—  SHORT COURSE  —

A PRACTICAL APPROACH TO GASTROINTESTINAL PATHOLOGY


ADENOMAS AND MALIGNANT COLORECTAL POLYPS

Robert E. Petras, M.D.





Case 10 - Invasive Moderately Differentiated Adenocarcinoma Arising in Association with a Tubular Adenoma
Case 10 - Invasive Moderately Differentiated Adenocarcinoma Arising in Association with a Tubular Adenoma
Case 10 - Invasive Moderately Differentiated Adenocarcinoma Arising in Association with a Tubular Adenoma

The nomenclature of colorectal adenomas, dysplasia and malignant polyps can be confusing. Unfortunately, there is no universally accepted nomenclature.1  Most surgical pathologists appear to apply variations of the 1989 World Health Organization (WHO) terminology.2  According to this system, the terms dysplasia, adenocarcinoma in-situ, intramucosal adenocarcinoma and invasive adenocarcinoma are acceptable and each has a precise meaning when applied to colorectal polyps. Appropriate patient care requires that the endoscopist, surgeon and surgical pathologist understand the significance of each of these terms.

I have used the following criteria which are based on the 1989 WHO. All adenomas have at least low-grade epithelial dysplasia. Without dysplasia, adenoma could not be distinguished from normal colonic mucosa. Low-grade dysplasia is characterized by a slight decrease in the amount of intracellular mucin, mild nuclear enlargement with hyperchromasia with some nuclear stratification, and an increased number of mitoses figures. With increasing degrees of dysplasia (low-grade to high-grade) there is progressive loss of intracellular mucin, progressive nuclear enlargement with stratification and a loss of nuclear polarity. Adenocarcinoma in-situ describes the next step in the dysplasia-carcinoma sequence. Here, the atypical glands assume a complex cribriform or back-to-back gland configuration but the basement membrane remains intact. Some consider carcinoma in-situ as part of the spectrum of high-grade dysplasia and report both under the same term.3  When carcinoma cells infiltrate into the lamina propria and/or muscularis mucosae only, terms such as high-grade dysplasia and carcinoma in-situ are technically no longer accurate (because both require an intact basement membrane) and the term intramucosal adenocarcinoma may be used.2  Finally, when carcinoma cells have invaded the submucosa (or beyond) the lesion is labeled invasive adenocarcinoma. Invasion of the submucosa is invariably associated with an infiltrative pattern to neoplastic glands associated with tumor desmoplasia. This tumor desmoplasia is extremely helpful in recognizing invasion (defined as infiltration of at least the submucosa) especially in small biopsy specimens.

The nomenclature controversy principally centers on the observation that in the colon, infiltrating carcinoma cells do not become clinically significant (i.e. able to metastasize) until they have invaded the submucosa.4,5  Only polyps containing invasive adenocarcinoma require a therapeutic decision on the part of the clinician. Adenoma (adenoma with dysplasia), adenocarcinoma in-situ and even intramucosal adenocarcinoma lack metastatic capability and are adequately treated by polypectomy alone.5  As a result, some pathologists advocate modification of the nomenclature to account for clinical behavior.1  Although the 1989 WHO accepted and defined two (low-grade, high-grade) or three grades (mild, moderate, severe) of dysplasia, carcinoma in-situ and intramucosal carcinoma, they recommended a similar behavior-based modification for intramucosal carcinoma stating; "… intramucosal adenocarcinoma of the colon has not been shown to metastasize, and for this reason ' carcinoma in-situ' is more appropriate. To prevent potential confusion, the term 'intramucosal carcinoma' is best avoided in the large bowel".2 

The 2000 version of the WHO classification added little clarification and introduced additional and in my opinion, confusing terminology. The authors state that the defining feature for colorectal adenocarcinoma is invasion through the muscularis mucosae into the submucosa. This is something all can agree upon. However, once adenocarcinoma has been so defined, worrisome lesions not fulfilling the criteria for adenocarcinoma become difficult to describe. For example, the WHO attempts to define adenocarcinoma in-situ and intramucosal adenocarcinoma as lesions with morphologic characteristics of "adenocarcinoma" confined to the epithelium or that "invade" the lamina propria alone and lack invasion through the muscularis mucosae. The WHO goes on to state that these lesions have virtually no risk of metastasis. According to the WHO, " … high-grade intraepithelial neoplasia is more appropriate than adenocarcinoma in-situ and that intramucosal neoplasia is more appropriate than intramucosal adenocarcinoma". The WHO believes that use of these terms will help avoid overtreatment. The problems I have with this classification are many. The loose use of the term "invasion" to describe lesions that are not by definition invasive adenocarcinoma is confusing. The lesser lesion of high-grade intraepithelial neoplasia sounds worse than the term used to describe intramucosal adenocarcinoma (intramucosal neoplasia). Furthermore, adenoma, strictly speaking is an intraepithelial or intramucosal neoplasm.

The latest effort to reach consensus (largely between Eastern [Japanese] and Western Pathologists)7,8,9  resulted in the Vienna classification of the gastrointestinal neoplasia7  presented in Table 1.

TABLE 1
VIENNA CLASSIFICATION OF GASTROINTESTINAL NEOPLASIA

Category Definition
1Negative for neoplasia/dysplasia
2Indefinite for neoplasia/dysplasia
3Non-invasive low-grade neoplasia (low grade adenoma/dysplasia)
4Non-invasive high-grade neoplasia
- High grade adenoma/dysplasia
- Non-invasive carcinoma (CIS)
- Suspicious for invasive carcinoma
5Invasive neoplasia
- Intramucosal carcinoma
- Submucosal carcinoma or beyond

Problems with the Vienna system include: a) the loose application of the word invasion, b) category IV "non-invasive" high-grade neoplasia including potentially dangerous lesions (suspicious for invasive adenocarcinoma) and; c) Category V "invasive neoplasms" including intramucosal adenocarcinoma which is known to be clinically benign in the colon. It is unlikely that a numerical system without clinical correlation will ever gain widespread acceptance.

A Biased and Pragmatic View on Nomenclature and Reporting
As modified from the 1989 WHO classification described above, low-grade dysplasia, high-grade dysplasia, adenocarcinoma in-situ, intramucosal adenocarcinoma exist and can be recognized by pathologists. This nomenclature is attractive because it can be applied throughout the entire gastrointestinal tract. If one chooses to diagnose high-grade dysplasia, adenocarcinoma in-situ, and intramucosal adenocarcinoma in colorectal biopsy specimens, I specifically mention that these lesions lack metastatic potential in the report.

Since infiltrating carcinoma cells in a colorectal polyp do not become clinically significant (i.e., able to metastasize) until they have invaded the submucosa,4,5,11-38  only a polyp containing invasive adenocarcinoma should be considered malignant. Only invasive adenocarcinoma requires a therapeutic decision. Therefore the presence or absence of invasive adenocarcinoma should be specifically mentioned in the pathology report.

Mistakes pathologists make in reporting colorectal adenoma, dysplasia, and malignant polyps occur in three major categories: 1) the pathology report is not clear. (Nonspecific or noncommittal terms are used or the presence or absence of invasion is not clearly stated); 2) mispositioned glands (pseudocarcinomatous invasion) are misinterpreted as invasive carcinoma. 3) the margin of excision is either not identified or not commented upon.

Differential Diagnosis
A common problem concerns differentiating invasive carcinoma in a colorectal polyp from pseudocarcinomatous invasion. Pseudocarcinomatous invasion describes a situation in which neoplastic glands of the adenoma have been mispositioned, presumably by trauma, into or beneath the muscularis mucosae.39-47  Pseudocarcinomatous invasion is relatively common, having been reported in 3% to 10% of resected polyps .39,41,42  Distinguishing this epithelial misplacement from invasive adenocarcinoma can be difficult. Reported series of "malignant polyps" have included, and indeed even illustrated polyps with pseudocarcinomatous invasion as examples of invasive carcinoma associated with adenoma.40  Histologic features favoring pseudoinvasion include; lack of an infiltrative pattern, lack of tumor desmoplasia, presence of lamina propria around the mispositioned glands, lack of increased atypia in mispositioned epithelium as compared to the surface epithelium of the adenoma, and the presence of hemorrhage and/or hemosiderin deposits in nearby connective tissue.

Occasionally, the misplaced glands of pseudocarcinomatous invasion can become cystic and can be associated with dissection of mucus into the connective tissue of the polyp stalk. Here the distinction between mucinous adenocarcinoma and misplaced glands can be difficult.39-48  Table 2 illustrates histologic features that aid in this differential. Remember that examination of additional sections can help in difficult cases because almost all mucinous adenocarcinomas contain foci of typical nonmucinous-type adenocarcinoma.

TABLE 2
DIFFERENTIAL FEATURES BETWEEN DISSECTING MUCUS
OF PSEUDOCARCINOMATOUS INVASION & INVASIVE
MUCINOUS ADENOCARCINOMA

Feature Pseudoinvasion Invasive Mucinous Adenocarcinoma
Shape of mucous pools Rounded Irregular, infiltrating
Location of epithelium Periphery of pool Floating in pool
Configuration of epithelium Single often discontinuous layer, basal polarity of nucleiCellular piling up, complex glandular proliferation, gland in gland configuration
Cytologic features Dysplasia similar to surface adenoma Atypia more pronounced
Tumor desmoplasia Absent Usually present
Hemorrhage and hemosiderin depositionUsually present Usually absent
Supporting lamina propria Sometimes present Absent

Malignant Colorectal Polyps - Patient Management
A rational decision concerning subsequent management of a patient with an endoscopically removed malignant colorectal polyp (one containing invasive carcinoma) is based upon weighing the chances of finding residual or metastatic cancer during a follow-up surgical resection (whom do I help?) against the risk of surgical mortality and morbidity (whom do I hurt?).

Some investigators have advocated subsequent major surgical resection for all patients.49  Currently, however, almost all surgeons and gastroenterologists have embraced a more conservative approach using a number of gross and histologic features as "indications" for follow-up colectomy.47  These include sessile growth,5,12  residual villous adenoma, a short stalk (less than 3 mm),13  "stalk" invasion,14  "level 4" invasion,5, 12  lymphatic or vascular permeation,34  lack of residual adjacent adenoma (polypoid carcinoma), poor differentiation15,16,32,34  and invasive carcinoma at or near the margin of resection.15,16,32 

The Cleveland Clinic experience16,32  parallels that of St. Mark's Hospital in London15,19  in that use of two features appears to identify a group of patients likely to avoid an adverse outcome (residual or metastatic carcinoma in a subsequent colectomy specimen or in five-year clinical follow-up). In these studies, the group of patients in the "favorable histologic group" (well or moderately differentiated adenocarcinoma with a 2 mm tumor-free margin of resection in the polypectomy specimen) experienced no adverse outcome and are considered as adequately treated by polypectomy alone. Similar therapeutic recommendations have been adopted by the American College of Gastroenterology.18  Their guidelines consider colonoscopic polypectomy definitive treatment for a patient with a malignant polyp if the following criteria are fulfilled: 1) the polyp is considered completely excised at endoscopy, 2) the specimen is properly processed in the pathology laboratory, 3) the cancer is not poorly differentiated, 4) there is no histologic evidence of vascular or lymphatic involvement, and 5) the resection margin is not involved by carcinoma.

The presence of lymphatic or venous invasion has been proposed as an indication for follow-up colectomy .5,13,17,24,25,29,34,49  However, only a few malignant polyps with these features have been reported and almost all have had positive margins, grade-III invasive carcinoma, or both.35  We believe that lymphatic/venous invasion is not a reliable criterion because the distinction of true invasion from retraction artifact is frequently difficult. Cooper et al. encountered significant interobserver variation in assessing this feature.34  Furthermore, no guidelines exist that establish the extent to which a pathologist must go to diagnose lymphatic/venous invasion (e.g., number of sections or use of immunostains). We believe that you can successfully group patients into high-risk and low-risk groups without examining lymphatic or venous invasion by using margin status and grade of invasive adenocarcinoma.32 

Therapeutic recommendations are difficult to make based upon any one or two studies because of low patient numbers. Therefore, Table 3 lists the major studies of endoscopic polypectomy for malignant polyps. Note that in the "favorable histologic group" the chance of finding residual or metastatic cancer in a subsequent colon resection specimen or during the follow-up period is about one percent. Weighing this against the published operative mortality rates for colectomy, that range between 2% to 8%,28-30  it would appear that subsequent major surgery should be avoided in this "favorable histologic group." At the Cleveland Clinic we consider these patients adequately treated by polypectomy alone.16,32 

TABLE 3
ENDOSCOPICALLY REMOVED MALIGNANT COLORECTAL POLYPS:
FREQUENCY OF CANCER AT SURGERY OR FOLLOW-UP

Cancer at Surgery or Follow-up
1st Author # of PolypsTotal ## in the
Favorable
Histologic Group*
Chantereau (29) 59 3 0
Christie (19) 70 3 0
Cooper (13) 56 9 0
Cranley (16) 40 10 0
Coverlizza (38)a 50 5 0
Cunningham (31) 36 2 0
Fried (20) 22 0 0
Geraghty (17)b 21 1 0
Hackelsberger (35)86 8 1****
Kodaira (21) 7 2 1**
Kyzer (12) 42 4 0
Langer (22) 23 6 1***
Moore (36) 54 5 1****
Morson (15) 60 3 0
Netzer (33) 32 5 0
Nivatvongs (14) 23 2 0
Richards (27) 80 10 3****
Rossini (23) 31 0 0
Shatney (24) 28 1 0
Speroni (26) 30 2 0
Volk (32) 47 10 0
Whitlow (37) 59 4 1
Wolff (25) 46 8 2**
  1,002 103 (10%) 10 (1%)

* Excision complete plus (a) cancer not close (< 2 mm) to margin, (b) not grade III.
** Assessment of resection line not stated.
*** Grade and distance from margin not stated.
****Distance from margin not clearly stated and may have been less than 2 mm.
(a) Some may have been reported in reference 23.
(b) Report details experience with 81 malignant polyps but
60 were previously reported in Reference 15.

If a decision for subsequent colon resection is made, we perform the standard recommended cancer operation because cancer was the indication for surgery. In Table 3, notice that the highest chance of an adverse outcome is approximately 10%. Therefore, finding no residual carcinoma in a follow-up resection specimen is the rule rather than the exception and does not indicate an inadequate dissection by the pathologist.

Specimen Handling and Reporting
Since evaluation of the resection line is so critical to patient management, proper handling of the polypectomy specimen is of utmost importance.16,30,32  We advocate placing the polyp immediately is adequate. We prefer Hollande's solution because it better preserves the morphology and penetrates the surface more quickly, allowing trimming after as little as 20 minutes of fixation. For polyps with a stalk, we trim on either side of the stalk as illustrated in Fig. 1. The polyp with stalk can be embedded in a block, maintaining the correct anatomic relationship. For polyps without stalks (sessile growths or those where the stalk has retracted), we look for the effects of the diathermy burn on the gross specimen. This will appear as a lighter-colored area or defect on the external surface of the polyp. We carefully trim on either side of this defect (Fig. 2). We then microscopically examine a minimum of two to three step-sections stained with hematoxylin and eosin from each block.

In the pathology report, the presence or absence of invasive carcinoma must clearly be stated. Tumor differentiation must be noted. The resection line must be identified and assessed, and the status of that resection line must be clearly stated in the final report.

While it is clear that there is no place for crusading zeal in advising surgery, the treating physician must individualize the decision for follow up colectomy weighing the estimated cancer recurrence risk against the predicted operative morbidity and mortality while considering the patients own attitudes.30  Furthermore, advancements in laparoscopic segmental resection of the colon could drastically reduce the morbidity and mortality of operative resection which now constitutes the major contraindication for surgery. This new surgical technique may require future reassessment of our current management recommendations.

* Figures 1 and 2 reproduced with permission from: Cranley JP, Petras RE, Carey WD, Paradis K, Sivak MV. When is endoscopic polypectomy adequate therapy for colonic polyps containing invasive carcinoma- Gastroenterology 91:419-427, 1986.

REFERENCES

  1. Lewin KJ. Nomenclature problems with gastrointestinal epithelial neoplasm. Am J Surg Pathol 22:1043-1047, 1998.
  2. Jass JR, Sobin LH. Histologic typing of intestinal tumours. 2nd Ed Springer-Verlag, Berlin 1989, pg 30.
  3. Pascal RR. Dysplasia and early carcinoma in inflammatory bowel disease and colorectal adenoma. Human Pathol 25:1160-1171, 1994
  4. Fenoglio CM, Kaye GI, Lane N. Distribution of human colonic lymphatics in normal, hyperplastic, and adenomatous tissue: its relationship to metastasis from small carcinomas in pedunculated adenomas, with two case reports. Gastroenterology 64:51-66, 1973.
  5. Haggitt RC, Glotzbach RE, Soffer EE, Wruble LD. Prognostic factors in colorectal carcinomas arising in adenomas: Implications for lesions removed by endoscopic polypectomy. Gastroenterology 89:328-336, 1985.
  6. Hamilton SR, Rubio CA, Vogelstein B, et al. Tumours of the colon and rectum. In: Hamilton SR, Aaltonen LA (ed). Tumours of the Digestive System. IARC Press, Lyon 2000, page 109.
  7. Schlemper RJ, Itabarni M, Kato Y, et al. Differences in the diagnostic criteria used by Japanese and Western Pathologists to diagnose colorectal carcinoma. Cancer 82:60-69, 1998.
  8. Riddell RH, Iwafuci M. Problems arising from Eastern and Western classification systems for gastrointestinal dysplasia and carcinoma: are they resolvable- Histopath 33:197-202, 1998.
  9. Jass JR. Discrepancies between East and West. Cancer 88:969-970, 2000.
  10. Schlemper RJ, Riddell RH, Kato Y, et al. The Vienna classification of gastrointestinal epithelial neoplasia. Gut 47:251-255, 2000.
  11. Morson BC, Dawson IMP, Day DW, Jass JR, Price AB, Williams GT. Morson and Dawson's Gastrointestinal Pathology. Third Edition, Oxford: Blackwell Scientific Publications, 1990.
  12. Kyzer S, Begin LR, Gordon PH, Mitmaker B. The care of patients with colorectal polyps that contain invasive adenocarcinoma; Endoscopic polypectomy or colectomy- Cancer 70:2044-2050, 1992.
  13. Cooper HS. Surgical pathology of endoscopically removed malignant polyps of the colon and rectum. Am J Surg Pathol 7:613-623, 1983.
  14. Nivatvongs S, Goldberg S. Management of patients who have polyps containing invasive carcinoma removed via colonoscope. Dis Colon Rectum 21:8-11, 1978.
  15. Morson BC, Whiteway JE, Jones EA, Macrae FA, Williams CB. Histopathology and prognosis of malignant colorectal polyps treated by endoscopic polypectomy. Gut 25:437-444, 1984.
  16. Cranley JP, Petras RE, Carey WD, Paradis K, Sivak MV. When is endoscopic polypectomy adequate therapy for colonic polyps containing invasive carcinoma- Gastroenterology 91:419-427, 1986.
  17. Geraghty JM, Williams CB, Talbot IC. Malignant colorectal polyps: venous invasion and successful treatment by endoscopic polypectomy. Gut 32:774-778, 1991.
  18. Bond JH. Position paper polyp guideline: Diagnosis, treatment, and surveillance for patients with nonfamilial colorectal polyps. Ann Intern Med 119:836-843, 1993
  19. Christie JP. Malignant colon polyps ? cure by colonoscopy or colectomy? Am J Gastroenterol 79:543?547, 1984.
  20. Fried GM, Hreno A, Duguid WP, Hampson LG. Rational management of malignant colon polyps based on long-term follow-up. Surgery 96:815-821, 1984.
  21. Kodaira S, Ono S, Purri P et al. Endoscopic polypectomy of the large bowel: Management of cancer-bearing polyps. International Surgery 66:311-314, 1981.
  22. Langer JC, Cohen Z, Taylor BR, Stafford S, Jeejeebhoy KN, Cullen JB. Management of patients with polyps containing malignancy removed by colonoscopic polypectomy. Dis Colon Rectum 27:6-9, 1984.
  23. Rossini RP, Ferrari A, Coverlizza S. Colonoscopic polypectomy in diagnosis and management of cancerous adenomas: An individual and multicentric experience. Endoscopy 14:124-127, 1982.
  24. Shatney CH, Leber PH, Gilbertsen VA, Sosin H. The treatment of pedunculated adenomatous colorectal polyps with focal cancer. Surg Gynecol Obstet 139:845-850, 1974.
  25. Wolff WI, Shinya H. Definitive treatment of "malignant" polyps of the colon. Ann Surg 182:516-525, 1975.
  26. Speroni AH, Meiss RP, Calzona C et al. Early colorectal cancer ? follow?up after endoscopic polypectomy. Endosc 20:18?20, 1988.
  27. Richards WO, Webb WA, Morris SJ et al. Patient management after endoscopic removal of the cancerous colon adenoma. Ann Surg 205:665-672, 1987
  28. Welch CE, Malt RA. Abdominal surgery. N Engl J Med 300:705-712, 1979.
  29. Chantereau MJ, Faivre J, Boutron MC, et al. Epidemiology, management, and prognosis of malignant large bowel polyps within a defined population. Gut 33:259-263, 1992.
  30. Williams CB, Saunders BP. The rationale for current practice in the management of malignant colonic polyps. Endoscopy 25:469-474, 1993.
  31. Cunningham KN, Mills LR, Schuman BM, Mwakyusa DN. Long-term prognosis of well-differentiated adenocarcinoma in endoscopically removed colorectal adenomas. Dig Dis Science 39:2034-2037, 1994.
  32. Volk EE, Goldblum JR, Petras RE, Carey W, Fazio VW. Management and outcome of patients with invasive carcinoma arising in colorectal polyps. Gastroenterology 109:1801-1807, 1995.
  33. Netzer P, Binek J, Hammer B, Lange J, Schmassmann A. Significance of histologic criteria for the management of patients with malignant colorectal polyps and polypectomy. Scand J Gastroenterol 32:910-916,1997.
  34. Cooper HS, Deppisch LM, Gourley WK, et al. Endoscopically removed malignant colorectal polyps: Clinicopathologic correlations. Gastroenterology 108:1657-1665, 1995.
  35. Hackelsburger A, Fruhmorgen P, Weiler H, Heller T, Seeliger H, Junghanns K. Endoscopic polypectomy and management of colorectal adenomas with invasive carcinoma. Endoscopy 27:155-158, 1995
  36. Moore JWE, Hoffmann DC, Roursland R. Management of the malignant colorectal polyp: The importance of clinicopathological correlation. Aust NZ J Surg 64:242-246, 1994.
  37. Whitlow C, Gathright JB Jr, Hebert SJ et al. Long-term survival after treatment of malignant colonic polyps. Dis Colon Rectum 40:929-934,1997
  38. Coverlizza S, Risio M, Ferrari A, Fenoglio-Preiser CM, Rossini FP. Colorectal adenomas containing invasive carcinoma: Pathologic assessment of lymph node metastatic potential. Cancer 64:1937-1947, 1989.
  39. Muto T, Bussey HJR, Morson BC. Pseudocarcinomatous invasion in adenomatous polyps of the colon and rectum. J Clin Pathol 26:25-31, 1973.
  40. Fechner RE. Adenomatous polyp with submucosal cysts. Am J Clin Pathol 59:498-502, 1973.
  41. Greene FL. Epithelial misplacement in adenomatous polyps of the colon and rectum. Cancer 33:206-217, 1974.
  42. Qizilbash AH, Meghji M, Castelli M. Pseudocarcinomatous invasion in adenomas of the colon and rectum. Dis Colon Rectum 23:529-535, 1980.
  43. Cheung DK, Attiyeh FF. Pseudocarcinomatous invasion of colonic polyps. Dis Colon Rectum 24:399-401, 1981.
  44. Wayte DM and Helwig EB. Colitis cystica profunda. Am J Clin Pathol 48:159-169, 1967.
  45. Petras RE. Adenomas and malignant polyps of the colon and rectum: invasive carcinoma versus pseudocarcinomatous invasion. Mod Pathol 2:250-251, 1989.
  46. Pascal RR, Hertzler G, Hunter S, Goldschmidt S. Pseudoinvasion with high-grade dysplasia in a colonic adenoma: Distinction from adenocarcinoma. Amer J Surg Pathol 14:694-697, 1990.
  47. Cooper HS, Deppisch LM, Kahn EI, Lev R, Manley PN, Pascal RR, Qizilbach AH, Rickert RR, Silverman JF, Wirman JA. Pathology of the malignant colorectal polyp. Hum Pathol 29:15-26,1998.
  48. Silver H, Stolar J. Distinguishing features of well-differentiated mucinous adenocarcinoma of the rectum and colitis cystica profunda. Am J Clin Pathol 51:493-500, 1969.
  49. Colacchio TA, Forde KA, Scantleburg VP. Endoscopic polypectomy: inadequate treatment for invasive colorectal carcinoma. Ann Surg 194:704-707, 1981.