Accuracy of Endorectal Ultrasonography in ... - Springer Link

Accuracy of Endorectal Ultrasonography in Preoperative Staging of Rectal Tumors Julio Garcia-Aguilar, M.D., Ph.D., Johan Pollack, M.D., Suk-Hwan Lee, M.D., Enrique Hernandez de Anda, M.D., Anders Mellgren, M.D., Ph.D., W. Douglas Wong, M.D., Charles O. Finne, M.D., David A. Rothenberger, M.D., Robert D. Madoff, M.D. From the Division of Colon and Rectal Surgery, Department of Surgery, University of Minnesota, and the University of Minnesota Cancer Center, Minneapolis, Minnesota

R

ectal cancer has traditionally been treated by radical en bloc excision of the rectum and its lymphatic drainage, either with a sphincter-preserving anastomosis or an abdominoperineal resection.1 In recent years, transanal local excision has become an accepted surgical alternative in selected patients with rectal tumors still localized to the rectal wall.2 Preoperative chemoradiation has become the standard adjuvant therapy in patients with advanced rectal cancer.3 Therapeutic strategies that attempt to tailor treatment by tumor stage require a precise knowledge before the operation of the depth of tumor invasion in the rectal wall and of the presence of tumor in the regional lymph nodes. During the last decade, endorectal ultrasonography (ERUS) has become the most common diagnostic modality for local staging of rectal cancer.4, 5 However, its accuracy, reliability, and validity are still controversial. Most studies include fewer than 100 patients and often represent only the initial institutional experience with the technique.6–15 At the University of Minnesota, ERUS has been used in rectal cancer staging for more than ten years.16 The aim of the present study was to review our experience with ERUS and to determine its accuracy in discriminating between early and advanced rectal tumors.

PURPOSE: Preoperative staging of rectal tumors is considered essential to tailor treatment for individual patients. The aim of the present study was to evaluate the accuracy of endorectal ultrasonography in preoperative staging of rectal tumors. METHODS: Eleven hundred eighty-four patients with rectal adenocarcinoma or villous adenoma underwent endorectal ultrasonography evaluation at a single institution during a ten-year period. We compared the endorectal ultrasonography staging with the pathology findings based on the surgical specimens in 545 patients who had surgery (307 by transanal excision, 238 by radical proctectomy) without adjuvant preoperative chemoradiation. Comparisons between groups were performed using chi-squared tests and logistic regression analysis. RESULTS: Overall accuracy in assessing the level of rectal wall invasion was 69 percent, with 18 percent of the tumors overstaged and 13 percent understaged. Accuracy depended on the tumor stage and on the ultrasonographer. Overall accuracy in assessing nodal involvement in the 238 patients treated with radical surgery was 64 percent, with 25 percent overstaged and 11 percent understaged. CONCLUSION: The accuracy of endorectal ultrasonography in assessing the depth of tumor invasion, particularly for early cancers, is lower than previously reported. The technique is more precise in distinguishing between benign tumors and invasive cancers and between tumors localized to the rectal wall and tumors with transmural invasion. Differences in image interpretation may in part explain discrepancies in accuracy between studies. [Key words: Rectal neoplasm; Rectal tumor; Staging; Endorectal ultrasonography] Garcia-Aguilar J, Pollack J, Lee S-K, Hernandez de Anda E, Mellgren A, Wong WD, Finne CO, Rothenberger DA, Madoff RD. Accuracy of endorectal ultrasonography in preoperative staging of rectal tumors. Dis Colon Rectum 2002;45: 10–15.

PATIENTS AND METHODS We reviewed the charts of 1,184 patients whose rectal adenocarcinomas or villous adenomas were staged by ERUS and who were operated on by the Division of Colon and Rectal Surgery faculty at the University of Minnesota-affiliated hospitals during a ten-year period. Our analysis excluded 639 patients for various reasons (Table 1). The remaining 545 patients are the focus of this study.

Supported by a limited grant from the Minnesota Colon and Rectal Foundation. Present address of Drs. Pollack and Mellgren: Department of Surgery, Karolinska University Hospital, Stockholm, Sweden. Present address of Dr. Wong: Division of Colon and Rectal Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York. Address reprint requests to Dr. Garcia-Aguilar: Division of Colon and Rectal Surgery, Mayo Mail Code 450, 420 Delaware Street SE, Minneapolis, MN 55455.

10

Vol. 45, No. 1

STAGING OF RECTAL TUMORS

11

Table 1. Exclusion Criteria Number Associated pathology (Crohn’s disease, fistula-in-ano, familial adenomatous polyposis) Tumor fulguration Incomplete clinical information Incomplete evaluation (inaccessible/stenotic) Recurrent cancer Endocavitary radiation Tumor in a snare polypeptomy specimen Preoperative radiation therapy Total

Five board-certified, experienced colorectal surgeons performed the ERUS examinations. Two of the surgeons performed only 3 percent of the examinations (n ⫽ 17) and were excluded from our analysis of accuracy by surgeon. The other three surgeons (A, B, C) performed 9 percent, 40 percent, and 48 percent of the examinations. Each patient was examined with a triad of diagnostic modalities: digital examination, proctoscopy, and ERUS. The patient was prepared for the examination with a Fleets® enema one hour before the ERUS (C.B. Fleet Company, Inc., Lynchburg, VA). The patient was examined in a left lateral position. The tumor, if reachable, was palpated with the finger to assess mobility and morphology and then inspected through the rigid proctoscope. A 7-MHz or a 10-MHz endosonographic probe (Type 3535, Bru¨ el & Kjaer, Naerum, Denmark) was introduced through the proctoscope. The probe, covered with a rubber balloon, was carefully passed from the anal verge to the upper rectum. The balloon was then filled with variable amounts of water to achieve optimal contact with the rectal wall, and the probe was slowly retracted to the level of the tumor. Results were documented on a standard form and fed into a database. During ERUS, rectal wall penetration was assessed using a modification of the TNM classification based on a five-layer rectal wall model proposed by Hildebrandt and Feifel.17 Pathologic lymph nodes were defined as circular or slightly ovalshaped structures, often with an irregular border, and with an echogenicity similar to the tumor, as proposed by Beynon et al.18 The referring surgeon decided the treatment strategy based on tumor and patient characteristics. The surgical specimens were sent for pathologic examination and staging according to the pTNM classification. The perirectal lymph nodes

17 22 35 37 37 99 122 270 639

were classified as free from metastatic disease (pN0) or as positive for metastatic spread (pN1). The ultrasound images were compared with the postoperative pathologic stage. The tumor rectal wall penetration (T stage) was compared in all patients, whereas the nodal status (N stage) was compared in patients undergoing radical surgery. For purposes of our analysis, patients were divided into three groups depending on whether the examination was performed between 1987 and 1990 (n ⫽ 106), between 1991 and 1995 (n ⫽ 219), or between 1996 and 1999 (n ⫽ 203). Depending on the distance of the lower tumor margin from the anal verge, tumors were considered low (⬍6 cm; n ⫽ 224), intermediate (6–10 cm; n ⫽ 288), or high (⬎10 cm; n ⫽ 33). Comparisons between groups were performed using chi-squared tests and logistic regression analysis.

RESULTS Of the 545 patients in our study, 238 underwent radical surgery (abdominoperineal resection or low anterior resection) and 307 underwent transanal local excision of their rectal tumors. There were 273 males and 272 females; the mean age was 67 (range, 25–98) years. The mean tumor diameter was 4.3 cm; the mean distance from the anal verge was 7.4 cm. Overall accuracy of ERUS in assessing the depth of tumor penetration in the rectal wall (T stage) was 69 percent, with 18 percent of the tumors overstaged and 13 percent understaged (Table 2). The positive predictive value was 72 percent, and the negative predictive value was 93 percent. The accuracy of ERUS in the diagnosis of lymph node metastasis in our 238 patients who underwent radical surgery without preoperative radiation ther-

GARCIA-AGUILAR ET AL

12

Dis Colon Rectum, January 2002

Table 2. Rectal Wall Invasion: ERUS (uT) vs. Pathologic Examination (pT)

uT0 uT1 uT2 uT3 uT4 Total

pT0 (n)

pT1 (n)

pT2 (n)

pT3 (n)

pT4 (n)

Total (n)

129 34 4 0 0 167

14 49 20 0 0 83

3 21 104 37 2 167

2 1 25 92 2 122

0 0 0 2 4 6

148 105 153 131 8 545

Overstaged (%)

Understaged (%)

Accuracy (%)

13 21 16 2

87 47 68 70 50 69

32 16 28 50 18

13

Table 3. Lymph Node Metastasis: ERUS (uN) vs. Pathologic Examination (pN)

uN0 uN1 Total

pN0 (n)

pN1 (n)

Total (n)

Accuracy (%)

124 27 151

58 29 87

182 56 238

68 52 64

Overstaged (%)

Understaged (%) 32

48 25

11

Figure 1. Accuracy of ERUS staging by operator. * P ⬍ 0.05 by chi-squared test.

apy was 64 percent (Table 3). The sensitivity for nodal involvement was 33 percent, but the specificity was 82 percent. The positive predictive value was 52 percent, and the negative predictive value was 68 percent. Overall accuracy of ERUS by surgeon (limited to the 3 who performed more than 50 examinations) is shown in Figure 1. Differences between surgeons for T staging accuracy, but not for N staging, were statistically significant. Accuracy was not influenced by such factors as patient age, gender, tumor size, or distance of the tumor from the anal verge. We found no difference in accuracy over time for either T or N staging. However, only one surgeon performed ERUS during all three periods studied, so the periods and surgeons were highly associated. By logistic regression analysis, only the T stage of the tumor and the surgeon performing ERUS were independent factors influencing overall accuracy in assessing rectal wall penetration.

DISCUSSION This study represents the largest institutional experience with preoperative staging of rectal tumors by ERUS. A superficial interpretation of our data seems to indicate that overall accuracy of ERUS in assessing the depth of invasion and nodal involvement is lower than previously reported.6–15 However, at our institution most patients with advanced rectal cancers (ultrasound Stage II and III) underwent preoperative radiation therapy and so were excluded from our analysis. Our experience suggests that the accuracy of ERUS in assessing the depth of rectal wall invasion varies with tumor stage: it is more accurate in benign villous adenomas and advanced tumors that penetrate the perirectal fat. Our exclusion of a significant proportion of the tumors in which ERUS is more accurate may have resulted in an underestimation of its overall accuracy. The same limitation may apply to lymph node metastasis. The risk of lymph node metastasis

Vol. 45, No. 1

STAGING OF RECTAL TUMORS

increases with the depth of rectal wall invasion so the use of neoadjuvant therapy in advanced rectal cancer may have resulted in an underestimation of the accuracy of ERUS in diagnosing lymph node metastasis. A significant proportion of rectal tumors diagnosed by endoscopic biopsy as villous adenomas contain foci of invasive adenocarcinoma that are not clinically evident. Accurate preoperative staging of such lesions may determine the choice between a submucosal or a full-thickness excision or even a radical resection. In our study, ERUS was most accurate in diagnosing benign villous adenomas—a finding that contradicts other authors’ opinions that ERUS cannot differentiate between villous adenomas and invasive cancers.19–21 This contradiction most likely reflects differences in the interpretation of the ultrasonic images of the rectal wall. Authors who suggest that uT0 and uT1 tumors should be classified together believe that neither the muscularis mucosa nor the submucosa is sonographically definable and that the first hypoechoic layer corresponds anatomically with the mucosa and submucosa.19–21 Such authors propose, because both villous adenomas and T1 carcinomas manifest as a broadening of the first hypoechoic layer, that both should be classified together. We disagree with this interpretation. Instead, we follow the criteria of Beynon et al.22 and consider the first hypoechoic layer as the mucosa and muscularis mucosa and the middle hyperechoic layer as the submucosa. Consequently, for us, lesions that expand the inner hypoechoic layer and are surrounded by a uniform middle hyperechoic layer are considered villous adenomas and lesions that expand the inner hypoechoic layer and have distinct echo defects of the middle hyperechoic layer are considered uT1 tumors. Kuntz et al.,23 using criteria similar to ours, have also reported a high degree of accuracy discriminating between villous adenomas and early invasive cancers. In our study, ERUS correctly staged most villous adenomas but less than half of T1 tumors (Table 2). The probability of an uT1 tumor being correctly staged was less than 50 percent, with an almost equal probability of overstaging and understaging. Other authors use low accuracy of ERUS in diagnosing uT1 tumors to argue against its use in selecting rectal tumors for local excision.19 However, all but one of our tumors that were staged as uT1 were contained within the rectal wall. Therefore, we consider such information clinically relevant, because the probability of transmural extension of the cancer, which would result in leaving residual tumor if the primary

13

lesion is removed by transanal excision, is lower than 1 percent. The poor endosonographic diagnosis of T2 tumors, only 68 percent, is worse than previously reported, except for the study of Sailer et al.21 that reported an accuracy of 41 percent in T2 tumors. Overstaging (16 percent), often attributed to peritumoral tumor inflammation,11, 24, 25 was the most common type of error. The low accuracy of the ERUS in the characterization of T2 tumors emphasizes the need to base final treatment after local excision on the pathologic and not on the ultrasound stage. Two important factors in selecting the treatment modality for patients with rectal tumors are whether the tumor is localized to the rectal wall or penetrates into the perirectal fat and whether it has already metastasized to the regional lymph nodes. In our study, ERUS correctly identified 194 (77 percent) of 250 patients whose tumors were localized to the rectal wall and who were potential candidates for local excision and 100 (78 percent) of 128 of pT3 or pT4 tumors that were candidates for preoperative chemoradiation therapy. Despite our exclusion of the most advanced tumors, the accuracy of ERUS in diagnosing transmural invasion was high, with a positive predictive value of 72 percent and a negative predictive value of 87 percent. Therefore, we still find ERUS reliable in preoperative assessment of rectal wall invasion. The impact of the tumor distance from the anal verge on the accuracy of ERUS staging has been controversial. Sentovich et al.26 reported higher accuracy in the staging of rectal tumors located within 6 cm from the anal verge, but most studies have reported higher accuracy for proximal compared with distal tumors.11, 21 The differences in accuracy have been attributed to difficulties in achieving uniform contact between the water balloon and the rectal wall and to the suboptimal delineation of the rectal wall immediately above the anorectal ring. However, the number of patients in these studies was too small to reach meaningful conclusions. Our results suggest that, for examinations considered adequate by the surgeon, the distance of the tumor from the anal verge did not influence accuracy. Our data confirm the observation that the accuracy of ERUS in the staging of rectal cancers is surgeon dependent.27, 28 The accuracy of individual surgeons was different for wall penetration and for diagnosis of lymph node metastasis. Assuming that the technical skills for imaging wall invasion and

GARCIA-AGUILAR ET AL

14

nodal metastasis are similar, the differences between surgeons in accuracy of staging each variable can be attributed to differences in interpretation.

CONCLUSION ERUS is useful in staging rectal tumors preoperatively. It identifies transmural invasion and reliably distinguishes between benign tumors and invasive cancers. Our data indicate that the accuracy of ERUS in preoperative staging of rectal tumors varies by tumor stage and by operator.

ACKNOWLEDGMENTS We thank Claudia Genung and Robin Bliss for their generous assistance with the Colorectal Cancer Database and Mary E. Knatterud for reviewing the manuscript.

REFERENCES 1. MacFarlane JK, Ryall RD, Heald RJ. Mesorectal excision for rectal cancer. Lancet 1993;341:457– 60. 2. Garcia-Aguilar J, Mellgren A, Sirivongs P, Buie D, Madoff RD, Rothenberger DA. Local excision of rectal cancer without adjuvant therapy: a word of caution. Ann Surg 2000;231:345–51. 3. Anonymous. Improved survival with preoperative radiotherapy in resectable rectal cancer. Swedish Rectal Cancer Trial. N Engl J Med 1997;336:980 –7. 4. Solomon MJ, McLeod RS. Endoluminal transrectal ultrasonography: accuracy, reliability, and validity. Dis Colon Rectum 1993;36:200 –5. 5. Phang PT, Wong WD. The use of endoluminal ultrasound for malignant and benign anorectal diseases. Curr Opin Gastroenterol 1997;13:47–53. 6. Hildebrandt U, Feifel G, Schwarz HP, Scherr O. Endorectal ultrasound: instrumentation and clinical aspects. Int J Colorectal Dis 1986;1:203–7. 7. Beynon J, Roe AM, Foy DM, Channer JL, Virjee J, Mortensen NJ. Preoperative staging of local invasion in rectal cancer using endoluminal ultrasound. J R Soc Med 1987;80:23– 4. 8. Holdsworth PJ, Johnston D, Chalmers AG, et al. Endoluminal ultrasound and computed tomography in the staging of rectal cancer. Br J Surg 1988;75:1019 –22. 9. Milsom JW, Graffner H. Intrarectal ultrasonography in rectal cancer staging and in the evaluation of pelvic disease: clinical uses of intrarectal ultrasound. Ann Surg 1990;212:602– 6.

Dis Colon Rectum, January 2002

10. Katsura Y, Yamada K, Ishizawa T, Yoshinaka H, Shimazu H. Endorectal ultrasonography for the assessment of wall invasion and lymph node metastasis in rectal cancer. Dis Colon Rectum 1992;35:362– 8. 11. Herzog U, von Flue M, Tondelli P, Schuppisser JP. How accurate is endorectal ultrasound in the preoperative staging of rectal cancer? Dis Colon Rectum 1993;36: 127–34. 12. Nielsen MB, Qvitzau S, Pedersen JF, Christiansen J. Endosonography for preoperative staging of rectal tumours. Acta Radiol 1996;37:799 – 803. 13. Nishimori H, Sasaki K, Hirata K, Natori H. The value of endoscopic ultrasonography in preoperative evaluation of rectal cancer. Int Surg 1998;83:157– 60. 14. Massari M, De Simone M, Cioffi U, Rosso L, Chiarelli M, Gabrielli F. Value and limits of endorectal ultrasonography for preoperative staging of rectal carcinoma. Surg Laparosc Endosc 1998;8:438 – 44. 15. Adams DR, Blatchford GJ, Lin KM, Ternent CA, Thorson AG, Christensen MA. Use of preoperative ultrasound staging for treatment of rectal cancer. Dis Colon Rectum 1999;42:159 – 66. 16. Orrom WJ, Wong WD, Rothenberger DA, Jensen LL, Goldberg SM. Endorectal ultrasound in the preoperative staging of rectal tumors: a learning experience. Dis Colon Rectum 1990;33:654 –9. 17. Hildebrandt U, Feifel G. Preoperative staging of rectal cancer by intrarectal ultrasound. Dis Colon Rectum 1985;28:42– 6. 18. Beynon J, Mortensen NJ, Foy DM, Channer JL, Rigby H, Virjee J. Preoperative assessment of mesorectal lymph node involvement in rectal cancer. Br J Surg 1989;76: 276 –9. 19. Heintz A, Buess G, Frank K, Kuntz C, Strunck H, Junginger T. Endoluminal ultrasonic examination of sessile polyps and early carcinomas of the rectum. Surg Endosc 1989;3:92–5. 20. Mosnier H, Guivarc’h M, Meduri B, Fritsch J, Outters F. Endorectal sonography in the management of rectal villous tumours. Int J Colorectal Dis 1990;5:90 –3. 21. Sailer M, Leppert R, Bussen D, Fuchs KH, Thiede A. Influence of tumor position on accuracy of endorectal ultrasound staging. Dis Colon Rectum 1997;40: 1180 – 6. 22. Beynon J, Foy DM, Temple LN, Channer JL, Virjee J, Mortensen NJ. The endosonic appearances of normal colon and rectum. Dis Colon Rectum 1986;29:810 –3. 23. Kuntz C, Glaser F, Buhr HJ, Herfarth C. [Endorectal ultrasound in diagnosis and therapy planning of broadbase rectal adenomas. In German]. Chirurg 1993;64: 290 – 4. 24. Kruskal JB, Kane RA, Sentovich SM, Longmaid HE. Pitfalls and sources of error in staging rectal cancer with endorectal US. Radiographics 1997;17:609 –26.

Vol. 45, No. 1

STAGING OF RECTAL TUMORS

25. Schnall MD, Furth EE, Rosato EF, Kressel HY. Rectal tumor stage: correlation of endorectal MR imaging and pathologic findings. Radiology 1994;190:709 –14. 26. Sentovich SM, Blatchford GJ, Falk PM, Thorson AG, Christensen MA. Transrectal ultrasound of rectal tumors. Am J Surg 1993;166:638 – 41.

15

27. Burtin P, Rabot AF, Heresbach D, et al. Interobserver agreement in the staging of rectal cancer using endoscopic ultrasonography. Endoscopy 1997;29:620 –5. 28. Roubein LD, Lynch P, Glober G, Sinicrope FA. Interobserver variability in endoscopic ultrasonography: a prospective evaluation. Gastrointest Endosc 1996;44:573–7.

The five-year cumulative index for Volumes 36 through 40 (1993–1997) of Diseases of the Colon & Rectum is available online at www.lww.com/DCR