The Interpretation of Magnifying Endoscopy for the Diagnosis of Colorectal Lesions

Endoscopic management strategies have become the standard treatment of superficial colonic lesions. These minimally invasive techniques have demonstrated significant benefits, including reduced mortality rates, fewer complications, and quicker patient recovery times. Accurate endoscopic diagnosis is pivotal in determining the optimal treatment strategy for colorectal lesions. These strategies include cold snare polypectomy (CSP), endoscopic mucosal resection, and endoscopic submucosal dissection, each tailored to specific clinical indications [1‒3].

The cornerstone of endoscopic diagnosis is white light imaging (WLI). However, WLI can sometimes be insufficient for detailed visualization, necessitating the use of advanced imaging techniques of image-enhanced endoscopy such as narrow band imaging (NBI) and blue laser/light imaging (BLI). Additionally, magnified observation of pit patterns becomes essential when higher diagnostic precision is required [2].

These advanced diagnostic tools allow for precise differentiation between various types of colorectal lesions, such as adenomas, T1 cancers, and sessile serrated lesions (SSLs). For instance, NBI enhances the visualization of blood vessels and mucosal surface structures, aiding in the identification of neoplastic changes. Similarly, BLI provides high-contrast images that facilitate the detection of subtle findings.

Recent innovations in endoscopic systems have markedly improved image clarity and resolution, enabling clinicians to make more precise diagnoses. Enhanced image quality not only supports accurate lesion characterization but also informs the selection of the most appropriate therapeutic intervention.

This review provides a comprehensive overview of the role of magnified endoscopy in the diagnosis of colorectal lesions. We will explore the roles of WLI, NBI, BLI, and pit pattern observation in the diagnostic process.

NBI is a light observation technique developed by Sano, Muto, and their colleagues, along with the Olympus Corporation, in 2006 to enhance lesion detection and diagnosis {EVIS LUCERA SPECTRUM (CV-260SL [video system]/CLV-260SL [light source]) in Asia, or EVIS EXERA II (CV-180) in the West, Olympus Co., Tokyo, Japan} [4]. The principle of NBI involves altering the spectral characteristics of a xenon light source endoscope to match the absorption wavelengths of oxidized hemoglobin (415 nm and 540 nm) using a specialized filter, thereby enhancing the visualization of mucosal surface structures and blood vessels. To further improve image quality, the EVIS LUCERA ELITE (CV-190 [video system]/CLV-290 [light source]) (Asia) and EXERA III (CV-290) (the West) systems were subsequently introduced. In July 2020, Olympus Co. launched the EVIS X1 (CV-1500, Olympus Co., Tokyo, Japan), a new endoscopic system utilizing an LED light source instead of a xenon light source in some countries including Japan [5]. The system supports both the simultaneous imaging scope (CF-EZ-1500D, a dedicated colonoscope) and the conventional sequential imaging scope (CF-XZ1200, a dedicated colonoscope). The former features a dual-focus zoom, while the latter employs a manual zoom mechanism. The system also introduces two new observation modes: texture and color enhancement imaging (TXI) and red dichromatic imaging (RDI), and the potential for advancing endoscopic diagnosis is anticipated with TXI and RDI. Additionally, extended depth of field (EDOF) is adopted as a new technology and precise observations through continuous broad focus in endoscopic images and seamless magnification are achieved with EDOF.

On the other hand, Fujifilm Co. initially produced flexible spectral imaging color enhancement (FICE) as image-enhanced endoscopy in 2005 and FICE could display color images with RGB components that have been assigned selected spectra in real time for lesion enhancement [6]. Then, BLI was developed for improving image qualities in 2012 and BLI is a narrow-band light observation technique using a laser light source endoscope (VP-4450HD [processor]/LL-4450 [light source]: LASEREO, Fujifilm Co., Japan) [7‒9]. The technique employs a 410-nm laser beam to enhance the visualization of blood vessels and mucosal structures, and a 450-nm laser beam that excites phosphors, ensuring brightness and resulting in clear images. BLI has two modes: a high-contrast mode used with magnification (BLI mode) and a bright mode (BLI bright mode), which provides a brighter image than BLI contrast, enabling enhanced visualization from a far field of view. However, laser endoscopes were not available in Europe and the USA (only in Asia and some other countries). Instead, LED light source endoscopic systems (VP-7000 [processor]/BL-7000 [light source]: ELUXEO, Fujifilm Co.) were released in those areas and are more prevalent now worldwide. BLI, utilizing LED light for narrow-band light observation, was made possible through multi-light technology, which uses four types of LED light (blue-violet, blue, green, and red) to enhance blood vessels and mucosal structures. The usefulness of this technology in detecting and diagnosing colorectal lesions has been suggested. The endoscope’s magnification can be selected between stepwise and manual zooming, with a magnification scale displayed on the top-right side of the screen for convenience. In a multicenter study, we demonstrated the non-inferiority of WLI and linked color imaging (LCI) of laser and LED endoscopic systems in visualizing lesions in the colorectum [10]. LCI is a mode primarily used for lesion detection. The difference between the laser and LED modes was minimal: the LED endoscope was significantly brighter, while the laser endoscope had a more reddish image. Furthermore, comparable visibility in vessel and surface patterns for the LED endoscope was reported for BLI in 95.4% and 95.9% of cases, respectively [11]. These results suggest that LED can be effectively used for WLI, LCI, and BLI, similar to laser. Moreover, a less expensive, and compact, tricolor LED endoscope system (ELUXEO Lite: 6000 series, EP-6000, Fujifilm) with an integrated processor and light source is available on the market. This system has been reported to enable BLI observation with accuracy comparable to that of the laser-magnifying scope [12]. In 2024, the EP-8000 (Fujifilm) was introduced in Japan. Similar to the VP-7000 system, it uses an LED light source and offers further improvements with triple noise reduction and extended-dynamic range image processing (E-DRIP). Both triple noise reduction and E-DRIP help reduce noise in endoscopic observation and enhance brightness in distant views. These enhancements result in image quality superior to that of previous systems (online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000543996). Especially, the brightness and contrast of images are improved.

The basic principle for lesion diagnosis is performed with WLI (Fig. 1) [9]. The diagnostic accuracy for distinguishing between neoplastic and nonneoplastic polyps is reported to be 83.2% for lesions <10 mm and 98.3% for lesions ≥10 mm [13]. Recently, many endoscopists began to use NBI/BLI, with or without magnification, for lesion detection and characterization. There is sufficient evidence supporting the diagnosis of small, typical adenomas and hyperplastic polyps (HP) using NBI/BLI. The reported accuracy for differentiating neoplastic and nonneoplastic polyps ranges from 89.0% to 90.0% for NBI without magnification and from 94.1% to 96.1% for NBI with magnification [14‒17]. These findings suggest that diagnosis using NBI/BLI alone, skipping WLI, is feasible. However, for diagnosing Tis and T1 cancers, endoscopic findings with WLI are particularly helpful. For such lesions, a combination of WLI and NBI/BLI should be considered for optimal lesion characterization. In particular, characteristic WLI findings of T1b cancer include (1) expansive appearance (protrusion and overextension of the lesion and/or surrounding normal mucosa, such as a submucosal tumor) tumor; (2) a depressed surface; (3) a rough appearance (rough surface without shine); (4) a sessile elevation in the depressed area [2, 18]. The diagnosis of colorectal tumors is based on identifying the irregularities by WLI and then using NBI and BLI of the whole area including the irregular area to make a diagnosis using the NBI International Colorectal Endoscopic (NICE) classification if the lesion is observed without magnification, or using the Japan NBI Expert Team (JNET) classification if the lesion is magnified (Fig. 1) [19, 20]. In the colorectum, NBI and BLI are rarely used at maximum magnification due to the difficulty in focusing on the lesions and the potential for bleeding caused by contact. Thus, JNET classification can be used at moderate magnification. When it is difficult to diagnose lesions with NBI and BLI, it is recommended to use pit pattern classification by additional chromoendoscopy with indigo carmine or crystal violet [21, 22]. The treatment methods can then be determined based on the estimated pathological diagnosis and lesion size.

The JNET classification using NBI and BLI categorizes lesions into four types based on surface and vessel patterns for diagnostic purposes (Fig. 1; online suppl. Fig. 2) [20]. Type 1 corresponds to histopathology findings of HP and SSLs. Type 2A denotes adenomas with low to moderate dysplasia, while type 2B encompasses adenomas with high-grade dysplasia and T1a cancer. Type 3 specifically identifies T1b cancers. This classification is crucial for determining treatment strategies. The diagnosis of type 3 is particularly important, characterized by an amorphous surface pattern and a vessel pattern that shows either no vessels (loose vascular area) or the localized presence of large vessels lacking a network pattern, as it is outside the scope of local endoscopic management according to the current guidelines [1]. To become proficient with the JNET classification, it is helpful to first understand the surface pattern, which is analogous to the pit pattern, before progressing to the vessel pattern. This step-by-step approach facilitates a more comprehensive understanding of the classification system.

Type 2A corresponds to type IIIL, IV, and IIIs pit patterns, characterized by a well-defined and regular pattern. Type 2B aligns with type V_I pit patterns, exhibiting irregularity, while type 3 is analogous to V_N pit patterns, showing a destroyed pattern.

In JNET classification, the diagnostic accuracy for T1b cancer is relatively high, exceeding 90% in most reports, but the sensitivity remains low (Table 1) [23‒28]. In a study by Iwatate et al. [28], the positive diagnosis rate dropped to 81.3% without incorporating WLI information. This study was based on a web-based analysis of 100 magnified NBI images, evaluated by 25 expert endoscopists from various representative institutions in Japan. It is important to explain the diagnosis to the patient and develop a treatment strategy, considering the risks of overdiagnosis and underdiagnosis. In a report comparing NBI observation of the surface pattern and pit pattern diagnosis using a dye with magnifying observation simultaneously performed, it was shown that lesions with regular structures, the fine surface patterns (pit-like structures) observed using NBI magnification, showed almost perfect agreement with pit pattern findings obtained through dye-based magnifying observation [29]. However, in lesions with irregular structures, the pit pattern diagnosis based on dye-magnifying observation provided more detailed irregular findings compared to NBI-magnifying observation. Additionally, a report indicated that combining BLI and NBI observations, which reveal irregularities, with pit pattern classification using crystal violet staining, significantly improved diagnostic accuracy [30]. Therefore, the combined diagnostic approach utilizing NBI/BLI, pit pattern analysis, and WLI is anticipated to enhance diagnostic accuracy, especially for T1 cancers exhibiting irregular/destroyed surface and vessel patterns identified by NBI/BLI.

Pit pattern classification utilizes indigo carmine and crystal violet to categorize patterns into type I to type V [21, 35] (online suppl. Fig. 3). Type I pits exhibit a normal circular shape, type II pits are star-shaped and slightly larger than normal, type IIIL pits present a tubular pattern, type IIIs pits resemble a smaller circular pattern, and type IV pits display a gyrus-shaped or dendritic pattern. Type V pits include three subtypes: type V_I pits feature irregular polygons and are subclassified as mildly or severely irregular, while type V_N pits show a pattern where the structure has disappeared.

These patterns correlate with different pathological findings: type I corresponds to normal mucosa, type II pits indicate HP or SSL, types IIIL, IIIs, and IV suggest adenoma, type V_I pits with mild irregularity are indicative of Tis and T1a lesions, and type Vi pits with severe irregularity are associated with Tis, T1a, and T1b lesions, with 30–50% of these cases progressing to T1b. Type V_N pits are almost exclusively associated with T1b lesions and exhibit high specificity [31]. The diagnostic accuracy of type V_N pits and type V_I pits with severe irregularity for detecting T1b cancer is reported to be as high as 84.2–98.8%. However, their sensitivity ranges from 50.0% to 98.7%, indicating the need for careful interpretation (Table 1) [31‒34]. When irregular findings are detected in NBI/BLI, it is advisable to assess the degree of irregularity using pit pattern classification. However, in some cases, the presence of mucus or hemorrhage (which can be caused by contact with the scope or flushing the dye or water directly on the lesion) can obscure the structured findings, posing challenges for pit pattern classification. In these cases, NBI/BLI magnification may assist, highlighting the importance of proficiency in both pit pattern and NBI/BLI observations. Recently, the World Health Organization (WHO) has recommended restricted use of crystal violet in line with national regulations and guidelines of academic societies, despite the absence of human-specific reports [36].

The diagnosis of SSL involves identifying specific characteristics under WLI and chromoendoscopy, such as a cloud-like surface, indistinct borders, irregular shape, and the presence of a mucus cap. Additionally, features including open type II pit patterns, serrated pit patterns, dilated glandular ducts, and dilated blood vessels observed in NBI or BLI findings have been associated with pathologically confirmed serrated glandular ducts and their dilatations [31‒34, 36‒42].

In the JNET classification, type 1 lesions encompass both HP and SSL. Dilated vessels and glandular ducts serve as valuable criteria for differentiating between these lesions (Table 1) [37‒40]. For the diagnosis of SSL with dysplasia (SSLD), intralesional granules, erythema, and depression are important findings under WLI, while irregular network vessels are particularly informative under NBI [41‒46]. NBI has demonstrated a diagnostic performance with 46.2% sensitivity, 97.3% specificity, and 93.3% positive predictive value [42]. Similar interpretations of irregular network vessels can be made with BLI as with NBI [43]. However, 4.6% (14/306) of patients diagnosed with SSL using WLI exhibited SSLD, even in flat lesions lacking granules or depression [42]. This highlights the critical role of magnified observation using NBI/BLI in detecting minute vascular changes that cannot be identified with WLI, emphasizing its importance in determining appropriate treatment strategies. While CSP remains the standard for SSLs smaller than 10 mm, recent studies suggest that piecemeal CSP may also be acceptable for larger lesions [1, 47, 48]. However, SSLD is occasionally associated with high-grade dysplasia or carcinoma, necessitating en bloc resection whenever feasible.

The utility of magnified endoscopic observation in diagnosing colorectal tumors was extensively discussed, encompassing the JNET classification using NBI and BLI, alongside pit pattern classification via chromoendoscopy. Incorporating these observations with WLI enhances the accuracy of diagnosis and facilitates appropriate treatment decisions.

We thank all members of the Department of Molecular Gastroenterology and Hepatology, Kyoto Prefectural University of Medicine, for making this paper.

Yoshida N. is an editor of Digestion. Yoshida N. and Dohi O. received a research grant from Fujifilm Co.

This study received no funding.

Yoshida N. made the initial version of this review paper and arranged it. Inoue K., Ghoneem E., Inagaki Y., Kobayashi R., Iwai N., Dohi O., Hirose R., and Itoh Y. reviewed the manuscript.