Extent of Mesorectal Tumor Invasion as a Prognostic Factor After Curative Surgery for T3 Rectal Cancer Patients

Abstract

Objective:

To determine the significance of the extent of mesorectal tumor invasion as a prognostic factor for T3 rectal cancer patients.

Summary Background Data:

There is controversy as to which primary lesion characteristics, other than regional lymph node involvement, in T3 rectal cancer are reliable prognostic factors.

Patients and Methods:

The extent of mesorectal tumor invasion was evaluated using 2 data sets comprising 196 and 247 patients undergoing curative surgery at separate institutes. When the outer aspect of the muscular layer was not identifiable, an estimate was obtained by drawing a straight line between the 2 break points of the muscular layer.

Results:

We selected 6 mm as the optimal value for subclassification of T3 rectal patients into 2 groups, based on the extent of mesorectal invasion, using the first data set. The overall 5-year survival rate was significantly higher in patients with <6 mm than in those with ≥6 mm of mesorectal invasion (72% versus 50%; P< 0.01). Similarly, in the second data set, the overall 5-year survival rates of patients with mesorectal invasion <6 mm and ≥6 mm were 59% and 37%, respectively (P < 0.01). In both data sets, multivariate analyses verified the extent of mesorectal invasion to be an independent prognostic factor, together with nodal involvement. Regarding positive nodal involvement and mesorectal invasion ≥6 mm as risk factors, the overall 5-year survival rates with none, one, and both of these factors were 84%, 61%, and 38%, respectively, in the first data set (P < 0.01). Prognostic results were similar for the second data set.

Conclusion:

Extent of mesorectal invasion, based on a 6-mm cutoff value, is useful for subclassification of T3 rectal cancer patients.

The extent of mesorectal tumor invasion was evaluated using 2 data sets comprising 196 and 247 patients at separate institutes to determine the significance of it as a prognostic factor for T3 rectal cancer patients. The extent of mesorectal invasion, based on a 6-mm cutoff value, is useful for survival subclassification of T3 rectal cancer patients.

Recurrence of rectal cancer is very difficult to control. Thirty to fifty percent of patients who undergo potentially curative surgery alone ultimately experience recurrence in Western countries,^1–3 and one third of those receiving curative resections will die of the disease.⁴ The most important factor impacting the survival of rectal cancer patients is tumor stage, determined by the depth of penetration through the bowel wall, and nodal involvement.⁵

Because of high failure rates, regardless of whether nodal involvement is present, postoperative adjuvant therapy has often been used to prevent local recurrence and improve survival rates for patients with transmurally invasive rectal cancers.^6–8 For example, in 1990, the National Institutes of Health Consensus Statement⁹ recommended that patients with resected stage II and III rectal cancers receive adjuvant treatments such as chemotherapy and/or irradiation. However, some investigators hold the view that postoperative adjuvant therapy should be based not only on stage, but also other clinicopathologic factors and that adjuvant therapy may not be necessary for a group of T3 rectal cancer patients at very low risk of recurrence.¹⁰

The extent of mesorectal tumor invasion has been shown to be an independent risk factor for recurrence in some studies.^10–12 However, the TNM staging system¹³ does not take into account of the extent of mesorectal tumor invasion, so long as there is no invasion of adjacent organs. Thus, the T3 tumor stage covers a spectrum encompassing minimal invasion beyond the muscularis propria (MP) to gross invasion into the mesorectum that stops short of invading an adjacent organ. Tumors at opposite ends of the T3 spectrum may have very different prognostic implications.

This study was designed to analyze factors impacting the survival of patients with T3 rectal cancer (stages II and III), and in particular to identify indicators for adjuvant therapy. The extent of mesorectal tumor invasion was a major focus of the present study.

MATERIALS AND METHODS

Patients

A total of 196 patients with primary rectal cancer located below the peritoneal reflection who underwent curative surgery at the Department of Surgery I, National Defense Medical College, Japan, between 1980 and 1997 constituted the first data set. No patients had evidence of distant metastasis, and all patients were confirmed to have undergone curative surgical resection, without circumferential resection margin involvement, based on histologic examination. We confined our analysis to 196 patients with T3 rectal cancer as determined by pathologic findings of surgically resected specimens, who had not received preoperative neoadjuvant therapy. In this period, there were few patients underwent systematic intravenous postoperative adjuvant therapy at the National Defense Medical College. Overall, 87% of our cases were given only oral anticancer drugs such as 5-fluorouracil, 5′-DFUR, and HCFU. The patients consisted of 134 males and 62 females, with an average age 59.5 years at the time of surgery (range, 25–85 years). Mean follow-up for all patients was 77 months (range, 1–231 months).

A total of 247 patients with primary rectal cancer located in the lower two thirds of the rectum who underwent complete surgical resection of their tumors between 1960 and 1969 at St. Mark's Hospital comprised the second data set. These tumors were confirmed to be T3 and did not have circumferential resection margin involvement by pathologic findings of surgically resected specimens. None of these patients had synchronous cancers or cancers complicating familial adenomatous polyposis or inflammatory bowel disease. There were 164 males and 83 females with an average age at the time of surgery of 61.4 years (range, 33–89 years). All patients selected were followed up for at least 5 years (average, 140 months; range, 61–263 months) or until death. It was exceptionally rare for adjuvant therapy to be used at St. Mark's in the 1960s.

For these tumors, morphologic features of known prognostic importance such as tumor diameter, tumor differentiation, and nodal involvement had previously been recorded in pathology reports.

Extent of Mesorectal Tumor Invasion

All surgically resected specimens were opened in the operating theater by cutting along the antimesenteric border, carefully avoiding the tumor unless growth was circumferential. The specimens were pinned to a cork board and fixed in 20% formalin. One or more large longitudinal sections of the whole tumor were obtained at the point of maximum extension through the bowel wall, as judged macroscopically. The slice containing the deepest invasion was selected for division into blocks, which were then embedded and routinely processed for hematoxylin and eosin staining. In these sections, the extent of mesorectal tumor invasion, ie, the distance between the outer border of the MP and the outermost part of the tumor, was measured when the outer aspect of the muscular layer was identifiable (type A tumors).^10,12 When the outer aspect of the muscular layer was not identifiable because of destruction by the tumor or extensive inflammation (type B tumors), an estimate was obtained by drawing a straight line between the 2 break points of the muscular layer (Fig. 1). Hematoxylin and eosin-stained sections, containing the deepest tumor extension, in rectal tumor cases with and without a discernible outer muscular layer, are presented in Figure 1C and D.

FIGURE 1. Method of measuring the depth of mesorectal tumor invasion. When the outer aspect of the muscular layer was identifiable, the distance between the outer border of the MP and the outermost portion of the tumor was measured (A, C). When the outer aspect of the muscular layer was not identifiable, an estimate was made by drawing a straight line between the 2 break points of the muscular layer (B, D).

Statistical Analysis

Statistical analyses were performed using StatView 4.11 software (Abacus Concepts, Berkeley, CA). Survival curves were drawn by the Kaplan-Meier method and were compared using the log-rank test. Overall survival was used for the survival analysis. Multivariate analyses to identify independent prognostic factors were performed using the Cox stepwise regression model.

RESULTS

Distribution of Extent of Mesorectal Invasion (First Data Set)

Tumor numbers were determined in 51 and 145 patients with type A and type B tumors, as shown in Fig. 1, respectively, for the first data set.

The distribution of the extent of mesorectal invasion (in millimeters) in 196 patients is shown in Fig. 2. The extent of mesorectal invasion in 55 patients (30% of all patients) ranged from 3 mm to 5 mm. The extent of mesorectal invasion was within 7 mm in 148 patients (75% of all patients), and the mean value for all patients was 5.2 mm (range, 0.1–22.1 mm).

FIGURE 2. Distributions of the extent of mesorectal invasion.

Cutoff Value of Extent of Mesorectal Invasion (First Data Set)

For subclassification of T3 rectal cancer patients into 2 groups based on the extent of mesorectal tumor invasion, we attempted to identify the most clinically relevant cutoff value for extent of mesorectal tumor invasion by focusing on hazard ratios for the first data set. Among the 9 cutoff values (2, 3, 4, 5, 6, 7, 8, 9, 10 mm) examined, 6 mm was identified as the optimal value based on having the smallest P value for different outcomes (Table 1). The 5-year survival rate of patients with mesorectal invasion ≥6 mm was 50% and that of patients with mesorectal invasion <6 mm was 71% (P < 0.01)(Fig. 3).

TABLE 1. Selection of the Optimal Cutoff Value for Extent of Mesorectal Invasion (EMI) Using the First Data Set

FIGURE 3. Overall survival curves by the extent of mesorectal invasion (EMI).

For both type A and B tumors, patients with mesorectal invasion ≥6 mm had significantly poorer survival than those with mesorectal invasion <6 mm. In 51 type A tumor patients, the overall 5-year survival rate was 36% for those with mesorectal invasion ≥6 mm and 74% for those with mesorectal invasion <6 mm (P = 0.0066). In 145 type B tumor patients, the overall 5-year survival rate was 53% for those with mesorectal invasion ≥6 mm and 72% for those with mesorectal invasion <6 mm (P = 0.0332).

Distribution of Extent of Mesorectal Invasion (Second Data Set)

The distribution of the extent of mesorectal invasion (in millimeters) for the second data set is shown in Fig. 2. The distribution is similar to that of the first data set. The mean value for extent of mesorectal invasion for data set 2 was 6.1 mm (range, 0.1–50.0 mm).

Cutoff Value for Extent of Mesorectal Invasion (Second Data Set)

When the 6-mm cutoff value was tested on the second data set, patients with mesorectal invasion ≥6 mm had significantly poorer survival than those with mesorectal invasion <6 mm (37% versus 59%, 5-year survival; P < 0.01) (Fig. 3).

Prognostic Significance of Extent of Mesorectal Invasion According to TNM Stage (Both Data Sets)

The prognostic significance of the extent of mesorectal invasion according to the TNM stage is shown in Figure 4. Patients with stage II tumors, in the second data set, could be categorized into 2 groups with different prognoses by the extent of mesorectal invasion (52% versus 73%, 5-year survival; P < 0.01). Stage III tumor patients could be categorized into 2 groups in both data sets (first data set: 40% versus 58%, 5-year survival; P < 0.05; second data set: 27% versus 40%, 5-year survival; P < 0.05).

FIGURE 4. Prognostic significance of extent of mesorectal invasion (EMI) by TNM stage.

Multivariate Prognostic Analyses (Both Data Sets)

The results of univariate and multivariate prognostic analyses using parameters including extent of mesorectal invasion, tumor size, tumor differentiation, and nodal involvement are presented as Table 2. In both data sets, nodal involvement and extent of mesorectal invasion were identified as independent prognostic factors.

TABLE 2. Univariate and Multivariate Prognostic Analyses

Nodal involvement and extent of mesorectal invasion had almost same impact for poor prognosis in second data set because each hazard ratio and P value were almost same (hazard ratio, 1.8, 1.9; P value: 0.0019, <.0001). Nodal involvement had stronger impact than extent of mesorectal invasion for poor prognosis in first data set (hazard ratio, 2.4, 1.8; P value: 0.0004, 0.0066).

Subclassification for Survival by Number of Risk Factors (Both Data Sets)

Taking nodal involvement and mesorectal invasion ≥6 mm as risk factors, the respective overall 5-year survival rates with none, one, and both of these factors were 84%, 61%, and 38% in the first data set (P < 0.01). The model was tested on the second data set and similar prognostic results were obtained (73%; 45%; 27% 5-year survival; P < 0.01).

DISCUSSION

Prognoses vary markedly in both stage II and stage III rectal cancer patients. To individualize the selection of optimal adjuvant therapy, subclassification of stages II and III rectal cancer patients based on other reliable prognostic factors may be needed.^14–23 The extent of mesorectal tumor invasion has been shown to be an independent risk factor in some studies.^10–12 However, in those studies, different cutoff values (2 mm or 4 mm) for the extent of mesorectal tumor invasion were used without detailed explanations as to how these cutoff values were determined.^10,12 Therefore, we attempted to identify the most clinically relevant cutoff value using our first data set. We examined 9 cutoff values (2, 3, 4, 5, 6, 7, 8, 9, and 10 mm), calculating a hazard ratio for each using the Cox proportional hazards model. Based on the results of this analysis, we concluded that 6 mm is the most useful cutoff value because the hazard ratio of 2.1 represented a sharp increase from the hazard ratio of 1.5 at 5 mm and the P value at 6 mm (0.0011) was the most significant of the 9 values examined.

Cawthorn et al used a 4-mm cutoff value in assessing the distribution of the extent of mesorectal spread in 167 patients.¹² However, they did not use a statistical analysis to determine this cutoff value, and their study included 50 T2 patients (absence of mesorectal invasion). Willett et al characterized tumors as showing minimum (through the MP with <2 mm invasion into perirectal fat), moderate (2–8 mm into perirectal fat), or extensive (≥8 mm into perirectal fat) invasion.¹⁰ None of these earlier studies presented statistical analyses of their data. Steel et al divided tumors into minimally invasive T3 (defined as invasion evident on microscopic examination only) and advanced T3 (defined as invasion discernible on macroscopic examination of the transected tumor).¹¹ They did not measure actual values and could not establish a cutoff value. Although we determined the best cutoff value to be 6 mm for the first data set of 196 patients, this value also fit the second data set of 247 patients well.

Cawthorn et al demonstrated poorer outcomes (overall 5-year survival rate) with extensive mesorectal invasion (25%) as compared with slight invasion (55%).¹² Willett et al, in a study of 117 patients, found a significant difference in local recurrence rates between patients with mesorectal invasion of ≥2 mm and those in whom invasion was less than 2 mm.¹⁰ Steel et al demonstrated the depth of invasion into the mesorectum (macroscopic versus microscopic) to apparently be an independent prognostic factor for local recurrence in a cohort of 222 patients undergoing surgical resection.¹¹ These 3 studies are summarized in Table 3.

TABLE 3. Studies on Prognostic Significance of Mesorectal Invasion in Rectal Cancer Patients

In 2 of the 3 studies, the extent of mesorectal invasion was shown to be a significant risk factor for local recurrence while having no impact on overall survival. In our first data set, the local recurrence rate is significantly higher for those with marked as compared with slight tumor invasion when the cutoff value is 7 mm or 8 mm. However, when the cutoff value is 6 mm, the value that is most significant statistically in terms of survival, the local recurrence rate does not differ markedly according the extent of mesorectal invasion. According to the recurrence-free survival curve in the first data set, patients with mesorectal invasion ≥6 mm have significantly lower recurrence-free survival rate than those with mesorectal invasion <6 mm (5-year recurrence-free survival rate: 55% versus 69%; P < 0.05). So, total recurrence rates significantly differ because of the extent of mesorectal invasion. Many previous studies found that the circumferential resection margin is very important for prediction of local recurrence.^24–29 The extent of mesorectal invasion appears to be more meaningful for survival than for local recurrence, however.

Only Cawthorn et al¹² found the extent of mesorectal invasion to be an independent factor predicting the overall survival rate by multivariate analysis. However, their study included 50 T2 rectal cancer patients and 45 patients who underwent palliative surgery. Thus, using 2 data sets comprising 196 and 247 patients, the present study more clearly shows the extent of mesorectal tumor invasion to be an independent factor predicting overall survival of T3 rectal cancer patients after curative surgery. Indeed, for both stage II and stage III tumors, patients in the second data set could be categorized into 2 groups with different prognoses based on the extent of mesorectal invasion. We examined the extent of mesorectal tumor invasion (cutoff value = 6 mm), tumor size, tumor differentiation, and nodal involvement in multivariate analysis using the Cox stepwise regression model. Extent of mesorectal invasion ≥6 mm and nodal involvement were found to be independent prognostic factors in T3 rectal cancer patients. Patients with T3 rectal cancer could clearly be categorized into 3 groups based on these risk factors (overall 5-year survival rates with none, one, and both risk factors: 84%, 61%, and 38%, respectively; P < 0.01 for the first data set, 73%, 45%, and 27%, respectively; P < 0.01 for the second data set) (Fig. 5).

FIGURE 5. Overall survival curves according to the number of risk factors: lymph node involvement, extent of mesorectal invasion ≥6 mm.

CONCLUSION

The extent of mesorectal tumor invasion, in addition to nodal involvement, is useful for subclassifying T3 rectal cancer patients. These 2 factors were statistically independent of each other. Adjuvant therapies may not be necessary for T3 rectal cancer patients with none of these risk factors. Moreover, determination of criteria for neoadjuvant therapy may be possible by the use of ultrasound in current day staging to determine depth on invasion.