Preoperative Parameters Expanding the Indication of Sphincter Preserving Surgery in Patients With Advanced Low Rectal Cancer

Abstract

Objective:

To clarify the preoperative parameters of the required distal margin that can be applied to the criteria of sphincter-preserving surgery in rectal cancer.

Summary Background Data:

Although aggressive sphincter-preserving surgery, including intersphincteric resection, is beginning to be applied to low rectal tumors, unexpected distal cancer spread might undermine local control in patients undergoing such a procedure. The ‘two-centimeter rule’ of distal clearance is predominant at present, whereas preoperative criteria to determine the individual required distal margin have not yet been established.

Methods:

First, by reviewing 556 rectal cancers, promising risk parameters of intramural distal spread (IM) were selected and, subsequently, such parameters were examined in regard to whether they could be evaluated preoperatively. Furthermore, 80 patients with lower rectal cancers located above the anal canal who were undergoing abdominoperineal resection were reviewed as to whether IM risk factors could be used as criteria to identify the low rectal cancer with or without anal canal involvement.

Results:

IM was observed in 10.6% (IM ≥ 10 mm: 2.3%) of the patients examined, and the incidence was higher in tumors with certain unfavorable histologic characteristics, including tumor “budding,” in their submucosal region at the distal edge (24.4%) than in those with no such histology (5.3%). Regarding such unfavorable histology as IM risk factor, together with 3/4 or more annularity and type 3 gross appearance, IM rates were 3.3% (IM ≥ 10 mm: 0.5%) in the no-risk group, 9.1% (IM ≥ 10 mm: 1.7%) in the one-risk group, and 29.1% (IM ≥ 10mm: 7.8%) in the multiple-risks group. These results were reproduced well even if such risk factors were evaluated endoscopically or histologically on preoperative biopsy specimens. Furthermore, no anal canal involvement was observed in 32 tumors without IM risk; however, microscopic cancer spread down to the anal canal, including that into outside of the internal sphincter muscle, was observed in 9.1% of tumors with one IM risk and in 26.7% of multiple-risk tumors.

Conclusions:

The preoperative evaluation of particular parameters related to IM enabled the accurate selection of rectal cancer to which the one-centimeter rule of distal clearance can be applied. This could allow us to expand the indication of sphincter preservation for very low rectal cancer patients.

The preoperative evaluation of 3 parameters (histologic characteristics in the submucosal region, annularity, and gross appearance of tumor, which were selected as independent parameters of intramural spread by reviewing 556 tumors) enabled the accurate selection of rectal cancer to which the “one-centimeter rule” of distal clearance can be applied. In addition, it was demonstrated that they could be used as criteria to identify the low rectal cancer with or without anal canal involvement.

In determining the surgical procedure against cancer, surgeons are often caught in a dilemma of antithetical concepts, that is, should the resection margin be extended for curability or should it be at its minimum, giving thought to the postoperative functional results? This is a scenario faced by physicians as they determine the distal cutting level for patients with low rectal cancer. Several reports have described the physical and psychologic burdens of patients undergoing abdominoperineal resection with permanent colostomy.^1–3 The most limiting factor of sphincter-preserving surgery has been the tumor’s lower edge, and intersphincteric resection has allowed us to obtain a certain extent of distal margin and preserve sphincter function.^4–6 Actually, this is thought to be the only method for patients with tumors found at operation to infiltrate the internal anal sphincter, allowing the avoidance of abdominoperineal resection; we are awaiting long-term prognostic and functional results to determine this method’s applicability.

However, for patients with low advanced rectal cancer located above the anal sphincter, there are no generally accepted guidelines as to which patients should undergo intersphincteric resection instead of conventional sphincter preserving surgery (ie, low anterior resection) or abdominoperineal resection, although we have the “two-centimeter rule” concerning the distal clearance, which has recently been pointed out to mislead us into selecting excessive surgery.^7–9 Generally, the level of distal tumor edge has been considered the only informal criterion for the selection of surgical procedure, and this criterion varies among surgeons. How much weight each surgeon gives to the unexpected distal cancer spread is thought to account for this variation. It is certain that accurate prediction of distal cancer spread beyond the macroscopic tumor’s distal end contributes not only to the selection between abdominoperineal resection and sphincter preserving surgery but also to the determination of the degree of resection of the sphincter muscle, which has been reported to affect the long-term outcome regarding fecal continence.¹⁰

Intramural distal spread has been shown to occur in 4–24% of patients who have undergone curative surgery.^9,11,12 Although the direct-type of intramural spread has been detected using transrectal ultrasonography,¹³ little is known regarding the kind of preoperative parameters we can use to predict distal spread, including discontinuous microscopic spread, which has long been known as distal spread’s dominant mode.^9,11 In the present work, we first analyzed the features of rectal cancer with intramural distal spread (IM) using surgically resected specimens and determined possible parameters that could be estimated preoperatively. As one of the parameters to predict IM, we analyzed the histologic characteristics of the submucosal lesion at the tumor edge, which could be evaluated preoperatively by biopsy and has been seen to be closely associated with the extent of extramesorectal cancer spread.¹⁴ In the following study, we verified whether such IM risk parameters could actually be evaluated preoperatively and correlated with the occurrence of IM. Finally, to determine the indication of sphincter function preservation with intersphincteric resection for patients with low rectal cancer located above the anal canal, the IM risk factors selected above were examined in regard to whether they can be used as criteria to judge the potential of anal canal involvement below the surface.

MATERIALS AND METHODS

Group 1 Patients: Intramural Distal Spread in Surgically Resected Specimens

Five hundred and sixty patients who underwent potentially curative surgery for advanced rectal cancer at the Department of Surgery I, National Defense Medical College, Japan, between 1981 and 1999 were included in the study. All tumors were evaluated preoperatively as being located above the anal canal, based on digital examination with and without voluntary anal contraction. This study was intended for 556 patients, excluding 4 diffusely infiltrating type cancers whose macroscopic border could not be apparently recognized. This patient group was composed of 335 males and 221 females; average age was 61.1 at operation (range, 31–88 years). Among these patients, 29 patients had distant metastases (liver, 18 cases; peritoneal dissemination, 2; nonregional nodes, 2; lung, 1; ovary, 1; liver + peritoneal dissemination, 2; liver + nonregional nodes, 2; peritoneal dissemination + nonregional nodes, 1) that had been completely removed surgically. No patients received preoperative adjuvant therapy; abdominoperineal resections were performed in 133 patients, anterior resections in 403, Hartmann’s procedure in 14, and total pelvic exenteration in 6. All distal ends of the surgically resected specimens examined, including the donuts in the stapler devices, were confirmed histologically to be free of cancer.

Handling of Resected Specimens

The tumor-bearing segments that were resected surgically were opened along the long axis of the rectum, and tumor size, annularity (the ratio of the largest length of the tumor perpendicular to the canal long axis to the circumference of the canal circle), and gross appearance of the tumor were evaluated in the fresh state and recorded. The gross appearance of advanced cancer was classified as follows: type 1, protuberant type; type 2, ulcerated type with clear margins; type 3, ulcerated type with infiltration; type 4, diffusely infiltrating type (not included in this study); and type 5, unclassified type.¹⁵ After fixation with 20% formalin, several section lines were made along the long axis of the intestine, which included the center of the tumor, according to the rules of the Japanese Classification of Colorectal Carcinoma.¹⁵ Before and after sectioning, photographs and sketches of the intestinal mucosa were made for mapping of the lesion and cutting lines. Pathologically diagnosed parameters, including depth of penetration, number of nodes involved, upward lymphatic spread (involvement of nodes around superior mesenteric artery), predominant histologic type, and degree of lymphatic and venous invasion, had been recorded in the pathologic reports, according to the definition and classification in the guideline noted above.¹⁵

Definition of Distal Intramural Cancer Spread (IM)

In detailing the tumor’s distal edge, 2 patterns should be reckoned with in measuring the length of IM: tumor developing without the accompanying normal mucosa, and tumor developing with the accompanying normal mucosa. In the present study, irrespective of the above pattern, IM length was defined as the distance between the beginning of macroscopic elevation on histologic slides and the most distal microscopic cancer deposits (or the most advanced point of direct distal spread; Fig. 1). This was measured basically on the specimen distally to the tumor edge that was located nearest to the anus by using a scale eyepiece.

FIGURE 1. Schema of distal intramural cancer spread (IM). IM length was measured between the beginning of macroscopic elevation on histologic slides and the most distal microscopic cancer deposits (or the most advanced point of direct distal spread). Upper, tumor developing without normal mucosa at its distal edge. Below, tumor developing accompanied with normal mucosa at its distal edge. White circle, actively invasive submucosal field examined histologically, presuming transanal submucosal biopsy. Black circle, discontinuous intramural cancer spread. Star, direct intramural cancer spread; M, muscular layer; SM, submucosal layer; MP, muscularis propria.

Assessment of Cancer Histology in the Submucosal Field at the Tumor’s Distal Edge

The submucosal field at the tumor’s distal edges (Fig. 1) had been examined histologically and recorded before the procedure of evaluating IM. The parameters examined in the present study were the following 3 histologic features in the submucosal invasive region: 1) tumor “budding,” 2) differentiation, and 3) vascular invasion. The estimation of these histologic parameters has been previously reported,¹⁴ that is, definition of a budding focus was an isolated single cancer cell or a cancer cluster composed of fewer than 5 cancer cells (Fig. 2). After choosing one field where the budding was the most intensive, a budding count was made under 200× magnification, with a count of less than 5 foci considered negative, and a count of 5 or more as positive. Judgment of budding in each case was made by one of the authors (H.U.), and the reproducibility of his judgment had been reported previously.¹⁶ With regard to tumor differentiation, grading was based on part of the actively invasive region. Lymphatic and venous invasion were examined and assessed together as vascular invasion. Positive tumor budding, the presence of vascular invasion, and poor differentiation were regarded as unfavorable characteristics in the submucosal invasive region at the tumor’s distal edge.¹⁴

FIGURE 2. Tumor budding. An isolated single cancer cell or a cancer cluster composed of fewer than 5 cancer cells observed in the actively invasive region was defined as tumor budding (hematoxylin and eosin, original magnification ×100).

Group 2 Patients: Preoperative Prediction for Intramural Distal Spread

One hundred and twenty advanced rectal cancer patients underwent transanal submucosal biopsy preoperatively between July 1995 and February 2001 at the institute mentioned above. Among these, 90 patients whose tumors were located above the anal canal and whose biopsy specimens were sufficient enough to allow histologic examination were studied in terms of the correlation between the histologic features of the submucosal biopsy specimens and IM observed on the surgical specimens obtained by subsequent operation. The types of operation performed were as follows: abdominoperineal resection in 14 patients, anterior resection in 72, Hartmann’s procedure in 2, and total pelvic exenteration in 2. The histologic parameters and categories examined were the same as those mentioned above. The process of making biopsy specimens used in this study has been previously reported,¹⁴ that is, biopsy specimens were extracted transanally from the tumor edge with the intention of including the actively invasive submucosal region. These were processed and sectioned after formalin fixation, then stained with hematoxylin and eosin.

Group 3 Patients: Anal Canal Involvement

Eighty pathologic specimens of low rectal cancers in cases that were resected by curatively intended abdominoperitoneal resection between 1981 and 1999 at the institute mentioned above were reviewed histologically. These included 8 patients who had distant metastases removed surgically, liver metastasis (4 patients), nonregional lymph node involvement (3), and peritoneal dissemination (1). In the same manner as in the 2 aforementioned data sets, no patients received preoperative adjuvant therapy, and the distal edges of these tumors were determined based on preoperative digital examination not to extend to the anal canal. It was further confirmed in the pathologic slides that the macroscopic anal edges of the tumors were all located above the height of the puborectal muscle. Reviewing the pathologic specimens made using the process mentioned above, anal canal involvement (ie, intramural or extramural spread down to the anal canal) by low rectal cancer was examined. In the present study, anal canal involvement was defined as cancer deposits found below the level of the upper edge of the puborectal muscle or direct distal spread beyond this level (Fig. 3).

FIGURE 3. A, Schema of anal canal involvement by rectal cancer above the anal canal. Dotted line shows the upper edge of the puborectal muscle, which is crossing or coming close to the longitudinal muscle. Based on histologic examination, cancer deposits below this line, and the direct distal spread beyond this line, was defined as anal canal involvement. Black circle, discontinuous intramural cancer spread. Star, direct distal cancer spread; IS, internal sphincter muscle; ES, external sphincter muscle; DL, dentate line. B and C, Anal canal involvement by lymphatic spread. Cancer deposits are observed in the lymphatic canal outside the internal sphincter muscle (hematoxylin and eosin, original magnification: B, ×3; C, ×17).

RESULTS

Incidence of IM

Distal intramural spread (IM) was observed in 59 cases (10.6%) with the longest length being 24 mm (Fig. 4); 27 cases showed IM in the submucosal layer, 19 cases in the muscularis propria, and 13 cases in both the submucosal layer and muscularis propria. Among the cases with IM, no difference of average IM length existed between the submucosal layer (average: 3.3 mm) and the muscularis propria (3.0 mm), although the mode of spread was slightly different (discontinuous spread such as lymphatic spread was the main mode of IM in the submucosal, whereas the direct spread type of IM was predominant in the muscularis propria).

FIGURE 4. The actual length of the spread of distal intramural cancer. Distal intramural spread (IM) was observed in 59 cases (10.6%) with the longest being 24 mm among 556 patients who underwent potentially curative surgery for advanced rectal cancer.

No differences in IM incidence were observed with regard to the location of the tumor’s anal edge or by the presence of distant metastases (Table 1). The parameters that significantly correlated with IM were as follows: maximum tumor diameter, annularity of tumor, gross tumor appearance, tumor depth of penetration, number of nodes involved, upward lymphatic spread (nodal involvement around superior mesenteric artery), predominant histologic type, degree of lymphatic invasion, and venous invasion (Table 1).

TABLE 1. Distal Intramural Spread (IM) and Other Clinicopathologic Parameters in Surgically Resected Specimens

Impact of Histology in the Submucosal Field at the Tumor’s Distal Edge on IM

Three histologic parameters in the actively invasive submucosal field at the tumor’s distal edge (ie, tumor budding, differentiation, and vascular invasion) were significantly correlated with IM, respectively (Table 2). Based on multivariate analysis, tumor budding and vascular invasion were the significant, independent parameters that had impact on IM, whereas tumor differentiation was a parameter that was almost independent of these 2 parameters (P = 0.08; Table 2). Based on these results, we concluded that it was acceptable to assess these 3 parameters together as IM risk parameters. The number of unfavorable histologic markers, that is, 1) positive tumor budding, 2) poor differentiation, and 3) the presence of vascular invasion, showed significant correlation with IM. Namely, IM-positive rates were 5.3% in tumors with no unfavorable histology, whereas they were 20.0% in those with one unfavorable histologic marker and 34.4% in those with multiple unfavorable histologic markers (Table 2). Likewise, the rate of IM of 10 mm or more increased in accordance with the number of unfavorable histologic markers. Based on these results, tumors with one or more unfavorable histologic marker were determined to indicate IM risk.

TABLE 2. Cancer Histology in Submucosal Layer at Distal Tumor Edge and Distal Intramural Spread (IM)

Selection of Preoperative IM Risk Factors

Considering the importance of preoperative evaluation, the following were chosen from parameters that related to IM as 3 possible parameters allowing a high degree of preoperative diagnostic accuracy, that is, gross appearance, annularity, and histology in the submucosal region at tumor edge. These 3 parameters were shown to have an independent impact on IM (Table 3). Regarding 1) type 3 appearance, 2) 3/4 or more annularity, and 3) unfavorable histology in the submucosal field as the risk factors for IM, the incidence of IM increased with number of risk factors: 3.3% in the no-risk group, 9.1% in the one-risk group, and 29.1% in the multiple-risks group (Table 3). There were no differences in these figures when patients were restricted to those without distant metastases (3.4% in the no-risk group; 9.6% in the one-risk group; and 28.6% in the multiple-risks group). Among the 211 tumors with no-risk factor, only one tumor, which was accompanied by the involvement of 13 nodes, had IM of 5 mm or more (17 mm of IM).

TABLE 3. Risk Factors of the Distal Intramural Spread (IM) and Actual Incidence of IM

Using the same method as the study of resected surgical specimens, the incidence and length of IM was examined based on 3 risk factors assessed preoperatively (ie, presence of unfavorable histology, 3/4 or more annularity, and type 3 gross appearance). IM was observed (length: 2 mm) in only one (3.7%) of 27 tumors that showed no-risk factors (Table 4). In contrast, tumors with IM were seen at 19.4% (7/36) in the one-risk factor group and in 22.2% (6/27) in the multiple-risk factors group. These 2 groups with risk factors include patients with tumors accompanied by IM at lengths of 5 mm or more.

TABLE 4. Number of Preoperative Risk Factors for Distal Intramural Cancer Spread (IM) and Actual Incidence of IM

Application of IM Risk Factors in the Selection of Patients With Anal Canal Involvement

Among the 80 cases that underwent abdominoperineal resection for rectal cancer above the anal canal, anal canal involvement was observed in 7 cases (all but one were the discontinuous spread type). No anal canal involvement was observed in 32 tumors without risk factors for IM (ie, unfavorable histology in submucosal region at tumor edge, 3/4 or more annularity, and type 3 gross appearance; Table 5). However, 3 tumors (9.1%) showed anal canal involvement in the one-risk factor group, and the rate reached 26.7% in the multiple-risk factors group. Although no tumor with extension below the dentate line was found, cancer deposits outside the internal sphincter muscle existed in 3 cases (Table 5).

TABLE 5. Risk Factors for Distal Intramural Cancer Spread and Anal Canal Involvement

DISCUSSION

Distal spread of rectal cancer can be divided into that with intramural spread (IM) and that spreading into the mesorectum. The latter pattern of distal cancer spread, which has been reported to be observed in around 20% of cases with advanced rectal cancer,¹⁷ was regarded recently as the major source of local recurrence after low anterior resection, and now, eradication of the cancer deposits in the distal mesorectum has become the basic principle of total mesorectal excision that has been shown to reduce the actual incidence of local recurrence.^18–22

However, intramural distal spread is a well-known mode of cancer growth that has been recognized for the last 10 decades.²³ Until only around 20 years ago, however, intramural distal spread was clinically irrelevant because most of low rectal cancer had been treated by abdominoperineal rectal resection according to the “five-centimeter rule,” whose necessity was argued particularly in the 1950s.^11,24 Even in the 1970s and 80s, “an evolutional era” in rectal surgery, when the technique of coloanal anastomosis was introduced,²⁵ circular stapling devices appeared in the surgical filed,^26,27 and the two-centimeter rule had become dominant,^12,28 distal intramural spread might still carry little importance in determining surgical procedure because intramural spread beyond 2 cm was rare in patients undergoing curative-intent surgery. However, it has often been recently argued that a distal margin of resection of 1 cm is sufficient for most cancers of the rectum because of the rarity of extensive intramural tumor spread^9,29 and the lack of survival difference between patients with and without 1 cm of resection margin.^7,8 During the last 10 years in particular, surgeons are beginning to avoid abdominoperineal resection, even in cases in which the distal margin length to be obtained is very short, for example, occasionally less than 1 cm.³⁰ Now, the time has arrived in which surgeons can not be unconcerned about the spread of intramural distal cancer, and as well, they are released from the restriction of the two-centimeter rule.

As one of the requirements for the success of attempt to perform fewer colostomies, criteria to identify patients who would suffer a relapse because of short distal margin are indispensable. Our present study supported the results of other studies, including our previous one based on 297 cases reported 10 years ago,³¹ showing that annularity¹¹ and gross appearance of tumor³² had a significant correlation with intramural distal spread. Using these 2 parameters together with another, the histology of the submucosal region, which could be evaluated preoperatively and be a useful parameter in predicting the extent of extramural spread,¹⁴ was thought to be promising criteria for the prediction of distal cancer spread, because of their independent association with intramural spread. Actually, by estimating intramural spread based on the 3 risk factors (ie, presence of unfavorable histology in submucosal region, 3/4 or more annularity, and type 3 gross appearance), patients who had little risk of intramural spread could be selected, that is, among 211 patients showing no risk factors, only 7 patients (3.3%) showed intramural spread. This group included only one patient with an intramural spread of 5 mm or more.

In investigating the criteria to determine the required distal margin, we attached importance to preoperative parameters that could be evaluated with certainty. Consequently, the parameters that can be diagnosed endoscopically with exactness (gross appearance of tumor and annularity) had been adopted as criteria, not nodal involvement and depth of penetration, both of which are not necessarily enough as to their accuracy of a preoperative diagnosis. It should be noted, however, that accurate information regarding nodal involvement and depth of penetration, which have been reported as the most important relevant factors of intramural cancer spread if assessed histologically,^9,33,34 must provide additional reliable information concerning the possibility of distal spread. It is noteworthy that the results of this study indicated that we can establish a category of patients in which none had an intramural spread of 5 mm or more, excluding patients in pTNM-N2 from the no-risk group, and furthermore, intramural spread was observed in only 1 (direct distal spread in the submucosa of just 2 mm) among 99 patients with tumors whose depth of penetration was defined pathologically within the muscularis propria.

One of the important findings of this study might be that the assessment of intramural distal spread could be used as the criteria to judge potential anal canal involvement. Although no tumor with distal spread beyond the dentate line was observed in patients examined in this study, there were patients with cancer foci outside the internal sphincter muscle that could remain in the operative field after intersphincteric resection, and all of these patients showed risk for intramural distal spread. On the other hand, no patients in the no-risk group showed anal canal involvement. Therefore, patients in the no-risk group could avoid abdominoperineal resection, whether or not resection of the upper part of the internal sphincter muscle was necessary so as to obtain a 1 cm margin.

Among the pathologic parameters of the primary tumor, tumor differentiation is the most frequently evaluated as the risk factor of intramural distal spread,^12,34 and Schiessel et al argued that poorly differentiated tumors should be excluded from intersphincteric resection.³⁵ However, endoscopic biopsies taken from a tumor show only very limited and superficial portions and do not necessarily represent the histology of the tumor as a whole,¹¹ and this parameter is not considered effective as a guide for determining an adequate margin.¹¹ Similarly, vascular invasion has also not been considered a predictive parameter because of the difficulty of its preoperative assessment, despite its strong correlation with intramural distal spread. We showed previously that by using the transanal submucosal biopsy technique, this valued parameter could be observed at a rate 6 times higher than that allowed by conventional endoscopic biopsy.¹⁴ In the present study, such association between vascular invasion and intramural distal spread was recognized even if vascular invasion was restricted to that observed in the submucosal region at the tumor’s edge.

The current trend mentioned above, that abdominoperineal resection is now less frequently performed for rectal cancer patients, might be confirmed by figures from our institute; the abdominoperineal resection rate decreased from 48% of patients in the first quarter period of this study (between 1981 and 1985) to 33% in the second quarter (1986–1990), to 19% in the third quarter (1991–1995), and finally to 10% in the last quarter (1996–1999), although rectal cancers whose macroscopic anal edges are located within 2 cm from the upper level of the anal canal have basically been treated by abdominoperineal resection. Based on the precise estimation of the required distal margin for individual tumors, which is essential to increase the sphincter-sparing procedure rate, we can prudently select cases of very low rectal cancer patients to whom the one-centimeter rule can be applied and, consequently, fewer colostomies in combination with intersphincteric resection would be performed. Further studies, including prospective evaluation of actual anastomotic recurrence, will be used to confirm or modify the parameters of the required distal margin selected in the present study.