Prevalence And Clinicopathological Of Superficial Colorectal Neoplasms: A Retrospective Analysis

Background and ObjectiveMost colon cancer in Western countries is believed to arise from polypoid adenomas as a result of a gradual accumulation of genetic alterations.Further,interruption of this sequence by removal of adenomas(i.e.,by polypectomy) has been shown to decrease the subsequent incidence of carcinoma. This well-established adenoma-carcinoma sequence reinforces the concept that colorectal cancers arise slowly from grossly visible polyps and forms the basis of current clinical recommendations for colon cancer screening and prevention .However, recent reports form Japan have raised the prospect of alternative pathways for colorectal cancer formation. It called the "de-novo" theory,claims that a carcinoma emerges directly from normal epithelium without going through a stage of adenoma. These studies describe a peculiar type of colorectal tumor,the so-called superficial lesion which , although smaller in size than its polypoid counterpart. may be associated with a higher incidence of high-grade duslasia.Furthermore,when cancer is found in these lesion,it shows rapid submucosal invasion.The incidence and significance of superficial tumors in Western countries is controversial. Some American experts consider flat tumors to be not distinct pathologic lesions but simply very small polypoid adenoma,the higher incidence of high-grsde dysplasia reported from Japan has been attributed to differences in diagnosic criteria used by American and Japanese pathologists. So our aim was to determine the incidence of superficial lesions in our patient population and to identify differences in their clinicopathological,if any,compared with more"typical" colorectal polyps.Laterally spreading tumors (LST) represent an important subtype of superficial colonic neoplasms, and are distinguished from other superficial lesions by their size. By definition, LST are greater than 10 mm in diameter with a low vertical axis and extend laterally along the luminal wall. Two distinct subtypes of LST have been described—granular-type (G-type) and nongranular-type (NG-type).G-type lesions are composed of superficial spreading nodular aggregates forming a flat, broad-based lesion(carpet-like lesions); NG-type lesions typically have a flat, smooth surface with an absence of granularity. Aside from the phenotypic differences which have been described,these two subtypes of LST have been found to have different clinicopathologic behaviors. F-type LST are typically located in the right colon and are smaller than G-type LST.In addition, F-type LST have also been found to have a higher incidence of carcinoma with submucosal invasion as compared to G-type LST. However, the biological differences between LSTs and polyps tumors(more than 10mm in size)that have areas are unknown. In the present study, we examined the clinicopathological differences between these two types of tumors in order to explore whether they should be treated as two separate diseases.Regarding exophytic or polypoid lesions, high magnification chromoscopic colonoscopy(HMCC) is not required, as there is an established and validated correlation between size and neoplastic risk, which is the major consideration when choosing endoscopic snare polypectomy or surgical resection. However, superficial colorectal lesions do not conform to this basic rationale, where therapeutic decisions are highly dependent on the detailed morphological appearance, including the pit pattern. Indeed, some authors propose such lesions may favour a de novo pathogenic pathway where early submucosal invasion and risk of associated lymph node disease can occur. Conflicting data concerning the sensitivity, specificity, and overall accuracy of HMCC have additionally become apparent in the setting of routine clinical practice. Variability in these data is multifactorial, being in part related to operatorexperience, chromoscopic technique, and East-West ambiguity in morphological and histopathological classification. We therefore prospectively examined the efficacy of HMCC for the diagnosis of neoplasia in superficial colorectal lesions using standardised morphological, pit pattern, and histopathological criteria.MethodsPatientsBetween January 1,2001, and December 31,2005,all adult patients undergoing colonoscopies for detailed examinations after drinking 2500-3000ml wanter with 131g hesuang(complex PEG).The exclusion cariteria were familial adenomarous ployposis(FAP), hered itary nonpolyposis colorectal cancer(HNPCC),inflammatory bowel disease(IBD),presences of coagulopathy.ProceduresAll patients using the Olympus CF-Q240ZI magnify colonoscopy and Olympus CF-Q240I chromoendoscopy. Chromoendoscopy with indigocarmine (0.1%) dye spraying was performed in every patient with the slightest hint of a suspiciousmucosal change or tiny lesion. Resected lesions were fixed in 10% formalin solution for 24 h, then the mucosa were stained by using hematoxylin for 1 min. A stereomicroscopic view of each tumor was photographed. After a stereomicroscopic examination, the obtained materials were embedded in paraffin. A 4-mra cross-section was cut and chemically stained with hematoxylin and eosin after being deparaffinized.DifinitionsMaterials were morphologically divided into two types; polypoid and superficial types. The superficial type was further divided into three subtypes: flat-elevated(IIa),flat(IIb) and depressed types(IIc).We advocated a category "laterally spreading tumor (LST)", which is defined as lesions larger than 10mm in diameter but extending circumferentially rather than vertically.We have documented all resected lesions with pit-pattern findings. The pit patterns were classified according to the modified Kudo criteria I, II, IIIs (small), IIIL(large), IV, Vi (irregular), and Vn (nonstructural). Histological diagnosis was made according to the World Health Organization.Statistical analysis,Statistical analysis was performed usingchi-square test,.Any P value less than 0.05 was considered to be significant. All data were expressed as mean±SD.Results1. There were found 94 LST and 485 polypoid tumors which were large than 10mm in size. The average age of LST lesions were older than comparable polypoid tumors (59.56±11.60years old vs. 54.52±13 years old, P≤0. 05).2. The average size of LST was significantly large than comparable polypoid tumors (25.5±14.66mm vs. 17.8±11.03mm,P≤0.05).3. Incidence rate on rectum,ascending colon and cecum of LST tumors and polypoid tumors were identified in 42.6% vs.27.4%, 14.7% vs. 9.7% and 4.4%VS.2.4%, respectively. Incidence rate on rectum,descending colon, ascending colon and cecum of G-LST tumors and NG-LST tumors were identified in 53.1% VS.37.1%,6.3%VS. 4.8%,15.5% VS.14.5%and 6.3%VS.3.3%, respectively. (P≤0.05)4. The pit pattern of the most LST tumors were type IV,which was about 45.74%,while among the polypoid tumors the pit pattern of III_L is 61.44%. Incidence rates of tubulo-villous adenoma,villous adenoma cancer and serrate adenoma were observed in 25.5% VS.12.8%, 25.5% VS.23.3%, 14.9% VS.9.1%, 6.4%VS.5.6% (P≤0.05).5. There were 5 early cancers of 14 cancer in LST, while 7 of 44 in polypoid cancers.6. Superfical lesions were found in 4.36% of patients,and these fewer likely to be found than polypoid lesions(19.04%). The average age of Superfical lesions was significantly large than comparable polypoid lesions (53 years old VS.50 years old, P≤0.05). The average size of Superfical lesions was significantly smaller than comparable polypoid lesions(6.9 VS.8.3mm) (P≤0.05) .7. The average size of superfical with early cancer was significantly smaller than comparable polypoid lesions(24 VS.55mm) (P=0.05) .the invasive rates in superficial adenomas were 12.2% for those of 10mm~19mm,and 21.7%for those of more than 30mm,with these rates being even higher than those for polypoid lesions.8. The majority of superfical and polypoid lesions were located in the left colon and rectum,. Incidence rate on rectum of superfical and polypoid lesions were identified in 42.5% vs.25.9%.9. The detection rates of tumors and early cancer were observed in 55.6% vs.37.4% and 2.1% vs. 2.7% , superfical and polypoid lesions,respectively (P≤0.005) .10. Incidence rates of tubulo-villous adenoma and cancer observed in IIa type lesions were higher than that of polypoid lesions(10.3% vs.5.6% , 7.1% vs.3.9%) . Incidence rates of villous adenoma, tubulo-villous adenoma and serreat adenoma observed in IIa+IIc type lesions were higher than that of polypoid lesions(23.7%vs.5.6%, 28.9% vs.12.1%,7.9%vs.4.6%) . Incidence rates of cancer observed in IIc+IIa type lesions were higher than that of polypoid lesions(75% vs.3.9%) .11. The detection rates of severe dysplasia were observed in 12.2% vs.5.1%, Superfical and polypoid tumor,respectively (P≤0.005) .12. There were 94 patients with LST, 445 patients withⅡa and 24 patients withⅡb , and 9 patient withⅡc were found in totally 5721esions using HMCC.13. The finding of the pit patterns by magnifying endoscope were highly consistent with those by steromicroscope in these patients.Of 556 Superfical lesions,175were diagnosed pathologically as neoplastic lesions (adenomas and carcinomas).14. In 159 neoplastic lesions,150 demonstrated III_L,IIIs,IV,Vi and Vn pit pattern, with 88 moderated atypis and sever atypia. 14Vi and Vn pit patterns which includes maglignant change in 11 case ,and 2 moderate dysplasia and 1 severe dysplasia.9 demonstrated II pit patern with only 7 mild atypia and 2 moderated atypis and no severe atypia.15. The sensitivity and specificity of HMCC in distinguishing non-neoplastic from neoplastic lesions were 83.25% and 97.07%, respectively.Conclusions:1. It seems appropriate that these LSTs should be distinguished from polypoid tumors and regarded as constituting a new clinical entity.2. We conclude that there is a significant prevalence of superficial colonrectal lesions in this country and that these lesions have significant different clinical and clinicopathological features than polypoid lesions.3. HMCC has a high overall accuracy at discriminating neoplastic from non-neoplastic lesions but is not 100% accurate. HMCC is a useful diagnostic tool in vivo but presently is not a replacement for histology. Requirements for further education and training in these techniques need to be address.