Abstract Surgery for rectal cancer continues to develop to-wards improving local control and overall survival, main-taining quality of life and preserving sphincter, genitouri-nary and sexual function. The multidisciplinary approach integrated in a team of different specialists ensures an in-dividualised treatment for each patient with rectal cancer. Thus, the role of the pathologist has acquired an important relevance, not only in diagnosis, management and evalua-tion of the surgical specimen, but also for selection of the best adjuvant treatment. Parameters such as macroscopic quality of the mesorectum, status of the circumferential margin and lymph node harvest are considered basic cri-teria by current guidelines. Additionally, consistency in reporting based on the histologic classifi cation proposed by the World Health Organization (WHO) is mandatory, along with inclusion into the pathologic report of current criteria for tumour node metastasis (TNM) staging, assess-ment of response to neoadjuvant chemoradiation therapy and clinically relevant molecular studies. Detection of de-fects in mismatch repair genes and mutational analysis of

specifi c genes should be included as predictive markers for therapy.

Keywords Rectal cancer · Pathology · Update ·Diagnosis · Quality control

Introduction

During the past two decades, important changes in thera-peutic management of rectal cancer have been incorporat-ed. In the 1980s, Heald introduced the surgical technique called total mesorectal excision (TME) [1] based on a com-plete excision of the mesorectum along the avascular tissue between the mesorectal fascia and the pelvic sidewall, in-cluding lymph and blood vessels, lymph nodes and nerves. This procedure ensures that all pathways for tumour dis-semination are included in the specimen. This technique has become the gold standard of the surgical treatment of infi ltrating rectal carcinomas located in both middle and inferior thirds [2], obtaining a dramatic decrease of local recurrences to 4% compared with 30–40% of relapses pre-viously reported.

The current guidelines advise a multidisciplinary ap-proach to colorectal cancer, with an integrated team of specialists that ensures individualised treatment for each patient. In addition, new scientifi c evidence regarding mo-lecular pathogenesis of colorectal cancer and the steady introduction of new, targeted drugs in adjuvant therapy in recent years have added new complexities to the pathologi-cal evaluation and specimen reporting.

As a consequence of these changes, the role of the pa-thologist has acquired an important relevance, not only for diagnosing and evaluating the surgical specimen, but also for clinical management of the disease and the decision-making process. The pathologist must elaborate a complete

A. Rojo (�) · G. Toledo · J.F. GarcíaDepartment of PathologyM.D. Anderson International EspañaC/ Arturo Soria, 270ES-28033 Madrid, Spaine-mail: arojo@mdanderson.es

P. Sancho · S. EncinasDepartment of Medical OncologyM.D. Anderson International EspañaMadrid, Spain

Ó. AlonsoDepartment of Surgical OncologyM.D. Anderson International EspañaMadrid, Spain

S P E C I A L A R T I C L E

Update on the surgical pathology standards on rectal cancer diagnosis, staging and quality assessment of surgery

Alejandro Rojo · Pilar Sancho · Óscar Alonso · Sara Encinas · Gemma Toledo · Juan F. García

Received: 23 April 2010 / Accepted: 17 May 2010

Clin Transl Oncol (2010) 12:431-436DOI 10.1007/s12094-010-0530-8

432 Clin Transl Oncol (2010) 12:431-436

report, including all parameters of prognostic value, and evaluate the quality of surgery. In turn, the pathologist report is audited by the oncologist, who will base on it fur-ther clinical and therapeutic decisions [3].

Surgical techniques and anatomic structures

The transition from sigmoid to rectum is marked by fusion of the tenia coli of the sigmoid to form the circumferential longitudinal muscle of the rectal wall approximately 12–15 cm from the dentate line. The rectum is defined clini-cally as the distal large intestine starting opposite the sacral promontory and ending at the anorectal ring, which cor-responds to the proximal border of the puborectalis muscle palpable on digital rectal examination. When measuring below with a rigid sigmoidoscope, it extends 16 cm from the anal verge; thus, a tumour is classifi ed as rectal if its in-ferior margin lies <16 cm from the anal verge or if any part of the tumour is located at least partly within the supply of the superior rectal artery. A tumour is classifi ed as rectosig-moid when differentiation between rectum and sigmoid is not possible.

Anterior resection (AR) is the most frequent surgi-cal procedure (70–90% in the referral centres for rectal cancer surgery). Anatomically, the surgeons include one rectal–sigmoid segment, with the surrounding mesorectum, in front of the peritoneal refl ection and the rectoprostatic fascia. The surgeon must ensure a distal margin of 2 cm, except in undifferentiated carcinomas, in which the mar-gin should be 5 cm. Moreover, the surgeon should avoid involvement of the circumferential margin, considered to be the main cause of local relapse [4]. In the neoplasms of the upper third, subtotal mesorectal excision provides satis-factory oncologic results, allowing a distal margin of 5 cm and lesser morbimortality. In the tumours of the middle and low third of the rectum, the procedure is anterior ultralow resection, being the objective the total mesorectal excision.

Abdominoperineal resection (APR) includes excision of the anus and elevator muscles. Nowadays, it is consid-ered on regression because of the tendency to preserve sphincters. Moreover, for specialised centres, APR results are not different from AR regarding relapse-free and over-all survival [5].

Preoperative staging and neoadjuvant therapy

Although pathologic analysis is considered the gold stan-dard in the diagnosis of rectal cancer, preoperative staging is necessary to establish therapeutic management, select the surgical procedures, and consider the possibility of neoadjuvant treatment [6].

When necessary, administration of neoadjuvant therapy modifi es the macroscopic aspect of the tumour and makes

managing the specimen diffi cult from the pathologist point of view, thus infl uencing sampling, lymph node harvest and microscopic analysis.

Pathology report

Macroscopic assessment of mesorectum quality

Routine analysis of mesorectum quality could improve the prognostic value of the pathology report as well as the quality of the surgical procedure. It was recently included in the recommendations of the College of American Pa-thologists (CAP) and is considered a basic criterion recom-mended for multidisciplinary team audit [7, 8].

Immediately after surgery, the pathologist should evalu-ate the fresh specimen and mesorectum quality and take photographs from the anterior and posterior side before opening the specimen. It is very important that surgeon and pathologist work together to ensure delivery of a high-quality rectal specimen and optimal pathologic assessment. Macroscopic pathologic assessment of completeness of me-sorectal specimen, scored according to the worst area, accu-rately predicts both local recurrence and distant metastasis:

– Complete excision: Intact mesorectum with minor irregularities of a smooth mesorectal surface. No defect is deeper than 5 mm, and there is no coning towards the distal margin. There is a smooth, cir-cumferential resection margin on slicing. In cases of abdominoperineal amputation, the specimen presents a complete circumferential component of striated muscle in the zone of elevator muscle insertion.

– Nearly complete: Moderate bulk to the mesorectum, but irregularity of the mesorectal surface. Moder-ate coning of the specimen is allowed. At no site is the muscularis propria visible, with the exception of insertion of elevator muscles. In the cases of abdomi-noperineal amputation, there is no striated muscle, and the resection margin is formed by muscular propria.

– Incomplete: Little bulk to mesorectum with defects down to the muscularis propria and/or very irregular circumferential resection margin. In cases of abdomi-noperineal amputation, even perforations of the wall and superfi cial presence of tumour are evident.

Sampling and analysis of the circumferential margin

In addition to addressing the proximal and distal margins, the circumferential (radial) margin must be assessed for tu-mour involvement. The radial margin represents the adven-titial soft tissue margin closest to the deepest tumour pene-tration and is created surgically by blunt or sharp dissection of the retroperitoneal or subperitoneal aspect, respectively. Multivariate analysis suggests that tumour involvement of

Clin Transl Oncol (2010) 12:431-436 433

the circumferential (radial) margin is the most critical fac-tor in predicting local recurrence [9].

After gross evaluation of the quality of mesorectal excision, the pathologist opens the specimen, paints the mesorectum tissue with black ink and fi xes the specimen in 10% formalin. After fi xation, sampling consists of trans-versal seriated sections that should include representative areas of the tumour and the surrounding mesorectum for microscopic analysis.

Tumour measurements, location, macroscopic appear-ance and distance to longitudinal margins are reported. In the microscopic report, the pathologist should record the status of the circumferential margin, giving the distance in millimetres from the tumour. When circumferential tissue is involved, it should be specifi ed whether there is a con-tinuous tumoural front, a focal tumoural impact or a lymph node with capsular breakdown. The circumferential margin is regarded as being involved when the distance between the malignant cells and the margin is 1 mm; such involvement may be through direct continuity with the main tumour, by tumour in veins, by lymphatics or lymph nodes or by tu-mour deposits discontinuous from the main growth [7].

Lymph node harvest

Lymph node harvest refl ects the quality of the pathologic report and is an important point of sampling requested by the multidisciplinary team for staging. A minimum of 12 lymph nodes should be dissected to avoid infrastaging of the patient, and recent reports recommended 15–22 lymph nodes, depending of the tumour staging [10]. This is a care-ful and time-consuming process. The pathologist should identify the maximum number of lymph nodes, a diffi cult situation in cases that have received preoperative radiother-apy. If <12 lymph nodes are found, re-examining the speci-men for additional lymph nodes, with or without visual en-hancement techniques, should be considered. The pathology report should clearly state the total number of lymph nodes examined and the total number involved by metastases. Routine examination of regional lymph nodes is limited to conventional pathologic techniques (gross assessment and histologic examination). There are no data to recommend special measures to routinely detect micrometastasis or iso-lated tumour cells. Thus, neither multiple levels of paraffi n blocks nor the use of special/ancillary techniques such as immunohistochemistry (IHC) are performed [7, 11].

An important and diffi cult issue in routine microscopic examination is to distinguish between mesorectal tumour impact and lymph node metastatic deposits. Current rec-ommendations [7] indicate that tumour nodules in the pericolonic/perirectal fat without histologic evidence of residual lymph node tissue are classifi ed as tumour depos-its (peritumoural deposit or satellite nodule) and are not considered as positive lymph nodes. In the absence of un-equivocal lymph node metastases, these perirectal tumour deposits are recorded as N1c.

Histological types

For consistency in reporting, the histologic classifi cation proposed by the World Health Organization (WHO) is recommended. Adenocarcinomas are the most frequent primary colorectal tumours and are graded on the basis of extent of glandular appearances (see below). Identifi cation of other histological types and rare variants could be clini-cally meaningful: mucinous adenocarcinoma (>50% of the lesion is composed of mucin), signet-ring-cell carcinoma (tumour cells with prominent intracytoplasmic mucin), adenosquamous carcinoma, medullary carcinoma, spindle cell, small cell, and others. By convention, signet-ring cell carcinomas, small cell carcinomas, and undifferenti-ated carcinomas are high-grade tumours. The only histo-logic types of colorectal carcinoma that have been shown to have adverse prognostic signifi cance independent of stage are signet-ring-cell [12] and small cell carcinoma (high-grade neuroendocrine carcinoma) [13]. The term carcinoma, not otherwise specifi ed is not part of the WHO classifi cation.

Medullary carcinoma is a distinctive histologic type strongly associated with high levels of microsatellite insta-bility (MSI-H), indicative of defects in normal DNA repair gene function. Medullary carcinoma may occur either sporadically or in association with hereditary nonpolyposis colon cancer (HNPCC). This tumour type is characterised by solid growth in nested, organoid or trabecular patterns, with numerous tumour-infi ltrating lymphocytes and with-out immunohistochemical evidence of neuroendocrine dif-ferentiation [14].

Histologic grade

A number of grading systems for colorectal cancer have been proposed, but a single accepted and uniformly used standard for grading is lacking. For adenocarcinomas, the WHO recommends using four-grade system: well dif-ferentiated (grade 1; lesions exhibit glandular structures in >95% of the tumour), moderately differentiated (grade 2; 50–95% glands), poorly differentiated (grade 3: 5–50% glands) and undifferentiated (grade 4; <5% of tumour are gland structures). Grading could be subjective, and in the majority of studies, documenting the prognostic power of tumour grade and the number of grades has been collapsed to produce a two-tiered stratifi cation for data analysis, as follows [7]: low-grade (well- and moderately differentiated tumours 50% gland formation) versus high-grade (poor and undifferentiated, <50% gland formation) tumours. Grading should be based on the worst area, even if this does not predominate. There is wide variation in grading concerning the stratification of low-grade tumours into well- or moderately differentiated categories, whereas in-terobserver variability in diagnosing high-grade carcinoma is relatively small.

434 Clin Transl Oncol (2010) 12:431-436

Presence of lymphatic, venous, and perineural invasion

These fi ndings have been shown by multivariate analysis to be independent indicators of poor prognosis in rectal cancer [15]. The pathologist must report the presence of neoplastic nests inside vascular and lymphatic lumen, paying special attention to the retraction artefact due to tissue processing. In some cases, IHC methods to identify endothelial cells such CD31, CD34 or D2–40 can be used. Additionally, the invasion of extramural veins in particular has been shown to be an independent indicator of unfavourable outcome and increased risk of occurrence of hepatic metastasis.

Postneoadjuvant therapy tumour analysis

Tumours in the advanced stage are treated with neoadju-vant chemoradiotherapy. This procedure has been shown to improve outcomes and reduce local recurrences [16].

Tumour regression detected by pathologic examination of the resected specimen is associated with a signifi cantly better prognosis; thus, specimens from patients receiving neoadjuvant chemoradiation should be thoroughly sectioned. Minimal residual disease has been shown to have a bet-ter prognosis than gross residual disease. Although several grading systems for tumour response have been proposed, a three-point tumour regression grade (TRG) has been shown to provide better interobserver reproducibility with similar prognostic signifi cance compared with other schemas [7]:

• TRG 0 (complete response): no viable cancer cells; the pathologist must do an exhaustive sampling of the suspicious zones and, if necessary, use immunohis-tochemical markers to rule out viable malignant cells

• TRG 1 (moderate response): single cells or small groups of cancer cells

• TRG 2 (minimal response): residual cancer outgrown by fi brosis

• TRG 3 (poor response) minimal or no tumour kill; extensive residual cancer

Complete response with tumour disappearance occurs in about 15% of cases, so the pathologist must make an exhaustive sampling from the fi brosis zone and a care-ful study of the slices to identify residual malignant cells. Tumour regression must be assessed only in the primary tumour; lymph node metastases should not be included in the assessment.

TNM staging/prognostic groupings

Surgical resection remains the most effective therapy for colorectal carcinoma, and the best estimation of prognosis is derived from the pathologic fi ndings on the resection specimen. The anatomic extent of disease is by far the most important prognostic factor in colorectal cancer. The

Dukes and Bussey modifi cation of the original Dukes clas-sifi cation of resection specimens (subdividing the tumours into A, B, C1 and C2 stages) was recommended in previ-ous guidelines. However, current protocols recommend the TNM staging system of the American Joint Committee on Cancer (AJCC) and the International Union Against Cancer (UICC) [7, 11] but do not preclude the use of other staging systems. By AJCC/UICC convention, the designation pT, pN and pM refers to primary tumours that have not been previously treated and is based on gross and microscopic examination of the surgical specimen or biopsy adequate to evaluate the highest measure of the tumour (pT), the microscopic evaluation of adequate number of nodes (pN), and microscopic examination of distant lesions (pM).

Microsatellite instability (MSI) status in colorectal tumours

Identification of MSI-H status in colorectal tumours is important, as mismatch repair defi ciency may serve as a prognostic marker of patient outcome, a predictive marker of response to chemotherapy and a screening tool for HNPCC–Lynch syndrome. Revised Bethesda guidelines for HNPCC detection recommend testing colorectal tumours for microsatellite instability in the following situations [17]:

1. Colorectal cancer diagnosed in patients younger than 50 years

2. Presence of synchronous, metachronous or other HNPCC-associated tumours (endometrial, stomach, ovar-ian, pancreas, ureter and renal pelvis, biliary tract, small bowel and brain and sebaceous adenomas and keratoacan-thomas), regardless of age

3. Colorectal cancer with MSI-H histology (presence of tumour-infi ltrating lymphocytes, Crohn-like lymphocytic reaction, mucinous/signet-ring-cell differentiation or med-ullary growth pattern) in patients younger than 60 years

4. Colorectal cancer in one or more fi rst-degree rela-tives with an HNPCC-related tumour, with one of the can-cers being diagnosed in a person younger than 50 years

5. Colorectal cancer diagnosed in two or more fi rst- or second-degree relatives with HNPCC-related tumours, re-gardless of age

Histopathologic features suggestive of microsatelliteinstability

Tumour-infi ltrating lymphocytes and medullary architecture are closely associated with MSI-H status and should be distinguished from Crohn-like peritumoural infi ltrates (lym-phoid aggregated or follicles are the tumour edge, not asso-ciated with pre-existing lymph node) [18]. Other pathologic features associated with MSI-H in colorectal carcinomas include right-sided location, high-grade histology and lack of dirty necrosis. Also, distinction should be made between

Clin Transl Oncol (2010) 12:431-436 435

traditional adenomas (tubular or villous) and the newly de-scribed sessile serrated adenomas [19]. Sessile serrated ade-noma may be the precursor lesion for colorectal carcinomas with MSI-H, are more commonly found in the right colon and are characterised by serrated architecture with bulbous dilatation of deep crypts and lack of overt nuclear atypia.

MSI testing

Examination of tissue for defective DNA mismatch repair is recommended for cases that present any of the criteria in the revised Bethesda guidelines. In addition, emerging data suggest that MIS-H in sporadic colon cancers are associ-ated with better outcome and may serve as a predictor of response to 5-fl uorouracil (5-FU)-based chemotherapy [20], although these latter indications for testing are not clearly established and have not been accepted as standard of care.

Loss of DNA mismatch repair gene expressionTumours with an MSI-H phenotype may have loss of ex-pression of at least one of several DNA mismatch repair (MMR) genes/proteins (e.g. MLH1, MSH2, MSH6 or PMS2), which can be easily detected by IHC with com-mercially available antibodies. Patients with tumours that do not exhibit an MSI-H phenotype are very unlikely to have HNPCC. However, loss of MMR expression could refl ect a germline mutation of MMR genes in patients with HNPCC or, in approximately 15% of cases of sporadic colorectal cancer, it could be due to somatic mutations or epigenetic phenomenon of gene inactivation. Thus, MMR immunohistochemical testing has high sensitivity but not necessarily high specifi city for HNPCC.

Any positive reaction in the nuclei of tumour cells is considered as intact expression (normal). Interpretation of expression loss should be made only if positive reaction is seen in internal control cells, such as the nuclei of stromal, infl ammatory or nonneoplastic epithelial cells. Intact ex-pression of all four proteins indicates that MMR proteins tested are intact but does not entirely exclude Lynch syn-drome, as approximately 5% of families may have a mis-sense mutation (especially in MLH1) that can lead to a nonfunctional protein with retained antigenicity.

Detection of MSI by PCRRecent data suggest that dinucleotide repeats may have lower sensitivity and specifi city for identifying tumours with an MSI-H phenotype. As a consequence, there has been a move towards including more mononucleotides and fewer dinucleotides in MSI testing panels. Most laborato-ries use fi ve mono- and dinucleotide markers for MSI test-ing recommended by the National Cancer Institute (NCI) [17]. To isolate DNA, pathologists should identify areas of the tumour that have at least 40% of tumours cells. MSI testing is frequently done in conjunction with IHC testing

for DNA MMR protein expression. If the results of MMR IHC and MSI testing are discordant, then the laboratory should make sure that the same sample was used for MSI and IHC testing and that there was no sample mixup. Ide-ally, the results of MMR IHC and MSI testing should be incorporated into the surgical pathology report. Also, exter-nal audit and profi ciency testing surveys are desirable.

KRAS and BRAF mutation analyses in colorectal tumours

The presence of the KRAS gene mutation has been shown to be associated with lack of clinical response to therapies targeted at the epidermal growth factor receptor (EGFR), such as cetuximab [21] and panitumumab [22]. Whereas clinical guidelines for KRAS mutational analysis are evolv-ing, current provisional recommendations from the Ameri-can Society for Clinical Oncology (ASCO) are that all patients with stage IV colorectal carcinoma who are can-didates for anti-EGFR antibody therapy should have their tumour tested for KRAS mutations. Anti-EGFR antibody therapy is not recommended for patients whose tumours show mutation in KRAS codon 12 or 13. The detection can be easily done by quantitative real-time polymerase chain reaction (RT-PCR) or direct DNA sequencing from DNA extracted from representative tumour samples. Although many diagnostic tools have been developed for KRAS mu-tation analysis, validated methods and standardised testing procedures are lacking. Given the important clinical impli-cations of the test, external audit is desirable to ensure that the laboratory assays are working as expected. Guideline recommendations of KRAS analyses and a European qual-ity assurance program for KRAS mutation testing in col-orectal carcinoma have been recently proposed [23].

Mutations in the BRAF gene have been found in 4–15% of colorectal cancers. This frequency increases to 70% in colorectal cancers with an MSI-H phenotype due to hyper-methylation of the MLH1 promoter. Cell lines with RAS/BRAF mutations are highly resistant to cetuximab in vitro compared with wild-type cells. However, there are no data available on the role of BRAF mutations in predicting clini-cal response to anti-EGFR agents. Lievre et al. screened 30 colorectal cancer patients receiving cetuximab for several mutations, including BRAF but none of these patients had a tumour with a BRAF mutation or an MSI phenotype [21].

Additionally, analysis for somatic mutations in the V600E hot spot in BRAF may be indicated for tumours that show MSI-H, as this mutation has been found in sporadic MSI-H tumours but not in HNPCC-associated cancers, independently of KRAS mutations [24]. Use of BRAF mu-tational analysis as a step before germline genetic testing in patients with MSI-H tumours may be a cost-effective means of identifying patients with sporadic tumours for whom further testing is not indicated.

Q1 Confl ict of interest The authors declare that they have no confl ict of interest relating to the publication of this manuscript.

436 Clin Transl Oncol (2010) 12:431-436

References

1. Heald RJ (1988) The ‘Holy Plane’ of rectal sur-gery. J R Soc Med 81:503–508

2. Nelson H, Sargent DJ (2001) Refi ning multimodal therapy for rectal cancer. N Engl J Med 345:690–692

3. García-Granero E, Faiz O, Muñoz E et al (2009) Macroscopic assessment of mesorectal excision in rectal cancer: a useful tool for improving qual-ity control in a multidisciplinary team. Cancer 115(15):3400–3411

4. Hall NR, Finan PJ, al-Jaberi T et al (1998) Cir-cumferential margin involvement after mesorectal excision of rectal cancer with curative intent. Predictor of survival but not local recurrence? Dis Colon Rectum 41:979–983 5. d e n Dulk M, Putter H, Collette L et al (2009) The abdominoperineal resection itself is associated with an adverse outcome: the European experi-ence based on a pooled analysis of fi ve European randomised clinical trials on rectal cancer. Eur J Cancer 45:1175–1183

6. Salerno G, Daniels IR, Moran BJ et al (2006) Clarifying margins in the multidisciplinary man-agement of rectal cancer: the MERCURY experi-ence. Clin Radiol 61:916–923

7. College of American Pathologists, Cancer Proto-cols 2010. Available at http://www.cap.org/apps/docs/committees/cancer/cancer_protocols/2009/Colon_09protocol.pdf

8. Quirke P (2003) Training and quality assurance for rectal cancer: 20 years of data is enough. Lan-cet Oncol 4:695–702

9. Quirke P, Durdey P, Dixon MF, Williams NS et al (1986) Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopatho-logical study of lateral tumour spread and surgical excision. Lancet 2:996–999

10. Mekenkamp LJ, van Krieken JH, Marijnen CA et al (2009) Lymph node retrieval in rectal can-cer is dependent on many factors—the role of the tumour, the patient, the surgeon, the radio-therapist, and the pathologist. Am J Surg Pathol 33(10):1547–1553

11. Edge SB, Byrd DR, Carducci MA, Compton CC (eds) (2009) AJCC cancer staging manual. 7th edn. Springer, New York

12. Kang H, O’Connell JB, Maggard MA et al (2005) A 10-year outcomes evaluation of mucinous and signet-ring cell carcinoma of the colon and rec-tum. Dis Colon Rectum 48(6):1161–1168

13. Bernick PE, Klimstra DS, Shia J et al (2004) Neu-roendocrine carcinomas of the colon and rectum. Dis Colon Rectum 47(2):163–169

14. Wick MR, Vitsky JL, Ritter JH et al (2005) Spo-radic medullary carcinoma of the colon: a clinico-pathologic comparison with nonhereditary poorly differentiated enteric-type adenocarcinoma and neuroendocrine colorectal carcinoma. Am J Clin Pathol 123:56–65

15. Tsai HL, Chu KS, Huang YH et al (2009) Predic-tive factors of early relapse in UICC stage I–III colorectal cancer patients after curative resection. J Surg Oncol 100:736–743

16. Sauer R, Becker H, Hohenberger W et al (2004) Preoperative versus postoperative chemoradio-therapy for rectal cancer. N Engl J Med 351:1731–1740

17. Umar A, Boland CR, Terdiman JP et al (2004) Revised Bethesda guidelines for hereditary non-polyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 96(4):261–268

18. Greenson JK, Bonner JD, Ben-Yzhak O et al (2003) Phenotype of microsatellite unstable colorectal car-cinomas. Am J Surg Pathol 27(5):563–570

19. Snover DC, Jass JR, Fenoglio-Preiser C, Batts KP (2005) Serrated polyps of the large intestine: a morphologic and molecular review of an evolv-ing concept (see comment). Am J Clin Pathol 124(3):380–391

20. Ribic CM, Sargent DJ, Moore MJ et al (2003) Tumour microsatellite-instability status as a pre-dictor of benefi t from fl uorouracil-based adjuvant chemotherapy for colon cancer. N Engl J Med 249(3):247–257

21. Lievre A, Bachet J-B, Le Corre D et al (2006) KRAS mutation status is predictive of response to cetuximab therapy in colorectal cancer. Cancer Res 66(8):3992–3995

22. Amado RG, Wolf M, Peters M et al (2008) Wild-type KRAS is required for panitumumab effi cacy in patients with metastatic colorectal cancer. J Clin Oncol 26(10):1626–1634

23. van Krieken JH, Jung A, Kirchner T et al (2008) KRAS mutation testing for predicting response to anti-EGFR therapy for colorectal carcinoma: pro-posal for an European quality assurance program. Virchows Arch 453(5):417–431

24. Bessa X, Balleste B, Andreu M et al (2008) A prospective, multicenter, population-based study of BRAF mutational analysis for Lynch syndrome screening. Clin Gastroenterol Hepatol 6(2):206–214