Endoscopic visualization of cancer and dysplasia in patients with ulcerative colitis following sensitization with oral 5‐aminolevulinic acid

Objective Early diagnosis of colitis‐associated cancer and dysplasia through surveillance endoscopy is vital for patients with ulcerative colitis (UC). This study aimed to evaluate the efficacy of autofluorescence endoscopy (AFE) using 5‐aminolevulinic acid (ALA) and to investigate the fluorescence signal localization pattern following 5‐ALA administration in tumorous lesions diagnosed as colitis‐associated cancer and dysplasia. The sensitivity and specificity of tumorous lesions detected by white light endoscopy (WLE) with and without AFE were evaluated. Methods Overall, 13 endoscopic procedures were performed in 11 patients with UC using WLE and AFE following the oral administration of 5‐ALA. The biopsied lesions detected via endoscopy and resected specimens from cases underwent colectomy were assessed histopathologically. The sensitivity and specificity of detecting tumorous lesions by WLE with and without AFE were evaluated. Results Of the 68 lesions detected and biopsied, 63 were detected via WLE, and five were detected via AFE alone. The sensitivity of detecting colitis‐associated cancer and dysplasia via WLE combined with AFE was 36.4%, and the specificity, positive predictive value and negative predictive value were 94.2%, 57.1%, and 87.5%, respectively. Tumorous lesions displayed three types of fluorescence patterns on AFE. Conclusions AFE using 5‐ALA can detect colitis‐associated cancer and dysplasia in patients with long‐standing UC and lesions that could not be detected via WLE. The distinctive fluorescence patterns in lesions may permit qualitative diagnoses of colitis‐associated cancer and dysplasia.


| INTRODUCTION
Ulcerative colitis (UC) is a chronic inflammatory bowel disease of unknown etiology that affects the large intestine. Patients with long-standing UC are at an increased risk of adverse outcomes, including colorectal cancer. Currently, the risk of colorectal cancer in UC patients has been recognized to increase with disease duration [1][2][3][4] and UC may progress to a precancerous stage (dysplasia).
Dysplasia, which can be further classified into high-grade (HGD) and low-grade (LGD) types, plays a particularly important role in the development of colitic cancer. 5 The management of dysplaisa is similar to that of colon cancer. Therefore, early diagnosis of both colitis-associated cancer and dysplasia (CC and D) via surveillance endoscopy in UC is vital. 6 Endoscopic surveillance of CC and D is commonly performed using a high-definition white light endoscopy (WLE) and chromoendoscopy (CE), both of which are highly recommended. 7 Additionally, image-enhanced endoscopy using narrow-band imaging (NBI), which shows the crypt surface structure or pit pattern, has been found to improve the efficacy of surveillance. [8][9][10][11][12] Reports of surveillance using NBI with equivalent detection to CE observation of colitis-asscoiated neoplasia 13 are intermixed with reports of its better results than conventional WLE. 14 Before using endoscopy with NBI, it is necessary to detect suspicious lesions by WLE first. Subsequently, observation of the surface mucosa and microvasculature is needed to differentiate between tumorous and non-tumorous lesions. This is time-consuming and requires the clinicians to be highly skilled.
Autofluorescence endoscopy (AFE) relies on selective autofluorescence of non-cancerous tissues and has been reported to be particularly effective for detecting colonic tumorous lesions with a flat surface. 15 AFE has the advantage of being simple to perform, and lesions can be easily identified as switching between fluorescent images is possible, resulting in an easy differentiation of tissue types. However, mucosal inflammation in UC also reduces autofluorescence [16][17][18] and leads to a false-positive detection of CC and D in UC. 12,19 AFE with exogenous sensitization using substances with an affinity to tumors such as 5-aminolevulinic acid (5-ALA) has also been attempted to detect CC and D in UC, but the results are inconclusive. [20][21][22] Tumor cells metabolize 5-ALA into protoporphyrin IX (PpIX), which exhibits fluorescence. 23 We have previously reported the effect of PpIX localization on the diagnosis of tumorous lesions after the administration of 5-ALA in a mouse model of colitis-associated dysplasia. 24 The aims of this study were to investigate the fluorescence signal localization pattern following the administration of 5-ALA in tumorous lesions diagnosed histopathologically as CC and D in patients with long-standing UC and to evaluate the additional efficacy of AFE using 5-ALA.

| Endoscopy
After a 12-hour fasting, 5-ALA 20 mg/kg (Alabel; Nobelpharma, Tokyo, Japan) dissolved in 250 mL of distilled water was administered orally, followed by pretreatment with polyethylene glycol (Niflec; EA Pharma, Tokyo, Japan) that was performed one hour later. An endoscope equipped with the AFE function (CF-FH260AZI; Olympus Medical Systems, Tokyo, Japan) was used.
The endoscope was inserted as far as the ileocecal region and the colon was visualized in five sections: the cecum and ascending colon, the transverse colon, descending colon, the sigmoid colon, and the rectum. For each segment, WLE was performed first using CE with indigo carmine (0.08%). All elevated, flat and depressed lesions with boundaries and regional redness detected via WLE were checked for green fluorescence at 615 nm via AFE. 27 All lesions detected by either WLE or AFE were biopsied for histopathological assessment. A target biopsy method was used where biopsies were obtained from suspicious lesions observed on WLE/CE. Random biopsies, where biopsies were obtained randomly at an interval of every 10 cm, were performed as recommended. However, recent study has reported that equivalent results can be obtained with targeted biopsies. 28 All lesions that were detected on WLE/CE or AFE with 5-ALA were biopsied (eg, by target biopsy) and checked histopathologically. Resected specimens from colectomy cases were observed macroscopically and assessed histopathologically.

| Histopathological diagnosis
The histopathological diagnoses were performed by consensus between two expert gastrointestinal pathologists using the Vienna criteria. 29 Neoplasia was defined as noninvasive low-grade neoplasia (adenoma or dysplasia), noninvasive high-grade neoplasia (noninvasive carcinoma, high-grade adenoma or dysplasia, or suspicion of invasive carcinoma), and invasive neoplasia (submucosal carcinoma, intramucosal carcinoma, or beyond). Lesions that were indefinite for neoplasia or dysplasia were not treated as neoplastic lesions. It was difficult to differentiate between sporadic adenoma and CC and D the diagnosis was determined using p53 immunostaining. Specimens obtained through endoscopic treatment or colectomy were assessed for high-grade intraepithelial neoplasia and adenocarcinomas.

| Follow-up
All nonsurgical cases were followed up annually with a surveillance endoscopy yearly for 3 to 5 years. The absence of new lesions during the follow-up period confirmed that CC and D had not been overlooked during the endoscopy.

| Statistical analysis
Statistical analyses were performed by using the STATA software ver- 3 | RESULTS

| AFE fluorescence patterns and histopathological findings of tumorous lesions
Of the 12 lesions that showed strong green fluorescence on AFE, five were histopathologically diagnosed as CC and D (one as HGD, one as LGD, two as low-grade tubular adenomas, and one as adenocarcinoma; Figure 2). The following three distinct types of PpIX fluorescence signal patterns were observed in CC and D lesions:   This study also showed that tumorous lesions could display at least three distinctive fluorescence patterns on AFE.

| DISCUSSION
LGD lesions display a strong green fluorescence pattern at the tumor margins, similar to that previously identified in a mouse model, 22 13,14,34 Therefore, the added precision that AFE provides to CME/CE is preferable, as this technique is easier to perform and can be judged more accurately even by less experienced analysts.
There were some limitations to this study. First, we only compared WLE/CE using indigo carmine and AFE using 5-ALA and did not investigate its effectiveness using other sensitizing agents and imaging modalities, such as NBI and confocal laser endoscopy. 35 Second, this was a single-center study with a small number of subjects. Therefore, further investigation with large sample sizes in multiple facilities is required to obtain more robust results.