Advertisement

Endoscopic Scoring System for T2 Invasion in Colorectal Cancer

Open AccessPublished:November 19, 2021DOI:https://doi.org/10.1016/j.tige.2021.11.005

      Abstract

      Background and Aims

      : The depth of tumor invasion in colorectal cancer (CRC), especially T1b or T2, is crucial in treatment decision-making. However, their differences are not well-characterized. Thus, this study aimed to investigate the predictive endoscopic findings in tumor invasion of CRC.

      Findings

      : T2 invasion was predicted by 6 characteristics. The risk scoring system was developed using the regression coefficient values of the above variables. The area under the ROC curve was 0.894. Cases with a score ≥4 had a high risk of T2.

      Implications for patient care

      : our scoring system is a valid tool for predicting tumor invasion and distinguishing between T1b and T2. The indications for ESD may be expanded especially in elderly people by these findings.

      Methods

      : Data from patients with T1b or T2 CRCs resected endoscopically or surgically were reviewed retrospectively. The patients were divided into two groups: T1b (n=298) and T2 (n=267) tumor invasion. A scoring system was established based on the endoscopic findings in each group, and the accuracy of the system was assessed using a receiver-operating-characteristic (ROC) curve analysis.

      Results

      : T2 invasion was predicted by tumor size, irregular bottom of depression, existence of depression, expansion appearance, convergency of folds, and erosion or white coat. The risk scoring system was developed using the regression coefficient values of the above variables. The area under the ROC curve was 0.894 (95% confidence interval, 0.868–0921). Cases with a score ≥4 had a high risk of T2 (sensitivity, 84.5%; specificity, 78.9%).

      Conclusion

      : Our scoring system was useful for the diagnosis of T1b and T2, and a score ≥4 could predict T2 invasion. Additional studies are warranted to confirm these results before our scoring system can be applied clinically.

      Keywords

      Introduction

      Colorectal cancer (CRC) is the third most common cancer globally, accounting for more than 1.8 million new cases and almost 900,000 deaths each year.
      • Bray F
      • Ferlay J
      • Soerjomataram I
      • et al.
      Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
      Endoscopic resection (ER) for early-stage CRC is recommended, and it is a widely accepted type of treatment.
      • Zauber AG
      • Winawer SJ
      • O'Brien MJ
      • et al.
      Colonoscopic polypectomy and long-term prevention of colorectal-cancer deaths.
      ,
      • Tamegai Y
      • Saito Y
      • Masaki N
      • et al.
      Endoscopic submucosal dissection: a safe technique for colorectal tumors.
      The curability of the tumor can be estimated using pathological factors according to the Japanese Society for Cancer of the Colon and Rectum Guidelines.

      Japanese Society for Cancer of the Colon and Rectum (ed.). Japanese Classification of Colorectal Carcinoma, Appendiceal, and Anal Carcinoma, 3rd English edn Tokyo, Japan: Kanehara. 2019.

      If the lesion is invasive, negative vertical and horizontal margins and the following four factors are required for resection to be considered curative: depth of submucosal invasion <1000 μm, well and/or moderately differentiated adenocarcinoma, negative lymphovascular invasion, and a budding grade of 1.
      In contrast, T1b, depth of submucosal invasion ≥1000, colorectal cancer is not included as an indication for ER because of the possibility of lymph node metastasis.
      • Kitajima K
      • Fujimori T
      • Fujii S
      • et al.
      Correlations between lymph node metastasis and depth of submucosal invasion in submucosal invasive colorectal carcinoma: a Japanese collaborative study.
      Therefore, an additional surgical procedure is required when the resected lesion reveals submucosal invasion histologically, since submucosal invasion suggests the presence of lymph node metastasis.
      • Kitajima K
      • Fujimori T
      • Fujii S
      • et al.
      Correlations between lymph node metastasis and depth of submucosal invasion in submucosal invasive colorectal carcinoma: a Japanese collaborative study.
      ,
      • Coverlizza S
      • Risio M
      • Ferrari A
      • et al.
      Colorectal adenomas containing invasive carcinoma. Pathologic assessment of lymph node metastatic potential.
      However, several patients with T1b CRC, especially elderly patients, are considered for ER even when they do not wish to undergo surgery or there is no indication for surgery. In fact, ER for T1b CRC is occasionally selected as one of the treatments.
      T2 CRC is defined when there is invasion into the muscularis propria. It is generally considered that T2 CRC cannot be resected completely by ER. It is recommended that patients with T2 CRC undergo complete tumor resection.
      Thus, it is critical for patients with T2 CRC to be distinguished from patients with T1b so that complete tumor resection is performed. Previous studies have indicated that T1a can be distinguished from T1b using endoscopic findings.
      • Horie H
      • Togashi K
      • Kawamura YJ
      • et al.
      Colonoscopic stigmata of 1 mm or deeper submucosal invasion in colorectal cancer.
      ,
      • Hurlstone DP
      • Cross SS
      • Adam I
      • et al.
      Endoscopic morphological anticipation of submucosal invasion in flat and depressed colorectal lesions: clinical implications and subtype analysis of the kudo type V pit pattern using high-magnification-chromoscopic colonoscopy.
      However, the endoscopic difference between T1b and T2 invasion remains unclear. Therefore, in this study, we assessed the characteristic factors of T2 invasion and aimed to create a scoring system for the diagnosis of T2 invasion.

      Methods

       Patients

      A retrospective study was performed to examine the predictive factor of tumor invasion in patients with CRC (T1b or T2). We conducted an image evaluation study of digital files containing endoscopic images of localized neoplastic lesions in patients who had been treated with ER, including endoscopic mucosal resection, endoscopic submucosal dissection (ESD), or surgical operation between April 2008 and April 2020. The exclusion criteria were pedunculated lesions, neoplastic lesions associated with inflammatory bowel disease and familial adenomatous polyposis, neoplastic lesions for which patients had undergone previous chemotherapy or chemoradiotherapy, and recurrence of lesions after treatment procedures. The clinical and clinicopathological information was collected from the hospital records and was reviewed retrospectively.
      This study protocol was approved by the Institutional Review Board at the Tokyo Metropolitan Cancer and Infectious Diseases Center at Komagome Hospital and followed the principles outlined in the Declaration of Helsinki.

       Endoscopic assessment

      The primary objective of our study was to develop a scoring system that differentiates T1b from T2 CRC. Initially, experienced colonoscopists (AS and KK) interpreted the images obtained from 653 cases. Histological diagnoses were reviewed by a pathologist (SH) with expertise in the field of colorectal tumors. Histopathology was defined according to the World Health Organization classification system.
      • Nagtegaal ID
      • Odze RD
      • Klimstra D
      • et al.
      The 2019 WHO classification of tumours of the digestive system.
      Tumors from the cecum to the transverse colon were defined as right-sided cancers. Tumors located in the rectosigmoid junction or within the rectum were considered rectal cancers. Tumors were classified into several types: 0 (including I p, I sp, I s, II a, II b, IIc, I s + II c, II a + II sp, II a + II c, and II c + II a), 1, 2, 3, and 4, according to the Japanese Classification of Colorectal, Appendiceal, and Anal Carcinoma.

      Japanese Society for Cancer of the Colon and Rectum (ed.). Japanese Classification of Colorectal Carcinoma, Appendiceal, and Anal Carcinoma, 3rd English edn Tokyo, Japan: Kanehara. 2019.

      The pathological types of tumors are defined as follows: tub1, well-differentiated adenocarcinoma; tub2, moderately differentiated adenocarcinoma; and por, poorly differentiated adenocarcinoma. In addition, mucinous carcinoma and signet-ring cell carcinoma were classified into other types of carcinomas.
      Endoscopic findings were assessed as follows: tumor size, unclear lobulation, deep depression surface, irregular bottom of depression, existence of depression, expansion appearance, convergency of folds, and erosion or white coat with reference to a previous study.
      • Saitoh Y
      • Obara T
      • Watari J
      • et al.
      Invasion depth diagnosis of depressed type early colorectal cancers by combined use of videoendoscopy and chromoendoscopy.
      Endoscopic characteristics are shown in Figure 1. Tumor size was recorded as the maximum horizontal tumor diameter, as measured from formalin-fixed tumor samples. Unclear lobulation was defined as loss of normal lobulation in the tumor surface. Deep depression surface was defined as deep depressed areas on the surface of tumor, 3 mm or deeper from the edge of the tumor. Irregular bottom of depression was defined as bumpy surface in the depressed area of the tumor. Existence of depression was described as a depressed demarcation within the lesion that was evident without indigocarmine spray. Expansion appearance was determined as a firm, hardly deformable appearance, filled to bursting. Fold convergence was defined as the convergence of at least four mucosal folds toward a central lesion and erosion or white coat as larger than the quarter of area covered with an apparent surface bleeding or white coat.
      Figure 1
      Figure 1Examples of typical colonoscopic findings are shown. Unclear lobulation (A) was defined as loss of normal lobulation in the tumor surface. Deep depression surface (B) was defined as deep depressed areas on the surface of tumor, 3mm or deeper from the edge of the tumor. irregular bottom of depression (C) was defined as bumpy surface in the depressed area of the tumor. Existence of depression (D) was defined as a depressed demarcation within the lesion that was evident without indigocarmine spray. Expansion appearance (E) was defined as a firm, hardly deformable appearance, filled to bursting. Fold convergence (F) was defined as the convergence of at least four mucosal folds toward a central lesion. Erosion or white coat (G) was defined as an area covered with an apparent surface bleeding or white coat.
      The evaluation of these endoscopic characteristics was independently performed by two colonoscopists. The third reviewer further evaluated the endoscopic characteristics when the opinion of the two colonoscopists were divided.

       Statistical analysis

      Statistical comparisons of baseline characteristics between T1b and T2 invasion cases were performed using the χ2 test or Fisher's exact test for categorical data and the Student's t test or Mann–Whitney's test for continuous data. All statistical analyses were performed with 5% alpha risk or 95% confidence intervals using SPSS version 25 (IBM, Chicago, IL). Multivariate analysis was conducted to adjust the odds ratio using endoscopic findings which showed difference between T1b and T2 invasion.
      ROC curves and the area under the curve (AUC) were analyzed to determine the accuracy of the scoring system. An AUC of 1.0 was an error-free prediction of cancer in patients, whereas an AUC of 0.50 represents a half likelihood of an accurate prediction of cancer invasion. The higher the AUC-ROC, the bigger the discriminatory power of the scoring system.

      Results

       Patient characteristics

      Among 639 patients with CRC and 653 lesions treated with endoscopy and surgery between April 2008 and April 2020, 88 cases were excluded because of the above criteria. Thus, in total, 552 patients with 565 lesions were eligible for inclusion in this study (Fig. 2).
      Table 1 shows the patient characteristics. The median patient age was 69.0 (range, 61.0–76.0) years, and 326 patients (57.7%) were males. Tumor location was not significantly different between the two groups. Regarding histology, the frequency of tub2 (moderately differentiated adenocarcinoma) was higher in the T2 group than in the T1b group (59.6% vs. 30.9%, P<0.001). Vascular invasion was significantly higher in the T2 group than in the T1b group (63.1% vs. 39.6%, P<0.001). Conversely, there were no significant differences in lymphatic invasion and lymph node metastases between the two groups.
      Table 1Patient characteristics in the both group
      featuresT1b (n = 298)T2 (n = 267)P value
      Age, median69.068.500.78
      Male, n (%)169 (56.7)157 (58.8)0.67
      Location
       Cecum17 (5.7)14 (5.2)0.03
       Ascending colon39 (13.1)26 (9.7)
       Transverse colon30 (10.1)20 (7.5)
       Descending colon19 (6.4)3 (1.1)
       Sigmoid colon86 (28.9)74 (27.7)
       Rectum107 (35.9)130 (48.7)
      Location
      Right side86 (28.9)60 (22.5)0.102
      Left side212 (71.1)207 (77.5)
      Histology
       Tub1202 (67.8)102 (38.2)<0.001
       Tub292 (30.9)159 (59.6)
      por2 (0.7)1 (0.4)
       other2 (0.7)5 (1.9)
      Morphology
      0-Ⅰp29 (9.7)1 (0.4)<0.001
       0-Ⅰsp64 (21.5)15 (5.7)
       0-Ⅰs108 (36.2)36 (13.7)
       0-Ⅱa39 (13.1)3 (1.1)
      0-Ⅱb0 (0)0 (0)
      0-Ⅱc2 (0.7)0 (0)
      0-Ⅰs+Ⅱc19 (6.4)6 (2.3)
      0-Ⅱa+Ⅰsp0 (0)1 (0.4)
      0-Ⅱa+Ⅱc24 (8.1)11 (4.2)
      0-Ⅱc+Ⅱa1 (0.3)0 (0)
      12 (0.7)55 (20.9)
      29 (3.0)131 (49.8)
      31 (0.3)3 (1.1)
      40 (0)0 (0)
      Primary treatment
       Endoscopic resection187 (62.8)4 (1.5)<0.001
      Surgery111 (37.2)263 (98.5)
      Lymphatic invasion86 (28.9)69 (25.7)0.450
      Vascular invasion118 (39.6)169 (63.1)<0.001
      Lymph node metastases (503)36 (15.3)59 (22.1)0.053

       Risk factors for T2 invasion

      The univariate and multivariate predictive powers of T2 risk factors are shown in Table 2. Univariate comparison of the endoscopic characteristics revealed that tumor size, unclear lobulation, deep depression surface, irregular bottom of depression, expansion appearance, convergency of folds, and erosion or white coat were significantly different in the T1b and T2 groups. In addition, multiple logistic regression analysis demonstrated six independent variables that were significantly associated with T2 CRC: tumor size, irregular bottom of depression, existence of depression, expansion appearance, convergency of folds, and erosion or white coat.
      Table 2endoscopic findings in the both group
      featuresT1b (n = 298)T2 (n = 267)Univariate analysesMultivariate analyses
      Odds Ratio (95% CI)P valueOdds Ratio (95% CI)P value
      Tumor size, ≥ median, n (%)102 (34.2)183 (68.5)4.19 (2.94, 5.95)<0.0014.85 (2.95, 8.00)<0.001
      Unclear lobulation203 (68.1)237 (88.8)3.70 (2.35, 5.81)<0.0011.78 (0.96, 3.30)0.069
      Deep depression surface (≥ 3mm)14 (4.7)78 (29.2)8.37 (4.60, 15.2)<0.0011.89 (0.85, 4.23)0.12
      Irregular bottom of depression8 (2.7)59 (22.1)10.3 (4.81, 22.0)<0.0013.36 (1.35, 8.36)0.009
      Existence of depression48 (16.1)168 (63.2)8.93 (6.00, 13.3)<0.0015.15 (2.94, 9.03)<0.001
      Expansion appearance256 (85.9)252 (94.4)2.76 (1.49, 5.10)0.0012.32 (1.00, 5.37)0.049
      Convergency of folds74 (24.8)205 (76.8)10.0 (6.80, 14.7)<0.0018.77 (5.39, 14.3)<0.001
      Erosion or white coat111 (37.2)164 (61.4)2.68 (1.91, 3.77)<0.0011.97 (1.22, 3.19)0.006
      Based on the method used to evaluate the six independent factors, the following formula was obtained:
      Risk Score of T2 invasion = (2 × tumor size) + (2 × existence of depression) + (2 × Convergency of folds) + (1 × irregular bottom of depression) + (1 × expansion appearance) + (1 × erosion or white coat)
      This formula uses the coefficients of the regression analysis reported in Table 2. A tumor size ≥ median, existence of depression, convergency of folds, irregular bottom of depression, expansion appearance, and erosion or white coat were attributed with scores of 2, 2, 2, 1, 1, and 1, respectively. According to the above formula, the risk score for a single patient can be within the range of 0–9. The discriminate validity for the risk score was assessed (P<0.001) and ROC curve analysis was used for the evaluation of sensitivity and specificity of the scores for all T1b and T2 groups. It revealed that cases with a score ≥4 had a high risk of T2 (sensitivity, 84.5%; specificity, 78.9%; and AUC, 0.894) (Fig. 3).
      Figure 3
      Figure 3The ROC curve was constructed by SPSS. Area under curve was 0.894 (95% CI, 0.868-0.921). P<0.001.

      Discussion

      In this study, we evaluated the efficacy of a scoring system based on endoscopic findings in patients with CRC. To our knowledge, this is the first report of the accuracy of such a scoring system for the prediction of T2 CRC. The strong efficacy of this tool for the prediction of T2 revealed by this study may be due to the cumulative effect of several endoscopic characteristics (such as tumor size, existence of depression, convergency of folds, irregular bottom of depression, expansion appearance, and erosion or white coat). The sensitivity and specificity of this scoring system based on a cut-off score of 4 points have been estimated to be 84.5% and 78.9%, respectively. A previous study reported that experienced endoscopists can accurately estimate the depth of invasion of about 80% of lesions on conventional endoscopy alone.
      • Haruki S
      • Kobayashi K
      • Yokoyama K
      • et al.
      Comparison of diagnostic accuracies of various endoscopic examination techniques for evaluating the invasion depth of colorectal tumors.
      Our study demonstrated that the accuracy of our scoring system was higher than conventional endoscopy assessed by experienced endoscopists. Additionally, the ROC curve for risk score revealed an AUC of 0.894 for distinguishing T2 invasion from T1b, which provides evidence that our scoring system is a useful indicator for differentiating tumor invasion.
      According to the guidelines, additional surgery is recommended for all patients after ESD when the current curative criteria, including a depth of submucosal invasion <1000 μm for colorectal cancer, are not met. However, sometimes patients, especially elderly patients, do not wish to undergo additional surgery. Moreover, the physicians decide not to perform a surgical procedure in consideration of the patient's characteristics, such as age, performance status, and comorbidities. In fact, there is only a 10%–15% risk of lymph node metastasis when colorectal cancer involves submucosal invasion of 1000 μm or deeper.
      • Kitajima K
      • Fujimori T
      • Fujii S
      • et al.
      Correlations between lymph node metastasis and depth of submucosal invasion in submucosal invasive colorectal carcinoma: a Japanese collaborative study.
      In addition, the postoperative mortality rates are reported to be 1.4% and 1.3% for right and left hemicolectomy, respectively.
      • Hinojosa MW
      • Konyalian VR
      • Murrell ZA
      • et al.
      Outcomes of right and left colectomy at academic centers.
      Therefore, it is not a contraindication for patients with CRC to undergo ER if the tumor shows submucosal invasion of 1000 μm or deeper. However, ER for T2 CRC is commonly a contraindication because the tumor grows into the muscularis propria. In this study, 81.3% of cases achieved pathological complete resection by ER in the T1b group, whereas no case achieved pathological complete resection by ER in the T2. The results show that it is crucial to distinguish T1b from T2.
      Endoscopic ultrasonography (EUS) has been considered a useful technique for the diagnosis of invasion depth in patients with colorectal cancer. In a previous study, the sensitivity of EUS was 90% and the specificity was 87%.
      • Mukae M
      • Kobayashi K
      • Sada M
      • et al.
      Diagnostic performance of EUS for evaluating the invasion depth of early colorectal cancers.
      However, EUS-based diagnosis requires high-cost equipment and can be challenging even for experienced endoscopists. In this context, our scoring system can be useful in estimating the depth of tumor invasion as it simply divides cases into T1b and T2.
      It is critical to note the limitations of our study. First, this was a retrospective study of endoscopic data from a single institution with a limited sample size, which has a potential for bias. Second, each endoscopic characteristic was assessed by only two physicians. Therefore, it is possible that there was bias in the assessment of endoscopic findings. Third, there were several types of lesions, such as 0-I p, 0-I s, 0-II a, etc., assessed by endoscopic examination in this study. Tumor invasions may be different depending on endoscopic findings. Given these limitations, the current study only generates a hypothesis, and a more detailed system needs to be further evaluated.
      In conclusion, our scoring system is a valid tool for predicting tumor invasion and distinguishing between T1b and T2. These findings deserve further investigation in a larger cohort to validate the efficacy of this scoring system.

      Ethical Statement

      The corresponding author, on behalf of all authors, jointly and severally, certifies that their institution has approved the protocol for any investigation involving humans or animals and that all experimentation was conducted in conformity with ethical and humane principles of research.

      CRediT authorship contribution statement

      Akinori Sasaki: Conceptualization, Data curation, Formal analysis, Methodology. Ryoko Shimizuguchi: Data curation, Project administration, Visualization, Writing – original draft, Writing – review & editing. Akinari Takao: Data curation. Satomi Shibata: Data curation. Souichiro Natsume: Data curation. Shin-ichiro Horiguchi: Data curation. Daisuke Nakano: Data curation. Tatsuro Yamaguchi: Data curation. Koichi Koizumi: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Visualization, Writing – review & editing.

      Declaration of Competing Interests

      All authors have no conflicts of interests or financial disclosures relevant to the manuscript.

      Funding

      No source of funding

      References

        • Bray F
        • Ferlay J
        • Soerjomataram I
        • et al.
        Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.
        CA Cancer J Clin. 2018; 68: 394-424
        • Zauber AG
        • Winawer SJ
        • O'Brien MJ
        • et al.
        Colonoscopic polypectomy and long-term prevention of colorectal-cancer deaths.
        N Engl J Med. 2012; 366: 687-696
        • Tamegai Y
        • Saito Y
        • Masaki N
        • et al.
        Endoscopic submucosal dissection: a safe technique for colorectal tumors.
        Endoscopy. 2007; 39: 418-422
      1. Japanese Society for Cancer of the Colon and Rectum (ed.). Japanese Classification of Colorectal Carcinoma, Appendiceal, and Anal Carcinoma, 3rd English edn Tokyo, Japan: Kanehara. 2019.

        • Kitajima K
        • Fujimori T
        • Fujii S
        • et al.
        Correlations between lymph node metastasis and depth of submucosal invasion in submucosal invasive colorectal carcinoma: a Japanese collaborative study.
        J Gastroenterol. 2004; 39: 534-543
        • Coverlizza S
        • Risio M
        • Ferrari A
        • et al.
        Colorectal adenomas containing invasive carcinoma. Pathologic assessment of lymph node metastatic potential.
        Cancer. 1989; 64: 1937-1947
        • Horie H
        • Togashi K
        • Kawamura YJ
        • et al.
        Colonoscopic stigmata of 1 mm or deeper submucosal invasion in colorectal cancer.
        Dis Colon Rectum. 2008; 51: 1529-1534
        • Hurlstone DP
        • Cross SS
        • Adam I
        • et al.
        Endoscopic morphological anticipation of submucosal invasion in flat and depressed colorectal lesions: clinical implications and subtype analysis of the kudo type V pit pattern using high-magnification-chromoscopic colonoscopy.
        Colorectal Dis. 2004; 6: 369-375
        • Nagtegaal ID
        • Odze RD
        • Klimstra D
        • et al.
        The 2019 WHO classification of tumours of the digestive system.
        Histopathology. 2020; 76: 182-188
        • Saitoh Y
        • Obara T
        • Watari J
        • et al.
        Invasion depth diagnosis of depressed type early colorectal cancers by combined use of videoendoscopy and chromoendoscopy.
        Gastrointest Endosc. 1998; 48: 362-370
        • Haruki S
        • Kobayashi K
        • Yokoyama K
        • et al.
        Comparison of diagnostic accuracies of various endoscopic examination techniques for evaluating the invasion depth of colorectal tumors.
        Gastroenterology research and practice. 2012; 2012621512
        • Hinojosa MW
        • Konyalian VR
        • Murrell ZA
        • et al.
        Outcomes of right and left colectomy at academic centers.
        Am Surg. 2007; 73: 945-958
        • Mukae M
        • Kobayashi K
        • Sada M
        • et al.
        Diagnostic performance of EUS for evaluating the invasion depth of early colorectal cancers.
        Gastrointest Endosc. 2015; 81: 682-690