COLORECTAL NEOPLASMS

I. Precursors to colorectal adenocarcinoma

1. Adenomatous polyps

1. Adenomatous polyps are found in 5% of all barium enemas

1. 50% are in recto-sigmoid
2. 50% are multiple

2. Diff morphological types:

1. Tubular (adenomatous)--7% become malignant
2. Tubulovillous (villoglandular)--20% become malignant
3. Villous--33% become malignant
4. Tubular & tubulovillous are us. pedunculated; villous us. sessile

3. Do all distal adenomatous polyps seen on sigmoidoscopy need to be f/u'd with colonoscopy?
   1. 401 pts > 50yo with negative fecal occult blood tests and no risk factors for colon Ca who had adenomatous polyps on flex sig. were studied. 301 subsequently underwent colonoscopy. The likelihood of colonoscopy revealing an "advanced" proximal polyp (tubulovillous, villous, or cancerous), was 6% in pts who on sigmoidoscopy had had either multiple small (1-5mm) tubular adenomas, any tubular adenoma 6-10mm in diameter, or "advanced" polyps, and 0% (95% CI 0-4%) in pts who had had a single small (1-5mm) tubular adenoma at sigmoidoscopy (Ann. Int. Med 129:273, 1998--JW)

4. Prevention of adenomata

1. Calcium--930 pts with recently removed adenomatous polyps randomized to CaCO3 3g/d (1.2g elemental Ca) vs. placebo. Followup colonoscopy at 4y showed RR of adenoma of 0.85 (sig.) (NEJM 340:101, 1999--JW)
2. Aspirin
   1. In a case-control study of 379 pts undergoing colonoscopy, regular use of ASA or NSAIDs was ass'd with RR 0.56 for having an adenoma on colonoscopy, after adjustment for a variety of confounders (sig.; Gastroent. 114:441, 1998--JW)
   2. 635 pts with prior colorectal Ca randomized to ASA 325 mg PO QD vs. placebo.  Over avg. 31mo f/u (during which time most had 1-2 colonoscopies), incidence of recurrent adenomata was sig. lower in ASA recipients (17% vs. 27%) (NEJM 348:883, 2003--JW)
   3. 1121 pts with prior adenomatous colonic polyp randomized to ASA 325mg/d, ASA 81mg/d, or placebo.  Over mean 33mo f/u, recurrent adenomata dx'd in 45%, 38%, and 47%, respectively (only diff. between 81mg/d and placebo groups was sig.). (NEJM 348:891, 2003--JW)
   4. 272 pts with h/o adenomatous colorectal polyps randomized to lysine acetylsalicylate 160mg/d or 300mg/d vs. placebo.  Prevalence of adenomatous polyps on repeat colonoscopy at 1y was not sig. diff. between the two groups. ("APACC" Trial; Gastroent. 125:328, 2003--AFP)

2. Hyperplastic polyps
   1. Tend to occur on the right side of the colon
   2. Traditionally thought to have no malignant potential but that has been called into question based on the presence in some hyperplastic polyps of microsatellite instability, an abnormality in non-protein-coding portions of DNA also found in some colorectal cancers (J. Nat. Ca. Inst. 93:1307, 2001--JW)

II. Screening

1. Fecal Occult Blood Testing ("Hemoccults")

1. Annual fecal occult-blood testing reduced mortality from colorectal Ca in tested subjects by 33% at 13y c/w controls in a randomized trial (NEJM 328:1365, 1993--JW); every 2y screening was not as effective (J. Natl. Ca. Inst. 3:91, 1999--JW)

2. Lancet 348:1467, 1996-JW

1. Danish study randomized 62,000 people 45-75yo to every-other year screening with Hemoccult II with dietary resrictions, without rehydration for 10y vs. no screening
2. If positive for occult blood, offered colonoscopy
3. 481 vs. 483 cases of colorectal Ca in screening vs. non-screening groups
4. 205 vs. 249 deaths from colorectal Ca in screening vs. nonscreening groups (RR 0.82; significant)

3. Lancet 348:1463, 1996-JW

1. UK study randomized 153,000 people 45-74yo to same method & schedule as above study; median f/u 7.8y
2. 360 vs. 420 deaths attributable to colorectal Ca in screening vs. control groups (sig.)
3. Note that 75% of the diagnosed colorectal Ca cases in the screening group were NOT diagnosed as a result of screening, i.e. presented by some other means!

2. Flexible sigmoidoscopy

3. Colonoscopy--Not common for screening in general population as of 2002 though evidence mounting to support

1. 3121 pts 50-75yo underwent screening colonoscopy. 37.5% had any neoplasm, 10.5% had "advanced disease" (tubular adenoma 1cm or greater, villous adenoma, adenoma w/high-grade dysplasia, or invasive carcinoma). 4.1% had advanced disease proximal to the descending colon. Of pts with no adenomas distal to splenic flexure, 2.7% had advanced proximal disease. 10 pts had "serious" complications from colonoscopy. (NEJM 343:162, 2000--JW)
2. 1994 pts > 50yo underwent screening colonoscopy; 2.5% had an "advanced" proximal neoplasm (polyp or polypoid lesion w/villous features, high-grade dysplasia, or Ca), including 1.5% of the pts with no distal lesions. 1 pt had colonic perforation (NEJM 343:162, 2000--JW)
3. 2885 pts all had FOBT then screening colonoscopy; 306 had "advanced neoplasia" (adenoma with > 10mm duration, villous features, or high-grade dysplasia; or invasive Ca); of those, only 76% had either positive FOBT or findings that would have been noted on flex sig (lesions in rectum and sigmoid colons) (NEJM 345:555, 2001--JW)

4. CT colonography
   1. In a study of 205 pts who underwent both virtual colonoscopy using spiral CT w/oral contrast and conventional colonoscopy, virtual colonoscopy had sens 62% and spec 71% for identifying lesions seen on colonoscopy, BUT had neg predictive value of 99% for lesions of 10mm diameter or larger (Gastroent. 125:304, 2003--abst)
   2. In a study of 1233 asymptomatic average-risk adults (mean age 58yo), all of whom underwent CT "virtual colonoscopy" and conventional colonoscopy, sensitivity for "advanced lesions" (polyps > 9mm or smaller but with advanced histology) was 92% with CT vs. 88% with colonoscopy (NEJM 349:2191, 2003--JW)
5. Magnetic Resonance Colonography--With MR-contrast enema
   1. In one study comparing MRC w/colonoscopy in 70 pts requiring colonoscopy, sensitivity/specificity was 96%/93%, respectively for identifying polyps of > 1cm in diameter (Gatroent. 119:300, 2000--JW)
   2. In another such study with 177 pts, sensitivity/specificity was 93%/99%, respectively for identifying polyps > 1cm in diameter (Radiol. 216:383, 2000--JW)

6. Note that race and gender may affect site of occurrence of colon Ca--in a retrospective study of 38,391 pts with colorectal Ca, incidence of proximal (up to and including descending colon) colorectal Ca was higher in black than in white population and distal colorectal Ca was higher for male than for female pts in both races (Cancer 80:193, 1997--AFP)

7. Risk factors for proximal neoplasias in pts found to have adenomas on screening sigmoidoscopy: Age > 65, first-degree relative with colorectal Ca, multiple adenomas on sigmoidoscopy, or villous features in an adenoma found on sigmoidoscopy (JAMA 281:1611, 1999--JW)

8. Surveillance after diagnosis & removal of adenomatous polyps
   1. Most common practice is colonoscopy Q3-5y
   2. Barium enema missed many adenomas seen on colonoscopy in one study of 580 pts with h/o adenomatous polyps (NEJM 342:1766, 2000--JW)

III. Prevention

1. Aspirin/NSAIDs for prevention

1. 22,000 adult males in the Physicians' Health Study randomized to ASA 325mg QOD vs. placebo; after 5y subjects chose whether or not to take ASA. After 12y f/u, no sig. diff. in incidence of colorectal Ca between the 2 original groups or between men choosing to take ASA > 2x/wk vs. men taking ASA less frequency (Ann Int. Med. 128:713, 1998--JW)
2. In a retrospective analysis of 100,000 people > 65yo found that > 2y of cumulative NSAID use in prior 5y ass'd with RR of 0.49 of being dx'd with colorectal Ca compared with nonusers; no reduction in risk for < 3mos of use (no association w/dosage seen) (Arch. Int. Med. 159:161, 1999--JW)

2. Postmenopausal HRT for prevention

1. 59,000 postmenopausal women in Nurse's Health Study followed prospectively x 14y. Current use of HRT (75% on estrogen-only) ass'd with RR 0.65 for development of colorectal Ca; duration of use didn't affect (Ann. Int. Med 128:705, 1998--JW)

3. Folic acid supplementation for prevention

1. The Nurses' Health Study prospectively studied 88,756 female nurses who were questioned about diet and use of supplements; of 14y f/u, after multivariate analysis, dietary folate intake (including from multivitamins) of > 400ug ass/d with RR 0.69 of colorectal Ca compared with < 200ug/d; statistically sig. after 15y of high folate intake. No effect of Calcium or vitamins A, E, C, or D. (Ann. Int. Med. 129:517, 1998--JW)

4. Dietary fiber for prevention of colon Ca--may not be effective

1. In a prospective study of 90,000 women in the Nurses' Health Study followed for > 15y, dietary fiber intake was not ass'd with risk of colorectal Ca or adenoma; still no ass'n on multivariate analysis including analyses of subgroups defined by type of fiber (NEJM 340:169, 1999--JW)
2. In a randomized trial of high-fiber, low-fat diet vs. "usual" diet in 2079 pts with h/o colonic adenomas, there was no difference in adenoma recurrence over 4y f/u (NEJM 342:1149, 2000--JW)
3. In a randomized trial of 1429 pts with h/o colonic adenomas, addition of a high-fiber cereal to usual diet was not ass'd with a sig. difference in risk for adenoma recurrence over 3y f/u (NEJM 342:1156, 2000--JW)

IV. Surveillance after diagnosis

1. In a randomized study of 325 pts s/p curative resection of colon ca, standard surveillance (hx/px, FOBT, CEA, and LFT's Q3mos x 2y then Q6mos) vs. "intensive" surveillance (the same plus yearly colonoscopy, liver CT, and CXR) showed no sig. diff. in survival at 5y (Gastroent. 114:7, 1998--JW)
2. In a meta-analysis of 5 RCT's with total 1342 pts of intensive f/u after curative resection for colon Ca (involved various uses of colonoscopy, CEA, liver imaging, and CXR), intensive surveillance ass'd with sig. reduced all-cause mortality (RR 0.81). The 4 trials using CT and frequent CEA measurements ass'd with even lower RR (0.73) for all-cause mortality. (BMJ 324:813, 2002--JW)

V. Treatment

1. Surgery
2. Chemotherapy (usually 5-FU + either levamisole or folinic acid)--beneficial in stage III but not stage II colon Ca
   1. Response to chemotherapy was sig. greater in pts with "Microsatellite Instability" (a type of DNA change) in their tumors in a nonrandomized trial of 656 pts with Dukes' C colorectal Ca (Lancet 355:1745, 2000--JW)
   2. Hepatic arterial chemotherapy for hepatic mets (after resection of hepatic mets)--Ass'd with sig. higher 2y survival (86% vs. 72%) in a randomized trial of 156 pts (NEJM 341:2039, 1999--JW)
   3. In a meta-analysis of 13 randomized trials involving 1365 pts studying chemotherapy (9 trials involved fluorouracil) vs. no chemo for metastatic colon Ca, chemo ass'd with sig. better median survival (11.7mo vs. 8mo) and better quality of life (BMJ 321:531, 2000--JW)
3. "Active specific immunotherapy" (I'm a little unclear on how this actually works--improves survival in stage II colon Ca--Lancet 353:345, 1999--JW)
4. Resection of hepatic metastases--predictors of recurrence after resection; 5y survival after resection was 60% with none and 14% with all 5 in a series of 1,001 pts (Ann. Surg. 230:309, 1999--JW)
   1. > 1 hepatic met
   2. CEA > 200 ng/ml
   3. Any met > 5cm
   4. Node-positive primarytumor
   5. Disease-free interval of < 12mos between initial dx of colon Ca and detection of a met