Conservative Management of Rectal Cancer With Local Excision and Adjuvant Therapy

Introduction

The standard surgical treatment for cancer of the rectum is primarily abdominoperineal resection or low anterior resection. If the tumor is transmural or lymph node positive, adjuvant therapy with fluorouracil(Drug information on fluorouracil) (5-FU)-based chemotherapy and radiation is indicated.[1,2] Standard management of lesions closer to the dentate line often mandates use of abdominoperineal resection, due to the need for an adequate distal margin. Depending on the exact location, abdominosacral resection, stapled low anterior resection, or proctectomy with coloanal anastomosis may be feasible in selected patients.[3,4] These radical approaches can be associated with complications and functional consequences.

Postoperative Complications and Recurrence Patterns

In some series, postoperative complication rates range from 30% to 46%, and postoperative mortality from 2% to 6%.[5,6] Neurogenic bladder with inability to void is common after extensive pelvic dissection.[6] Sexual dysfunction is also common, particularly in men. Retrograde ejaculation, related to sympathetic nerve dysfunction, is seen in 15% to 100% of men. Erectile impotence resulting from damage to the pelvic parasympathetic plexus is seen in 3% to 39% of men.[7-9] Even after a low anterior resection, up to 50% of patients report decreased rectal sphincter function, and up to 30% report urinary retention or sexual dysfunction.[9,10] Finally, the psychological and social consequences of colostomy in patients who require it cannot be underestimated.

With standard surgery for early-stage rectal cancer (T1/2, N0), local recurrence rates range from 0% to 10%, and 5-year survival rates from 78% to 100%.[11-14] Sticca et al evaluated 71 patients (T1: 20, T2: 51) with stage I rectal adenocarcinoma who were treated with an abdominoperineal resection. At a median follow-up of 81 months, the local control rate was 100% for T1 tumors, and 90% for T2 tumors. Of the 7 recurrences, 2 were local, 4 were distant, and 1 was local and distant. For patients with T2 disease, the 5-year disease-free survival rate was 88%.[13]

Willett et al reported the recurrence patterns of stage I rectal cancer treated by abdominoperineal resection. The local control and disease-free survival rates at 6 years were both 100% for 12 patients with T1 disease. For the 52 patients with T2 disease, the local control and actuarial disease-free survival rates were 84% and 80%, respectively.[14] These numbers represent the standard to which any proposed conservative treatment must be compared. However, functional outcome with conservative treatment approaches must be evaluated, because a poorly functioning sphincter may have a more detrimental effect on quality of life than a well-functioning colostomy.

Surgical Techniques

Local options for surgical management include transanal local excision, transsacral excision (Kraske approach), transsphincteric resection (Bevan or York-Mason approach), transanal endoscopic microsurgery, and electrocoagulation. Transanal local excision should be performed as a full-thickness excision where the perirectal fat serves as the deep plane of dissection.

The Kraske excision is used for lesions too large or proximal for transanal local excision, and allows for removal of some perirectal lymph nodes. A perineal incision is made just above the anus, the coccyx is removed, and the fascia is divided. The rectum is mobilized through this incision, and a wide local excision or sleeve resection can be performed.

The transsphincteric procedure is identical to the Kraske procedure, except that the anal sphincter is divided posteriorly in the midline. The anus is reconstructed at completion of the operation, with little risk of functional impairment.[15]

In transanal endoscopic microsurgery, an operating rectoscope is used to perform a full-thickness disc excision of the rectum, with primary closure. Fulguration is completed in multiple steps with the patient under general or regional anesthesia, and involves charring of the tumor, then scraping with a curette. There is a 10% to 20% risk of delayed hemorrhage due to sloughing of the scar at 7 to 10 days.[16-21]

Full-thickness local excision is preferred over destructive approaches because the complete specimen is available for assessment of depth of invasion, margins, and pathologic features known to be prognostic for lymph node involvement.

Selection Criteria

Local treatment has historically been applied to patients with medical contraindications to radical surgery—for example, severe cardiopulmonary disease or patient blindness (which makes colostomy care difficult). More recently, the use of local treatment has been explored as an alternative to abdominoperineal resection. Studies suggest that small (< 3 cm), exophytic, well-differentiated tumors limited to the submucosa warrant this approach.

Local treatment approaches have also been used for selected tumors invading the muscularis propria. Tumors that would be suitable for a local approach should be chosen carefully, based on clinical, pathologic, and radiologic factors.

Clinical Criteria

An evaluation of clinical criteria in the series examining local excision is difficult due to variations in patient selection, the definition of clinical and pathologic features, and length of follow-up. Moreover, the series examining local excision alone were limited by the use of univariate analysis, rather than multivariate analysis. It is, therefore, difficult to make recommendations based solely on clinical criteria. However, tumors that are small, mobile, and involve less than 40% of the circumference of the rectal wall could most likely be excised with negative margins by a full-thickness, nonfragmented excision.

Pathologic Criteria

The impact of various pathologic factors on local recurrence and the overall outcome of rectal cancer has been studied extensively. Regional lymph node metastases that are unrecognized at the time of local excision are thought to account for the majority of local failures after conservative treatment. Factors that are known to have an impact on the incidence of lymph node involvement include the presence of high-grade features, lymphatic or blood vessel invasion, colloid histology, or penetration of tumor through the bowel wall.

Tumors without any of these features comprise only 3% to 5% of all rectal cancers. These select cancers have a low enough incidence of local failure (5% to 10%) and positive nodes (< 10%) so as not to require adjuvant therapy.[20] However, once adverse pathologic features manifest or the tumor invades the muscularis propria, the local failure rate approaches 20% and the incidence of positive pelvic or mesorectal lymph nodes is 10% to 15%.[22]

Tumor Grade/Differentiation: Minsky et al retrospectively studied various clinical and pathologic features in 168 patients who underwent a potentially curative surgery for rectosigmoid or rectal cancer at the New England Deaconess Hospital. They found that the incidence of positive lymph nodes increased with the grade of adenocarcinoma. None of the well-differentiated tumors had lymph node involvement, while 30% of moderately differentiated, and 50% of poorly differentiated tumors had lymph node involvement (P = .07).[23]

Brodsky et al also found that grade significantly affected the incidence of lymph node metastases in 154 patients with pT1 or pT2 rectal cancer treated by radical resection. The risk of lymph node metastases was 0% for well-differentiated tumors, but rose to 24% for moderately and poorly differentiated tumors (P = .04).[24]

Lymphatic/Blood Vessel Invasion: It is established that the incidence of lymph node metastasis increases significantly with lymphatic or blood vessel invasion. Investigators from Memorial Sloan-Kettering Cancer Center analyzed 154 patients with pT1 or pT2 tumors and found that 17% of T1/T2 tumors without lymphatic or blood vessel invasion had lymph node metastasis, while 31% of T1/T2 tumors with lymphatic or blood vessel invasion had lymph node metastases (P = .04). Furthermore, within the group of patients with T1 tumors, 0 of 15 without vascular invasion had lymph node metastases, while 3 of 9 patients with vascular invasion had lymph node metastases (P < .05).[24]

Margins: Adam et al evaluated 141 specimens from resections thought by the surgeon to be curative. The closest point of the tumor to the circumferential margins was measured microscopically, and any specimen with tumor ≤ 1 mm from the circumferential margin of excision (25% of specimens) was recorded as having positive margins. At a median follow-up of 5 years, the authors found that local recurrence occurred in 78% of patients with involvement of the circumferential margin, compared with 10% of those without such involvement (P < .001).[25]

T Stage: Increasing T stage also significantly increases the risk of lymph node metastasis. A study from Basingstoke District Hospital evaluated 454 rectal excision specimens. Of 109 patients, 22 (20%) with tumor confined locally to the bowel wall had metastases in local lymph nodes. Among 27 patients (11%) with tumors that did not penetrate the submucosa, 3 (11%) had lymph node metastases (P = .28).[26] Minsky et al also found that increasing T stage was associated with increasing risk of lymph node metastases. In their series, lymph node involvement was seen after radical surgery in 0% of T1 tumors, 28% of T2 tumors, 36% of T3 tumors, and 53% of T4 tumors (P = .04).[23]

Use of these pathologic criteria in selecting "early" carcinomas of the rectum has been criticized. Nelson et al retrospectively reviewed 76 cases of adenocarcinoma of the lower rectum treated by abdominoperineal resection. They found that 50% of patients with tumors ≤ 4.0 cm had lymphatic invasion. Overall, 53% of patients who had well- or moderately differentiated tumors had lymphatic metastases, and 25% of those with only partial invasion of the muscular layer had lymphatic metastases. The authors concluded that the commonly used criteria of size and histologic grade are inadequate and do not allow confident selection of patients with lesions suitable for local treatment.[27]

Recent data from Memorial Sloan-Kettering Cancer Center suggest that the overall risk of lymph node metastasis in patients selected on the basis of small (£ 4.0 cm) tumor size and early T stage (T1/T2) is 15% (T1:10%, T2:17%). On subset analysis, while T stage, degree of differentiation, and lymphatic vessel invasion influenced the risk of lymph node metastasis, only blood vessel invasion reached statistical significance as a single predictive factor of lymph node metastases (blood vessel invasion-negative = 13%; blood vessel invasion-positive = 33%; P = .04). Even in patients in the most favorable group (T1 cancers with no adverse pathologic features), 7% had lymph node metastases.[28]

Other studies have confirmed that T1 and T2 tumors with unfavorable histologic features have decreased rates of survival and local control, whether treated by local excision or abdominoperineal resection. Willett et al found that tumor size greater than 3 cm, high-grade histology, invasion of the muscularis propria, vascular invasion, and positive margins were associated with a local failure rate of at least 20%, as well as an increase in distant metastases.[29,30] In summary, once adverse pathologic factors are present (high grade, lymphatic or blood vessel invasion, colloid histology, signet ring cell or tumoral invasion of the muscularis propria), local therapy alone is inadequate.