Accuracy of radiographic assessment for the ... - BIR Publications

The British Journal of Radiology, 76 (2003), 611–616 DOI: 10.1259/bjr/63403347

E

2003 The British Institute of Radiology

Accuracy of radiographic assessment for the diagnosis of invasion depth in small invasive colorectal cancer 1

T MATSUMOTO, MD, 2M ESAKI, MD, 1K HIZAWA, MD, 1K KURAHARA, MD, 1K HIRAKAWA, MD, T YAO, MD and 1M IIDA, MD

3 1

Departments of Medicine & Clinical Science and 3Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Maidashi 3-1-1, Higashi-ku, Fukuoka 812-8582 and 2Department of Endoscopic Diagnostics & Therapeutics, Kyushu University Hospital, Fukuoka, Japan

Abstract. Radiographic findings of 113 submucosally invasive colorectal cancers (CRCs) ¡20 mm were investigated by two observers to examine whether barium radiography is predictive of the invasion depth in small colorectal cancer. Smooth surface, central barium fleck, fold convergency and eccentric rigidity were chosen as the radiographic determinants. Predictive values of the determinants for deep submucosal invasion were compared between 45 CRCs ¡10 mm and 68 CRCs .10 mm. 47 CRCs had invaded the submucosa superficially while 66 CRCs had invaded deeply. The concordance rate in the assessment of radiology was 90.3% in smooth surface, 92.9% in central barium fleck, 90.3% in fold convergency and 79.6% in eccentric rigidity. Positive predictive value of central barium fleck for deep submucosal invasion was significantly higher in CRCs.10 mm than in those ¡10 mm (0.86 vs. 0.64, p50.034). In contrast, negative predictive values of smooth surface (0.89 vs. 0.34, p,0.00001), central barium fleck (0.87 vs. 0.38, p50.0002), fold convergency (0.64 vs. 0.33, p50.0023) and eccentric rigidity (0.79 vs. 0.48, p50.023) were significantly higher in CRCs ¡10 mm than in those .10 mm. These findings suggest that barium radiography is a procedure which can give useful information in prediction of invasion depth in CRCs ¡10 mm in size.

With the widespread application of colonoscopy, it has been reported that small, but invasive colorectal cancers (CRCs) do exist in the colon and rectum [1]. Endoscopic removal of intramucosal cancers is established, and is now also regarded as an appropriate treatment for cancers with the invasion limited to the superficial layer of the submucosa, provided such cancers are free from unfavourable histology [2–5]. Recent advances in endoscopic procedure have suggested that endoscopic ultrasound (EUS) [6–10] and chromoscopy with magnifying instruments [11–14] can predict invasion depth in CRC. However, such procedures are of limited practical value because of interobserver variations [15–17]. In the field of radiology, double-contrast barium enema examination (BE) has been regarded as a screening and diagnostic procedure for the diagnosis of CRC. Furthermore, BE has been shown to depict the nature of colorectal neoplasia and enables the assessment of the invasion depth [18, 19]. Watari et al [19] reported that converging folds, semilunar deformity and obvious barium fleck were sensitive radiographic signs for the prediction of submucosal invasion in CRC. In their article, however, the size of the tumour was also an independent predictor for the submucosal invasion. The present investigation was conducted to determine whether BE can be utilized for assessment of invasion depth in small invasive cancers.

Materials and methods This study was retrospectively conducted on patients with an established diagnosis of submucosally invasive Received 14 October 2002 and in revised form 10 March 2003, accepted 13 June 2003.

The British Journal of Radiology, September 2003

CRC at our institutions during a period from August 1995 to April 2001. During the period, patients with CRC detected by colonoscopy were subsequently examined by BE with informed consent that BE is used as a procedure to determine whether endoscopic or surgical therapy is appropriate. Among 207 CRCs with submucosal invasion, 160 CRCs in 160 patients measured less than 20 mm in its largest dimension. Among these patients, 132 patients underwent BE. 19 patients were excluded from the following analysis, because the CRC was not depicted by BE. The remaining 113 patients were the subjects of the present investigation.

Double-contrast barium enema examination BE was performed with our standard preparation (low residue diet, 7.5 mg sodium picosulphate, and 34 g of magnesium citrate on the day before examination and 10 mg bisacodyl suppository 2 h before examination) and double-contrast technique by using 70–80% wt/vol barium sulphate [20]. The radiographs were reviewed by two observers (TM and ME) who were blind to the colonoscopic and pathological findings of the lesions. In accordance with an article by Watari et al [19], the following items were reviewed by the observers: 1) surface nature of the protrusion; 2) the central barium fleck; 3) convergency of the haustral folds; 4) eccentric rigidity on the profile view. Obvious in-drawing of the profile luminal line was categorized as eccentric rigidity. The surface nature of the protrusion was determined to be smooth or nodular (Figure 1). The other criteria were determined to be positive or negative (Figures 2 and 3). 611

T Matsumoto, M Esaki, K Hizawa et al

(b)

(a)

Figure 1. Examples of determination in surface features and fold convergency. (a) A 1.5 cm cancer with deep submucosal invasion in the transverse colon. The surface of the lesion is determined to be nodular. There is a fold convergency towards the protrusion (arrows). (b) A 0.9 cm cancer with deep submucosal invasion in the sigmoid colon. There is a smooth protrusion with fold convergency (arrows).

and stained by haematoxylin and eosin. Under light microscopy, the depth of invasion was graded to be deep (sm-d), when carcinoma cells had invaded level 3 or 4 of the criteria reported by Haggitt et al [4], the upper one-third or deeper in the submucosa [21], or they were present at the resected margin. Otherwise, the invasion depth was determined to be superficial (sm-s). The following histologic findings were assessed as risk factors for unfavourable outcome [2–5, 22]: 1) tumour differentiation at the invasive area as determined by the size of nuclei of carcinoma cells and glandular structure (well or moderate to poor); 2) lymphatic invasion as defined by tumour cells within a true endothelial-lined channel in the absence of red blood cells; 3) venous invasion as defined by tumour cells in an endothelial-lined channel surrounded by a smooth muscle wall; 4) growth pattern as determined by the presence or absence of carcinoma cells at the margin of the tumour (either polypoid or non-polypoid); and 5) lymph node metastasis in cases treated by surgery.

Predictive values Figure 2. An example of smooth surface feature and central barium fleck. A 0.8 cm cancer with deep invasion in the sigmoid colon. There is a smooth protrusion with an obvious barium fleck in the centre.

Each radiographic determinant was regarded to be positive when both of the two observers reported a positive result. Otherwise, the radiographic determinant was regarded to be negative. Based on the result, positive (PPV) and negative predictive values (NPV) for the diagnosis of deep invasion were calculated.

Histology The size and the depth of invasion was determined in the resected specimen by a pathologist (TY), who was blind to the radiographic findings. After the lesions were removed by endoscopic mucosal resection or by surgery, each resected specimen was cut in slice with 2 mm intervals 612

Statistics McNemar’s test was used for the comparison of determination in radiographic findings between the observers. Pathologic findings were compared among the groups, using the chi-squared test. When any significant difference was The British Journal of Radiology, September 2003

Radiography for small invasive colorectal cancer

(b)

(a)

Figure 3. Examples of positive central barium fleck and eccentric deformity. A 1.2 cm cancer with deep submucosal invasion in the transverse colon. (a) There is a barium fleck accompanied by nodular protrusion. (b) On lateral view, eccentric deformity is also seen (arrow).

detected, the values were compared between any two groups, using Fisher’s exact probability test. For the comparison of predictive values, Fisher’s exact probability test was used. Probabilities less than 0.05 were considered to be significant.

Results Clinicopathologic features of colorectal cancers The CRCs were classified into four groups, according to the size and the invasion depth (17 CRCs in group A, sm-d

¡10 mm; 28 CRCs in group B, sm-s ¡10 mm; 49 CRCs in group C, sm-d .10 mm; 19 CRCs in group D, sm-s .10 mm). Histologic findings of the four groups of CRCs are summarized in Table 1. As indicated in the table, neither the site, nor the frequencies of vascular invasion and moderate differentiation were different among the four groups. In contrast, there were significant differences in the frequencies of adenoma and non-polypoid growth among the four groups. CRCs in group A were less frequently accompanied by adenoma (12%) than those in group D (63%). Non-polypoid growth was more frequent

Table 1. Comparison of clinicopathologic findings among the four groups of colorectal cancer Group A (sm-d¡10 mm) [n517] Location Rectum Sigmoid colon Proximal colon Adenoma Present Absent Lymphatic permeation Present Absent Venous permeation Present Absent Grade of cancer Well Well.moderate Well,moderate Growth pattern Polypoid Non-polypoid a,b,c,d

Group B (sm-s ¡10 mm) [n528]

Group C (sm-d .10 mm) [n549]

Group D (sm-s .10 mm) [n519]

2 8 7

4 12 12

14 18 17

5 10 4

p50.45

2 15

9 19

8 41

12a 7a

p,0.0005

5 12

3 25

16 33

2 17

p50.057

3 14

3 25

16 33

2 17

p50.064

15 1 1

28 0 0

39 2 8

16 2 1

p50.086

2 15

18b 10b

21c 28c

17d 2d

p,0.0001

Significance

The frequency is significantly different when compared with group A (ap50.0019; bp50.0006; cp50.0018; dp,0.0001).


613


in group A (88%) than in groups B (36%), C (57%) and D (11%).

Reproducibility of radiographic findings Table 2 indicates a comparison of determination in radiographic findings between the two observers. The overall concordance rates were 90.3% in surface feature, 92.9% in central barium fleck, and 90.3% in fold convergency. For each radiographic criterion, the determination was not significantly different between the two observers. The rate of nodular surface nature was the highest in group D (94.7%) and those of central barium fleck and fold convergency were the highest in group B (100% and 92.9%). However, there was no significant difference in the concordance rates of the three radiographic findings among the four groups. There were CRCs in which the eccentric rigidity was regarded to be undetermined. The overall concordance rate in the determination of eccentric rigidity was 79.6%. There was no significant difference in the concordance rate among the four groups.

Predictive value of radiographic findings PPV and NPV of concordant radiographic findings for the diagnosis of sm-d CRCs are indicated in Table 3. In CRCs ¡10 mm in size, PPV for deep invasion was the highest in eccentric rigidity, followed by smooth surface. In contrast, PPV of fold convergency was the highest in CRCs .10 mm. PPVs of smooth surface and fold convergency were higher in CRC .10 mm in size than in those ¡10 mm, but the differences did not reach statistical significance. PPV of central barium fleck was significantly higher in CRCs .10 mm (0.86) than in those ¡10 mm (0.64, p50.034). In contrast to PPV, there were obvious differences in NPVs of radiographic determinants for sm-d CRCs. As has been indicated in Table 3, NPV of all the four determinants were significantly different between CRCs .10 mm and those ¡10 mm. Table 4 compares predictive values of combined radiographic items for sm-d CRCs. In CRCs .10 mm, NPVs for sm-d CRCs increased when two of three items were combined. Whereas PPVs and NPVs of smooth surface plus central barium fleck and smooth surface plus fold convergency were not statistically different between CRCs ¡10 mm and .10 mm, NPV of central barium fleck plus

Table 2. Concordance rate of determination in radiographic findings between two observers Group A (sm-d ¡10 mm) [n517]

Group B (sm-s ¡10 mm) [n528]

Group C (sm-d .10 mm) [n549]

Group D (sm-s .10 mm) [n519]

Total [n5113]

Surface nature Smooth Nodular Total (%) [non-concordance (%)]

14 2 16 (94.1) [5.9]

4 21 25 (89.3) [10.7]

18 25 43 (87.8) [12.2]

3 15 18 (94.7) [5.3]

39 63 102 (90.3) [9.7]

Central barium fleck Present Absent Total (%) [non-concordance (%)]

14 2 16 (94.1) [5.9]

8 20 28 (100.0) [0]

25 18 43 (87.8) [12.2]

4 14 18 (94.7) [5.3]

51 54 105 (92.9) [7.1]

Fold convergency Present Absent Total (%) [non-concordance (%)]

2 13 15 (88.2) [11.8]

1 25 26 (92.9) [7.1]

13 31 44 (89.8) [10.2]

1 16 16 (89.5) [10.5]

17 85 102 (90.3) [9.7]

Eccentric rigidity Undetermined Absent Present Total (%) [non-concordance (%)]

2 1 10 13 (76.5) [23.5]

8 12 1 21 (75.0) [25.0]

11 4 25 40 (81.6) [18.4]

4 9 3 17 (84.2) [15.8]

25 26 39 90 (79.6) [20.4]

Table 3. Predictive values of radiographic findings for deep invasion

Smooth surface Central barium fleck Fold convergency Eccentric rigiditya

Groups A and B (¡10 mm) [n545]

Groups C and D (.10 mm) [n568]

Significance

PPV

NPV

PPV

NPV

PPV

NPV

0.78(14/18) 0.64(14/22) 0.67(2/3) 0.91(10/11)

0.89(24/27) 0.87(20/23) 0.64(27/42) 0.79(19/24)

0.86(18/21) 0.86(25/29) 0.93(13/14) 0.89(25/28)

0.34(16/47) 0.38(15/39) 0.33(18/54) 0.48(12/25)

0.41 0.034 0.33 0.69

,0.0001 0.0002 0.0023 0.023

PPV, positive predictive value; NPV, negative predictive value. Colorectal cancers, in which the eccentric rigidity was classified to be undetermined, were excluded.

a

614


Radiography for small invasive colorectal cancer Table 4. Predictive values of combined radiographic findings for deep invasion Groups A and B (¡10 mm) [n545]

Groups C and D (.10 mm) [n568]

Significance

PPV

NPV

PPV

NPV

PPV

NPV

0.83(29/35) 0.67(28/42) 0.64(27/42)

0.88(15/17) 0.78(7/9) 1.0(10/10)

0.92(47/51) 0.80(47/59) 0.84(49/58)

0.59 0.55 0.23

0.16 0.11 0.018

Smooth surface+central barium fleck 0.85(11/13) Smooth surface+fold convergency 1.0(3/3) Central barium fleck+fold convergency 0.67(2/3)

PPV, positive predictive value; NPV, negative predictive value.

fold convergency was significantly higher in CRCs .10 mm than in CRCs ¡10 mm.

Discussion Increasing evidence suggests that there are CRCs that invade the deep submucosal layer whilst small in size. In a large survey by Tanaka et al [23], 113 of 470 CRCs with submucosal invasion were under 20 mm in size. As small CRCs are candidates for endoscopic removal, pre-therapeutic determination of the invasion depth in such CRCs is desirable in order to avoid inappropriate choice of therapy. Recent advances in colonoscopic procedures revealed that EUS and magnifying colonoscopy can provide detailed information for the assessment of invasion depth in CRCs [24]. Comparisons of clinicopathologic features in our CRCs indicated that non-polypoid growth and negative adenomatous component were histologic features predominant in sm-d CRCs ¡10 mm. Of note is the fact that nonpolypoid growth was more frequent in sm-d CRCs ¡10 mm than in any other groups of CRCs. This result is consistent with an observation reported by Kurisu et al [25], in which the size of non-polypoid submucosally invasive CRCs were significantly smaller than polypoid ones. In addition, since non-polypoid growth has been proven to be an unfavourable pathological feature predictive of lymph node metastasis [26, 27], there seems to be a need for special attention for the diagnosis of invasion depth in CRCs ¡10 mm in size. Based upon the result of the present investigation, deep invasion in CRCs ¡10 mm may be predicted by BE, according to the scheme applied for the non-polypoid CRCs [19]. In contrast to endoscopy, the value of BE for the diagnosis of the invasion depth in CRC has not been investigated. While BE is a conventional procedure available for screening and diagnosis of CRCs, there have been few descriptions of the precise analysis of the depicted findings [19, 28]. However, Watari et al [19] have shown recently that in non-polypoid cancers, deep depression, fold convergency and eccentric deformity are radiographic findings specific to CRC with submucosal invasion. The investigators reported an accuracy of 85% in BE for the diagnosis of submucosally invasive CRCs. The observation by Watari et al [19] led us to examine possible application of BE for small CRCs. Our results indicated that BE was likely to be more predictive of invasion depth for CRCs ¡10 mm than for those .10 mm, since NPV in each radiographic sign was significantly higher in the former. Since the determination in each radiographic finding is not completely objective, we initially assessed interobserver variations. Although the difference was not statistically significant, there were CRCs in which the determination The British Journal of Radiology, September 2003

dissociated. This was especially the case for eccentric rigidity, which has been shown to be diagnostic of invasion depth in CRC regardless of the configuration [18, 19]. The small size of the CRC in the present investigation seems to have further contributed to the dissociation. In addition, the profile view of CRC, which is necessary for the assessment of the eccentric deformity, could not always be obtained by BE [19]. Although eccentric deformity had the highest PPV and NPV for the diagnosis of deep invasion in CRCs in which the sign could be assessed, it is not a criterion uniformly applicable on all occasions. We have shown previously that by means of BE, small non-polypoid neoplastic lesions were depicted as a radiolucent area with central barium fleck [20]. Because the barium flecks could be depicted even in lesions under 5 mm in size [20], it seems reasonable that BE was able to demonstrate the depression in larger and invasive CRCs. In the present investigation, the PPV of central barium fleck for deep invasion was higher in CRCs .10 mm than in those ¡10 mm. This observation suggests that the central barium fleck under BE does not necessarily indicate deep invasion, but rather it is representative of nonpolypoid growth. Furthermore, the higher NPV of the central barium fleck for deep invasion in CRCs ¡10 mm seems to suggest that polypoid lesions rarely invade deeply whilst small in size. In contrast to central barium fleck, the PPV of smooth surface in CRCs ¡10 mm was equivalent to and the NPV was higher than in CRCs .10 mm in size. Because the nonpolypoid growth was predominant in deeply invasive CRCs ¡10 mm, the smooth surface seems to be representative of surrounding non-neoplastic mucosa. Furthermore, colonoscopic and stereomicroscopic investigations revealed that the surface of deeply invasive CRCs was characterized by distorted, or amorphous pattern [11, 14, 24, 29]. It thus seems likely that the smooth surface on barium radiography is indicative of either the non-polypoid growth, deep invasion, or both. Although Watari et al [19] reported that fold convergency is a radiographic sign specific to submucosal invasion, the sensitivity of the radiographic finding has been shown to be 29% in flat CRCs and 50% in depressed CRCs. In accordance with this previous description, the PPV was the highest and the NPV was the lowest for the diagnosis of deep invasion in CRCs ¢10 mm in size. Furthermore, fold convergency was positive in only 3 of the 45 CRCs ¡10 mm. These observations suggest that fold convergency is insufficient for the diagnosis of invasion depth in small CRCs, whereas the sign is more specific for deep invasion. In conclusion, our results indicated that an analysis of BE with respect to smooth surface, central barium fleck, fold convergency and eccentric deformity has diagnostic 615


value for invasion depth in CRCs ¡10 mm, which was equivalent to or even superior to that in CRCs .10 mm. It was also suggested that BE is predictive of invasion depth in CRCs .10 mm, because the diagnostic values improved when any radiographic criterion was combined with other ones. The greater significance of a radiographic finding for the diagnosis of deep invasion in CRCs ¡10 mm than in CRCs .10 mm seems to have been attributed to the less heterogeneity in histology of deeply invasive cancer in the former group. While there is a need for a comparison of accuracy and cost-effectiveness between colonoscopy and BE, the latter still remains to be a procedure complementary to the former for pre-therapeutic assessment in small invasive CRCs, which are apparently candidates for endoscopic removal.

References 1. Takahashi H, Mitomi H, Igarashi M, Katsumata T, Okayasy I. Markedly elevated cell turnover is characteristic of small, deeply invasive carcinomas of the colorectum. Cancer 1999; 85:796–802. 2. Cooper HS, Deppisch LM, Gourley, Kahn EI, Lev R, Manley PN, et al. Endoscopically removed malignant colorectal polyps. Clinicopathologic correlations. Gastroenterology 1995;108:1657–65. 3. Nivatongs S, Rojanasakul A, Reiman HM, Dozois RR, Wolff BG, Pemberton JH, et al. The risk of lymph node metastasis in colorectal polyps with invasive adenocarcinoma. Dis Colon Rectum 1991;34:323–8. 4. Haggitt RC, Glotzbach RE, Soffer EE, Wirbule LD. Prognostic factors in colorectal carcinomas arising in adenomas: implications of lesions removed by endoscopic polypectomy. Gastroenterology 1985;89:328–36. 5. Kyzer S, Begin LR, Gordon PH, Mitmaker B. The care of patients with colorectal polyps that contain invasive adenocarcinoma: endoscopic polypectomy or colectomy? Cancer 1992;70:2044–50. 6. Cho E, Nakajima M, Yasuda K, Ashihara T, Kawai K. Endoscopic ultrasonography in the diagnosis of colorectal cancer invasion. Gastrointest Endosc 1993;39:521–7. 7. Hizawa K, Suekane H, Aoyagi K, Matsumoto T, Nakamura S, Fujishima M. Use of endosonographic evaluation of colorectal tumor depth in determining the appropriateness of endoscopic mucosal resection. Am J Gastroenterol 1996; 91:768–71. 8. Saitoh Y, Obara T, Einami K, Nomura M, Taruishi M, Ayabe T, et al. Efficacy of high-frequency ultrasound probes for the preoperative staging of invasion depth in flat and depressed colorectal tumors. Gastrointest Endosc 1996;44: 34–9. 9. Hu¨nerbein M, Totkas S, Ghadimi BM, Schlag PM. Preoperative evaluation of colorectal neoplasms by colonoscopic miniprobe ultrasonography. Ann Surg 2000;232:46–50. 10. Harada N, Hamada S, Kudo H, Oda S, Chijiiwa Y, Kabemura T, et al. Preoperative evaluation of submucosal invasive colorectal cancer using a 15-MHz ultrasound miniprobe. Endoscopy 2001;33:237–40. 11. Kudo S, Tamura S, Nakajima T, Yamano H, Kusaka H, Watanabe H. Diagnosis of colorectal tumorous lesions by magnifying endoscopy. Gastrointest Endosc 1996;44:8–14. 12. Kato S, Fujii T, Koba I, Sano Y, Fu KI, Parra-Blanco A, et al. Assessment of colorectal lesions using magnifying colonoscopy and mucosal dye spraying: can significant lesions be distinguished? Endoscopy 2001;33:306–10.

616

13. Togashi K, Konishi F, Ishizuka T, Sato T, Senba S, Kanazawa K. Efficacy of magnifying colonoscopy in the differential diagnosis of neoplastic and non-neoplastic polyps of the large bowel. Dis Colon Rectum 1999;42:1602–8. 14. Nagata S, Tanaka S, Haruma K, Yoshihara M, Sumii K, Kajiyama G, et al. Pit pattern diagnosis of early colorectal carcinoma by magnifying colonoscopy: clinical and histological implications. Int J Oncol 2000;16:927–34. 15. McClave SA, Jones WF, Woolfolk GM, Schrodt GR, Wiersema MJ. Mistakes on EUS staging of colorectal carcinoma: error in interpretation or deception from innate pathologic features? Gastrointest Endosc 2000;51:682–9. 16. Catalano MF, Sivak Jr MV, Bedford RA, Falk GW, van Stolk R, Presa F, et al. Observer variation and reproducibility of endoscopic ultrasonography. Gastrointest Endosc 1995;41: 115–20. 17. Roubein LD, Lynch P, Glober G, Sinicrope FA. Interobserver variability in endoscopic ultrasonography: a prospective evaluation. Gasrointest Endosc 1996;44:573–7. 18. Matsui T Yao T, Yao K, Takenaka K, Sakurai T, Iwashita A, et al. Natural history of superficial depressed colorectal cancer. Retrospective radiographic and histologic analysis. Radiology 1996;201:226–32. 19. Watari J, Saitoh Y, Obara T, Fujiki T, Taniguchi M, Nomura M, et al. Early nonpolypoid colorectal cancer. Radiographic diagnosis of depth of invasion. Radiology 1997;205:67–74. 20. Matsumoto T, Iida M, Kohrogi N, Tada S, Kuwano Y, Yao T, et al. Minute nonpolypoid adenomas of the colon depicted with barium enema examination. Radiology 1993;187:377–80. 21. Kikuchi R, Takano M, Takagi K, Fujimoto N, Nozaki R, Fujiyoshi T, et al. Management of early invasive colorectal cancer. Risk of recurrence and clinical guidelines. Dis Colon Rectum 1995;38:1286–95. 22. Shimoda T, Ikegami M, Fujisaki J, Matsui T, Aizawa S, Ishikawa E. Early colorectal carcinoma with special reference to its development de novo. Cancer 1989;64:1138–46. 23. Tanaka S, Haruma K, Ohe H. Conditions of curability after endoscopic resection for colorectal carcinoma with submucosally massive invasion. Oncol Rep 2000;7:783–8. 24. Matsumoto T, Hizawa K, Esaki M, Kurahara K, Mizuno M, et al. Comparison of EUS and magnifying colonoscopy for the assessment of invasion depth in small colorectal cancer. Gastrointest Endosc 2002;56:354–60. 25. Kurisu Y, Shimoda T, Ochiai A, Nakanishi Y, Hirata I, Katsu K. Histologic and immunohistochemical analysis of early submucosal invasive carcinoma of the colon and rectum. Pathology Int 1999;49:608–16. 26. Minamoto T, Mai M, Ogino T, Sawaguchi K, Ohta T, Fujimoto T, et al. Early invasive colorectal carcinomas metastatic to the lymph node with attention to their nonpolypoid development. Am J Gastroenterol 1993;88:1035–9. 27. Tanaka S, Haruma K, Teixeira CR, Tatsuta S, Ohtsu N, Hiraga Y, et al. Endoscopic treatment of submucosal invasive colorectal carcinoma with special reference to risk factors for lymph node metastasis. J Gastroenterol 1995;30:710–7. 28. McCarthy PA, Rubesin SE, Levine MS, Langlotz CP, Laufer I, Furth EE, et al. Colon cancer: morphology detected with barium enema examination versus histopathologic stage. Radiology 1995;197:683–7. 29. Tada S, Iida M, Yao T, Matsumoto T, Tsuneyoshi M, Fujishima M. Stereomicroscopic examination of surface morphology in colorectal epithelial tumors. Hum Pathol 1993;24:1243–52.
