Histologic Risk Factors and Clinical Outcome in Colorectal Malignant

Histologic Risk Factors and Clinical Outcome in Colorectal Malignant Polyp: A Pooled-Data Analysis Cesare Hassan, M.D.,1 Angelo Zullo, M.D.,1 Mauro Risio, M.D.,2 Francesco P. Rossini, M.D.,3 Sergio Morini, M.D.1 1

Department of Gastroenterology and Digestive Endoscopy, “Nuovo Regina Margherita” Hospital, Rome, Italy Department of Pathology, Institute for Cancer Research and Treatment, Candiolo–Turin, Italy 3 Department of Gastroenterology, Ospedale San Giovanni Battista, Turin, Italy 2

PURPOSE: The malignant polyp carries a significant risk of lymphohematic metastasis and mortality. Clinical usefulness of histologic risk factors is still controversial. The study was designed to compute the association between the main histologic risk factors and the occurrence of unfavorable outcomes in patients with malignant polyps. METHODS: A MEDLINE search regarding malignant polyps was performed. Three histologic risk factors (positive resection margin, poor differentiation of carcinoma, vascular invasion) and five (residual disease, recurrent disease, lymph node metastasis, hematogenous metastasis, mortality) unfavorable clinical outcomes were evaluated. Further analysis was performed by subgrouping polyps in high-risk and lowrisk groups. RESULTS: Thirty-one studies enrolling 1,900 patients with malignant polyp were selected. Positivity of resection margin was significantly predictive of the presence of residual disease (odds ratio, 22; P < 0.0001), poorly differentiated carcinoma was associated with an increased mortality (odds ratio, 9.2; P < 0.05), and vascular invasion with a higher lymph node metastasis risk (odds ratio, 7; P < 0.05). Patients with high-risk polyps showed a significantly worse outcome than those with low-risk, especially for mortality (odds ratio, 11; P < 0.05). Surgical-related death was as low as 0.8 percent. CONCLUSIONS: All three histologic risk factors are significantly associated with the clinical outcome. Classification in low-risk and high-risk patients may be regarded as a meaningful staging procedure. [Key words: Malignant polyp; Histologic risk factors]

Correspondence to: Sergio Morini, M.D., Ospedale Nuovo Regina Margherita, Gastroenterologia ed Endoscopia Digestiva, Via Morosini 30, 00153, Roma, Italia, e-mail: [email protected] Dis Colon Rectum 2005; 48: 1588–1596 DOI: 10.1007/s10350-005-0063-3 © The American Society of Colon and Rectal Surgeons Published online: 26 May 2005

T

he malignant polyp, namely an adenoma containing a carcinoma invading the submucosa but not the muscular layer, represents the vast majority of early colorectal cancer, carrying a small but significant risk of both lymphohematic metastasis and mortality.1 Its prevalence has been steadily rising during the past decades—mainly because of the widespread use of colonoscopy and the improvement of polypectomy techniques—accounting for as much as 11 percent of all endoscopically removed polyps.2,3 In clinical practice, malignant polyps often are an unexpected histologic finding that leaves both patient and physician with the understandable emotional reaction of a diagnosis of malignancy, despite an apparent radical polypectomy. The main therapeutic dilemma is whether endoscopic treatment is adequate or a surgical resection is needed.4 Although the former is highly effective in the local removal of lesions down to the submucosal layer, it carries a substantial risk of residual cancer when malignant tissue has penetrated deeper into the bowel wall or infiltrated the pericolic nodes.5 Moreover, in those patients in whom piecemeal polypectomy is technically unavoidable, meaningful histologic assessment is heavily compromised ensuing in uncertainty for management. Although surgery allows accurate staging and treatment of both local and lymphnodal disease, it does harbor a definite risk of morbidity and mortality, especially in elderly patients or in those with rectal dis-

1588

Vol. 48, No. 8

CLINICAL OUTCOME IN COLORECTAL MALIGNANT POLYP

ease.6 Moreover, it does not always prevent a dismal hematogenous metastasis. Therefore, patients run the risk of residual disease after an apparently radical polypectomy or undergo unnecessary surgery after a previously successful endoscopic procedure. To reduce such risks, many endoscopic and histologic variables have been evaluated. In detail, some putative risk factors, such as assessment of both resection margin and differentiation grade of carcinoma, and evaluation of vascular neoplastic invasion, have all been widely applied, whereas others, i.e., depth and mass of invasion, have been less frequently addressed.7 Unfortunately, the heterogeneous application of such criteria in previous studies originating from different settings, together with the small sample size of each of these studies, undermines a comprehensive interpretation of the available data. Consequently, although such topics have been debated for the last 20 years, official guidelines are still controversial, thus clouding the decision-making process.8 We conceived a systematic review of the relevant published studies to evaluate any association between the proposed histologic risk factors and the occurrence of unfavorable outcomes in patients with invasive malignant polyps.

METHODS We attempted to identify all reported studies regarding the treatment of colonic malignant polyps through a MEDLINE search and a survey of related bibliographies. The search was limited to full-paper publications in the English language. These studies were included in this review when dealing with patients described as having a diagnosis of malignant polyp—defined as infiltration of malignant cells into the submucosa of an adenomatous polyp9—and when at least one of the following histologic findings was clearly reported: 1) carcinoma in correspondence of the resection margin after endoscopic polypectomy, 2) poor differentiation of the invasive adenocarcinoma, and 3) vascular invasion. Studies were excluded when enrolling patients with flat or depressed early colorectal cancer or when the number of treated and cured patients could not be extracted. A comprehensive search of the literature was conducted, starting in January 1980 and ending in June 2003. Publications identified as duplicates were excluded. Two investigators performed the assessment of each article for inclusion in this review. Any dis-

1589

agreements between the two observers were resolved before the paper was available for the subsequent analysis. Data collection was performed by pooling each single histologic risk factor into five different clinical outcomes through a form specially prepared for this analysis. Two different reviewers independently analyzed each study and recorded the data on this form. Both forms were compared, and one form was used to correct errors and disagreements. Criteria for a dichotomized definition of each variable were introduced before pooling the data. In detail, the status of the resection margin was defined as positive when regarded as positive or doubtful in each study, as suggested elsewhere,10 and negative when a cancer-free edge of the submucosal transsection point was reported. Differentiation of carcinoma was graded as well/moderate and poor,6 according to the presence or absence of areas of poor differentiation. Vascular invasion was taken as positive when presence of cancer in lymphatic channels or venous vessels was stated, and it was defined as negative when cancer invasion of these structures was clearly excluded. Endoscopic and surgical data as well as complications and follow-up information also were collected from each study. In detail, the configuration of the polyp morphology (sessile or pedunculated)— based on the gross macroscopic examination or the gross description from the pathology or endoscopic reports—and the polyp site (rectum, left, or right colon) were separately analyzed. For clinical outcomes, the following five issues were carefully evaluated: 1) residual disease; 2) recurrent disease; 3) lymph node metastasis; 4) hematogenous metastasis; 5) mortality. Adverse outcomes were defined as above to simulate the actual clinical scenario faced by a physician when an endoscopically removed polyp is diagnosed as malignant, and a further surgical treatment is to be decided. In detail, residual disease was defined as present when residual cancer was described at the polypectomy site at postsurgical pathologic evaluation, and as negative when excluded at definitive pathology evaluation. Recurrent disease was regarded as the reappearance of cancer after successful endoscopic treatment. The total of residual and recurrent cases was regarded as local disease. The lymph node metastasis rate was only evaluated in the group of patients who underwent surgical resection. Hematogenous metastases were assessed where clearly stated. For each of the three histologic risk factors, data were collected and correlated with each of the five clinical outcomes. In addi-

HASSAN ET AL

1590

tion, further analysis regarding the selected clinical outcomes was performed by subgrouping patients according to the simultaneous absence of all the three histologic risk factors (low-risk malignant polyp) or the presence of at least one of them (high-risk malignant polyp), as suggested elsewhere.5 Finally, because the importance of vascular invasion has been recently questioned,11 the clinical outcome of lowrisk malignant polyps was evaluated after including in this group those polyps with only such a risk factor.

Statistical Analysis Statistical analysis was performed using Chisquared test and Fisher exact test as appropriate. Odds ratios and relative 95 percent confidence intervals also were calculated. Differences were considered significant at 5 percent probability level.

RESULTS Descriptive Analysis A total of 38 studies were analyzed. For this review, 31 papers matching the study criteria and including 1,900 patients with a diagnosis of invasive malignant polyp were selected.12–42 Seven studies were excluded from our analysis for not matching the definition of malignant polyp (n = 1), including duplication of the patients (n = 1), not including any histologic risk factor (n = 2), or dealing with flat or depressed early colorectal cancer (n = 3). All the included studies were retrospective. The number of patients enrolled in each study ranged from 9 to 151 (median, 50.5). Fifteen (48.4 percent) were based in the United States, 10 (32.2 percent) in Europe, and 6 (19.4 percent) in Asia. Fourteen (45.2 percent) came from surgical departments, five (16.1 percent) from endoscopic or clinical units, and four (12.9 percent) from pathology; the remaining eight (25.8 percent) involved more than one of these specialties. In three studies, only patients who underwent surgical treatment were enrolled; in one study, only those who underwent endoscopic polypectomy, whereas both therapeutic approaches were taken into account in the remaining 28 series. Overall, 1,020 (53.7 percent) patients were reported to have undergone surgical resection after previous polypectomy, whereas the remaining 880 (46.3 percent) were treated exclusively by endoscopy. In 23 series,14–22,24–27,29,31–35,38,40–42 clinical or endoscopic follow-up data were available, ranging from 12 to 90 (median, 60) months, and accounting

Dis Colon Rectum, August 2005

for 1,535 (80.1 percent) of the included polyps. Regarding endoscopic data, polyp site was clearly specified in 17 series, accounting for 1,461 polyps.14,20–22,24,26,27,29,20–37,39 In such studies, the rectum was involved in 19.6 percent of the cases, the left colon in 71 percent, and the right colon in the remaining 9.4 percent. Polyp morphology was available in 19 series, enrolling 1,145 polyps.12,14–16,20–23,25,26,28,29,32,33,35–37,40,41 Malignant polyps were reported to be pedunculated in 753 (65.7 percent) and sessile in the remaining 392 (34.3 percent) cases. Regarding the occurrence of unfavorable outcomes, presence of residual disease and lymph node metastasis at postsurgical examination was reported in 28 studies, accounting for 950 operated patients.12–22,24–34,36–38,40–42 In detail, residual disease was identified in 134 (14.1 percent) and lymph node involvement in 82 (8.6 percent) patients. Postsurgical staging of residual disease was only described in 7 studies, in which 11 (42 percent) of 26 patients were classified as Dukes A, 8 (31 percent) as Dukes B, 6 (23 percent) as Dukes C, and 1 (4 percent) as Dukes D.15,16,18,27,36,37,42 Recurrent disease at endoscopic follow-up was reported to have occurred in 23 (2.7 percent) of 845 nonoperated patients.13–19,21,22,24,26,27,29,31–36,39–42 Overall, local disease was reported in 157 (8.7 percent) of 1,795 patients. On the other hand, in 30 series, hematogenous metastasis occurred in 54 (3 percent) of 1,782 patients, without any significant difference between the surgical (30/982, 3 percent) and the endoscopic (24/800, 3 percent) groups.12–37,39–42 Finally, cancerrelated death was reported in 56 (3.6 percent) of those 1,535 patients for whom follow-up data were available, with an interval from the diagnosis ranging from 5 to 101 (median, 32) months. Death because of surgical complications, inferred from 28 studies, was reported in 8 (0.8 percent) of 951 operated patients.12–21,24–34,36–42

Resection Margin Evaluation of the resection margin was clearly described in 20 studies, including a total of 980 polyps.12–14,16–19,21,27–34,37,40–42 According to our protocol, the resection margin was reported to be positive in 325 (33.2 percent) and negative in the remaining 655 (66.8 percent) cases. In those studies for which polyp morphology also could be inferred, a positive resection margin was significantly more frequent in sessile (25/44) than in pedunculated (34/182) polyps

Vol. 48, No. 8


1591

Table 1. Relationship Between Histologic Risk Factors and Clinical Outcomes Risk Factor Margin of resection Positive Negative Odds ratio 95% CI Poor differentiation Positive Negative Odds ratio 95% CI Vascular Invasion Positive Negative Odds ratio 95% CI

Residual Disease

Recurrent Disease

Lymph Node Metastasis

55/181 (30.4)a 4/142 (2.8) 15 (5.3–42.7)

13/77 (16.8)a 4/357 (1.12) 17.9 (5.7–56.7)

10/56 (17.8%) 29/324 (9%) 2.2 (1–4.8)

— —

13/56 (23.2)a 23/324 (7.1) 3.9 (1.9–8.4)

11/14 (9.6)a 40/1,520 (2.6) 3.9 (2–7.9)

14/96 (14.6)a 27/1,487 (1.8) 9.2 (4.7–18.3)

6/34 (17.6%) 17/111 (15.3%) 1.2 (0.4–3.3)

— —

12/34 (35.3)a 8/111 (7.2) 7 (2.6–19.2)

13/250 (5.2) 38/1,279 (3) 1.8 (0.9–3.4)

7/210 (3.3) 28/1,194 (2.3) 1.4 (0.6–3.3)

13/181 (7.2) 13/142 (9.2) 0.8 (0.3–1.7)

Hematogenous Metastasis 30/325 (9.2)a 8/655 (1.2) 8.2 (3.7–18.2)

Mortality 26/325 (8)a 9/655 (1.4) 6.2 (2.9–13.5)

CI = confidence interval. Data are numbers with percentages in parentheses unless otherwise indicated. a P < 0.05.

(56.8 vs. 18.7 percent; P < 0.0001).12,14,28,32,33 Local disease was much more frequent in patients with a positive (68/258) than in those with a negative (8/499) margin (26.4 vs. 1.6 percent; odds ratio, 22; 95 percent confidence interval (CI), 10.3–46.6; P < 0.0001; Table 1). Although lymph node localization was similar in the two groups, hematogenous metastasis and cancer-related mortality were significantly more frequent in patients with a positive margin than in those without this risk factor.

from 0 to 56.8 percent in the different series.12–14,16,18–23,25,26,28–31,33,35,39,40–42 This risk factor tended to be more frequent in sessile (30/214) compared with pedunculated (30/327) polyps (14 vs. 9.2 percent, respectively; P = 0.08).12,20,21,26,28,33,35 Lymph node metastasis was much more common in polyps with such a risk factor than in those without it, whereas no substantial difference in residual disease, hematogenous metastasis, or cancer-related mortality was detected (Table 1).

Grade of Differentiation

Low-risk and High-risk Groups

The grade of differentiation was clearly reported in 27 studies.12–14,16–23,25–33,35,37–42 Poor differentiation was described in 116 (7.2 percent) of 1,612 polyps. Subgrouping for morphology, poor differentiation was more frequently detected in sessile (25/214) than in pedunculated (21/308) polyps (11.7 vs. 6.8 percent; P = 0.05).12,20,21,26,28,32,33,35 From 13 studies, it was apparent that the majority of patients with such a risk factor (56/60) underwent surgical resection compared with only 336 of the 739 without it (93.3 vs. 45.4 percent; P < 0.0001).12,17,18,20,26–29,33,37,38,41,42 In addition, patients with areas of poor differentiation had a substantially higher risk of both metastatic disease and cancer-related mortality compared with their counterparts (Table 1).

In 13 studies, 375 (58.3 percent) of 643 patients showed absence of all three risk factors and were classified—according to our protocol—into a lowrisk group, whereas the remaining 268 (41.7 percent), in which at least one risk factor was present, were categorized into a high-risk group.12,14,18,21,26,28,29,31,33,37,40–42 In detail, 190 of 268 high-risk patients were treated surgically after endoscopic polypectomy compared with only 80 of 375 low-risk patients (70.9 vs. 21.3 percent; P < 0.0001). All five unfavorable clinical outcomes were significantly more prevalent in the high-risk compared with the low-risk group (Table 2). In detail, local disease was reported in 1 of 295 low-risk patients compared with 40 of 230 high-risk cases (0.3 vs. 17.4 percent; odds ratio, 61.9; 95 percent CI, 8.4–454.2; P < 0.0001). In the high-risk group, the 190 patients treated by surgery compared with the 78 cases treated by endoscopy alone tended to have a lower risk of both he-

Vascular Invasion Vascular invasion appeared to be present in 254 (17.6 percent) of 1,438 polyps, ranging widely

HASSAN ET AL

1592


Table 2. Comparison Between Low-Risk and High-Risk Groups Histologic Evaluation Low-risk High-risk Odds ratio 95% CI

Residual Disease a

1/80 (1.2) 33/154 (21.4) 21.5 (2.9–160.8)

Recurrent Disease


a

a

0/215 (0) 7/76 (9.2) 46.5 (2.6–179.6)

4/78 (5.1) 20/178 (11.2) 2.3 (0.8–28.5)

Hematogenous Metastasis a

1/375 (0.3) 19/268 (7.1) 28.5 (3.8–214.6)

Mortality 2/287 (0.7)a 17/236 (7.2) 11 (2.5–48.4)


matogenous metastasis (5.8 vs. 8.9 percent; P = 0.5) and cancer-related mortality (6.6 vs. 10.1 percent; P = 0.5), although the difference was not statistically significant. Overall, 83 (30.1 percent) of 268 high-risk patients had vascular infiltration as the only histologic risk factor. By including these patients into the lowrisk group, a new sample population was created. Noteworthy, such group did not show any significant difference compared with the original low-risk group regarding all the considered outcomes (Table 3).

Polyp Morphology In those studies in which polyp morphology was related to therapeutic options, it was found that 158 (85.4 percent) of 185 patients with sessile lesions underwent surgical treatment compared with only 124 (46.3 percent) of 268 patients with pedunculated polyps (P < 0.0001).12,15,20,26,28,33 Moreover, patients with sessile polyps were reported to have a worse clinical outcome compared with those with pedunculated lesions for all the considered outcomes, but the lymph node metastasis rate (Table 4). Local disease was cumulatively more frequent in sessile polyps (22/222) than in pedunculated (5/362) lesions (9.9 vs. 1.4 percent; odds ratio, 7.9; 95 percent CI, 2.9–21; P = 0.0001).

DISCUSSION According to the recognized adenoma-carcinoma sequence, colorectal neoplasia is a single, indivisible continuum that begins within the mucosa as mild epithelial dysplasia and can progress through dysplasia of increasing severity until finally invasion across the muscularis mucosae occurs, at which time the term carcinoma applies.43 Because of the absence of lymphatic vessels in the mucosal layer, the intramucosal neoplasia behaves like a benign adenoma.44 For this reason, polyps harboring “in situ” or “intramucosal”

cancer are not generally regarded or treated as “invasive malignant” polyps.45 Indeed, the definition of “invasive malignant” polyp implies the penetration of cancer cells—beyond the muscolaris mucosae—into the submucosa. Invasion of the submucosa opens the way to metastasis via the lymphatic and blood vessels.27 Consistently, a substantial cumulative risk of local and metastatic disease with consequent cancerrelated death is recognized in these patients.23 Regrettably, the available literature on this topic mainly consists of miscellaneous, retrospective series performed in heterogeneous settings. Nonetheless, at least three histologic prognostic indicators—positivity at the resection margin, presence of poorly differentiated carcinoma, vascular invasion—have been extrapolated.7 Unfortunately, the limited size of these studies has led to divergent conclusions regarding their usefulness, so that even the official guidelines are hesitant on this issue.8 The relationship between the histologic prognostic indicators of cancerized adenomas and the different clinical outcomes has never been systematically quantified, so that management of this disease is made on a case by case basis without reliable criteria being available—a highly unusual situation for an oncologic disease. To our knowledge, this is the first large systematic review dealing with all the clinical features of the invasive malignant polyps. This pooled analysis allowed us to evaluate the specific predictive value of the three selected risk factors simultaneously: positive resection margin, poor differentiation, and vascular invasion. Intriguingly, our analysis has shown that the different histologic risk factors are clearly linked with distinct clinical outcomes. In detail, a positive resection margin is largely predictive of local disease, the presence of poorly differentiated carcinoma is mainly associated with a higher cancer-related mortality, and vascular invasion with a higher risk of lymph node metastasis.

Vol. 48, No. 8


1593

Table 3. Comparison Between Low-Risk Groups With and Without Vascular Invasion Histologic Evaluation

Residual Disease

Recurrent Disease


Hematogenous Metastasis

Mortality

Low-risk without VI Low-risk with VI Odds ratio 95% CI

1/80 (1.2) 1/105 (1) 0.7 (0.04–12.3)

0/215 (0) 2/254 (0.8) 0.2 (0.01–4.9)

4/78 (5.1) 8/116 (6.9) 0.7 (0.21–2.5)

1/375 (0.3) 2/400 (0.5) 0.5 (0.04–5.8)

2/287 (0.7) 3/325 (0.9) 0.7 (0.12–4.5)

VI = vascular invasion; CI = confidence interval. Data are numbers with percentages in parentheses unless otherwise indicated. Table 4. Comparison Between Pedunculated and Sessile Polyps Polyp Morphology Pedunculated Sessile Odds ratio 95% CI

Residual Disease

Recurrent Disease a

4/124 (3.2) 18/158 (11.4) 3.8 (1.3–11.7)

Lymph Node Metastasis a

1/238 (0.4) 4/64 (6.2) 15.8 (1.7–144)

12/124 (9.7) 17/162 (10.5) 1.1 (0.5–2.4)

Hematogenous Metastasis a

5/595 (0.8) 13/357 (3.6) 4.5 (1.6–12.6)

Mortality 4/627 (0.6)a 18/369 (4.9) 7.9 (2.7–23.8)


These observations clearly suggest that after endoscopic polypectomy all three risk factors need to be simultaneously evaluated by the pathologist and that the classification of patients in low-risk and high-risk groups is clinically meaningful. A second relevant finding is to have definitely shown that low-risk and high-risk polyps do not only differ in lymphnodal disease rate, as previously reported,19 but also for distant metastasis and mortality rates. Such worse clinical outcome occurred despite the majority of the highrisk patients underwent surgical resection. This observation strengthens the usefulness of this classification not only for addressing the therapeutic choice, but also as a staging procedure. On the other hand, the favorable outcome observed in patients with low-risk polyps is tainted only by the unexpected relatively high prevalence of lymph node metastasis. It should be noted that the calculated node metastasis rate of 5 percent in these patients is entirely because of the four cases in the series by Colacchio et al.,12 being otherwise 0 percent in the remaining studies. This observation clearly suggests a possible bias in the study by Colacchio et al., as admitted by the same authors and underlined by others.26,46 The histologic risk factors are likely to have different prognostic values. Although no doubt has been advanced regarding the therapeutic implications of a positive resection margin or a poor differentiated carcinoma, the predictive value of vascular invasion has

been widely debated in the literature.29 In our investigation, no significant differences in terms of clinical outcome were found pooling low-risk adenomas containing invasive cancer with those with only vascular invasion. Such observation strengthens the previous observations that the latter did not seem to be an independent risk factor at multivariate analysis22,39 and should be probably downgraded.11 Considering polyp morphology, the sessile type is clearly associated with an unfavorable outcome compared with that pedunculated. Although patients with sessile polyps frequently underwent surgery (85 percent), their overall mortality remained roughly eight times higher compared with patients with pedunculated polyps. This seems to be mainly because of a significantly higher prevalence of all the histologic risk factors in this group rather than to a predefined biologically aggressive behavior, as previously suggested.5,28 In detail, a positive resection margin seemed to be by far the most crucial risk factor in sessile polyps, likely because of an inadequate endoscopic removal of these lesions. This confirms the previous analysis by Haggitt et al.39 in which the level of invasion, but not the sessile morphology, seemed to be an independent risk factor of an adverse outcome. Whether the new endoscopic procedures of polypectomy based on submucosal blebs may reduce such dismal risk factor in patients with sessile polyps needs to be evaluated.

HASSAN ET AL

1594

A further relevant finding of our data is that the overall surgical-related death is only 0.8 percent—a much lower rate than the 2 percent currently computed in the literature.6 Because surgical-related mortality seems to be a crucial counterpart in this field,47 the future decision-making process is likely to be affected by this observation. Conversely, grouping the surgical series, it was noted that no more than 30 percent of patients with high-risk polyps did actually have residual cancer at postsurgical examination. This is presumably because of the possible fulguration of the residual cancer by the polypectomy current.5 This should serve as a warning before hastening high-risk surgery patients to the operating room, especially if we take into consideration that more aggressive endoscopic therapies, such as laser or mucosectomy,48 were scarcely applied in the included series. An unavoidable limitation of the current analysis is to be entirely based on retrospective series, no prospective studies being available. Although the absolute values of the odds ratios we calculated in this study may not be completely accurate—as well as a possible “effect-size” bias cannot be ruled out—they nevertheless indicate an unquestionable trend that will be useful in clinical practice. Prospective studies could obviously clarify the issue, but a well-designed, prospective, large, randomized study is unlikely to be undertaken because of ethical implications. Finally, although these histologic risk factors are easily identifiable, an interobserver disagreement among pathologists has been reported.49 Such limitation could be overcome by the identification of biomarkers predictive of an invasive/metastatic behavior in the early stage of colorectal cancer. Unfortunately, differently from advanced colorectal cancer, no significant relationship between proliferative index or DNA ploidy and the metastatic potential of malignant polyps has been demonstrated.50,51 This could be because of the possibility that the malignant polyp represents the morphologically indistinguishable outcome of two different genetic pathways, the former blocking the growth of early cancer, the latter allowing its fast progression toward advanced colorectal cancer.52

CONCLUSIONS The therapeutic implications of our analysis are to confirm the validity of a conservative approach (i.e., endoscopic polypectomy) in low-risk polyps because


of the small risk of local and distant disease, and to reassure patients with high-risk polyps about the low mortality from surgery. Chemotherapy should probably be pioneered in patients with poorly differentiated cancer in their polyps because of the high risk of metastatic disease and mortality rate despite a radical surgery. Patients at high-risk from surgery can be successfully managed by endoscopy in at least two-thirds of the cases, although local laparoscopic resection has been suggested as a promising option in patients unfit for conventional surgery.5 Obviously, whether any uncertainty on the accuracy of endoscopic or histologic assessment is present, surgery should be advised. The importance of sharing the ultimate therapeutic decision with the patient has always been stressed,46 as usually advised in all the oncologic diseases. We believe that our evaluation of the distant metastasis and mortality rates as well as the odds ratios of each single risk factor for all the main clinical outcomes can substantially ease this process.

REFERENCES 1. Lipper S, Kahn LB, Ackerman LV. The significance of microscopic invasive cancer in endoscopically removed polyps of the large bowel. Cancer 1983;52:1691–9. 2. Hermanek P, Fruhmorgen P, Guggenmoos-Holzmann I, Altendorf A, Matek W. The malignant potential of colorectal polyps—a new statistical approach. Endoscopy 1983;15:16–20. 3. Nusko G, Mansmann U, Partzsch U, et al. Invasive carcinoma in colorectal adenomas: multivariate analysis of patient and adenoma characteristics. Endoscopy 1997; 29:626–31. 4. Winawer SJ, Witt TR. Cancer in a colonic polyp, or malignant colonic adenomas—is polypectomy sufficient? Gastroenterology 1981;81:625–6. 5. Williams B, Saunders BP. The rationale for current practice in the management of malignant colonic polyps. Endoscopy 1993;25:469–74. 6. Waye JD, Haggitt RC. Controversies, dilemmas, and dialogues. When is colonoscopic resection of an adenomatous polyp containing a “malignancy” sufficient? Am J Gastroenterol 1990;85:1564–8. 7. Jass JR. Malignant colorectal polyps. Gastroenterology 1995;109:2034–5. 8. Anonymous. Guidelines for colorectal cancer screening and surveillance. Gastrointest Endosc 2000;51:777–82. 9. Riddell RH. Hands off “cancerous” large bowel polyps. Gastroenterology 1985;89:432–5. 10. Morson BC, Bussey HJ, Samoorian S. Policy of local excision for early cancer of the colorectum. Gut 1977; 18:1045–50.

Vol. 48, No. 8


11. Kikuchi R, Takano M, Takagi K, et al. Management of early invasive colorectal cancer. Risk of recurrence and clinical guidelines. Dis Colon Rectum 1995;38:1286–95. 12. Colacchio TA, Forde KA, Scantlebury VP. Endoscopic polypectomy: inadequate treatment for invasive colorectal carcinoma. Ann Surg 1981;194:704–7. 13. Fucini C, Wolff BG, Spencer RJ. An appraisal of endoscopic removal of malignant colonic polyps. Mayo Clin Proc 1986;61:123–6. 14. Hackelsberger A, Fruhmorgen P, Weiler H, Heller T, Seeliger H, Junghanns K. Endoscopic polypectomy and management of colorectal adenomas with invasive carcinoma. Endoscopy 1995;27:153–8. 15. Pines A, Bat L, Shemesh E, et al. Invasive colorectal adenomas: surgery versus colonoscopic polipectomy. J Surg Oncol 1990;43:53–5. 16. Conte CC, Welch JP, Tennant R, Forouhar F, Lundy J, Bloom GP. Management of endoscopically removed malignant colon polyps. J Surg Oncol 1987;36:116–21. 17. Volk EE, Goldblum JR, Petras RE, Carey WD, Fazio VW. Management and outcome of patients with invasive carcinoma arising in colorectal polyps. Gastroenterology 1995;109:1801–7. 18. Muller S, Chesner IM, Egan MJ, et al. Significance of venous and lymphatic invasion in malignant polyps of the colon and rectum. Gut 1990;30:1385–91. 19. Cooper HS, Deppisch LM, Gourley WK, et al. Endoscopically removed malignant colorectal polyps: clinico-pathological correlations. Gastroenterology 1995;108:1657–65. 20. Nivatvongs S, Rojanasakul A, Reiman HM, et al. The risk of lymph node metastasis in colorectal polyps with invasive adenocarcinoma. Dis Colon Rectum 1991;34: 323–8. 21. Kyzer S, Begin LR, Gordon PH, Mitmaker B. The care of patients with colorectal polyps that contain invasive adenocarcinoma: endoscopic polypectomy or colectomy? Cancer 1992;70:2044–50. 22. Masaki T, Muto T. Predictive value of histology at the invasive margin in the prognosis of early invasive colorectal carcinoma. J Gastroenterol 2000;35:195–200. 23. Park YJ, Kim WH, Paeng SS, Park JG. Histo-clinical analysis of early colorectal cancer. World J Surg 2000; 24:1029–35. 24. Richards WO, Webb WA, Morris SJ, et al. Patient management after endoscopic removal of the cancerous colon adenoma. Ann Surg 1987;205:665–72. 25. Russell JB, Chu DZ, Russell MP, Chan CH, Thompson C, Schaefer RF. When is polypectomy sufficient treatment for colorectal cancer in a polyp? Am J Surg 1990;160: 665–8. 26. Christie JP. Polypectomy or colectomy? Management of 106 consecutively encountered colorectal polyps. Am Surg 1988;54:93–9. 27. Morson BC, Whiteway JE, Jones EA, Macrae FA, Wil-

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

1595

liams CB. Histopathology and prognosis of malignant colorectal polyps treated by endoscopic polypectomy. Gut 1984;25:437–44. Coverlizza S, Risio M, Ferrari A, Fenoglio-Preiser CM, Rossini FP. Colorectal adenomas containing invasive carcinoma. Pathologic assessment of lymph node metastatic potential. Cancer 1989;64:1937–47. Geraghty JM, Williams CB, Talbot IC. Malignant colorectal polyps: venous invasion and successful treatment by endoscopic polypectomy. Gut 1991;32:774–8. Kawamura YJ, Sugamata Y, Yoshino K, et al. Endoscopic resection for submucosally invasive colorectal cancer: is it feasible? Surg Endosc 1999;13:224–7. Kitamura K, Taniguchi H, Yamaguchi T, Sawai K, Takahashi T. Clinical outcome of surgical treatment for invasive early colorectal cancer in Japan. Hepatogastroenterology 1997;44:108–15. Bartnik W, Butruk E, Orlowska J. A conservative approach to adenomas containing invasive carcinoma removed colonoscopically. Dis Colon Rectum 1985;28: 673–5. Sugihara K, Muto T, Morioka Y. Management of patients with invasive carcinoma removed by colonoscopic polypectomy. Dis Colon Rectum 1989;32:829–34. 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 1984;27:6–9. Pollard CW, Nivatvongs S, Rojanasakul A, Reiman HM, Dozois RR. The fate of patients following polypectomy alone for polyps containing invasive carcinoma. Dis Colon Rectum 1992;35:933–7. Rossini FP, Ferrari A, Spandre M, Coverlizza S. Coloscopic polypectomy in diagnosis and management of cancerous adenomas: an individual and multicentric experience. Endoscopy 1982;14:124–7. Dell’Abate P, Iosca A, Galimberti A, Piccolo P, Soliani P, Foggi E. Endoscopic treatment of colorectal benignappearing lesions 3 cm or larger: techniques and outcome. Dis Colon Rectum 2001;44:112–8. Chantereau MJ, Faivre J, Boutron MC, et al. Epidemiology, management, and prognosis of malignant large bowel polyps within a defined population. Gut 1992; 33:259–63. Haggitt RC, Glotzbach RE, Soffer EE, Wruble LD. Prognostic factors in colorectal carcinomas arising in adenomas: implications for lesions removed by endoscopic polypectomy. Gastroenterology 1985;89:328–36. Netzer P, Forster C, Biral R, et al. Risk factor assessment of endoscopically removed malignant colorectal polyps. Gut 1998;43:669–74. Cranley JP, Petras RE, Carey WD, Paradis K, Sivak MV. When is endoscopic polypectomy adequate therapy for colonic polyps containing invasive carcinoma? Gastroenterology 1986;91:419–27.

1596

HASSAN ET AL

42. Whitlow C, Gathright JB Jr, Hebert SJ, et al. Long-term survival after treatment of malignant colonic polyps. Dis Colon Rectum 1997;40:929–34. 43. Morson BC. The polyp-cancer sequence in the large bowel. Proc R Soc Med 1974;67:451–4. 44. 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 1973;64:51–66. 45. Muto T, Sawada T, Sugihara K. Treatment of carcinoma in adenomas. World J Surg 1991;15:35–40. 46. Wilcox GM, Anderson PB, Colacchio TA. Early invasive carcinoma in colonic polyps. A review of the literature with emphasis on the assessment of the risk of metastasis. Cancer 1986;57:160–71. 47. Wilcox GM, Beck JR. Early invasive cancer in adenomatous colonic polyps (“malignant polyps”). Evaluation

48.

49.

50.

51.

52.


of the therapeutic options by decision analysis. Gastroenterology 1987;92:1159–68. Karita M, Tada M, Okita K, Kodama T. Endoscopic therapy for early colon cancer: the strip biopsy resection technique. Gastrointest Endosc 1991;37:128–32. Cooper HS, Deppisch LM, Kahn EI, et al. Pathology of the malignant colonic polyp. Hum Pathol 1998;29: 15–26. Risio M, Rossini FP. Cell proliferation in colorectal adenomas containing invasive carcinoma. Anticancer Res 1993;13:43–8. Risio M, Geido E, Rapallo A. DNA ploidy analysis in selected regions of colorectal adenomas containing carcinoma. Int J Oncol 1993;3:941–7. Risio M, Casorzo L, Chiecchio L, De Rosa G, Rossini FP. Deletions of 17p are associated with transition from early to advanced colorectal cancer. Cancer Genet Cytogenet 2003;47:44–9.