Prediction of the invasion depth of superficial squamous cell carcinoma based on microvessel morphology: magnifying endoscopic classification of the Japan Esophageal Society

1Endoscopic resection (ER) can offer patients a curative and low-invasive treatment for superficial squamous cell carcinomas (SESCCs). Predicting invasion depth of SESCCs is crucial for determining the precise indication for endoscopic resection (ER) because the lymph node metastasis rate increases in proportion to the invasion depth of the SESCC. According to the Japanese guidelines for diagnosis and treatment of esophageal cancer, T1a-EP or T1a-LPM SCC is considered an absolute indication for ER; T1a-MM or T1b-SM1, a relative indication; and T1b-SM2, an investigative stage (functionally speaking, a contraindication). Subclassification of invasion depth of SESCC, the rate of lymph node metastasis, and the precise indication of ER are listed in Table 1 based on previous studies.

2Endoscopic prediction of the depth of invasion is essential in making decisions regarding the indication for ER in SESCCs. Although predicting the invasion depth of SESCC is possible by conventional endoscopy, magnifying endoscopy provides a more accurate prediction of the tumor depth, in particular for superficial and flat-type SESCCs. Inoue et al. and Arima et al. proposed magnifying endoscopic classifications based upon microvascular morphology and reported their diagnostic utility in predicting the histological invasion depth of SESCC as well as in deciding precise indications for ER. However, there are concerns that the multiplicity of classifications involving complicated criteria might confuse general endoscopists.

3Hence, the committee of the Japan Esophageal Society (JES) developed a simplified classification for the magnifying endoscopic diagnosis (characterization and predicting invasion depth) of SESCC based on the Inoue and Arima classifications. Subsequently, we, the committee members, conducted a prospective multicenter study to evaluate the JES classification for predicting the invasion depth of SESCC.

4Mainly focused on the abnormal microvessels, this magnifying endoscopic classification offers simplified criteria not only for characterization but also for prediction of the invasion depth of SESCC. This is essential for developing a treatment strategy for SESCC, in particular the indication for ER. Therefore, in this classification, morphological types of microvessels are classified into two categories of non-cancerous [normal epithelium, inflammation, and low grade intraepithelial neoplasia (LGIN)] and cancerous [high grade intraepithelial neoplasia (HGIN) and invasive SCC] lesions. The cancerous types of microvessels corresponding to histology of HGIN or invasive SCC are subclassified into three groups based upon an indication for ER as follows: an absolute indication type (HGIN, T1a-EP or T1a-LPM), a relative indication type (T1a-MM or T1b-SM1), and a contraindication type (T1b-SM2).

5Diagnostic criteria of the JES classification are based on the degree of microvascular irregularity in the target lesion observed by magnifying endoscopy. Intrapapillary capillary loops (IPCL) are a basic unit of microvasculature in the squamous mucosal surface. The microvascular irregularity is evaluated for the presence or absence of each of the following morphological factors: weaving (i.e., tortuosity), dilatation, irregular caliber, and different shape (i.e., various shapes).

6Microvessels are classified as type A if they have three or fewer factors and type B if they have all four. Type B is then subclassified into B1, B2, and B3 based on the running pattern or degree of dilatation of severely irregular microvessels. The JES classification criteria are summarized in Table 2 and mentioned in detail as follows:

7Type A: Normal IPCL (Fig. 1a) or abnormal microvessels without severe irregularity (Fig. 1b).

8Type B: Abnormal microvessels with severe irregularity or highly dilated abnormal vessels.

9B1 is defined as type B vessels with a loop-like formation (Fig. 2a, b). The B1 vessels normally appear as dot-like microvessels in a target area (e.g., a brownish area) under NBI endoscopic observation with low or no magnification.

10B2 is defined as type B vessels without a loop-like formation that have a stretched and markedly elongated transformation. The B2 vessels often show a multilayered arrangement or an irregularly branched/running pattern (Fig. 3a, b).

11B3 is defined as highly dilated abnormal vessels whose caliber appears to be more than three times that of the usual B2 vessels and the B3 vessels often appear green in color (Fig. 4a, b).

12When target lesions have only type B1 vessels, the histological invasion depth is predicted as T1a-EP or T1a-LPM. When B2 and B3 vessels are seen in target lesions, the histological invasion depth is predicted as T1a-MM or T1b-SM1 and T1b-SM2 or deeper, respectively.

13AVA was originally defined as a low or no vascularity area surrounded by stretched irregular vessels such as B2 or B3 vessels. The definition of AVA does not include the stretched irregular vessels in this classification. Thus, AVA is surrounded by all subtypes of type B microvessels including B1 vessels.. Since a diameter of AVA is positively correlated with the histological invasion depth of SESCC, the AVA was categorized into three types as follows: AVA-small (smaller than 0.5 mm in diameter: Fig. 5), AVA-middle (0.5 mm or between 0.5 and 3 mm; Fig. 6), and AVA-large (3 mm or larger; Fig. 7). Any types of AVA (small, middle, and large) surrounded by B1 vessels are suggestive of T1a-EP or T1a-LPM SCC. AVA-middle and AVA-large surrounded by B2 or B3 vessels are suggestive of T1a-MM or T1b-SM1 and T1b-SM2 invasive SCC, respectively.

14Reticular pattern vessels are defined as plexiform microvessels (Fig. 6). This vascular pattern is often found in invasive SCC or non-SCC types of malignant epithelial neoplasms (e.g., basaloid (-squamous) carcinoma, adenosquamous carcinoma, and endocrine cell carcinoma) with an infiltrative growth pattern (i.e., INFc) composed of single cells, small tumor nests, or a trabecular arrangement of tumor cells.

15Brownish epithelium between microvessels in a brownish area visualized by NBI-ME was defined as intervascular background coloration.

16From January to August 2011, we prospectively enrolled 211 consecutive patients with SESCC lesions scheduled for ER in five high-volume centers (Table 3). Prior to enrollment, we used pretreatment imaging (CT and/or MRI) to confirm no findings of lymph node metastasis or distant metastasis of SESCC among any of the 211 patients. The depth of tumor invasion of the all lesions was estimated prior to treatment using a magnifying endoscope (GIF-H260Z or GIF-Q240Z; Olympus Corporation, Tokyo, Japan) combined with NBI. The estimation of tumor invasion depth by magnifying endoscopy was made based upon type B vessels in the JES classification and recorded in case report form. The aim of this study was to evaluate the diagnostic values of type B vessels in the pretreatment estimation of invasion depth of SESCC.

17Informed consent was obtained from all participants. The informed consent includes information about the possible risks, benefits, and limits of the endoscopic procedure and treatment. We provided all patients with information about the possibility that additional treatments such as surgery or chemoradiotherapy will be required particularly if ER specimens would show SESCC invading the muscularis mucosa or deeper which is associated with increased risks of lymph node metastasis and distant metastasis. Especially for the patients with pretreatment NBI-ME diagnosis of B3 vessels, we provided the patients with sufficient explanation as follows: (1) B3 vessels will be a suggestive finding of tumor invasion of T1b-SM2 or deeper, (2) T1b-SM2 tumor are basically contraindication for ER because of considerably high risk of lymph node metastasis and distant metastasis, (3) additional therapy such as esophagectomy and chemoradiotherapy will be required if histology from ER would show tumor depth of T1b-SM2 or deeper.

18We performed ER on all enrolled patients. The lesions were removed by endoscopic mucosal resection or endoscopic submucosal dissection. Using the resected specimens, the histological tumor invasion depth was established by pathologists in each center with expertise in gastrointestinal cancer, according to the Japanese classification of esophageal cancer. None of the pathologists were aware of any endoscopic findings. We evaluated the diagnostic utility of the JES classification in predicting histological depth of invasion by comparing the pretreatment NBI-ME diagnosis with the histology from resected specimens.

19All of the SESCC tumors enrolled in this study had type B vessels. The numbers of SESCC tumors subclassified as T1a-EP or T1a-LPM, T1a-MM or T1b-SM1, and T1b-SM2 were 163, 28, and 20, respectively.

20The relationships between histology and pretreatment NBI-ME diagnosis of tumor invasion depth are listed in Table 4. One hundred fifty-nine (97.5 %) of the 163 T1a-EP or T1a-LPM tumors were accurately diagnosed by pretreatment NBI-ME, and four (2.5 %) were overestimated as T1a-MM or T1b-SM1. Twenty-one of 28 (75 %) T1a-MM or T1b-SM1 tumors were accurately diagnosed, and seven (25 %) were underestimated as T1a-EP or T1a-LPM. None of the T1a-EP or T1a-LPM/T1a-MM or T1b-SM1 tumors were overestimated as T1b-SM2. Eleven of 20 (55 %) T1b-SM2 tumors were accurately diagnosed. Six (30 %) and three (15 %) were underestimated as T1a-EP or T1a-LPM and T1a-MM or T1b-SM1, respectively. The overall accuracy of NBI-ME based upon the JES classification was 90.5 % (95 % confidence interval, 85.7–94.1 %).

21Table 5 shows the diagnostic values of B1, B2, and B3 vessels for estimating the depth of invasion of T1a-EP or T1a-LPM, T1a-MM or T1b-SM1, and T1b-SM2 tumors, respectively. The sensitivity/specificity/positive predictive value (PPV) and negative predictive value (NPV) of B1, B2, and B3 vessels were 97.5 %/72.9 %/92.4 % and 89.7 %; 75.0 %/96.2 %/75.0 % and 96.2 %; and 55 %/100 %/100 % and 95.5 %, respectively.

22Endoscopic estimation of invasion depth is a key factor in deciding the treatment policy of SESCCs. We, the JES committee members, have developed a simplified magnifying endoscopic diagnosis classification (i.e., the JES classification). We conducted a prospective multicenter study focused on evaluating the diagnostic values of type B vessels in the classification for predicting the depth of invasion of SESCC.

23The overall accuracy rate of the type B1, B2, and B3 microvessels was 90.5 %, which is of sufficiently high accuracy for clinical use. No T1a-EP or T1a-LPM or T1a-MM or T1b-SM1 tumor was overestimated as T1b-SM2. Although 45 % (9/20) of SM2 tumors did not have B3 vessels, B3 vessels may be highly suggestive of T1b-SM2 tumors because the PPV of B3 vessels was 100 %.

24The sensitivity and PPV of B1 vessels for T1a-EP or T1a-LPM tumors were 97.5 and 92.4 %, respectively. Thus, B1 vessels will provide optimal decision-making regarding the indication for ER. In contrast, the sensitivity and PPV of B2 vessels for T1a-MM or T1b-SM1 tumors were both a suboptimal 75.0 %, while the diagnostic accuracy of white light endoscopy showed a lower value at 66–74 %. Previous studies have indicated that T1a-MM or T1b-SM1 tumors without lymphovascular invasion had a significantly lower risk of lymph node metastasis and a good prognosis. Therefore, T1a-MM or T1b-SM1 tumors are potential candidates for curative treatment by ER. Esophagectomy involves severe problems related to its high mortality rate and a decrease in quality of life after surgery. Further study on the definition of B2 vessels is warranted to reduce the risk of unnecessary surgery for patients with T1a-MM or T1b-SM1 tumors.

25In the JES classification, we partially modified definitions of AVA, which will be useful to predict invasion depth of SESCC as previous studies showed. Although AVA was not assessed in this study, we intend to investigate the predictive values of AVA and whether evaluation of AVA has a positive effect on the predictive value based upon type B vessels (especially for B2 vessels) in an upcoming study.

26Arima et al. reported that reticular pattern (Type R) vessels were characteristic vascular structure reflecting a histologic infiltrative growth (i.e., INFc) of invasive SCC tumors or non-SCC types of malignant epithelial neoplasms. This study did not included non-SCC types of malignant epithelial neoplasms and the type R vessels were not seen in invasive SCC tumors included in this study. A validation study on type R vessels in predicting the histologic features of SCC and non-SCC tumors is warranted.

27Intervascular background coloration will be useful for predicting histology of HGIN and invasive SCC. A recent study indicated that intervascular background coloration might be caused by an extravascular component of hemoglobin that is produced within tumor cells. The other study showed that intervascular background coloration might be related to the thinning of the keratinous layer or the epithelium caused by neoplastic cell proliferation. In this study, we did not evaluated diagnostic value of intervascular background coloration to predict HGIN and invasive SCC. The diagnostic validity and reliability of intervascular background coloration should be assessed in further multicenter studies.

28This study had several limitations. First, the number of T1a-MM or T1b-SM1 and T1b-SM2 tumors was relatively small and no tumor resected by surgery was included. Further study including greater or adequate numbers of these tumors should be conducted to validate the results of the current study. Second, a central review on both the magnifying endoscopic findings and the histology was not conducted in this study. The endoscopic diagnosis and histology were left to each endoscopist and each pathologist in their respective hospitals. Thus, this study may lack diagnostic standardization in magnifying endoscopy and histology. Third, the endoscopists involved in this study are all experts. Additional study involving non-experts is needed to determine whether the results of this study on diagnostic values are applicable in general. Fourth, the reliability or reproducibility of the JES classification was not assessed in this study. It would be desirable to investigate inter- or intra-observer agreements in future studies.

29The newly developed JES classification is proposed to be useful in estimating invasion depth of SESCCs. Modification of the definition of B2 vessels will be necessary, and further studies are required to validate these study results as well as to improve the diagnostic utility of the JES classification.