WORKSHOP IN COLON'CANCER--RAPPORTEUR REPORT
Section Meeting on Pathology This staccato summation records the substance and conclusions of more lengthy documented presentations and deliberations. Dr. David Wood (University of California, San Francisco), Chairman, convened the session and oriented participants in this initial meeting. Members of the group reported on preassigned topics relating to pathology and cancer of the large intestine. Dr. Lauren Ackerman (State University of New York, Stonybrook) presented a survey of Precancerous Lesions of the Large Bowel. He related his experience in studying material acquired over a 10-year period from the Bantus while on a year's sabbatical in Africa. The incidence of large-bowel lesions was strikingly different from that in this part of the world. Retention (nonneoplastic) polyps were common, but there were only 6 adenomas and 96 cancers in the material studied at an institution which admitted 250 patients per day. No large-bowel polyps were found in 1,000 autopsies; in a comparable US series, he would have anticipated about 300. These findings point to the need for a study of the incidence of polyps and carcinomas in American blacks. Ulcerative colitis a recognized precancerous condition, was discussed with the notation that cancers do not arise in the pseudopolyps but in tlhe intervening mucosa. The precancerous lesions or cancers associated with ulcerative colitis as described by Morson were discussed, and it was recommended that the cancers and precancerous lesions of this disease be studied. Dr. Ackerman noted that recent studies indicate that patients with Crohn's disease have a 20 times greater than expected rate of colon cancer. It was suggested that perhaps some cases previously reported as carcinoma in ulcerative colitis were really in Crohn's disease. This is an area worthy of future study. Digestive Diseases, Vol. 19, No. 10 (October 1974)
Peutz-Jeghers syndrome was discussed. It was agreed that the intestinal polyps of this familial syndrome are not precancerous. However, it was noted that these people may develop cancer in uninvolved mucosa, in the ampulla of Vater, and in the pancreas. Villous adenoma a sessile, papillary filiform lesion, was noted to be precancerous. However, relative to the total incidence of colon cancer, this is a rare lesion. Dr. Nathan Lane (Columbia Presbyterian Medical Center, New York) discussed the problem of carcinomatous changes in adenomas (adenomatous polyps). He has had 3 cases with lymph node metastases when the carcinomatous extension was limited to the submucosa but not into the stalk or margin of resection. Otherwise, he has not seen lymph node metastases when the carcinoma did not penetrate the muscularis mucosa. He referred to his study of lymphatics in the colon published in the January 1973 issue of Gastroenterology. He described a capillarylymphatic dichotomy: there are no lymphatics in the lamina propria; they are in the submucosa and intertwine with the muscularis mucosa. Therefore, carcinoma arising in an adenoma must extend through the muscularis before penetrating lymphatics. Dr. Eleanor Deschner (Memorial Sloan-Kettering Cancer Center, New York) discussed Patterns of Cell Proliferation in Cancerous and Precancerous Lesions of the Large Bowel. Using an in vitro technique, tritiated thymidine (TdR3H) labeling of epithelial cells of the normal human colonic mucosa was limited to the lower two-thirds of the crypts. In contrast with normal mucosa, TdR3H studies of adenomas, whether occurring as a single isolated polyp in the general population or occurring in multiple or familial polyposis, showed labeling of epithe973
HUTTER
lial cells at the tips of the crypts as they open into the lumen. Apparently normal mucosa at a distance of 1-2 cm from the isolated polyp, as well as flat mucosa between excrescences, of polyposis patients revealed a high frequency of labeling at the upper surface of crypts. T h e flat intervening mucosa of familial polyposis patients had proliferation activity at the surface that arose in patches rather than uniformly throughout the mucosa. T h e colonic epithelium from "mucosal hyperplasia" also showed labeling at the tips of the crypts as did the normal mucosa from a patient with a previously excised in situ carcinoma. A discussion arose centering on whether this evidence of proliferative activity should be called only hyperplasia or whether it of itself is sufficient to be called a polyp. The colonic mucosa of 6 asymptomatic relatives with familial polyposis was studied, and 5 of the 6 manifested the surface labeling phenomenon. T h e mucosa of 10 people with no gastrointestinal symptoms or signs was examined, and 2 showed the abnormal proliferative activity. It might be concluded that (1) all polyps arise from epithelial cell proliferation at the surface and tips of crypts and form when cell birth rate is in excess of cell extrusion rate; (2) labeling at the surface in histologically normal crypts is an early sign of polyp formation; (3) the flat (normal) mucosa from people with familial polyposis and their symptom-free relatives, as well as some apparently normal people, manifest this surface labeling phenomenon; (4) this surface phenomenon is nonuniform and occurs in patches of mucosa which appear earlier and are more widespread in familial polyposis. T h e Pathologic Classification and Prognostic Factors in Colon Cancer were next reviewed by Robert V. P. Hutter (St. Barnabas Medical Center, Livingston, New Jersey). Reference was made throughout the discussion to the review by Buckwalter and Kent in Surgery, Gynecology, and Obstetrics of M a r c h 1973 on 974
prognosis and surgical pathology of colon cancer. Dr. Hutter noted that (1) prognostic factors can be related to the inherent characteristics of the cancer and certain morphologic manifestations of host response to the presence of the cancer and (2) clinical and pathological staging of the extent of disease are key factors in prognosticating. The Duke's system and the T N M system of the Union Internationale Contre le Cancer and The American Joint Committee were reviewed. T h e classification of gross types of cancer described by Wood was reviewed and included (1) polypoid, favorable type, (2) constrictive, unfavorable, (3) infiltrative, unfavorable, and (4) diffuse, very unfavorable. Various factors were tabulated in detail and were noted as to whether their presence was associated with a favorable or unfavorable prognosis: (a) Gross tumor margin: well circumscribed, favorable; poorly circumscribed, infiltrating, unfavorable. (b) T h e degree of extension through the bowel wall from mucosa to serosa is inversely related to survival. (c) T h e proportion of the circumference of the bowel wall involved is inversely related to survival. (d) T h e size of the primary cancer has no specific relationship to survival. Small anaplastic lesions have a much poorer prognosis than large polypoid cancers. (e) Margins of resection which contain cancer connote local recurrence. (f) Histologic types of cancer. This discussion was limited to the ordinary adenocarcinomas and the mucinous or colloid subtype. T h e latter represents 15% of adenocarcinomas and has a very poor prognosis. T h e 2% of patients with the signet-ring type of mucinous carcinoma rarely survive 2 years. (g) Histologic grading although still quite subjective is of value; anaplasia is inversely related to survival. (h) L y m p h node metastases may be considered in several ways. T h e mere presence of Digestive Diseases, Vol. 19, No. 10 (October 1974)
REPORT ON PATHOLOGY SECTION
modal metastases is inversely related to sur- Carcinoma by Dr. W. O. Russell (University of vival; 80% survival without and 30% survival Texas Cancer Center, Houston). Whole organ with metastases (Morson). The level of in- tissue sections, including the lesion, adjacent volvement is also important; metastases at the mucosa, and lymph nodes, made on "giant" vascular line of resection has a 14% salvage slides were demonstrated. This method procompared with 41% if the nodal metastases vides an excellent technique for studying the have not extended that far. The number of types of lesion and regional spread of disease. nodes involved is inversely related to survival. Dr. Russell distributed two work sheets on (i) Sinus histiocytosis present in one-third of colon cancer now being used at the M.D. Anresected lymph nodes bears further study. derson Hospital for data acquisition. One is (j) Blood vessel invasion occurs in 35% of being used in a retrospective study of 2,500 colon cancers. It is apparently of no significance cases and the other is being used in a prospecif limited to the submucosa, but indicates a poor tive study. The pathology section includes the prognosis when extramural veins are involved: recording of I3 factors; the first 5 are consid3~0% survival v s 55% without vascular invasion. ered of major importance: (k) Round cell infiltration of the primary tu(1) proportion of circumference of bowel mor probably indicates a favorable prognosis. wall involved by the cancer. Dr. Spjut (Baylor University Medical School, (2) maximum direct extension through the Houston) noted that an acute inflammatory wall. infiltrate with microabscess formation is rarely (3) vascular invasion. associated with nodal metastases and usually (4) the number of lymph nodes examined implies a favorable prognosis. and the number with metastases. (1) Peritoneal spread is a poor prognostic (5) the distance of the cancer from the marsign, particularly if also associated with histogins of resection, particularly if a margin logic anaplasia and when muhifocal. is less than 5 cm. Dr. Oscar Auerbach (East Orange VA HosDr. Harlan Spjut presented Newer Concepts pital, New Jersey) presented a review of Whole of Large-Bowel Cancer Based on Studies of Organ Studies done by himself and Dr. Robert Chemical Carcinogenesis and Human Cancers Rickert on unselected colons obtained at au- of the Large Bowel. The first portion of the pretopsy from patients over 20 years of age. Each sentation was devoted to a photographic demcolon was grossly examined, and any visible le- onstration of various gross and microscopic sion was charted, photographed, and removed types of carcinomas and whole organ preparafor separate histologic preparation. Then the tions. This was followed by animal data based entire specimen was serially embedded for mi- on the induction of lesions in colonies of Wistar croscopic examination. Preliminary data on 324 rats by dimethylaminobiphenyl. The research colons (with or without rectum) are as follows: was designed to answer many questions, but in (a) 324 Colons examined: 1% carcinoma (4 particular (1) can transition from a benign cases), 24% no significant lesion, 15% diver- polyp to carcinoma be demonstrated, and (2) do ticuli, 61% had some grossly visible lump or de- adenocarcinomas arise de n o v o ? Conclusions pression. under these experimental conditions indicate (b) Distribution of 511 lesions in 324 colons: that both adenocarcinomas and polyps ap8% cecum, 35% ascending colon, 25% trans- peared to arise independently and de novo. No verse colon, 12% descending colon, 13% sig- transition from polyp to carcinoma was obmoid, 7% rectum (not present in all specimens). served; however, Dr. Spjut conceded that this of This was followed by the discussion of Whole course did not mean that it does not occur. Organ Studies of Colons Surgically Resected for The findings in rat colons were further sumDigestive Diseases, Yol. 19, No. 10 (October 1974)
975
HUTTER
marized: (1) Carcinomas were most common in the distal half of the colon. (2) Polyps were uniformly distributed throughout the colon. (3) Decreasing the dose of carcinogen reduced the number of polyps and carcinomas. (4) If a segment of the colon were excluded from the fecal stream, no lesions developed in that segment. It was postulated that a metabolic product of the carcinogen is probably formed in the liver and excreted in the bile into the fecal stream. It was recommended that studies on the possible carcinogenic role of asbestos in colon cancer are worthy of further investigation. Dr. Nathan Lane discussed the Origin of Large-Bowel Adenomatous Hyperplasia. He noted that it is important in the presentation of data on frequency and distribution of colonic lesions that the tiny insignificant hyperplastic polyps be separated from adenomas, particularly from large adenomas which are definitely precancerous. In answer to Dr. Lane's question, " H a s anyone seen a minute de novo colon cancer arising in otherwise normal colon mucosa?", Dr. Weldon Bullock stated he had and promised to send slides to Dr. Lane. There followed a discussion on the origin of colon carcinoma: whether it invariably arises from adenomas or whether it may arise de novo. The conceptual differences may be in 'part due to the possibility that different pathologists may look at the same slides and see different things. The earlier discussion relating to cell kinetics of adenomatous hyperplasia was reexplored. It was noted that if the minimal changes in one crypt described and demonstrated by Dr. Deschner were interpreted as adenomas rather than as (adenomatous) hyperplasia, then by this definition any proliferative alteration in a crypt
976
which preceded carcinoma would thereby represent an antecedent adenoma. It would also follow that a carcinoma which involved only two crypts might have already obliterated the adenoma which spawned it. The presentation on Genetic Factors in Lesions of the Colon and Rectum by Dr. Oscar Auerbach included a historical review and specific significant points in relation to familial polyposis, an autosomal dominant. Other known syndromes were discussed: (1) Gardner's syndrome with polyposis coli and a variety of lesions, particularly osteomas of the head and facial bones, sebaceous cysts, and soft tissue lesions. (2) Peutz-Jegher's syndrome; perioral and mucosal pigmentation with retention or hamartomatous polyps which do not become malignant. (3) Turcot's syndrome; gastrointestinal polyposis, central nervous system tumors, and soft tissue lesions. Colon-cancer families were notea to be in an intermediate position in relation to colon cancer in the general population and in those with familial polyposis. The need for a registry for genetically determined polypoid lesions was discussed. It was suggested that if such a registry were to be developed, a medical geneticist and an interested pathologist would be important assets. In conclusion, I paraphrase an anonymous quotation: I know you understand what you heard me say; but I cannot be certain that the original authors meant exactly what you heard.
ROBERT V. P. HUTTER, MD St. Barnabas Medical Center Livingston, N e w Jersey
Digestive Diseases, Vol. 19, No. 10 (October 1974)