J Cancer Res Clin Oncol DOI 10.1007/s00432-015-1971-9
ORIGINAL ARTICLE – CLINICAL ONCOLOGY
Rapid discrimination of malignant lesions from normal gastric tissues utilizing Raman spectroscopy system: a meta‑analysis Huan Ouyang1 · Jiahui Xu2 · Zhengjie Zhu1 · Tengyun Long1 · Changjun Yu1
Received: 5 January 2015 / Accepted: 13 April 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Objectives To systematically analyze the diagnostic accuracy of Raman spectroscopy system (RAS) in the rapid diagnosis of gastric cancer with histopathology as the reference standard. Methods We searched a wide range of electronic databases for all published researches that assessed the diagnostic accuracy of RAS to detect gastric carcinoma. Full papers were obtained for potentially eligible studies and evaluated according to predefined criteria. The Quality Assessment of Diagnostic Accuracy Studies checklist was used to assess the quality of included studies. From each study, we extracted information on diagnostic performance of RAS. After exploring heterogeneity, we adopted a random effects model to pool related effect sizes.
Results The initial literature search identified 257 reference articles in which 15 relevant articles with 15 data sets were selected and reviewed. The pooled sensitivity and specificity of RAS in diagnosing gastric cancer were 0.89 (95 % CI 0.84–0.92) and 0.92 (95 % CI 0.88–0.95), respectively. The positive likelihood ratio, the negative likelihood ratio, and the area under the curve were 10 (95 % CI 6.5– 15.3), 0.13 (95 % CI 0.08–0.22), and 0.96 (95 % CI 0.94– 0.97), respectively. All the pooled estimates, calculated by random and fixed effect models, were similar. There was no evidence of considerable publication bias. Conclusions RAS is an objective and sensitive optical diagnostic technology for detecting gastric cancer and has advantages of being noninvasive to the body, real-time diagnosis, and ease of use. Consequently, it does deserve to be recommended.
Huan Ouyang and Jiahui Xu have contributed equally to this work.
Keywords Gastric cancer · Raman spectroscopy · Rapid diagnosis · Diagnostic trial · Meta-analysis
* Changjun Yu [email protected]
Introduction
Huan Ouyang [email protected] Jiahui Xu [email protected] Zhengjie Zhu [email protected] Tengyun Long [email protected] 1
Department of Gastrointestinal Surgery, The First Affiliated Hospital of Anhui Medical University, No. 218, Jixi Road, Hefei 230022, China
2
Department of Geriatric Medicine, Anhui Provincial Hospital, No. 17, Lujiang Road, Hefei 230001, China
Currently, the morbidity of gastric cancer is falling, but it is still the fourth most common malignancy and also the second leading cause of cancer deaths in humans worldwide (Milne et al. 2007; Yeoh 2007). It is estimated that a total of 989,600 new gastric carcinomas cases and 738,000 deaths occurred in 2008 around the world, accounting for 8 % of the total cancer cases and 10 % of total deaths (Jemal et al. 2011). Hence, a lot of efforts have been undertaken to improve survival of patients with gastric cancer, the most significant of which is the development of a screening system and early detection (Graham and Asaka 2010). It has been demonstrated that the early identification and
13
effective treatment of premalignant lesions are crucial to decrease the mortality of patients with gastric cancer (Clark et al. 2006). Furthermore, being screened at an appropriate frequency improves the probability of tumor control with immediate treatment and, thus, avoids invasion and metastasis and enables its complete surgical resection (Lopes et al. 2011; Sandler 2010). However, inchoate distinction of dysplasia in the stomach could be very difficult to detect by conventional diagnostic methods. At present, white-light endoscopy is the most important conventional tool to diagnose gastric carcinoma, nonetheless, it heavily relies on the visual observation of gross morphological changes in pathologic tissues, leading to a poor diagnostic accuracy and bringing patients unexpected complications. In recent years, many studies have suggested that early cancer cells can generate characteristic protein that was encoded by oncogenes activated and different from normal cells in structures and conformations. In particular, RAS, which measures inelastic light-scattering processes of molecular vibrations and provides specific spectroscopic features about biomolecular structures and conformations of tissues, has its special advantages in detecting subtle biomolecular alterations associated with neoplastic transformation, and it may be useful for us to further demonstrate the molecular mechanism of the cancer development process (Caspers et al. 2003; Gniadecka et al. 1997; Huang et al. 2003; Mahadevan-Jansen et al. 1998; Mahadevan-Jansen and Richards-Kortum 1996; Stone et al. 2000). Hence, RAS is a promising tool for the diagnosis and characterization of neoplastic progression of tissues with high diagnostic specificities at the bimolecular level (Liu et al. 1992). Moreover, it is also a technique which has advantages of being noninvasive to the body, real-time diagnosis, and ease of use. Recently, many studies (Bergholt et al. 2010, 2011a, b, c, 2013; Duraipandian et al. 2012; Hu et al. 2008; Huang et al. 2010; Jin and Mao 2014; Kawabata et al. 2008, 2011; Luo et al. 2013; Teh et al. 2008a, b, 2010) had shown that RAS might be a useful discriminative tool to identify the malignant lesions of the stomach through ex or in vivo tissues. However, these studies were inconclusive because of mono-centric, inadequate sample sizes and different diagnostic algorithms employed. We undertook a systematic review and meta-analysis of the literature published to systematically analyze the diagnostic performance of RAS on the detection of gastric cancer with histopathology as the reference standard.
Materials and methods Literature search Published articles were systematically searched using databases (PubMed, SDOS, CNKI, and CBM) up to October
13
J Cancer Res Clin Oncol
2014, to identify all clinical studies in English or Chinese language investigating RAS as a diagnostic tool for gastric carcinoma. The search was performed by the following key words: Raman, spectroscopy, spectra, spectrum, scattering, spectroscope, spectrometry, gastric, stomach, cancer, carcinoma, and neoplasm. The cited words and their combinations used were: Raman spectroscopy, Raman spectra, Raman spectrum, Raman scattering, Raman spectroscope, Raman spectrometry, gastric cancer, gastric carcinoma, stomach cancer, gastric neoplasm, and stomach neoplasm. The “related articles” function and references retrieved from articles were used to perform the search of all related studies, abstracts, and citations. Our literature search was restricted to studies conducted in humans. Study selection criteria After screening the titles and abstracts for all the related articles, we selected the potentially eligible researches and read the full texts to decide whether they were really eligible. Two reviewers (Huan Ouyang and Jiahui Xu) screened studies identified for inclusion and determined study eligibility, independently. Disagreements were adjudicated by a third reader (Changjun Yu). The studies were selected on the basis of the following criteria: (1) only human gastric tissues, no matter ex and in vivo, were detected by RAS with histopathology as the golden reference standard. (2) Studies were required to provide a fourfold table could be constructed for true positive (TP), false negative (FN), false positive (FP), and true negative (TN) values. It included a control group (healthy subjects or patients with other confirmed diagnoses) not having gastric cancer. (3) RAS was independently used to diagnose gastric cancer or combined with other procedures. Excluded criteria: (1) papers not providing relevant data on diagnostic performance were excluded as were reviews or duplicate reports. (2) Studies without a control group including case reports and case series were excluded, as were studies that involved nonhuman subjects. (3) Conference abstracts were eliminated as these typically do not undergo rigorous peer review and their results may not be final. Data extraction and quality assessment Methodological and technical data, the number of patients, criteria used to select control and case groups, number of TP, FP, TN, and FN were extracted from each study (Lean et al. 2009). The quality of each study was assessed by using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) guidelines (Schuetz et al. 2010), which is an established, evidence-based tool for systematic reviews of diagnostic studies designed for diagnostic accuracy
J Cancer Res Clin Oncol
tests. For quality assurance, from the QUADAS checklist, we choose the study in that the total score was greater than or equal to 9 points. The same two reviewers (Huan Ouyang and Jiahui Xu) extracted the data and assessed the study quality independently. The third reviewer (Changjun Yu) assessed all the discrepant items to resolve the disagreement. Statistical analysis Data analysis was implemented utilizing Meta-DiSc 1.4, Stata 12.0 and SPSS 16.0 statistic software. For all tests, P values
ORIGINAL ARTICLE – CLINICAL ONCOLOGY
Rapid discrimination of malignant lesions from normal gastric tissues utilizing Raman spectroscopy system: a meta‑analysis Huan Ouyang1 · Jiahui Xu2 · Zhengjie Zhu1 · Tengyun Long1 · Changjun Yu1
Received: 5 January 2015 / Accepted: 13 April 2015 © Springer-Verlag Berlin Heidelberg 2015
Abstract Objectives To systematically analyze the diagnostic accuracy of Raman spectroscopy system (RAS) in the rapid diagnosis of gastric cancer with histopathology as the reference standard. Methods We searched a wide range of electronic databases for all published researches that assessed the diagnostic accuracy of RAS to detect gastric carcinoma. Full papers were obtained for potentially eligible studies and evaluated according to predefined criteria. The Quality Assessment of Diagnostic Accuracy Studies checklist was used to assess the quality of included studies. From each study, we extracted information on diagnostic performance of RAS. After exploring heterogeneity, we adopted a random effects model to pool related effect sizes.
Results The initial literature search identified 257 reference articles in which 15 relevant articles with 15 data sets were selected and reviewed. The pooled sensitivity and specificity of RAS in diagnosing gastric cancer were 0.89 (95 % CI 0.84–0.92) and 0.92 (95 % CI 0.88–0.95), respectively. The positive likelihood ratio, the negative likelihood ratio, and the area under the curve were 10 (95 % CI 6.5– 15.3), 0.13 (95 % CI 0.08–0.22), and 0.96 (95 % CI 0.94– 0.97), respectively. All the pooled estimates, calculated by random and fixed effect models, were similar. There was no evidence of considerable publication bias. Conclusions RAS is an objective and sensitive optical diagnostic technology for detecting gastric cancer and has advantages of being noninvasive to the body, real-time diagnosis, and ease of use. Consequently, it does deserve to be recommended.
Huan Ouyang and Jiahui Xu have contributed equally to this work.
Keywords Gastric cancer · Raman spectroscopy · Rapid diagnosis · Diagnostic trial · Meta-analysis
* Changjun Yu [email protected]
Introduction
Huan Ouyang [email protected] Jiahui Xu [email protected] Zhengjie Zhu [email protected] Tengyun Long [email protected] 1
Department of Gastrointestinal Surgery, The First Affiliated Hospital of Anhui Medical University, No. 218, Jixi Road, Hefei 230022, China
2
Department of Geriatric Medicine, Anhui Provincial Hospital, No. 17, Lujiang Road, Hefei 230001, China
Currently, the morbidity of gastric cancer is falling, but it is still the fourth most common malignancy and also the second leading cause of cancer deaths in humans worldwide (Milne et al. 2007; Yeoh 2007). It is estimated that a total of 989,600 new gastric carcinomas cases and 738,000 deaths occurred in 2008 around the world, accounting for 8 % of the total cancer cases and 10 % of total deaths (Jemal et al. 2011). Hence, a lot of efforts have been undertaken to improve survival of patients with gastric cancer, the most significant of which is the development of a screening system and early detection (Graham and Asaka 2010). It has been demonstrated that the early identification and
13
effective treatment of premalignant lesions are crucial to decrease the mortality of patients with gastric cancer (Clark et al. 2006). Furthermore, being screened at an appropriate frequency improves the probability of tumor control with immediate treatment and, thus, avoids invasion and metastasis and enables its complete surgical resection (Lopes et al. 2011; Sandler 2010). However, inchoate distinction of dysplasia in the stomach could be very difficult to detect by conventional diagnostic methods. At present, white-light endoscopy is the most important conventional tool to diagnose gastric carcinoma, nonetheless, it heavily relies on the visual observation of gross morphological changes in pathologic tissues, leading to a poor diagnostic accuracy and bringing patients unexpected complications. In recent years, many studies have suggested that early cancer cells can generate characteristic protein that was encoded by oncogenes activated and different from normal cells in structures and conformations. In particular, RAS, which measures inelastic light-scattering processes of molecular vibrations and provides specific spectroscopic features about biomolecular structures and conformations of tissues, has its special advantages in detecting subtle biomolecular alterations associated with neoplastic transformation, and it may be useful for us to further demonstrate the molecular mechanism of the cancer development process (Caspers et al. 2003; Gniadecka et al. 1997; Huang et al. 2003; Mahadevan-Jansen et al. 1998; Mahadevan-Jansen and Richards-Kortum 1996; Stone et al. 2000). Hence, RAS is a promising tool for the diagnosis and characterization of neoplastic progression of tissues with high diagnostic specificities at the bimolecular level (Liu et al. 1992). Moreover, it is also a technique which has advantages of being noninvasive to the body, real-time diagnosis, and ease of use. Recently, many studies (Bergholt et al. 2010, 2011a, b, c, 2013; Duraipandian et al. 2012; Hu et al. 2008; Huang et al. 2010; Jin and Mao 2014; Kawabata et al. 2008, 2011; Luo et al. 2013; Teh et al. 2008a, b, 2010) had shown that RAS might be a useful discriminative tool to identify the malignant lesions of the stomach through ex or in vivo tissues. However, these studies were inconclusive because of mono-centric, inadequate sample sizes and different diagnostic algorithms employed. We undertook a systematic review and meta-analysis of the literature published to systematically analyze the diagnostic performance of RAS on the detection of gastric cancer with histopathology as the reference standard.
Materials and methods Literature search Published articles were systematically searched using databases (PubMed, SDOS, CNKI, and CBM) up to October
13
J Cancer Res Clin Oncol
2014, to identify all clinical studies in English or Chinese language investigating RAS as a diagnostic tool for gastric carcinoma. The search was performed by the following key words: Raman, spectroscopy, spectra, spectrum, scattering, spectroscope, spectrometry, gastric, stomach, cancer, carcinoma, and neoplasm. The cited words and their combinations used were: Raman spectroscopy, Raman spectra, Raman spectrum, Raman scattering, Raman spectroscope, Raman spectrometry, gastric cancer, gastric carcinoma, stomach cancer, gastric neoplasm, and stomach neoplasm. The “related articles” function and references retrieved from articles were used to perform the search of all related studies, abstracts, and citations. Our literature search was restricted to studies conducted in humans. Study selection criteria After screening the titles and abstracts for all the related articles, we selected the potentially eligible researches and read the full texts to decide whether they were really eligible. Two reviewers (Huan Ouyang and Jiahui Xu) screened studies identified for inclusion and determined study eligibility, independently. Disagreements were adjudicated by a third reader (Changjun Yu). The studies were selected on the basis of the following criteria: (1) only human gastric tissues, no matter ex and in vivo, were detected by RAS with histopathology as the golden reference standard. (2) Studies were required to provide a fourfold table could be constructed for true positive (TP), false negative (FN), false positive (FP), and true negative (TN) values. It included a control group (healthy subjects or patients with other confirmed diagnoses) not having gastric cancer. (3) RAS was independently used to diagnose gastric cancer or combined with other procedures. Excluded criteria: (1) papers not providing relevant data on diagnostic performance were excluded as were reviews or duplicate reports. (2) Studies without a control group including case reports and case series were excluded, as were studies that involved nonhuman subjects. (3) Conference abstracts were eliminated as these typically do not undergo rigorous peer review and their results may not be final. Data extraction and quality assessment Methodological and technical data, the number of patients, criteria used to select control and case groups, number of TP, FP, TN, and FN were extracted from each study (Lean et al. 2009). The quality of each study was assessed by using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) guidelines (Schuetz et al. 2010), which is an established, evidence-based tool for systematic reviews of diagnostic studies designed for diagnostic accuracy
J Cancer Res Clin Oncol
tests. For quality assurance, from the QUADAS checklist, we choose the study in that the total score was greater than or equal to 9 points. The same two reviewers (Huan Ouyang and Jiahui Xu) extracted the data and assessed the study quality independently. The third reviewer (Changjun Yu) assessed all the discrepant items to resolve the disagreement. Statistical analysis Data analysis was implemented utilizing Meta-DiSc 1.4, Stata 12.0 and SPSS 16.0 statistic software. For all tests, P values
