer was based on two observations concerning mixed polyps that comprise separate nondysplastic (hyperplastic) and dysplastic (adenomatous) components. First, the two components sometimes showed identical mutations in microsatellite markers. Second, the adenomatous component often was a serrated adenoma rather than a conventional adenoma . Chance collision is a highly unlikely explanation for these twin phenomena because (1) mutations are rare events and therefore are unlikely to occur by chance in contiguous lesions, and (2) serrated adenomas are much rarer than conventional adenomas.
Some dysplastic foci had multiple mutations indicative of MSI-H. It therefore was suggested that the serrated pathway could evolve into colorectal cancers with both low-level MSI (MSI-L) and MSI-H , but it still was assumed that most MSI-H colorectal cancers arise in conventional adenomas . After the investigation of mixed polyps and cancers occurring in the hyperplastic polyposis syndrome (HPS), it was argued that most, if not all, sporadic MSI-H colorectal cancers arise in hyperplastic polyps.
Hyperplastic Polyps, Serrated Adenomas, and the Serrated Pathway to Colorectal Cancer
For half a century colorectal epithelial polyps were assumed to comprise two main and nonoverlapping groups: hyperplastic polyps and adenomas. Reports of mixed or intermediate types of polyp had occasionally appeared in the pathologic literature, but it was not until 1990 that Longacre and Fenoglio-Preiser explored the concept of an intermediate type of polyp in a detailed survey of more than 18,000 colorectal polyps. Among these polyps, 110 (0.6%) were diagnosed as serrated adenomas. In the original pathologic reports most of these polyps had not been identified as having unusual histology: about one third had been classified as hyperplastic polyps, about one third as typical adenomas, and about one third (0.2% of all polyps) as intermediate lesions. This variability in classification indicates the difficulty in the histologic diagnosis of these rare intermediate lesions. Of note, 11% of the serrated adenomas showed severe dysplasia or intramucosal carcinoma, indicating that these lesions are precancerous and may be particularly prone to malignant transformation. Moreover, these serrated adenomas occurred more often in the proximal colon (35.4%) than did either conventional adenomas (23.3%) or hyperplastic polyps (7.5%). This pathologic study concluded that serrated adenomas were fundamentally adenomas that resembled hyperplastic polyps because the crypt epithelium had adopted a serrated or saw-tooth contour similar to that of hyperplastic polyps. This study also highlighted the existence of mixed hyperplastic and adenomatous polyps. Such lesions were interpreted as chance collisions between the two common types of colorectal polyp.
Although seminal, this study did not conclude that hyperplastic polyps could transform into serrated adenomas even though residual foci of a hyperplastic polyp were sometimes identified within serrated adenomas . The concept of a serrated pathway to colorectal cancer was based on two observations concerning mixed polyps that comprise separate nondysplastic (hyperplastic) and dysplastic (adenomatous) components. First, the two components sometimes showed identical mutations in microsatellite markers. Second, the adenomatous component often was a serrated adenoma rather than a conventional adenoma . Chance collision is a highly unlikely explanation for these twin phenomena because (1) mutations are rare events and therefore are unlikely to occur by chance in contiguous lesions, and (2) serrated adenomas are much rarer than conventional adenomas.
Some dysplastic foci had multiple mutations indicative of MSI-H. It therefore was suggested that the serrated pathway could evolve into colorectal cancers with both low-level MSI (MSI-L) and MSI-H , but it still was assumed that most MSI-H colorectal cancers arise in conventional adenomas . After the investigation of mixed polyps and cancers occurring in the hyperplastic polyposis syndrome (HPS), it was argued that most, if not all, sporadic MSI-H colorectal cancers arise in hyperplastic polyps.
Hyperplastic Polyposis and the Sessile Serrated Adenoma
Hyperplastic polyposis originally was perceived as lacking any cancer risk. Its clinical significance lay only in the potential for misdiagnosis as familial adenomatous polyposis . Occasional descriptions of malignancy were explained by the presence of coexisting adenomas. Case reports then appeared showing transitions from hyperplastic polyps through dysplasia (mixed polyps) to carcinoma . In a small series of hyperplastic polyposis, the dysplastic components of such mixed or progressing polyps were found to show both loss of expression of the DNA mismatch repair protein and the presence of MSI-H in microdissected DNA. Most, but not all, of the associated colorectal cancers showed MSI-H . Based on these and other data, it was argued that sporadic MSI-H colorectal cancers develop from hyperplastic polyps .
Based on analysis of a clinical series of hyperplastic polyposis complicated by colorectal cancer, Torlakovic and Snover proposed in 1996 that these polyps were not hyperplastic polyps but instead were a form of serrated adenoma. This was an important development. Torlakovic and Snover introduced the term “sessile serrated adenoma” in a detailed study of sporadic hyperplastic polyps in 2003. They proposed that there are two types of serrated adenoma, the “traditional” serrated adenoma (TSA), as highlighted in 1990, and a sessile serrated adenoma (SSA) that shows a marked predilection for the proximal colon and lacks the cytologic features of adenomatous dysplasia. The term “serrated adenoma” introduced in 1990 by Longacre and Fenoglio-Preiser , however, probably was being applied to the two main types of serrated adenoma as surmised from (1) the original classification of one third of the serrated adenomas as hyperplastic polyps, and (2) the marked predilection for serrated adenomas to be in the proximal colon. The term “TSA” therefore may be inappropriate. The work of Torlakovic and Snover greatly expanded the diagnostic limits of “hyperplastic-like” serrated adenomas by including in this category lesions with architectural, but not cytologic, atypia. A subset of SSAs (or hyperplastic-like serrated adenomas), however, is sufficiently atypical with respect to both architecture and cytology to warrant a label of dysplasia but is distinguishable from TSA.
There are important reasons for distinguishing two types of serrated adenoma. The TSA is more adenoma-like macroscopically and microscopically. It typically is pedunculated with a tubulovillous or villous architecture, and it occurs mainly in the distal colorectum . By contrast, the SSA is usually sessile (as implied by its name), has a tubular architecture, is the typical lesion in hyperplastic polyposis, and shows a strong predilection for the cecum and proximal ascending colon when it occurs sporadically . The distinction between the two types of serrated adenoma is well documented in the Japanese literature. Using dye-spray and magnification of the surface epithelial topography, some serrated adenomas were indistinguishable from hyperplastic polyps (type I serrated adenomas), whereas others had the cerebriform surface typical of tubulovillous or villous adenomas (type II serrated adenomas) . At the molecular level, most SSAs have mutation of BRAF and exhibit extensive DNA methylation . A subset may show MSI, particularly those with foci of dysplasia . By contrast, the molecular signature of TSA is more varied: some have mutation of BRAF , others have mutation of KRAS , and some have neither mutation . TSAs also show infrequent mutation of APC or p53 , or loss of heterozygosity; have no aberrant expression of beta-catenin; and rarely exhibit chromosomal instability and MSI .
Hyperplastic polyposis may be heterogeneous. It may have two phenotypes: (1) the presence of at least 30 (but not necessarily large) hyperplastic polyps in a pancolonic distribution, and (2) five hyperplastic polyps proximal to the sigmoid colon with at least two being 1 cm in diameter or larger . Some examples can meet both definitions. The proximal and large polyps occurring in the second type of hyperplastic polyposis are likely to be SSAs. Conceivably, the risks of cancer and the molecular pathway of carcinogenesis may differ, with MSI-H cancers linked more closely to the second category. More research is required into the phenotypic and genotypic diversity of hyperplastic polyposis .
Multiple Serrated Pathways to Colorectal Cancer
The existence of distinct types of serrated adenoma (SSA and TSA) raises the question of multiple pathways to what has been termed “serrated adenocarcinoma” . The SSA has been associated with proximal colorectal cancers, BRAF mutation, and extensive DNA methylation. These cancers include a subset of sporadic MSI-H colorectal cancer with methylation and inactivation of the DNA mismatch repair gene MLH1 . These colorectal cancers are more common in women . Contrariwise, TSA has been associated with mainly left-sided colorectal cancer that is more common in men and is characterized by low-level DNA methylation, MSI-L, mutation of KRAS , and methylation of the direct DNA repair gene O-6-methylguanine DNA methyltransferase .
At least 20% of colorectal cancers have molecular signatures that fit with an origin from either SSA or TSA. Although SSA and TSA have been largely ignored in polypectomy guidelines, circumstantial evidence now suggests that this omission may have led to the development of preventable cancers. In a large retrospective review, 91 MSI-H colorectal cancers were identified in patients who had an earlier diagnosis of a hyperplastic polyp within the same area of the proximal colon . Among 106 proximal hyperplastic polyps in this group of patients, all the polyps showed the features of SSA. Colonoscopic surveillance and polypectomy reduces but does not eliminate the risk of developing colorectal cancer, possibly because of cancer arising from small and/or flat adenomas, missed adenomas, or even de novo initiation of cancer . Interval colorectal cancers are three times more likely than noninterval cancers to occur in the proximal colon and are 3.7 times more likely than noninterval cancers to be MSI-H . These observations may be explained by missed cancers arising from failure to detect or remove prior SSAs. The management problem is less critical for TSAs because these typically present like and are managed like adenomas. SSAs, however, are relatively inconspicuous and often is overlooked by the pathologist, endoscopist, and virtual colonographer.
Diagnosis of Sessile Serrated Adenoma: Recognition and Nomenclature
As previously noted, SSAs originally were labeled as “hyperplastic polyps.” Based on systematic analysis of histologic features in sporadic hyperplastic polyps, Torlakovic and Snover showed that 18% of “hyperplastic polyps” satisfied the criteria for SSA as described in hyperplastic polyposis. Development of a universally acceptable nomenclature for SSA has been problematic because these lesions have subtle differences from hyperplastic polyps and do not display the features of conventional adenomas. SSAs are characterized by more exaggerated crypt serration, serration throughout the crypt length, hypermucinous epithelium, crypt dilatation, crypt branching, horizontal crypt extensions at the crypt base, and aberrant proliferation. They often exhibit slight cytologic changes and lack the nuclear changes that characterize adenomatous dysplasia . The term “SSA” dominates the literature, but the terms “sessile serrated polyp” , and “serrated polyp with abnormal proliferation” are also used . Many pathologists probably still label such lesions merely as “unusually large” or “atypical” hyperplastic polyps.
Hyperplastic polyps exhibit morphologic and molecular heterogeneity. Goblet cell hyperplastic polyps (GCHPs) typically are very small, show minimal deviation from normal, and most often occur in the distal colorectum . Most GCHPs have KRAS mutation and little DNA methylation . Microvesicular hyperplastic polyps (MVHPs) are characterized by columnar cells containing mucin-filled microvesicles with inconspicuous goblet cells, larger polyp size, and frequent proximal location . MVHPs usually have mutation of BRAF , and those occurring in the proximal colon are likely to show extensive DNA methylation . SSAs and GCHPs are relatively distinctive, but overlaps between SSAs and MVHPs suggest that these lesions represent a continuum. Although a proportion of SSAs and MVHPs are highly characteristic histologically, with high levels of diagnostic agreement among pathologists, some lesions exhibit intermediate features. Pragmatically, all proximal serrated polyps 1 cm in diameter or larger should be deemed clinically significant, but SSAs can be smaller than 1 cm and can occur in the distal colon . As precancerous lesions, SSAs need to be recognized and distinguished from clinically insignificant serrated polyps. Further research is required to determine diagnostic features that predict progression to dysplasia and cancer. It is important to develop uniformly accepted terminology and diagnostic criteria for SSA. The histologic spectrum of “serrated polyps” including GCHP, MVHP, SSA, and TSA is illustrated in Fig. 1 A–D.
Sessile Serrated Adenoma and Risk of Colorectal Cancer
Stratification of “hyperplastic polyps” into GCHP, MVHP, and SSA has occurred recently . Without this stratification, the increased risk associated with SSA would be diluted by the negligible risk associated with small, distal hyperplastic polyps . Most right-sided hyperplastic polyps (or SSAs) do not become malignant. In an autopsy study conducted in New Zealand, 43 right-sided hyperplastic polyps (12.9% prevalence) were detected among 333 subjects . Many of these lesions probably would be reclassified now as SSAs. The lifetime risk of developing colorectal cancer is approximately 5%, but only about 15% of sporadic colorectal cancers show MSI-H, the molecular characteristic of proximal serrated polyps, for an incidence of 1 in 132 of colorectal cancer with MSI-H (.15 × .05). If the incidence of right-sided hyperplastic polyps (or SSAs) is 1 per 8 patients, only 1 in 17 right-sided hyperplastic polyps develop into colorectal cancers with MSI-H (132/8).
There is confusion regarding the rapidity of malignant change in serrated polyps. SSAs rarely contain unequivocal foci of dysplasia . One of the mechanisms driving this transformation is methylation and the loss of expression of the DNA mismatch repair gene MLH1 . Once this rare but critical step has occurred, progression from dysplasia to malignancy may occur rapidly. This scenario fits with the important role of genetic instability in tumorigenesis. It is consistent with the concept of aggressive adenomas in Lynch syndrome . It explains the observed excess of interval cancers with MSI-H . It explains the rarity of SSAs caught in the act of malignant transformation . Until this or a similar transformation occurs, SSAs typically remain stable and benign.
Detection of Hyperplastic Polyps and Sessile Serrated Adenomas
Detection of hyperplastic polyps has been little analyzed because these lesions were considered to confer negligible risk of advanced colorectal neoplasia . Screening techniques to detect adenomas and early cancers include fecal tests (occult blood and DNA testing), endoscopic tests (flexible sigmoidoscopy and colonoscopy), and radiologic tests (barium enema and CT colonography [“virtual colonoscopy”]). Consideration of size, morphology, anatomic distribution, and pathologic and genetic characteristics of hyperplastic lesions narrows the choices. Hyperplastic polyps are much less likely than adenomas or carcinomas to show evidence of hemorrhage, rendering fecal occult blood testing unlikely to be effective, and some genetic markers used in fecal DNA tests (eg, KRAS ) are common in adenomas but rare in SSAs . Alternate genetic markers more specific to SSAs (eg, BRAF mutations) might overcome this limitation. The pilot data seem to be promising . CT colonography has superseded barium enema for colonic polyp detection . Flat lesions are more difficult to detect than polypoid lesions at CT colonography (see the article by Summerton and colleagues in this issue). Most hyperplastic polyps are minimally elevated, flat lesions . Flat hyperplastic polyps are more difficult to detect with CT colonography than flat adenomas , possibly because the use of air or carbon dioxide to distend bowel renders many hyperplastic polyps almost completely flat and impossible to detect with this imaging modality. No study has directly examined the ability of CT colonography to detect hyperplastic polyps and SSAs, but it seems likely that sensitivity for these polyps would be inferior to optical colonoscopy because of the limitations of CT technology . Flexible endoscopy is likely to be the most effective way to detect these lesions, but these lesions, because of their flat morphology and lack of hypervascularity, may be difficult to detect even under direct endoscopic vision ( Fig. 2 A, B). Many large and flat hyperplastic polyps are proximal and beyond the reach of a flexible sigmoidoscope . Colonoscopy supplemented by chromoendoscopy (see later discussion) is the diagnostic method of choice to detect these lesions, particularly SSAs.
Optimizing Colonoscopic Detection
Detection of hyperplastic polyps is difficult at colonoscopy, in part because, with the focus on pedunculated lesions, the importance of flat lesions has been appreciated only recently . The colonoscopic challenge is to detect a minimally elevated, mostly transparent lesion containing a meager, aberrant, spider-like superficial vascular meshwork in contrast to the hypervascular network of adenomas (see Fig. 2 B). The miss rate for all polyps in back-to-back studies of colonoscopy is 21% (95% confidence interval [CI], 14%–30%), but the miss rate for non-adenomatous polyps is higher, at 27% (95% CI, 19%–37%). In Harrison and colleagues’ study, the miss rate for hyperplastic polyps was 13 of 22 (59%). High-quality bowel preparation and adequate luminal distension is essential, as is slow colonoscopic withdrawal . A prominent coating of brown mucus that is difficult to wash off (see Fig. 2 A) can be a clue to a large underlying SSA because SSAs produce excessive mucus . Colonoscopists should be particularly vigilant for flat or inconspicuous polyps that may represent SSAs in high-risk groups such as elderly women, patients who have a family history of hyperplastic polyps or a personal history of SSA, and in regions such as the proximal colon .
In chromoendoscopy a contrast agent is sprayed onto the colonic surface during endoscopy. The most commonly used agent is indigocarmine, which accumulates in the pits (crypt openings) and innominate grooves of the colonic surface, highlighting the superficial topography, particularly flat lesions. It significantly improves polyp detection, particularly of small proximal hyperplastic polyps ( Table 1 ). A summary of four randomized trials indicates that chromoendoscopy roughly doubles the detection rate of hyperplastic polyps in the whole colon and in the proximal colon (see Table 1 ). These trials, however, did not report hyperplastic polyp size or histologic subtype, such as SSA . In a large series of asymptomatic American patients undergoing screening colonoscopy, 4% had a hyperplastic polyp removed in the proximal colon during colonoscopy without chromoendoscopy. This rate compared with rates of 9% in predominantly symptomatic patients undergoing colonoscopy without chromoendoscopy and 16% in symptomatic patients undergoing colonoscopy with chromoendoscopy in the summary of the four randomized trials. In a chromoscopic prospective hospital-based series, Spring and colleagues reported similar mean detection rates of hyperplastic polyps overall (0.81 versus 0.83) and in the proximal colon (0.23 versus 0.33). The chromoscopic detection rate of hyperplastic polyps in the entire colon also was similar in an autopsy study of New Zealanders of European origin, in which the mean rate was 0.83 per patient; in the same study however, the rates of hyperplastic polyps were significantly lower, 0.03 per patient, in the Maori/Polynesian population . An autopsy study of patients in Singapore also reported a lower rate of hyperplastic polyps than that reported in pancolonic chromoendoscopy studies conducted in predominantly European patients (7% versus 45%) . It seems, therefore, that the Spring and colleagues’ reports of the detection rate of a single, Japanese-trained endoscopist may be generalizable to Western endoscopists and populations. The indications for colonoscopy, as well as geographic location and racial distribution, affect the incidence, however.
Related posts:
Prevention of Colorectal Cancer: Diet, Chemoprevention, and Lifestyle
Implementation of Colon Cancer Screening: Techniques, Costs, and Barriers
Implementation of Colonoscopy for Mass Screening for Colon Cancer and Colonic Polyps: Efficiency with High Quality of Care
Surveillance of Patients at Increased Risk of Colon Cancer: Inflammatory Bowel Disease and Other Conditions
Surgical Therapy for Colorectal Adenocarcinoma
Colonoscopic Polypectomy
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
