Preoperative staging of rectal cancer by endoluminal ultrasoundvs. magnetic resonance imaging

PURPOSE: The aim of this study was to compare the value of endoluminal ultrasonography (ELUS) with magnetic resonance imaging (MRI) for preoperative staging of rectal carcinoma. METHODS: Thirty-seven consecutive patients were examined by ELUS and MRI. Imaging results were compared with pathohistologic studies. A tumor extending beyond the bowel wall was considered to be positive and one within the bowel wall was considered negative. Lymph node involvement was considered present if nodes equal to or greater than 5 mm in diameter were found in the perirectal tissue. For evaluating the differences between the two methods, the Mc Nemar test was performed. RESULTS: T-Staging was correct in 88.2 percent (30/34) of patients by ELUS and in 82.3 percent (28/34) by MRI (difference not significant). N-Staging was correct in 80 percent (20/25) by ELUS and in 60 percent (15/25) by MRI (difference of borderline significance). A comprehensive preoperative staging (T + N) was made correctly in 68 percent (17/25) by ELUS and in 48 percent only (12/25) by MRI (difference not significant). CONCLUSIONS: We suggest that ELUS and MRI must be evaluated within the framework of established parameters when treatment modalities such as preoperative radiation therapy and local or radical surgical approach must be decided.     

Predicting lymph node metastases in rectal cancer

For properly selected rectal cancers, local excision is a sphincter-saving alternative to abdominoperineal resection. If histologic assessment of a locally excised tumor reveals ominous features, further treatment with radical resection or irradiation may be necessary to treat potential lymph node metastases. PURPOSE: We wished to determine which features, if any, were predictors of nodal metastases. METHODS: Nine histologic and morphologic features of 62 radically excised rectal cancers were reviewed to determine which factors, if any, were associated with nodal disease. RESULTS: Using a chi-squared analysis, we found worsening differentiation (P=0.0001), increasing depth of penetration (P=0.026), a microtubular configuration of 20 percent or more (P=0.023), and the presence of venous (P=0.001) or perineural invasion (P=0.002) to significantly influence nodal disease. Lymphatic invasion was witnessed too infrequently to determine significance but, when present, was associated with nodal metastases in every case. Exophytic tumor morphology, mitotic count, and tumor size were not significant predictors. An analysis of variables determined that, of all factors or combination of factors examined, Broder's classification was the strongest predictor of nodal disease. CONCLUSIONS: If a rectal cancer is accessible and of small size to facilitate local excision, an in-depth histologic assessment is needed to determine if nodal metastases are likely on a statistical basis.This work was supported by the Bowman Research Fund.     

Prognostic value of positive lymph nodes in rectal cancer

Abdominal curative resections for rectal cancer in 109 patients with positive lymph nodes were prospectively studied. The best subdivision of patients for predicting outcome was into 1–3 and >3 positive lymph node groups. Comparison with patients with >3 positive lymph nodes demonstrated that patients with 1–3 positive nodes had less local (35.0 percent vs. 13.0 percent;P =0.007) and less distant recurrences (45.0 percent vs.26.0 percent;P =0.04) and had much better crude five-year survival (58.2 percent vs.17.0 percent; P < 0.0001). For predicting postsurgical outcome in patients with positive lymph nodes, the results justify subdividing patients into the following two prognostic subgroups: 1) those with 1–3 involved lymph nodes and 2) those with metastatic tumor in four or more lymph nodes.     

Evaluation and biopsy of recurrent rectal cancer using three-dimensional endosonography

PURPOSE: The value of endorectal ultrasonography for postoperative follow-up of rectal cancer is limited by the inability to distinguish recurrent malignancy from benign lesions,e.g., fibrotic tissue. This study was conducted to investigate the role of three-dimensional (3D) endosonography for evaluation and biopsy of recurrent rectal cancer. METHODS: Endorectal ultrasonography was performed in routine follow-up program after resection of rectal cancer. 3D volume scans were recorded using a bifocal multiplane 3D transducer (7.5/10 MHz) with a 100 longitudinal and a 360 transversal scan angle. For transrectal ultrasound-guided biopsy of pararectal lesions, a specially designed targeting device was attached to the endoprobe. RESULTS: Overall pararectal lesions were detected in 28 of 163 patients (17 percent) who were undergoing endorectal ultrasonography for follow-up after resection of rectal cancer. 3D image analysis facilitated assessment of suspicious pararectal lesions by contemporary display of three perpendicular scan planes or volume reconstructions of the scanned area. Ultrasound-guided biopsy was performed in all 28 patients with pararectal lesions identified by endorectal ultrasonography. Biopsy revealed recurrent disease or lymph node metastases in seven and two patients, respectively. Benign lesions explained the endosonographic findings in 17 patients. All patients with benign histology still have no evidence of recurrent disease after a median follow-up of seven months. Nonrepresentative material was obtained in only 2 of 28 patients (accuracy, 93 percent). Histology changed the endosonographic diagnosis in 28 percent of cases. CONCLUSIONS: 3D endosonography with ultrasound-guided biopsy improves the diagnosis of extramural recurrence after curative resection of rectal cancer. 3D image display allows precise control of the position of the biopsy needle within the target.     

Endosonography of pararectal lymph nodes

One hundred thirteen patients with carcinoma of the rectum were evaluated for lymph node metastases by endorectal ultrasound. With the use of 7.5 MHz and based on different echo patterns, two main groups of lymph nodes can be differentiated: hypoechoic and hyperechoic lymph nodes. Compared with pathologic findings, hypoechoic lymph nodes represent metastases, whereas hyperechoic lymph nodes are visualized due to unspecific inflammation. Lymph node metastases can be predicted with a sensitivity of 72 percent and inflammatory lymph nodes with a specificity of 83 percent. The physical basis of the differentiation of lymph nodes was assessed in vitro by the determination of ultrasound parameters (speed of sound, acoustic impedance, attenuation, and backscattered amplitude). The attenuation coefficient of benign lymph nodes [2.5 dB/(MHz×cm)] is significantly higher than the mean value of lymph node metastases [1.3 db/(MHz×cm)]. The results demonstrate that involved nodes can principally be differentiated from not involved nodes. Micrometastases, mixed lymph nodes, and changing echo patterns within inflammatory nodes explain the accuracy rate of 78 percent.Supported by a grant from the Deutsche Forschungsgemeinschaft Hi 385/1-1     

Endorectal ultrasonography for the assessment of wall invasion and lymph node metastasis in rectal cancer

Endorectal ultrasonography (ERUS) with a flexible-type radial scanner (Aloka Co. Ltd., Tokyo, Japan; 7.5 MHz) was applied to 120 patients with rectal cancer for the assessment of wall invasion and pararectal lymph node metastasis. Normal rectal wall was described as a five- or seven-layer structure excluding the lowest part within 3 cm from the anal verge. Loss of normal layers basically indicated the existence of cancer invasion. According to UICC classification, we divided the depth of wall invasion into four ultrasonographic levels (uT1–uT4), and results were correlated with histopathologic findings. Overall accuracy of the assessment was 92.0 percent (103/112). Overestimation occurred in 5 of 60 cases with T3 cancer (8.3 percent), and underestimation occurred in 1 of 19 cases with T2 cancer (5.3 percent) and 3 of 60 cases with T3 cancer (5 percent). Inflammatory cell infiltration was found around the cancer in a considerable number of cases. However, the assessment of wall invasion was hardly affected in our hands. Because the muscularis propria of the rectal wall was often recognized as a three-layer structure, uT2 cancer was subdivided into three subgroups of uPM1, uPM2, and uPM3. The assessment of invasion of sublayers in muscularis propria was possible in 14 of 19 cases (73.7 percent), and correct assessment was achieved in 57 percent of the cases. The ultrasonographic demonstration of pararectal lymph nodes was studied on 98 patients. No swollen lymph nodes were detected ultrasonographically in 35 of 98 cases (35.7 percent), but cancer metastasis was found histopathologically in 5 of these 35 cases (14.3 percent). The metastasis was observed more frequently in lymph nodes with a diameter of more than 5 mm (53.8 percent) and in those with a well-defined boundary and with an uneven and markedly hypoechoic pattern (72.3 percent). Although unable to detect minimal cancer foci, ERUS was considered a very useful tool for the assessment of the depth of cancer invasion in the rectal wall and pararectal lymph node metastasis.     

Incidence of metastases from rectal adenocarcinoma in small lymph nodes detected by a clearing technique

The incidence of metastases from primary adenocarcinoma of the rectum in lymph nodes smaller than 5 mm is not known. Lymph nodes measuring 5 mm usually are not detected by manual techniques of examination of the surgical specimen. This retrospective analysis describes the results when a lymph node clearing technique that identifies lymph nodes as small as 1 mm was used to treat surgical specimens from 27 consecutive patients with rectal adenocarcinoma who underwent abdominoperineal resection with a curative intent and for whom all pathologic data were retrievable. Nine hundred thirty lymph nodes were found, with an average of 34 lymph nodes per specimen (range 0–88). Seventy-two of the 345 lymph nodes found in patients with Dukes C tumors were found to have metastases. Fifty-six (78 percent) of these 72 lymph node metastases occurred in lymph nodes measuring 5 mm. Three lymph node metastases were found in the perianal zone, 53 in the perirectal zone, and 16 in the pericolonic zone. Lymph node metastases from rectal adenocarcinomas often will occur in lymph nodes smaller than 5 mm. We concluded that the use of lymph node clearing techniques discovers these metastases, thereby offering the potential for enhanced staging of primary rectal adenocarcinomas.Read at the XIIIth Biennial Congress of the International Society of University Colon and Rectal Surgeons, Graz, Austria, June 24 to 28, 1990.     

Prognostic profile in rectal cancer

PURPOSE: The purpose of this article was to review the prognostic factors of significance in rectal cancer. METHODS: This is a retrospective review of various reports and an examination of our data. RESULTS: Current imaging techniques with endorectal ultrasound, magnetic resonance imaging with a rectal probe, and possibly tagged monoclonal antibodies allow for presurgical assessment of the invasion of the rectal wall and detection of lymph node involvement. CONCLUSIONS: Tumor penetration of the rectal wall and lymph node metastases are the most important prognostic indicators in rectal cancer and predict local and distant recurrences. When lymph node metastases occur, they are more common in small lymph nodes (<5 mm). Patients with lymph node metastases always need to be considered for multimodality therapy. Tumors larger than or equal to T3 or with associated lymph node metastases should not be treated with sphincter-saving surgical procedures alone.     

Colon and rectal cancer in pregnancy

Colorectal carcinoma presenting in pregnancy is an uncommon disease that is reported to be associated with an extremely poor prognosis. To better characterize this disease, we surveyed the membership of the American Society of Colon and Rectal Surgeons by mailed questionnaire and reviewed the literature. Forty-one new cases of women with large bowel cancer who presented during pregnancy or the immediate postpartum period were identified. The mean age at presentation was 31 years (range, 16–41 years). Tumor distribution was as follows: right colon-3, transverse colon-2, left colon-2, sigmoid colon-8, and rectum-26. Dukes stage at presentation was A=0, B=16, C=17, and D=6 (two patients were unstaged). Average follow-up was 41 months. Stage for stage, survival was found to be similar to patients with colorectal tumors in the general population. Large bowel cancer coexistent with pregnancy presents in a distal distribution (64 percent of tumors in the current series and 86 percent of those reported in the literature were located in the rectum) and presents at an advanced stage (60 percent were Stage C or more advanced at the time of diagnosis). While patient survival is poor, it is no different stage for stage from the general population with colorectal tumors.Read at the meeting of The American Society of Colon and Rectal Surgeons, Boston, Massachusetts, May 12 to 17, 1991.     

DNA index as a significant indicator of lymph node metastasis and local recurrence of rectal cancer

To confirm the prognostic significance of the DNA index (DI) in cases of rectal cancer, the nuclear DNA content of tumor cells was examined in 184 cases of rectal cancer treated with curative surgery, and the incidence of lymph node metastasis and recurrence of the cancer was analyzed. The incidence of lymph node metastasis was 43.9 percent in cases with aneuploidy (DI above 1.5), being statistically different from the 18.0 percent incidence in cases with diploidy (P <0.001). Although the extent of lymph node metastasis was limited to adjacent lymph nodes in cases with diploidy, distant lymph node metastases were frequent in cases with aneuploidy, especially in those with a DI above 1.5. Furthermore, the incidence of recurrence of cancer, and especially of local recurrence, was significantly higher (P <0.001) in cases with aneuploidy (DI above 1.5) than in cases with diploidy and aneuploidy (DI below 1.4). These findings indicate the significant value of the DNA index for the prediction of lymph node metastasis and local recurrence in patients with rectal cancer.     

Feasibility study for use of brush cytology as a complementary method for diagnosis of rectal cancer

Brush cytology has previously been described as a feasible method for accurately diagnosing colorectal cancer. PURPOSE: This study was designed: 1) to determine the sensitivity and specificity of brush cytology for the diagnosis of rectal cancer; 2) to prospectively assess the extent of interobserver variability with this technique; 3) to prospectively examine the cost impact of the addition of brush cytology as a routine method of confirming the diagnosis of rectal cancer. PATIENTS AND METHODS: Three hundred fiftyseven patients who attended a rectal clinic and who were found to have a lesion between January 1990 and March 1996 were assessed. Each patient underwent rigid proctoscopy, followed by brush cytology and tissue biopsy. Results were compared with the final histologic diagnosis in each patient. The brushings from the last 92 consecutive patients in this series were independently examined by four cytologists and a pathologist to determine the rate of interobserver variability. RESULTS: Rectal adenocarcinoma was confirmed from surgically resected specimens in 303 patients. Brush cytology accurately diagnosed 278 of them. Of the remaining 25 patients, two had brushings that were insufficient for diagnosis. There was one false-positive case. Forceps biopsy correctly identified cancer in 260 patients, with no false-positive interpretations. Brush cytology accurately identified 53 of 54 adenomas as being benign, and forceps biopsy correctly identified all as benign. Sensitivity of brush cytology in this series was 92 percent, with a specificity of 92 percent. Interobserver agreement was 84 percent. Actual costs incurred with this method was an additional $17.00 per patient. CONCLUSIONS: Brush cytology can accurately diagnose rectal cancer in a high proportion of patients. Interobserver variation is low and compares favorably with other forms of cytologic interpretation. The additional cost remains a concern but can be kept within acceptable proportion.Read at the meeting of The American Society of Colon and Rectal Surgeons, Seattle, Washington, June 9 to 14, 1996.     

Endorectal ultrasound in the preoperative staging of rectal tumors

The preoperative staging of rectal cancer has important implications for treatment as local therapies become increasingly utilized. Seventy-seven patients underwent preoperative staging using endorectal ultrasonography. All patients had complete pathologic staging and none had preoperative radiotherapy. Depth of invasion of the tumor was accurately predicted in 75 percent of cases in the entire group, with 22 percent overstaged and 3 percent understaged. Accuracy improved greatly over the study period, and in the past six months, 95 percent have been accurately staged for depth of invasion with 5 percent overstaged. Lymph nodes have been properly classified into positive and negative groups in 88 percent of cases in the past year, with a specificity of 90 percent and a sensitivity of 88 percent. Endorectal ultrasound is an accurate preoperative staging modality. Accuracy is improved greatly with increased experience and it has been found that the 5-layer anatomical model facilitates accurate staging. Introduction of the ultrasound probe through a previously placed proctoscope ensures complete scanning of the entire lesion and should be used for the majority of examinations.Read at the meeting of The American Society of Colon and Rectal Surgeons, Toronto, Canada, June 11 to 16, 1989.     

How accurate is endorectal ultrasound in the preoperative staging of rectal cancer?

In rectal cancer, depth of infiltration and metastatic involvement of lymph nodes are important prognostic factors. The correct choice of operative treatment depends on the extent of the disease. In a prospective study, the value of endorectal ultrasound in staging rectal cancer was evaluated, and factors affecting the method's accuracy are discussed. The overall accuracy in staging depth of infiltration was 89 percent. Overstaging occurred in 10.2 percent, understaging in 0.8 percent. Tumors of the lower rectum are incorrectly staged in 16.7 percent, whereas tumors of the middle and upper rectum had an incorrect staging in 6.3 percent (P <0.001). Compared with computed tomography, endorectal sonography is the more accurate staging method (74.7 vs.90.8 percent). In staging lymph nodes, the overall accuracy was 80.2 percent, sensitivity was 89.4 percent, specificity was 73.4 percent, positive predictive value (PPV) was 71.2 percent, and negative predictive value (NPV) was 90.4 percent. The staging accuracy depends on the size of the node. Endorectal ultrasound is a safe, inexpensive, and accurate staging method, in the assessment of both depth of infiltration and nodal status. The results are strongly related to the experience of the investigator.     

Endorectal ultrasonic detection of malignancy within rectal villous lesions

PURPOSE: The ability of endorectal ultrasonography (EU) to detect the presence of a malignant focus within rectal villous adenomas was studied. METHODS: Clinical charts were reviewed of 62 consecutive patients undergoing EU of rectal villous adenomas, in whom Histologic confirmation was available. RESULTS: Twelve lesions were found to contain cancer, of which only two demonstrated clinical signs of induration. Positive predictive value of EU for detecting a malignant focus was 66.7 percent, negative predictive value was 88.7 percent, sensitivity was 50 percent, and specificity was 94 percent. There was moderate overall agreement between pathologic and ultrasound staging (kappa statistic, 0.48). When an optimal image was obtained, all cancers that penetrated the submucosa were detected. Sensitivity of the technique was compromised in some large exophytic lesions and those at the level of the anal sphincter because of artefacts produced in the ultrasonographic image. CONCLUSION: A clear EU image can detect a malignant focus within a villous adenoma and direct the surgeon to the appropriate plane of surgical resection. In lesions with an ambiguous image, a malignancy cannot be excluded.All work was performed at the Division of Colon and Rectal Surgery, University of Minnesota, Minneapolis, Minnesota.     

Prognostic significance of location of lymph node metastases in colorectal cancer

The prognostic value of stage of lymph node metastases was evaluated in 357 patients who underwent curative resection for colorectal cancer. Subdivision of Dukes C patients according to the number of positive nodes revealed that the five-year disease-free survival rate (5DFS) was 63 percent in the patients with one to three nodes and 53 percent in those with four or more nodes (not significantly different). Classification according to the location revealed that 5DFS was 70 percent in those who had only local node metastases (n1+), compared with 40 percent in those who had distant node metastases along the major vessels (n2+) (P <0.001). Twelve of 38 n2+ patients had only one distant node metastasis with no local node involvement (skip metastasis). They had lower 5DFS than the n1+ patients who had three or more positive local nodes (35 percent vs. 57 percent). We conclude that the location, rather than the number, of nodal metastases has a higher impact on prognosis in colorectal cancer patients.Read at the meeting of The American Society of Colon and Rectal Surgeons, Boston, Massachusetts, May 12 to 17, 1991.     

Accuracy of transrectal ultrasound in predicting pathologic stage of rectal cancer before and after preoperative radiation therapy

Transrectal ultrasound (TRUS) and CT scan staging of rectal cancers before, and TRUS staging after, 45 Gy of irradiation were compared with the pathologic stage of the resected specimen in 19 patients. Accuracy of TRUS before and after irradiation, and of CT scan before irradiation, was 32 percent, 63 percent, and 53 percent, respectively. CT scan before and TRUS after irradiation predicted lymph node involvement in 79 percent and 68 percent of cases, respectively. Positive predictive value for lymph node involvement before irradiation was 60 percent for CT scan and 37.5 percent for TRUS; after irradiation, it was 50 percent for TRUS. Negative predictive value was 100 percent for CT scan and TRUS before radiation and 88 percent for TRUS after irradiation. Preoperative radiation therapy makes TRUS and CT scan less effective as staging techniques. The absence of lymph nodes on TRUS and CT scan before and after irradiation is reliable.Read in part at the Tripartitate Meeting, Birmingham, England, June 19 to 22, 1989.     

Locally advanced rectal cancer: Resection and intraoperative radiotherapy using the flab method combined with preoperative or postoperative radiochemotherapy

PURPOSE: Surgery often fails to achieve local control in advanced rectal cancer. Additional measures are necessary to prevent local recurrence. The aim of this study was to evaluate intraoperative radiation therapy (IORT) (flab technique) combined with preoperative or postoperative radiochemotherapy. PATIENTS/METHODS: IORT is performed using a flexible flab containing hollow plastic tubes that are connected to a multichannel afterloading device with a 370 Gbq-192-Ir source. Patients receive an intraoperative dose of 15 Gy. Target volumes were measured in a cadaver experiment. From 1989 to 1993, 38 patients were included in this study. Nineteen patients were staged as T3 tumors by preoperative endosonography (Group I) and 19 as T4 tumors (Group II). Patients in Group I underwent resection (abdominoperineal resection (APR), 16; anterior resection, 3) and IORT, followed by postoperative radiochemotherapy (50 Gy/5-fluorouracil), whereas patients in Group II received preoperative radiochemotherapy (40 Gy/5-fluorouracil) followed by resection (APR, 18; anterior resection, 1) and IORT. Mean follow-up was 25.5 months. RESULTS: Operative radicality in Group I was R0 (13), R1 (3), and R2 (3), and in Group II it was R0 (14), R1 (3), R2 (2). R2 resections were attributable to preoperative undetected distant metastases. Perioperative mortality was 0 percent in Group I and 10.5 percent (n=2) in Group II. Postoperative morbidity was 53 percent (n=10) in Group I and 84 percent (n=16) in Group II with delayed sacral wound healing being the predominant problem. Stenosis of the ureter occurred in two patients (Group II). Late or persistent therapy-related complications were seen in two patients in Group I and in six patients in Group II. Local recurrence developed in three patients in Group I (15.8 percent) and in two patients in Group II (10.5 percent). Survival data do not reach statistical significance between the two groups because of small numbers but show a favorable trend for the preoperative radiochemotherapy group. When compared with a matched historical control group of patients receiving resection only, adjuvant/neoadjuvant radiotherapy with resection/IORT improves survival significantly. CONCLUSION: The flab method is a simple but especially practical technique for IORT in the pelvis. Adjuvant/neoadjuvant therapy combined with resection/IORT is associated with high morbidity but acceptable mortality. Preliminary survival data are encouraging and call for a controlled prospective randomized trial.Read at the meeting of The American Society of Colon and Rectal Surgeons, Montreal, Quebec, Canada, May 7 to 12, 1995.     

Flow cytometric examination of p53 protein in primary tumors and metastases to the liver and lymph nodes of colorectal cancer

PURPOSE AND METHODS: To confirm prognostic significance of overexpression of p53 in cases of colorectal cancer, expression of p53 protein was examined by flow cytometry in 113 cases of colorectal cancer and its metastasis to the liver and lymph nodes. RESULTS: Overexpression of p53 was found in 44 (39 percent) of the 113 primary tumors. There were no significant correlations among the level of p53 protein in the primary tumor, clinicopathologic features, and prognosis of colorectal cancer. Overexpression of p53 protein was detected in 72 percent (18/25) of liver metastases and in 40 percent (10/25) of lymph node métastases. Frequency of samples that were positive for p53 was significantly higher for liver metastases than for primary tumors and lymph node metastases (P<0.01). By comparing overexpression of p53 in primary tumors with that in corresponding secondary tumors, a decrease of more than 5 percent in the fluorescence index, compared with primary tumor, was not found in liver metastasis but was found in 20 percent of lymph node metastases. Incidence of cases with lower level expression of p53, compared with primary tumor, was significantly higher in lymph node metastases (32 percent) than in liver metastases (8 percent;P<0.05). CONCLUSIONS: From these results, it seems possible that overexpression of p53 may not be a good prognostic indicator of colorectal cancer and may be influenced by environments of the tumor.Presented at the meeting of the Japanese Gastroenterological Surgery, Fukui City, Japan, July 20 and 21, 1995.     

Progress in the assessment of lymphatic spread in rectal cancer Rectal endoscopic lymphoscintigraphy

Rectal endoscopic lymphoscintigraphy was performed in 10 control subjects and in a series of 85 patients with adenocarcinoma of the rectum as a prospective study to evaluate lymphatic drainage of the rectum and lymphatic spread in rectal cancer. Complete cranial drainage was demonstrated in all control subjects, and internal iliac nodes were also visible in 50 percent of cases. Results were correlated with histologic node examination in all patients operated upon for rectal cancer. Rectal endoscopic lymphoscintigraphy was assessed for sensitivity (85 percent), specificity (68 percent), overall accuracy (76 percent), positive predictive value (71 percent), and negative predictive value (83 percent). False-negative and false-positive results are discussed. Rectal endoscopic lymphoscintigraphy represents the only method currently available for evaluation of lymphatic spread in rectal cancer.     

Prospective comparative study of magnetic resonance imaging versus transrectal ultrasound for preoperative staging and follow-up of rectal cancer

The efficiency of magnetic resonance imaging (MRI) and that of transrectal ultrasound (TRUS) were compared in preoperative staging of 15 patients with rectal cancer and in postoperative follow-up of 12 patients. Thirteen of the 15 patients evaluated for preoperative staging were operated on. Preoperative staging and pathologic finding were identical in 11 patients (84.6 percent) examined by TRUS and in 10 patients (76.9 percent) examined by MRI. Recurrent cancer was detected in 3 of 12 patients in the follow-up group. MRI was able to diagnose correctly 10 of 12 patients (83.2 percent), one patient was misdiagnosed, and in one patient the MRI could not distinguish between fibrous tissue and recurrent cancer. TRUS diagnosed correctly only 5 of 12 patients (41.6 percent). One was falsely diagnosed, and, in 6 patients (50 percent), this examination could not differentiate between fibrous tissue and recurrent tumor. According to our results, both MRI and TRUS have a place in the preoperative staging of patients with rectal cancer. The main differences between the two methods were in the differential diagnoses of fibrous tissue and recurrent cancer. MRI being more specific in detection of recurrence.Read at the XIIIth Biennial Congress of the International Society of University Colon and Rectal Surgeons, Graz, Austria, June 24 to 28, 1990.