Natural history of untreated colonic polyps

GASTROENTEROLOGY 1987;93:loo9-13

Natural History of Untreated Colonic Polyps STEVEN J. STRYKER, BRUCE G. WOLFF, CLYDE E. CULP, SUSAN D. LIBBE, DUANE M. ILSTRUP, and ROBERT L. MA&ARTY Sections of Colon and Rectal Surgery and Medical Research Statistics, and the Department Diagnostic Radiology, Mayo Clinic and Mayo Foundation, and the Department of Nursing, Mary’s Hospital, Rochester, Minnesota

The natural history of untreated colonic polyps is uncertain. A retrospective review of Mayo Clinic records from a 6-yr period just before the advent of colonoscopy identified 226 patients with colonic polyps ~10 mm in diameter in whom periodic radiographic examination of the colon was elected over excisional therapy. In all patients, follow-up of polyps spanned at least 12 mo (mean, 68 mo; range, 12-229 mo) and included at least two barium enema examinations (mean, 5.2; range, 2-17). During the follow-up period, 83 polyps [37%) enlarged. Twenty-one invasive carcinomas were identified at the site of the index polyp at a mean follow-up of 108 mo (range, 24-225 mo). Actuarial analysis revealed that the cumulative risk of diagnosis of cancer at the polyp site at 5, IO,and 20 yr was 2.5%, i3%,and 24% respectively. In addition, 11 invasive cancers were found at a site remote from the index polyp during the same follow-up period. These data further support the recommendation for excision of all colonic polyps 210 mm in diameter. Periodic examination of the entire colon is recommended in this group of patients to identify neoplasms arising at a site remote from the index polyp. Although this study has limitations inherent to any retrospective analysis, comparable prospective data are unlikely to be available in the future because of the current widespread availability of colonoscopy. Carcinoma of the colon and rectum ranks as the second most common cause of cancer-related death in the United States, accounting for -60,000 deaths Received June 16, 1986. Accepted May 18, 1987. Addess requests for reprints to: Bruce G. Wolff, M.D., Section of Colon and Rectal Surgery, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905. C.E. Culp is Emeritus Consultant, Section of Colon and Rectal Surgery, Mayo Clinic and Mayo Foundation, Rochester, Minnesota. 0 1987 by the American Gastroenterological Association 0016-5085/87/$3.50

of Saint

projected for 1985 (1). The single most important prognostic feature remains the pathologic stage at the time of diagnosis (2,3). Therefore, major efforts are being made to devise an effective screening program in an attempt to identify “early” cancers or, preferably, preinvasive lesions. Any discussion of screening strategy, however, must include consideration of the origin of colorectal carcinoma. Although there has been much debate over the malignant potential of colonic polyps, the relative importance of polyps as precursors of “type ordinaire” colon cancer remains controversial. At one extreme are those investigators who, citing the absence of residual adenomatous tissue in excised malignancies, believe that seldom, if ever, do colonic cancers arise from benign adenomatous polyps (4,5). Taking an opposing viewpoint, many clinicians and pathologists assume that most large bowel malignant lesions arise from benign adenomas. Their conclusions are based on numerous studies suggesting an “adenoma-carcinoma sequence,” a phrase first used by Jackman and Mayo (6) in 1951. These studies have shown that residual benign adenomatous tissue can be found in small carcinomas (7,8), malignant foci are commonly observed in larger polyps (9,10), and the distributions of colonic polyps and cancers are similar (6). Although these static observations provide indirect evidence of the malignant potential of colonic polyps, few studies detail the transformation of adenomas to carcinomas. Rare observations of a benign-appearing polyp developing into an invasive carcinoma have been reported over the past 50 yr (10-15). Heretofore, studies attempting to document the natural history of untreated colonic polyps have been of short duration or have involved few patients, or both (16,17). The purpose of this study was to determine the incidence of colonic carcinoma in a group of persons

1010

GASTROENTEROLOGY

STRYKER ET AL.

with colonic polyps for many years.

followed

up

radiographically

Methods A computer search of Mayo Clinic records revealed patients who were diagnosed as having colorectal polyps by radiographs or surgical resection between January 1965 and December 1970. For the purposes of this study, “polyp” is defined as a projection into the lumen of the colon demonstrated radiographically or endoscopitally. This 6-yr period immediately preceded the introduction of colonoscopy at this institution. Criteria for inclusion in this polyp follow-up study were [a) radiographic demonstration of a polyp 210 mm in diameter, (b) demonstration of this polyp on at least two consecutive barium studies, (c) a period of radiographic surveillance extending beyond a year, and (d) polyp location beyond the reach of the rigid proctosigmoidoscope. Patients with polyposis coli or inflammatory bowel disease were excluded from the study. Two hundred twenty-six patients met these selection criteria. The reasons for exclusion and number of patients excluded were as follows: polyp resected within 1 yr of diagnosis, 2469; polyp
Vol. 93, No. 5

in greatest diameter of >25%, corresponding to an approximate doubling of volume. All patients were examined by the single-contrast fluoroscopic barium enema technique used at the Mayo Clinic. An -25% wtivol solution of barium sulfate filled the entire colon in several minutes. Each segment of colon was examined fluoroscopically; careful, graded manual compression distributed the barium evenly over the mucosal surface. Polyps were seen as filling defects within the barium column. Compressing the barium column under fluoroscopic control with the lead-gloved hand allowed the examiner to establish and characterize a polyp’s attachment to the mucosal surface and to differentiate polyps from nonattached filling defects, e.g., air bubbles and fecal matter. Compression spot films of polyps were obtained by use of a wooden device with a lead marker divided into centimeter gradations. This device allowed partial compensation for radiographic magnification when polyp size was estimated and permitted accurate comparison from one examination to the next. Above the level of proctoscopy, this technique has been shown to be as accurate as the double-contrast technique for detection of polyps and cancers (18,19) and comparable to colonoscopy for detection of lesions 21 cm in size (Lewis BD, Carlson HC, Schroeder KW, Spencer RJ, Taylor WF, Bergstralh EJ: unpublished data). The duration of polyp follow-up was the interval between the date of polyp discovery and the date of the most recent radiographic examination showing the polyp or the date of polyp removal, whichever occurred later. For this period of follow-up, pertinent clinical, radiographic, pathologic, and surgical data were abstracted from the clinic records. All original histologic sections were reviewed in patients in whom invasive colonic carcinoma was ultimately diagnosed. Information obtained from death certificates and follow-up correspondence proved useful in some cases in which patients had stopped making annual visits to the clinic after various periods of time. Total duration of follow-up was defined as the interval between polyp diagnosis and death or most recent correspondence. The cumulative incidences of malignancy at the site of its index polyp and at any site in the colon were estimated as a function of time past polyp diagnosis by the survivorship method of Kaplan and Meier (20).

Results The maximum duration of polyp observation was 229 mo (range, 12-229 mo; mean ? SE, 68 -+ 5 mo). One hundred thirty polyps were observed for 12-60 mo, 59 polyps for 60-120 mo, and 37 polyps for >I20 mo. Patients underwent from 2 to I7 barium enema examinations during the period of polyp observation (mean, 5.2). The recommended interval of 12 mo between examinations was adhered to during the early years of a patient’s follow-up; as time progressed, however, stable polyps were reexamined less frequently. The total duration of follow-up as measured from initial diagnosis to

NATIJKAL

Neoplasms

Pwxnt

I______ Tubular adftnoma Villous adenuma Carcinoma in situ

NO.

66 9 8

62 8

Invasive carr.inoma I’nknown”

21

20

3

3

”

Remnved rIsewhere;

pathologic

7

report unavailable

most recent contact (clinic visit, written questionnaire, or death certificate) was approximately twice as long #as the period of polyp observation (range, 12-242 mo; mean ? SE, 140 -+ 4 mo) because several polyps were removed during the period of surveillance. The most recent contact involved a clinic visit in 133 patients, written correspondence in 37, and information obtained from death certificates in 56. During the period of polyp surveillance, growth was detected in 83 (37%) of the polyps. Ultimately, lOi’ polyps (47%) were excised, mostly because of growth. Some were simply excised after the advent of colonoscopy, which allowed safe removal without celiotomy. The pathologic types of the 107 excised neoplasms are shown in Table 1. Eight polyps (4%) could not be demonstrated later after two or more radiographs confirmed their presence. Adenocarcinoma invading through the muscularis mucosae was identified at the site of the index polyp in 21 patients at a mean period of observation of 108 mo (mean t SE, 108 -C 13 mo; range, 24-225 mo). Eighteen of the 21 carcinomas had exhibited polyp growth, which prompted removal. Our modified Astler-Caller staging (21) of these carcinomas was as follows: stage A (invading muscularis mucosae), five tumors: BI (into muscularis propria), six: BZ (through

HISTORY

OF‘ 1INTREATEIl

0.80

0.60

0.40

0.20

0 0

5

10

15

20

Years Figure

1. Cumulative

risk of diagnosis

of invasive

carcinoma

1011

muscularis propria), four: C, (B, lesion with nodal involvement), one; Cz (Bp lesion with nodal involvement), two; and D (distant metastasis), three. Of the three polyps that did not exhibit growth but contained invasive carcinoma on excision, two were stage A and one was stage B1. The median interval from last barium enema to diagnosis of carcinoma was 14 mo (range, 6-118 mo). Four of the 6 patients with stage C or D disease had been examined within 24 mo. The risk of diagnosis of carcithe preceding noma in the index polyp is depicted actuarially in Figure 1. Thus, the cumulative risk of malignancy at the site of the index polyp at 5, 10. and 20 yr was 2.5%, 8%, and 24%, respectively. Anatomic distribution of the cancer-associated polyps did not differ greatly from that in the entire group [Table 2). Although a greater percentage of cancer-associated polyps were ~2 mm [7 of 27 (26%) compared with 14 of 199 (7%); p < 0.011, most were lo-14 mm in diameter (Table 2). Three of 2 1 patients (14%) with cancer-associated polyps had previously had a benign colorectal polyp excised. None had a previously diagnosed colorectal cancer. Ten of the 21 cancer-associated polyps (48%) were pedunculated at the time of discovery. The original histologic slides from 21 patients with cancer-associated polyps were reviewed. Residual benign adenomatous components in continuity with invasive cancer were identified in 10 of 21 specimens. The modified Astler-Coller staging in these carcinomas with residual benign components was as follows: stage A, four: B,, two: B,, one: C,, one; and D, two. During the follow-up period, adenocarcinoma was discovered at a site remote from the index polyp in 11 patients. Thus, the risk of discovering carcinoma

1.00

Probability

POLYPS

at site of index

polyp

during

follow-up

surveillancx~

1012

STRYKER

ET AL.

(;ASTKOEN’TEKO1,O(;Y

Probability

\:ol

93. No. 5

(3j?/o) 8

0

5

10

15

20

Years

at any 4%,

site within the colon at 5, 10, and 20 yr was and 35%. respectively (Figure 2).

140/o,

Discussion The main finding of this study is that if observed radiographically for several months or years, some colonic polyps 210 mm in diameter eventually are shown to harbor invasive carcinoma. This risk approaches 25% at 20 yr. What is even more sobering about this observation is that many of these tumors are already at an advanced stage, with regional or distant metastases. at the time of diagnosis. It becomes apparent that prolonged observation of polyps of this size is unsound if our goal is to identify and treat premalignant or “early” malignancies of the colon and rectum. The incidence of colonic carcinoma at any site during the period after diagnosis of a colonic polyp in this series parallels that in a group of patients seen at St. Mark’s Hospital in London. For a group of 1678

Table

2.

Localion

and Size of Pofy~~s

patients, Morson (22) calculated the risk of metachronous colonic polyps or carcinomas developing after excision of a benign polyp. Separate risk curves were generated for patients with index polyps (excised at the time of diagnosis) of < 10 mm or 2 10 mm in diameter. Of interest is the lo<%, risk of carcinoma at 15 yr after excision of benign polyps 2 10 mm in diameter in the St. Mark’s group. \Ye chose to study only polyps 210 mm in diameter because of their reliable demonstration on serial radiographs. In a recently completed study of a stable population group by Lotfi et al. (23). colorectal carcinoma developed in 6.2% of 323 patients with large (210 mm) colorrctal polyps, most often at a site different from that of the removed polyp. In the current series, at 20 vr of follow-up, the risk of malignancy at a site remote from the index polyp (that is. risk at any site minus risk at site of index polyp) is 10%. Few studies of the natural history of untreated colonic polyps have been reported previously. Muto et al. (10) had information on only four polyps that cvcre untreated during an 11-yr period when 25.2~ benign colonic neoplasms were excised. Knoerrlschild (17) did sigmoidoscopic folloiz-up on 257 rectal polyps for 3---s yr. These polyps were ~~-5 mm in diameter initially. and two contained invasive carcinoma on excision after growth was detected. The diminutive size of most of these pol~~ps suggests that many were hyperplastic rather than adenomatous and thus not at risk for malignant c.hange. Perhaps the most comprehensive information avnilable on growth patterns of colonic neoplasms is found in the study conducted by the radiolog! department in Malmii, Sweden (16). A doublecontrast technique was used for colon examinations in 16,177 citizens of Malmii. including 303 \vho had two or more consecutive studies shoIv_ing the same polypoicl lesion. Their finding Ihat polyps -’ 111mm

November

1987

NATIJKAL

HISTOKY

OF IINTKEATED

POLYPS

found to grow on serial examination are likely to harbor malignancy is similar to that in the current study. The mean follow-up in the Malmii group was of 30 mo, with a maximum of 128 mo. Mean duration polyp follow-up in the current study was 68 mo, and 39 of the 226 lesions (17%) were observed for periods exceeding 10 yr. These prolonged observations seem necessary to better characterize the natural history of untreated polyps, as Muto et al. (10) have postulated that an adcnoma-carcinoma transition can take 10 yr or longer. Although it is well known that the incidence of malignancy in excised polyps is most closely correlated to the size of the polyp (10). in the current study any polyp 210 mm in diameter was at risk. Despite the increased percentage of associated carcinoma in polyps initially measuring 220 mm, most malignant tumors were found in lesions originally measuring lo--l4 mm. The Malmii study confirms this finding because some of the polypoid lesions ultimately proved to be malignant were as small as 7 (16). mm when first detected Some questions about the origins of colorectal cancer cannot be answered by a retrospective study. The length of the polyp-cancer sequence can be only crudely approximated, as the date of origin of the benign component is unknown, and the diagnosis of early malignant invasion is fortuitous, even in polyps under serial surveillance. In addition, one cannot exlclude the possibility that invasive carcinoma was present in the index polyps in this series at the time of original diagnosis (24). Most of these polyps had typically benign features at diagnosis, so that serial observation was a reasonable option. Even if initial biopsy fails to demonstrate malignancy, sampling error makes it impossible to exclude the presence of invasive cancer. For this reason, we have chosen to avoid the phrase “malignant transformation” in discussing our results. Rather, we have investigated the practical problems of what outcome is to be expected if polyps 110 mm are to be observed radiographically over many years without interruption of their natural history by surgical intervention. Although this study possesses limitations inherent to any retrospective analysis, because of the current widespread availability of colonoscopy, it is unlikely that comparable prospective data will be available in the future. Because the prevalence of colonic polyps exceeds the prevalence of carcinoma by severalfold, it appears that few polyps eventually progress to clini-

diameter. The most appropriate treatment smaller polyps has yet to be determined.

cally apparent carcinoma. Nonetheless, the risk of this occurrence far outweighs the risks of treatment now attending colonoscopic polypectomy. This study further supports the already prevalent opinion 210 mm in about excision of all colonic polyps

23.

1013

for

References 1. Silverberg 2. Copeland

3.

4. 5. 6. 7. 8.

9.

10. 11.

12.

13.

14.

15. 16.

17. 18. 19.

20. 21.

22.

24.

E. Cancer statistics, 1985. CA 1~8~:35.19-35 EM, Miller LD. Iones KS. Prognostic: factors in of the colon and rectum. Am J Surg 1968:

carcinoma 116:875-81. Newland KC. Chapuis PH. Pheils MT. i%c:Pherson JG. The relationship of survival to staging and grading of colorectal carcinoma: a prospective studv of 503 (.ases. (Iancer 1981: 47:1424-Y Spratt IS Ir. Ackerman LV, Mover CA. Kelationship of polyps of the colon to c,oIonic cancer. Ann Surg lY58:1J8:682-96. Castleman B. Krickstein HI. Do adenomatous ~~olyps of the colon become malignant? N Engi ] Med 1962.267:469-75. Jackman KJ, Mayo C’I’V.The adenoma-carc.inoma sequence in cancer of the colon. Surg Gynecol Obstet 1951;93:327-30. Morson BC. Factors influencing the prognosis ot early c,anct:r of the rectum. Proc R Sot Med 19fi6;59:607-8. Turell R, Brodman HR. Adenomas of the colon and rectum. In: Turell R. ed. Diseases of the col~ln and anortxtum. Philadelphia: WI3 Saunders, 1959,326. Grinnell KS. Lane N. Henign and malignant adenomatous polyps and papillary adenomas of the colon and rectum: an analysis of 1,856 tumors in 1.335 patients. [nt Ahstr Surg 1958:106:519-38, Muto T. Bussey HJR. Morson BC. The evolution of cancer of the colon and rectum. Cancer 1975:36:2251--70 Buie LA. Brust JCM. Solitary adenomata of the rectum and lower sigmoid. Trans Am Proct Sot 1935:36:57-67. Scarborough RA. The relationship between polyps and carcinoma of the r,olon and rectum. Dis Colon Rectum 1960: 3:336&42. Muto T, Kamipa J. Kusama S. IJmeda N Kapid evolution of colon cancer from a colonic polyp. report of two cases. Am 1 Proctol Gastroenterol Colon Rectal Surg 1982:33:15. 22-24. Smith TR, Maeir DM, Metcalf W. Kaplowitz IS. Transformation of a pedunculated colonic polyp to adenocarcinnma? Report of a case. Dis Colon Rectum 1970;13:3R2-A. Figiel LS. Figiel SJ, Wietersen FK. Is surgical removal of every colonic polyp necessary? AJK 1962;88:721-32. Welin S. Youker J, Spratt JS. et al. The rates and patterns of growth of 375 tumors of the large in&tint> and rec.tum observed serially by double contrast enema study (Malmii technique). AJR 1963;90:673--87. Knoernschild HE. Growth rate and malignant potential of colonic polyps: early results. Surg Forum 1963;14:137-8. Teefey SA. Carlson HC. The fluoroscopic barium enema in colonic polyp detection. AJR 1983;141:1279-81. Johnson CD, Carlson HC. Taylor M’F, Weiland LP. Barium enemas of carc.inoma of the colon: sensitivity of double- and single-contrast studies. AIR 1983: 140:1143-9. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat ASSUI. 1!$58:53:457-81. Astler VB, Collar FA. The prognostic significance of direct extension of carcinoma of the colon and ret turn. Ann Surg 1934:139:846-51. Morson BC. The evolution of c:olorrc:ta1 carciur~ma. Clin Radio1 1984;35:425-31. Lotfi AM, Spencer RJ, llstrup DM, Melton L/ 111. Colorectal polyps and the risk of subsequent ccarcinoma. Mayo Clin Proc 1986:61:337-43. Lescher TC, Dockarty MB. Jackman RI. Beahrs OH. Histopathology of the larger colonic polo p. L)is Colon Rectum 1967; IO:1 18.-24.