CARCINOMA AND EPITHELIAL DYSPLASIA COMPLICATING ULCERA TIVE COLITIS

Transcription

1 GASTROENTEROLOGY 68: , by The Williams & Wilkins Co. Vol. 68, No.5, Part 1 Printed in U.S.A. CARCINOMA AND EPITHELIAL DYSPLASIA COMPLICATING ULCERA TIVE COLITIS M. G. CooK, M. R. C. PATH., AND J. C. GoLIGHER, M.S., F.R.C.S. University Department of Surgery, the General Infirmary at Leeds, Leeds, England The pathology of 19 specimens of carcinoma complicating ulcerative colitis, resected in the University Department of Surgery of the General Infirmary at Leeds, was reviewed with particular reference to the incidence of epithelial dysplasia. The carcinomas were found to be more frequently multiple, more evenly distributed in the large bowel, and much more often of atypical macroscopic appearances and of mucoid histological type than ordinary colorectal carcinomas, but the proportion of poorly differentiated tumours complicating ulcerative colitis was not as high as previously reported. Of the patients in our series 26% are alive and well at least 5 years after surgery. Unequivocal epithelial dysplasia was demonstrated in some part of the large intestine in 15 of 19 specimens with colitis carcinoma, but was also found in 4 of 14 specimens from a "control" series of patients with longstanding total colitis but without carcinoma. Clearly, therefore, the finding of dysplasia in a rectal biopsy of a patient with colitis is not a reliable guide to the presence of a frank carcinoma elsewhere in the bowel. Whether it indicates a special predisposition to the development of a growth in the future, as might be postulated from the analogy of similar changes in other organs, cannot be determined on the data of this study. The fact that epithelial dysplasia when present in cojitis is often patchy in distribution and frequently spares the rectum even in patients with definite carcinomas makes a negative rectal biopsy :garticularly unreliable in deciding on the absence of a tumour or the lack of predisposition to it. Multiple biopsies from different parts of the colon as well as the rectum would thus seem to be desirable if mucosal sampling is to be employed as a screening test. It is generally accepted that patients with ulcerative colitis are predisposed to the development of carcinoma of the large bowel. 1 ' 10 This risk is greater in those patients with a long history of extensive disease, particularly when it began in Received July 15, Accepted October 24, The work on which this paper is based was undertaken by one of us (M. G. C.) while in receipt of a grant from the Governors of the United Leeds Hospitals. We are grateful to Dr. D. J. Evans for his constructive comments, and to Mr. W. F. Hinkes for assistance with the photomicrography childhood; however, it is still not possible on clinical grounds to select from this group those who will develop colorectal cancer. Morson and Pang 11 noted that in patients with colitis who had already developed colonic cancers, rectal biopsy usually showed epithelial dysplasia (or precancer). They suggested that the detection of such dysplastic changes on rectal biopsy might be helpful in identifying patients with a special predisposition to the subsequent development of carcinomas. Others, although agreeing about the

2 1128 COOK AND GOLIGHER Vol. 68, No.5, Part 1 occurrence of epithelial atypia in chronic ulcerative colitis, emphasized that the changes of dysplasia can be difficult to distinguish from reactive hyperplasia and are patchy in their distribution. These points have recently been reemphasized by Yardley and Keren 13 whose report also endorses the need for exact definitions of precancer or epithelial dysplasia and its distinction from reactive hyperplasia. Over the past 15 years, in the University Department of Surgery in the General Infirmary at Leeds, 24 patients have been seen with carcinoma complicating ulcerative colitis. In 5 of these the lesion was advanced and quite inoperable at the time of presentation. The other 19 patients were all submitted to excisional surgery and it has been possible to review the pathological features of their carcinomas and of the underlying colitis in the operative specimens, with particular reference to the extent and degree of epithelial dysplasia. These changes were compared with those seen in a "control" group of specimens derived from patients who had had long histories of extensive colitis but without carcinomatous change. Material and Methods The 19 patients with carcinoma of the large bowel complicating ulcerative colitis were all treated by complete proctocolectomy between 1958 and The average age of the patients at the time of operation was 51.3 years (range 28 to 74 years), and the average length of their histories of colitis was 18 years (range 3 to 34 years). The resected specimens were fixed in 10% formalin after opening and pinning out in a manner resembling the anatomical arrangement. Blocks for histology were taken in every case from the cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid colon, and rectum, as well as at least two from the tumor and additional samples from areas of macroscopic interest. The average total was 15 blocks (range 13 to 18). The mesenteric lymph nodes were sought and mapped so that Dukes' staging could be given. The control patients were selected from a l~rge series of surgically treated colitis patients, simply on the basis that they had had a long history of total involvement of the colon by the disease and the pathological examination of their specimens had not shown any evidence of carcinoma. The same anatomical areas of the large bowel were sampled in these control cases with an average total of blocks of 14 (range 13 to 16). There were 14 patients in this group. They had an average length of history of 11.7 years (range 8 to 16). In this study each specimen in the primary series and in the control group was reexamined macroscopically for the number and sites of any carcinomas and microscopically for the grade of differentiation, the amount of mucin production, and the Dukes' staging of the growths. The inflammatory activity of the ulcerative colitis was also assessed in terms of severity and extent. Activity was indicated by the presence of an increased number of mononuclear cells and polymorphs in the lamina propria, and epithelial changes such as mucin depletion, crypt abscesses, and ulceration. A systematic search was conducted throughout the specimens for evidence of epithelial dysplasia. Histology of colorectal epithelial dysplasia. We based our definitions of epithelial dysplasia on that given for "precancer" by Morson and Pang. 11 The area of dysplasia may be unrecognizable macroscopically but in severe forms the mucosal surface appears velvety or even warty. Microscopically, these areas have striking features. The glands are lengthened, which tends to produce villous projections (fig. 1) responsible for the velvety appearance macroscopically. The glands in chronic ulcerative colitis are in any case almost invariably irregular and the surface epithelium frequently is uneven and nodular. Nevertheless such changes are readily distinguishable from dysplasia in that any projections are more bulbous or polypoid and are associated with attenuation of the intervening mucosa. In contrast, the villous projections in dysplastic mucosa are crowded without intervening mucosal thinning, rather a marked over-all thickening of the mucosa in the affected area. More importantly the cellular detail of epithelial dysplasia is distinctive (figs. 2, 3, and 4). As in other premalignant conditions the epithelial cells show a marked increased in the nuclear cytoplasmic ratio, but the nuclei vary in their degree of enlargement even in adjacent cells (fig. 2). In reparative hyperplasia the nuclei are enlarged in a uniform manner and not to the same degree as in dysplasia (fig. 5). In addition, the nuclear chromatin is more homogeneous than in dysplastic epithelium where there are variable

3 May 1975 CARCINOMA AND DYSPLASIA IN COLITIS 1129 FIG. 1. Colonic mucosa from a patient with chronic ulcerative colitis complicated by carcinoma. The features of severe epithelial dyspasia -present aref greatly lengthened, irregular mucin-depleted glands and a villous surface (H & E, x 135). chromatin patterns from dense hyperchromicity to coarsely granular chromatin (fig. 2). Not only are the nuclei enlarged to a variable degree in dysplasia but they are irregularly shaped or pleomorphic (figs. 2 and 4) with some severely distorted. Nucleoli can be prominent in both hyperplasia and dysplasia, but in the latter they are much larger, irregular, and often multiple (figs. 3 and 4). The nuclei in normal colorectal epithelial cells are somewhat flattened and basally situated. In reactive hyperplasia the nuclei acquire an orientation parallel to the cell and to each other (fig. 5) but are still basally situated, whereas in dysplasia the parallelism tends to be lost. The nuclei may also be found at various elevations from the base of the cell which gives the appearance of "pseudostratification" (fig. 3). Mitoses are numerous in both reactive and dysplastic epithelium, but abnormal forms can usually be found in the latter (fig. 3). In both epithelial changes the mucin content of the cells is diminished, usually more severely in dysplasia. However, the epithelium covering the tips of dysplastic villi may show less striking mucin depletion and nuclear atypia. The description of severe epithelial dysplasia can be summarized as comprising: (1) lengthening of glands tending to produce a villous surface; (2) marked but variable nuclear enlargement; (3) irregularly shaped nuclei or pleomorphism; (4) abnormal chromatin patterns; (5) loss of nuclear polarity; (6) "pseudostratification" of nuclei; (7) prominent nucleoli; and (8) abnormal mitoses. This condition of epithelial dysplasia is generally recognized by pathologists, although it may be given other names such as atypical hyperplasia, precancer, 11 or carcinoma in situ There is less agreement about milder forms of atypia and it is difficult to define what point on the spectrum of these changes represents the minimum that most pathologists would accept as dysplasia. We do not require all the features listed above to be present before reporting dysplasia and in practice make subjective divisions into mild and severe dysplasia. In mild dysplasia there need not be much lengthening of glands or villous pattern but there must be the positive features of muclear enlargement, pleomorphism, and some loss of polarity with nucleolar prominence beyond that which can be accepted in reactive epithelia. There is as yet no reliable way of measuring these changes so the distinction is inevitably subjective. The point at which we make the division between epithelial dysplasia and hyperplasia is illustrated by figs. 4

4 1130 COOK AND GOLIGHER Vol. 68, No.5, Part 1 FIG. 2. Part of figure 1 at a higher magnification to show the cytological details: marked nuclear enlargement and pleomorphism, loss of nuclear polarity, and coarsely granular chromatin (H & E, x 400). and 5. However, this may not be generally acceptable and our conclusions are therefore based on the findings of unequivocal or severe dysplasia in which all the listed features can be found. The incidence of the less severe forms of dysplasia as exemplified by fig. 4 are given in tables 2 and 3. Findings There were 13 patients with 1 carcinoma, 5 with 2 carcinomas and 1 patient with 3. Of these 26 carcinomas 13 were producing abundant mucin and the others were not. The grading, staging, and siting of these carcinomas are given in table 1. Only 5 of the carcinomas were typically nodular and ulcerated, 7 were plaque-like and superficia~ly ulcerated, 6 were forming fibrous stn~t~res, and 3 were polypoid. The remammg 5 were not recognized before microscopy. Two cases contained adenomataus polyps, one a simple adenoma, and the other a villous adenoma. Inflammatory polyps were only present in 4 cases. In the 19 patients with carcinomas, the changes of ulcerative colitis were seen throughout the whole large bowel in 15, but in the other 4 the disease was confined to the bowel distal to the splenic flexure. Uniform colitic activity was noted in 6 specimens, quiescence in 8, and in the 5 others activity was found distally with quiescent changes proximally. Seven of the 19 patients are alive and well for periods varying from 15 months to nearly 15 years after surgery, 5 of them having survived more than 5 years. Of the other 12 patients, 2 died immediately postoperatively. The survival of the rest varied from 2 months to nearly 5 years with 3

5 May 1975 CARCINOMA AND DYSPLASIA IN COLITIS 1131 FIG. 3. Unequivocal epithelial dysplasia in the rectal mucosa of a patient with chronic ulcerative colitis not complicated by carcinoma. There is nucleolar prominence and nuclear "pseudostratification" as well as enlargement. An abnormal tripolar mitosis is present above and to the left of center (H & E, x 400). patients living more than 3 years. Severe dysplasia was found in 14 of those cases complicated by carcinoma but its distribution varied considerably (table 2). Lesser degrees of dysplasia were also found in the rest of the cases. Since rectal biopsy through the sigmoidoscope is still far more commonly used than is colonic biopsy by means of the colonoscope-and in any case provides more satisfactory specimens-the assessment of dysplasia in the rectum was given particular attention and is summarized in table 3. Seven (36.8%) of the carcinoma cases showed colonic Paneth cell metaplasia; 6 of them were associated with severe dysplasia. In the control group suffering from ulcerative colitis without carcinoma, severe dysplasia was noted in 4 of 14 patients (see tables 2 and 3), and milder forms in 7 other patients. Six of the cases showed uniformly active ulcerative colitis, 7 were only distally active, and 1 was quiescent. Four of these control cases (28.5%) had some degree of colonic Paneth cell metaplasia, 2 of these being associated with severe dysplasia, but 1 without any associated dysplasia. Discussion Pathological characteristics of established colitis carcinomas. Counsell and Dukes 1 drew attention to the fact that colorectal carcinomas complicating colitis differ from ordinary large bowel cancers in that they may not cause a visible tumor or ulcer on the mucosal surface but may manifest themselves only as dense fibrous-

6 1132 COOK AND GOLIGHER Vol. 68, No.5, Part 1 FIG. 4. A rectal gland showing epithelial dysplasia from a patient with chronic ulcerative colitis complicated by carcinoma. The features of mild dysplasia, to be contrasted with those of severe reactive hyperplasia in figure 5, are marked and variable nuclear enlargement, nuclear pleomorphism and Joss of polarity, granular chromatin, and prominent, irregular and often multiple nucleoli (H & E, x 400). looking strictures with ill-defined boundaries, resembling an inflammatory rather than a neoplastic lesion. In our cases, of the 26 carcinomas, 18 were unusual in resembling fibrous strictures, flat plaques, or in being macroscopically unrecognizable. The minority were indistinguishable from ordinary colorectal cancers (table 1). It has also been pointed out that colorectal carcinomas complicating colitis are more frequently multiple than are those arising spontaneously in the large intestine, only 3% of which exhibit multiple primary lesions. 16 No less than 6 of the 19 specimens of colitis carcinomas in the present study contained more than one tumour (table 1). It is well known that the incidence of ordinary colorectal carcinomas is much greater in the distal part of the bowel, some 40% at least of the lesions being located in the rectum, 16 but a much more even distribution of colitis carcinoma throughout the large bowel has been observed, with only 25% or less being found in the rectal segment. Other authors have not observed much difference in the site incidence of colitis carcinomas and of ordinary colorectal cancers. 1 In the present series of surgically treated cases of colitis carcinomas, eight, or just under onethird of the 26 separate primary growths, were in the rectum (table 1). Ordinary large bowel cancer is a disease mainly of people of advancing years and the average age of such patients in the Birmingham area of England has been reckoned by Slaney and Brooke 3 to be 63

7 May 1975 CARCINOMA AND DYSPLASIA IN COLITIS 1133 FIG. 5. Reactive epithelial hyperplasia in a colonic gland of a patient with ulcerative colitis not complicated by carcinoma. The features of severe hyperplasia, to be contrasted with those of mild dysplasia in figure 4, are less marked nuclear enlargement without pleomorphism or loss of polarity. Single nucleoli are visible but without great prominence (H & E, x 400). Site TABLE 1. Pathological features of 26 colitis carcinomas occurring in 19 patients Gross Histological Dukes' appearance differentiation staging Cecum, 1 Indistinguishable from ordinary colorectal Well differentiated, 11 A,5 cancers, 5 Ascending colon, 5 B, 11 Hepatic flexure, 3 Transverse colon, 4 Of average differentiation, 7 C,, 8 Splenic flexure, 1 c2, 2 Descending colon, 1 Resembling fibrous strictures, 6 Poorly differentiated, 8 Sigmoid colon, 3 Rectum, 8 (Mucin producing, 13) Plaque-like, 7 Not recognizable macroscopically, 5 Polypoid, 3 years. By contrast the average age of patients with colitis carcinoma has been estimated by these authors to be 42.5 years and by Edwards and Truelove 5 to be 41 years. The average age of our patients with colitis carcinoma treated by radical surgery was 51.3 years at the time of operation. The differing character of colitis carci-

8 1134 COOK AND GOLIGHER Vol. 68, No.5, Part I TABLE 2. Distribution and severity of epithelial dysplasia in proctocolectomy specimens from colitis patients with and without carcinomas Presence and severity of dysplasia Patchy Severe (unequivocal) Milder changes.. Widespread Severe (unequivocal) Milder changes Localized: severe (unequivocal). Absent... Cases of ulcerative colitis complicated by carcinoma Control cases of ulcerative colitis without carcinoma TABLE 3. Presence and severity of epithelial dysplasia in rectum Presence and severity of dysplasia Severe (unequivocal)... Milder changes Absent... Cases of ulcerative colitis complicated by carcinoma Control cases of ulcerative colitis without carcinoma noma from ordinary colorectal carcinoma has been explained on the basis of its incidence in younger patients, 10 more malignant tumors being commoner in the young. 17 This does not explain the unusual macroscopic forms of tumor or the even distribution of colitic cancers in the large bowel. It may be that colonic carcinoma has a different behavior from that of the rectum 18 or that the different preceding condition of the bowel mucosa, namely being chronically inflamed, itself results in a preponderance of unusual forms of tumors. Prognosis of colitis carcinoma. The poor prognosis after surgical treatment for colitis carcinoma has been stressed by several authors For example, the average 5~year survival of patients with colitis carc_moma, based on a survey of the literature, 3 is 18.6%, which contrasts with approximately 50% in ordinary colorectal cancer. 20 These disappointing results are attributed to the relatively poor differentiation of tumors complicating ulcerative colitis 9 and to the younger age at which they occur. 10 Dukes 21 reported that 24% of ordinary colorectal cancers were poorly differentiated, whereas we found that 30.7% of our series of colitis carcinomas were poorly differentiated (table 1). Proportions of 21 and 50% have been reported in similar series. 7 9 The proportion of our cases of colitis carcinomas which were secreting abundant mucin was increased 4-fold over that of this type of carcinoma in ordinary colorectal tumours, 50% as compared to 12.4%. 21 The relatively better differentiation and less advanced stage, 61% not having lymph node metastases (table 1), of the carcinomas complicating colitis may be the reason for the somewhat improved 5-year survival (26%) and the 42% 3-year survival of our patients after their proctocolectomies. However, during the same period another 5 patients under our care with colitis carcinoma were hopelessly inoperable. Other reports also indicate that with early radical surgery the prognosis of established colitis carcinoma may not be as bad as was formerly imagined. Factors predisposing to the development of colitis carcinoma and the significance of epithelial dysplasia. Colitis cancer is generally held to be a complication of longstanding extensive colitis The average durat!on of the preceding colitis in the cases reported by Rosenqvist et al. 2 and Slaney and Brooke 3 was 15.4 and 14.8 years, respectively. Certainly the average length of history in our cases, 18 years (range 3 to 34 years), endorses these reports. However, we did not invariably find total colitis in these cases, 4 of the 19 having disease confined to the left colon and rectum. These are similar figures to those of Edwards and Truelove 5 who reported 5 of 21 cases having left-sid.ed. disease, whereas Morson and Pang 11 found carcinoma only as a complication of total colitis. Earlier reports suggested that inflammatory polyposis predisposed to the development of carcinoma, but this has

9 May 1975 CARCINOMA AND DYSPLASIA IN COLITIS 1135 been contradicted by Counsell and Dukes, 1 Rosenqvist et al., 2 Edling and EklOf, 25 de Dombal et al., 8 and our own findings, since polyps were absent in 15 of our cases and were only particularly numerous in the vicinity of a tumor in one case. Our finding of Paneth cell metaplasia in a proportion of those cases with dysplasia, as noted by Morson and Pang, 11 does not clarify its significance. These cells are rarely found in carcinomas and do not appear themselves to undergo malignant transformation; it seems therefore that this form of metaplasia is a reflection of chronic disease rather than being part of the premalignant changes. Svartz and Ernberg 26 and Rosenqvist et al. 2 have drawn attention to the fact that colitis has often been quiescent for long periods before a carcinoma is suspected and the patients have grown to accept the limitations imposed by the disease. In this context we found that 18 of the 26 tumors in our series were located in areas of the bowel where the colitis was no longer active as judged by the absence of polymorphs and lack of mucin depletion. Morson and Pang 11 described the features of epithelial dysplasia or precancer in cases of ulcerative colitis as being cytologically similar to premalignant changes in other organs such as the uterine cervix. They advocated regular rectal biopsy in those patients with a long history of extensive colitis to detect precancerous changes and thus help in the selection of patients for radical surgery. Yardley and Keren 13 rightly pointed out that in the presence of acute inflammation epithelial atypia must be regarded with some skepticism, but in our opinion if strict diagnostic criteria are rigorously applied epithelial dysplasia can be detected in actively inflamed mucosa. The same authors also seemed prepared to play down the histological evidence of severe dysplasia in patients with histories of ulcerative colitis shorter than 10 years. Our series of 19 patients with ulcerative colitis complicated by carcinoma included 3 with such short histories (3, 6, and 9 years). Thus, if it were accepted that the finding of epithelial dysplasia had value in the management of ulcerative colitis, it would be advisable to seek it in all patients irrespective of the duration of their disease. Our study confirms the association of colitis carcinoma and epithelial dysplasia in that 14 of the 19 specimens that harbored a carcinoma showed unequivocal dysplasia (table 2). This contrasted with an incidence of undoubted dysplasia of 4 of 14 patients with long histories of extensive ulcerative colitis but no carcinoma. Less severe dysplasia was found in all the other 5 cases with carcinomas and in 7 more of the control series. It is not clear what significance can be attached to the fact that the average length of history of colitis, in those patients of the control group with severe dysplasia, was years less than that of the patients with colitis carcinoma (13.25 versus U years). We do not know how many of th control cases might have developed carcinoma if they had been continued on conservative treatment instead of being submitted to proctocolectomy. Consequently, we are unable, on the basis of our own material to point to epithelial dysplasia as being an invariable precursor of carcinoma, and we are not aware of any published studies of large groups of colitis patients with and without dysplasia, who have been followed up over a long period on conservative management, to determine the relative incidence of carcinoma. The precise value of finding epithelial dysplasia in biopsy specimens thus remains sub judice. However, it could reasonably be argued from the analogy of similar cytological changes in the breast and uterine cervix, although admittedly in different types of epithelium, that epithelial dysplasia at least in severe form is a precancerous state in ulcerative colitis. The finding of such changes in a rectal biopsy might thus be held to afford a special indication of impending carcinoma formation somewhere in the large intestine. As to the significance of a negative rectal biopsy for epithelial dysplasia, it is to be noted that none of our 19 operative specimens containing colitis carcinomas failed to show epithelial dysplasia of some degree in some part of the intestine and only 5 did

10 1136 COOK AND GOLIGHER Vol. 68, No.5, Part 1 not contain severe dysplasia. It might be wondered whether a negative rectal biopsy could give any assurance that patients with long standing colitis might be continued indefinitely on conservative management so far as the risk of carcinoma is concerned. It must be emphasized, however, that dysplastic changes are often patchy in distribution; for example, the rectum was spared of unequivocal changes in 8 of our 19 cases with colitis carcinoma. Accordingly, if clinical reliance had been placed on negative rectal biopsies in these latter 8 patients, a wrong decision would have been reached regarding the propensity of their large bowel mucosa to undergo malignant change, even at the stage when a frank colon cancer had arready appeared. Morson and Pang, 11 Evans and Pollock, 12 and Morson and Dawson" have also stressed the unreliability of negative rectal biopsies. Certainly if mucosal biopsy is to be used in the management of long standing ulcerative colitis, as seems desirable, it would appear essential to extend the mucosal sampling process to include the colon by means of multiple colonoscopic biopsies, which are now so eminently practicable. REFERENCES 1. Counsell PB, Dukes, CE: The association of chronic ulcerative colitis and carcinoma of the rectum and colon. Brit J Surg 39: , Rosenqvist H, Ohrling H, Lagercrantz R, et al: Ulcerative colitis and carcinoma coli. Lancet 1: , Slaney G, Brooke BN: Cancer in ulcerative colitis. Lancet 2: , Edling NPG, Eklof 0: Distribution of malignancy in ulcerative colitis. Gastroenterology 41: , Edwards FC, Truelove SC: The course and prognosis of ulcerative colitis. IV. Carcinoma of the colon. Gut 5:15-22, MacDougall IPM: The cancer risk in ulcerative colitis. Lancet 2: , Goldgraber MG, Kirsner JB: Carcinoma of the colon in ulcerative colitis. Cancer 17: , de Dombal FT, Watts JM, Watkinson G, et a!: Local complications of ulcerative colitis: stricture pseudo-polyposis and carcinoma of colon and rectum. Brit Med J 1: , Hinton JM: Risk of malignant change in ulcerative colitis. Gut 7: , Hulten L, Kewenter J, Ahren C: Precancer and carcinoma in chronic ulcerative colitis. Scand J Gastroenterol 7: , Morson BC, Pang LSC: Rectal biopsy as an aid to cancer control in ulcerative colitis. Gut 8: , Evans DJ, Pollock DJ: In-situ and invasive carcinoma of the colon in patients with ulcerative colitis. Gut 13: , Yardley JH, Keren DF: "Precancer" lesions in ulcerative colitis. Cancer 34: , Morson BC, Dawson IMP: Gastrointestinal Pathology, Oxford, Blackwell Scientific Publications, 1972, p Bargen JA, Gage RP: Carcinoma and ulcerative colitis: prognosis. Gastroenterology 39: , Goligher, JC: Surgery of the Anus, Rectum and Colon. Second edition, London, Bailliere Tindall and Cassell, 1967, p Recio P, Bussey HJR: The pathology and prognosis of carcinoma of the rectum in the young. Proc Roy Soc Med 58: , Langman MJS: Epidemiology of cancer of the large intestine. Proc Roy Soc Med 59: , Dennis, C, Karlson KE: Cancer risk in ulcerative colitis: Formidability per patient-year of late disease. Surgery 50: , Slaney G: Modern Trends Surg. 3:69-89, Dukes CE: Cancer of the rectum: an analysis of 1000 cases. J Pathol Bacterial 50: , Brooke BN: Ulcerative colitis and carcinoma of the colon. J Roy Coli Surg Edinburgh 14: , Shands WC, Dokerty MB, Bargen JA: Adenocarcinoma of the large intestine associated with chronic ulcerative colitis. Surg Gynecol Obstet 94: , Bacon NE, Yang LMO, Carroll PT, et al: Nonspecific ulcerative colitis, with reference to mortality, morbidity complications, and long-term survivals following colectomy. Am J Surg 92: , Edling NPG, Eklof 0, Radiologic findings and prognosis in ulcerative colitis. Acta Chir Scand 121: , Svartz N, Em berg T: Cancer coli in cases of colitis ulcerosa. Acta Med Scand 135: , McDivitt RW, Hutter RV, Foote FW, et al: In situ lobular carcinoma. A prospective follow-up study indicating cumulative patient risks. JAMA 201:82-84, MacGregor JE, Fraser ME, Mann EM: Improved prognosis of cervical cancer due to comprehensive screening. Lancet 1:74-76, Kinlen LJ, Doll R: Trends in mortality from cancer of the uterus in Canada, and in England and Wales. Brit J Prev Soc Med 27: , 1973