1
Incidence of DVT after ESD may be not as high as that of abdominal surgery, but it is difficult to detect early because these patients are asymptomtic. CDdO1 ug/ml was expected as a predictor of DVT although operation timeS100min is a confounder. But ultrasonography should be required for early diagnosis of DVT after ESD. S1477 Is Autofluorescence Imaging Endoscopy Useful for Determining the Invasion Depth of Gastric Cancer? Yuji Amano, Aya Otani, Yuko Mishima, Yoshiko Takahashi, Kenji Koshino, Norihisa Ishimura, Shunji Ishihara, Yoshikazu Kinoshita Background and Aims: Newly developed autofluorescence imaging (AFI) endoscopy provides for detection of autofluorescence arising from the gastrointestinal mucosa. Lesions, such as a tumor or inflammation, block reflective autofluorescence, and the weakened autofluorescence is subsequently detected as a magenta color by the AFI endoscope instead of the light green of normal mucosa. This is valuable for effective diagnosis of the presence of a tumor or inflammation. In the normal gastrointestinal wall, autofluorescence is primarily generated by collagen fibers of the submucosal layer. Therefore, AFI endoscopy possibly provides for differentiation of tumor invasion depth. We evaluated whether it was possible to determine the depth of the tumor invasion by measuring the reflective autofluorescence. Patients and Methods: Endoscopic images were first processed by de-gamma software before numerical analysis of autofluorescent images. Then, images were split into 3 separate bands (R, G and B) and expressed as grayscale values. An AF index was defined as the ratio of the grayscale value for the G-band image to that of R-band in a preliminary study using porcine stomachs and a mimic lesion. From April 2007 to March 2008, AF indices were preoperatively calculated for 69 lesions of gastric cancer with elevated form from 66 patients and were statistically compared to invasion depth and tumor thickness. Lesion grades were histologically classified into the following 5 groups: ‘‘M’’, 29; ‘‘SM’’, 22; ‘‘MP’’, 5; ‘‘SS’’, 5; ‘‘SE’’, 8 after the resection. Gastric cancer with ulceration was excluded as the whitish exudate of ulceration disturbed the measurement of the autofluorescence intensity. Results: The mean AF indices for tumors at each depth stage were: 1.01 0.29, 0.86 0.18, 0.76 0.13, 0.72 0.17, and 0.62 0.08, respectively. Statistically significant differences were found between ‘‘M’’ and the other stages (p!0.05). The AF index also showed a significant negative correlation to the mean thickness of the tumor (rZ-0.024, p!0.05). Discussion: Determination of the tumor invasion depth plays an important role for therapeutic strategies, although it is not easily or reliably done for all cases. In the present study, we could differentiate the depth of ‘‘M’’ tumors from others via the analysis of autofluorescence intensity. This suggests a new simple diagnostic tool for the tumor depth evaluation. However, to demonstrate its validity, additional studies of a large number of different tumor types will be required. Conclusions: AFI endoscopy may be useful for determining tumor invasion depth in gastric cancer without ulceration. S1478 Magnifying Endoscopy with Narrow Band Imaging Achieves a Higher Diagnostic Accuracy As Compared to White Light Endoscopy in Superficial Gastric Depressed Lesion. A Prospective Comparative Study Masayuki Kato, Mitsuru Kaise, Jin Yonezawa, Toyoizumi Hirobumi, Noboru Yoshimura, Yukinaga Yoshida, Hisao Tajiri Background: Magnifying endoscopy combined with narrow band imaging (ME-NBI) has been used for differential diagnosis of various focal neoplasias. Conducting a retrospective study, we reported that the triad of ME-NBI findings (disappearance of fine mucosal structure, microvascular dilation and heterogeneity) was the best combination for gastric cancer diagnosis with a high accuracy (sensitivityZ70%, specificityZ85%) and moderate inter-observer accordance (0.47 of kappa value among 11 endoscopists), which were superior to those (sensitivityZ70%, specificityZ65%, kappa valueZ0.36) with white light endoscopy (WLE). Aim: To confirm the superiority of ME-NBI gastric cancer diagnosis with the triad in comparison to WLE in a prospective study. Patients and Method: Patients who underwent WLE followed by ME-NBI for a preoperative workup examination or for a post-ESD follow-up endoscopy at a referral academic center were eligible to this study. Superficial gastric depressed lesions, which were recognized with WLE as candidates for differential diagnosis between gastric cancers and benign lesions, were prospectively enrolled. All of the recruited lesions were biopsied, and the pathological diagnoses were used as the gold standard for this study. Gastric lesions that had already been proven pathologically as cancers were excluded. WLE and ME-NBI was serially performed using zoom endoscopes (GIF-240Z or GIF-H260Z) and an electric endoscopic system (EVIS 260 Spectrum; Olympus Medical Systems). Endoscopists reported WLE general diagnosis (benign or malignant), ME-NBI general diagnosis without any specified criteria, and ME-NBI diagnosis with the predefined triad. Result: 1) A total of 171 lesions from 91 patients (81 males and 10 females, mean age of 66 years) were enrolled. Mean diameter of the 171 lesions were 6.6mm. Nine of the 171 lesions were pathologically proven as gastric cancers, and others were benign depressed lesions. 2) The sensitivity (89%) and specificity (93%) of ME-NBI diagnosis with the triad were significantly (p!0.05) higher than the sensitivity (22%) and specificity (65%) of WLE general diagnosis, and were equivalent to the sensitivity (89%) and specificity (89%) of ME-NBI general diagnosis. Conclusion: The prospective study showed that ME-NBI with the triad achieves a significantly higher accuracy of cancer diagnosis as compared to WLE in superficial gastric depressed lesion. Therefore, ME-NBI can be a promising diagnostic modality for the population with a high risk of gastric cancer. S1479 One Year Follow-Up Results of Magnify NBI for Gastric Intestinal Metaplasia Surveillance Boonlert Imraporn, Sukprasert Jutaghokiat, Naruemon Wisedopas, Varocha Mahachai, Rungsun Rerknimitr, Sombat Treeprasertsuk, Pinit Kullavanijaya Background: Gastric intestinal metaplasia (GIM) can progress into gastric cancer. Narrow band imaging with magnification endoscopy (NBI-ME) can be used for more accurate targeted biopsy of GIM. To date, no data on the accuracy and progression of GIM detected by NBI-ME are available. Aim: To evaluate the utility of NBI-ME for a follow up of GIM. Methods: 16 patients with previously diagnosed with GIM were prospectively examined by NBI-ME (EVIS EXERA Olympus GIF Q160Z) on 2 occasions within 1 year interval. At each endoscopy, targeted biopsy (4 from antrum and 2 from incisura) was done according to endoscopic criteria including light blue crests, villous pattern and large long crests. Specimens were examined by one pathologist based on the updated Sydney classification. Results: There were 35/96 and 31/96 specimens with positive for GIM by histology at the 1st and 2nd endoscopies respectively. Results are shown in the table. Within 1 year GIM in 10/16 patients persisted, 4/16 regressed and 2/16 were steadily undetectable. Conclusion: NBI-ME can be effectively applied for follow up of GIM with high consistent of sensitivity, specificity, PPV, NPVand accuracy. However, one year follow-up is not long enough to detect the progression of GIM. Results of one year follow up for gastric intestinal metaplasia with magnify NBI Sensitivity Specificity PPV NPV Accuracy Criteria First Second First Second First Second First Second First Second LBC (%) 74.4 83.3 86.0 91.7 78.4 85.8 83.1 90.2 81.3 88.6 VP (%) 45.7 58.1 85.3 87.7 64.0 69.2 73.2 81.4 70.9 78.1 LLC (%) 25.0 38.5 91.1 88.6 66.7 55.6 63.0 79.5 63.6 75.0 All criteria (%) 74.3 90.3 88.5 89.2 78.8 80.0 85.7 95.1 83.3 89.6 LBCZlight blue crest VP Zvillous pattern LLCZlarge long crest LBC S1480 A Study On Emergency Gastro-Intestinal Endoscopic Hemostasis Related to Anti Platelet Drugs, Anti Thrombosis Drugs, and Non- Steroidal Anti-Inflammatory Drugs (NSAIDs) Jun Takada, Hiroshi Araki, Fumito Onogi, Takashi Ibuka, Hisataka Moriwaki Aim: We studied clinical characteristics of emergency gastro-intestinal endoscopic hemostasis related to anti platelet drugs, anti thrombosis drugs, and non-steroidal anti-inflammatory drugs (NSAIDs). Methods: From May 2004 to March 2006, we performed emergency gastro-intestinal endoscopic hemostasis in 113 cases, male 83 cases, female 30 cases. Mean age was 65.7. According to the use of anti platelet drugs and anti thrombosis drugs, we divided the cases into two groups. Group A is consisted of cases under medication with these drugs and group B is not under medication with them. We compared two groups for age, sex, background disease, symptom, diagnosis, location of lesion, number of lesion, use of anti-ulcer drugs, Abstracts AB182 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 5 : 2009 www.giejournal.org

Magnifying Endoscopy with Narrow Band Imaging Achieves a Higher Diagnostic Accuracy As Compared to White Light Endoscopy in Superficial Gastric Depressed Lesion. A Prospective Comparative

Embed Size (px)

Citation preview

Page 1: Magnifying Endoscopy with Narrow Band Imaging Achieves a Higher Diagnostic Accuracy As Compared to White Light Endoscopy in Superficial Gastric Depressed Lesion. A Prospective Comparative

Abstracts

Incidence of DVTafter ESD may be not as high as that of abdominal surgery, but it isdifficult to detect early because these patients are asymptomtic. CDdO1 ug/ml wasexpected as a predictor of DVT although operation timeS100min is a confounder.But ultrasonography should be required for early diagnosis of DVT after ESD.

S1477

Is Autofluorescence Imaging Endoscopy Useful for Determining

the Invasion Depth of Gastric Cancer?Yuji Amano, Aya Otani, Yuko Mishima, Yoshiko Takahashi, Kenji Koshino,Norihisa Ishimura, Shunji Ishihara, Yoshikazu KinoshitaBackground and Aims: Newly developed autofluorescence imaging (AFI)endoscopy provides for detection of autofluorescence arising from thegastrointestinal mucosa. Lesions, such as a tumor or inflammation, block reflectiveautofluorescence, and the weakened autofluorescence is subsequently detected asa magenta color by the AFI endoscope instead of the light green of normal mucosa.This is valuable for effective diagnosis of the presence of a tumor or inflammation.In the normal gastrointestinal wall, autofluorescence is primarily generated bycollagen fibers of the submucosal layer. Therefore, AFI endoscopy possibly providesfor differentiation of tumor invasion depth. We evaluated whether it was possible todetermine the depth of the tumor invasion by measuring the reflectiveautofluorescence. Patients and Methods: Endoscopic images were first processedby de-gamma software before numerical analysis of autofluorescent images. Then,images were split into 3 separate bands (R, G and B) and expressed as grayscalevalues. An AF index was defined as the ratio of the grayscale value for the G-bandimage to that of R-band in a preliminary study using porcine stomachs and a mimiclesion. From April 2007 to March 2008, AF indices were preoperatively calculatedfor 69 lesions of gastric cancer with elevated form from 66 patients and werestatistically compared to invasion depth and tumor thickness. Lesion grades werehistologically classified into the following 5 groups: ‘‘M’’, 29; ‘‘SM’’, 22; ‘‘MP’’, 5;‘‘SS’’, 5; ‘‘SE’’, 8 after the resection. Gastric cancer with ulceration was excluded asthe whitish exudate of ulceration disturbed the measurement of theautofluorescence intensity. Results: The mean AF indices for tumors at each depthstage were: 1.01 � 0.29, 0.86 � 0.18, 0.76 � 0.13, 0.72 � 0.17, and 0.62 � 0.08,respectively. Statistically significant differences were found between ‘‘M’’ and theother stages (p!0.05). The AF index also showed a significant negative correlationto the mean thickness of the tumor (rZ-0.024, p!0.05). Discussion:Determination of the tumor invasion depth plays an important role for therapeuticstrategies, although it is not easily or reliably done for all cases. In the present study,we could differentiate the depth of ‘‘M’’ tumors from others via the analysis ofautofluorescence intensity. This suggests a new simple diagnostic tool for the tumordepth evaluation. However, to demonstrate its validity, additional studies of a largenumber of different tumor types will be required. Conclusions: AFI endoscopy maybe useful for determining tumor invasion depth in gastric cancer withoutulceration.

S1478

Magnifying Endoscopy with Narrow Band Imaging Achieves

a Higher Diagnostic Accuracy As Compared to White Light

Endoscopy in Superficial Gastric Depressed Lesion. A

Prospective Comparative StudyMasayuki Kato, Mitsuru Kaise, Jin Yonezawa, Toyoizumi Hirobumi,Noboru Yoshimura, Yukinaga Yoshida, Hisao TajiriBackground: Magnifying endoscopy combined with narrow band imaging (ME-NBI)has been used for differential diagnosis of various focal neoplasias. Conductinga retrospective study, we reported that the triad of ME-NBI findings (disappearanceof fine mucosal structure, microvascular dilation and heterogeneity) was the bestcombination for gastric cancer diagnosis with a high accuracy (sensitivityZ70%,specificityZ85%) and moderate inter-observer accordance (0.47 of kappa valueamong 11 endoscopists), which were superior to those (sensitivityZ70%,specificityZ65%, kappa valueZ0.36) with white light endoscopy (WLE). Aim: Toconfirm the superiority of ME-NBI gastric cancer diagnosis with the triad incomparison to WLE in a prospective study. Patients and Method: Patients whounderwent WLE followed by ME-NBI for a preoperative workup examination or fora post-ESD follow-up endoscopy at a referral academic center were eligible to thisstudy. Superficial gastric depressed lesions, which were recognized with WLE ascandidates for differential diagnosis between gastric cancers and benign lesions,were prospectively enrolled. All of the recruited lesions were biopsied, and thepathological diagnoses were used as the gold standard for this study. Gastric lesionsthat had already been proven pathologically as cancers were excluded. WLE andME-NBI was serially performed using zoom endoscopes (GIF-240Z or GIF-H260Z)and an electric endoscopic system (EVIS 260 Spectrum; Olympus Medical Systems).Endoscopists reported WLE general diagnosis (benign or malignant), ME-NBIgeneral diagnosis without any specified criteria, and ME-NBI diagnosis with thepredefined triad. Result: 1) A total of 171 lesions from 91 patients (81 males and 10females, mean age of 66 years) were enrolled. Mean diameter of the 171 lesionswere 6.6mm. Nine of the 171 lesions were pathologically proven as gastric cancers,and others were benign depressed lesions. 2) The sensitivity (89%) and specificity

AB182 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 5 : 2009

(93%) of ME-NBI diagnosis with the triad were significantly (p!0.05) higher thanthe sensitivity (22%) and specificity (65%) of WLE general diagnosis, and wereequivalent to the sensitivity (89%) and specificity (89%) of ME-NBI generaldiagnosis. Conclusion: The prospective study showed that ME-NBI with the triadachieves a significantly higher accuracy of cancer diagnosis as compared to WLE insuperficial gastric depressed lesion. Therefore, ME-NBI can be a promisingdiagnostic modality for the population with a high risk of gastric cancer.

S1479

One Year Follow-Up Results of Magnify NBI for Gastric Intestinal

Metaplasia SurveillanceBoonlert Imraporn, Sukprasert Jutaghokiat, Naruemon Wisedopas,Varocha Mahachai, Rungsun Rerknimitr, Sombat Treeprasertsuk,Pinit KullavanijayaBackground: Gastric intestinal metaplasia (GIM) can progress into gastric cancer.Narrow band imaging with magnification endoscopy (NBI-ME) can be used formore accurate targeted biopsy of GIM. To date, no data on the accuracy andprogression of GIM detected by NBI-ME are available. Aim: To evaluate the utility ofNBI-ME for a follow up of GIM. Methods: 16 patients with previously diagnosedwith GIM were prospectively examined by NBI-ME (EVIS EXERA Olympus GIFQ160Z) on 2 occasions within 1 year interval. At each endoscopy, targeted biopsy (4from antrum and 2 from incisura) was done according to endoscopic criteriaincluding light blue crests, villous pattern and large long crests. Specimens wereexamined by one pathologist based on the updated Sydney classification. Results:There were 35/96 and 31/96 specimens with positive for GIM by histology at the 1stand 2nd endoscopies respectively. Results are shown in the table. Within 1 year GIMin 10/16 patients persisted, 4/16 regressed and 2/16 were steadily undetectable.Conclusion: NBI-ME can be effectively applied for follow up of GIM with highconsistent of sensitivity, specificity, PPV, NPV and accuracy. However, one yearfollow-up is not long enough to detect the progression of GIM.

Results of one year follow up for gastric intestinal metaplasia with magnify NBI

Sensitivity Specificity PPV NPV Accuracy

www.gie

Criteria

First S econd F irst Second F irst S econd First S econd F irst

journ

Second

LBC (%)

74.4 8 3.3 8 6.0 91.7 7 8.4 8 5.8 83.1 9 0.2 8 1.3 88.6 VP (%) 45.7 5 8.1 8 5.3 87.7 6 4.0 6 9.2 73.2 8 1.4 7 0.9 78.1 LLC (%) 25.0 3 8.5 9 1.1 88.6 6 6.7 5 5.6 63.0 7 9.5 6 3.6 75.0 All criteria(%)

74.3 9

0.3 8 8.5 89.2 7 8.8 8 0.0 85.7 9 5.1 8 3.3 89.6

LBCZlight blue crest VP Zvillous pattern LLCZlarge long crest

LBC

S1480

A Study On Emergency Gastro-Intestinal Endoscopic Hemostasis

Related to Anti Platelet Drugs, Anti Thrombosis Drugs, and Non-

Steroidal Anti-Inflammatory Drugs (NSAIDs)Jun Takada, Hiroshi Araki, Fumito Onogi, Takashi Ibuka,Hisataka MoriwakiAim: We studied clinical characteristics of emergency gastro-intestinal endoscopichemostasis related to anti platelet drugs, anti thrombosis drugs, and non-steroidalanti-inflammatory drugs (NSAIDs). Methods: From May 2004 to March 2006, weperformed emergency gastro-intestinal endoscopic hemostasis in 113 cases, male83 cases, female 30 cases. Mean age was 65.7. According to the use of anti plateletdrugs and anti thrombosis drugs, we divided the cases into two groups. Group A isconsisted of cases under medication with these drugs and group B is not undermedication with them. We compared two groups for age, sex, background disease,symptom, diagnosis, location of lesion, number of lesion, use of anti-ulcer drugs,

al.org