Experimental Helicobacter Pylori Infection in Humans

1220 COMMENTARIES

H pylori validate animal models for pathogenesis ...... studies. In addition to its use in studying the Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from pathogenesis of infectious diseases, Experimental Helicobacter pylori infection inducing challenge experiments have been used to evaluate the infection in humans: a multifaceted initial efficacy of vaccines before con- ducting large scale field tests for many challenge infectious diseases, including enteric pathogens.9 Typically, this step is under- P Michetti taken after basic research has provided data regarding potential protective anti- ...... gens, and allowed for a description of the host immune response. Then, Is there a scientific rationale for the use of an infection challenge ideally, animal models that mimic human model for Helicobacter pylori vaccine development in humans? infection and response are used to test efficacy before human studies are con- hallenge experiments have been the conditions required for exposure to sidered. Finally, candidate vaccine prep- arations should then be evaluated for an important method of studying H pylori to lead to chronic gastric safety and immunogenicity in humans, the pathogenesis of many infec- infection, and the early clinical and C outside of the challenge setting, to mini- tious diseases and of evaluating initial pathological events following colonisa- mise exposure of volunteers only to the efficacy of vaccines before large scale tion of the gastric mucosa.5 Indeed, the 12 most promising candidates. In this field tests are conducted. In challenge infection is typically acquired in child- instance, however, it is of central impor- experiments, infections are deliberately hood but most studies related to H pylori tance to determine whether the proposed induced under carefully controlled pathogenesis and host response are human challenge model is not only suited and monitored conditions to healthy conducted, for obvious ethical and to reproduce the natural infection but also research volunteers. Induced infections practical reasons, in adults. As young corresponds to the population that the are usually either self limiting or can be children are not always chronically vaccine is intended to protect. fully treated within a short period of infected following initial infection,6 a Vaccine development against H pylori time. Because physicians should be better understanding of the early phase started over 10 years ago, after a proof dedicated to alleviating disease and of the infection may prove useful in the of principle was established in mice avoiding harm to patients, this type of development of novel therapies and infected with Helicobacter felis.10 11 In this experiment may cause uncomfortable vaccines. Although it is not entirely model, and later in H pylori mouse symptoms and evoke serious moral clear that the early events in H pylori models, several protective antigens of concerns. It should be appreciated how- infection in adults are identical to those ever that clinical research commonly H pylori were identified. Urease was also in children during natural infection, shown to be effective in a vaccine involves risks to subjects that are not adult infection has been reported and http://gut.bmj.com/ outweighed by medical benefits but are 7 preparation administered to infected has led to typical chronic infection. A H animals.12 Urease, cytotoxin associated justified by the potential to acquire new pylori challenge model in adults may knowledge.3 In that regard, infection gene A (CagA), vacuolating cytotoxin thus present an unique opportunity to inducing challenges are not necessarily (VacA), and H pylori neutrophil activat- obtain information on the pathogenesis more ethically problematic than phase I ing protein (HpNAP) were also shown of this bacterium, which may be poten- trials aimed at determining maximum to be safe and immunogenic in humans tially useful in the development of tolerated doses of medications. Like any once their protective potential had been

new drugs or obtaining knowledge on 13 on September 30, 2021 by guest. Protected copyright. clinical research, challenge experiments established in mice. Although these the host response crucial for vaccine should be conducted by competent antigens are protective in mice, and development. investigators according to sound pro- elicit both humoral and cellular immune tocols that incorporate appropriate By inducing experimental H pylori responses, no clear immune correlate of infection in healthy volunteers, the protection could be identified in mice, safeguards to ensure the safety of 8 volunteers. Because these experiments report of Graham and colleagues in this which prevents the use of an immuno- may provide valuable information that issue of Gut has certainly contributed to logical test to evaluate the protective might not be otherwise obtained, lead to our understanding of this infection (see potential of a candidate vaccine in novel therapies, or speed up vaccine page 1235). This study provides valu- humans.14 Field trials thus represent development that will ultimately spare able information on the low inoculum the alternative, with protection as the morbidity or death from infectious dis- size required for infection, the impact of primary outcome. As H pylori infection is eases and reduce exposure of large some virulence factors, the symptoms acquired early in life, this would imply groups in field trials, challenge experi- associated with acute H pylori infection, that field trials would have to be ments may be justified.4 However, the the physiological changes induced by conducted in children, in areas of high scientific rationale should be carefully H pylori, as well as on the rapid devel- incidence (that is, on the less favoured examined for any given pathogen and opment of histological changes pre- children of most societies). If a pro- model. When such a rationale exists, viously thought to be linked to chronic phylactic vaccine is to be developed, then the question of risks and discom- infection. This harvest of valuable infor- such a field approach would certainly forts should be addressed. mation would have been very difficult elicit several serious ethical concerns. Twenty years after the first culture of to obtain by other means, especially if These concerns may be progressively Helicobacter pylori there are still many early infection is to be analysed in addressed if benefit can be shown for gaps in our knowledge of this world- children. These data represent a well children but the incentive to invest in wide infection. In particular, little infor- controlled body of information, useful the long research needed to follow that mation is available regarding the precise to the scientific community, to better approach would certainly be helped by mode of transmission of the infection, understand transmission and to further positive results in adults. As natural

www.gutjnl.com COMMENTARIES 1221 infection is rare in adults, a challenge associated with decreased eradication therefore needs to think carefully before model would then be required to pro- rates. Despite this, full confidence that deciding to use this model in humans gress substantially towards prophylactic H pylori infection, even with a fully for the development of a vaccine against Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from vaccine development. In this context, antibiotic sensitive strain, can be eradi- H pylori. the model developed by Graham and cated is lacking. Volunteers that might Gut 2004;53:1220–1221. 8 colleagues certainly represents an remain infected would then be at doi: 10.1136/gut.2004.042135 advance. The value of this adult model increased risk, albeit a small risk, of in predicting vaccine induced protection morbidity. In addition, as the precise Correspondence to: Professor P Michetti, in children however would further transmission mode of is as yet Division of Gastroenterology and Hepatology, H pylori BH10N-545, Centre Hospitalier Universitaire depend on the assumptions that both unclear, there was a small risk that the Vaudois, CH-1011 Lausanne, Switzerland; infection and vaccine induced protective experimental strain would pass to other [email protected] responses are similar in adults and individuals. In the course of their study, children. In addition, the model would Graham et al achieved 100% eradication REFERENCES not be appropriate for testing the rate, and no apparent transmission of 1 Sack DA, Tacket CO, Cohen MB, et al. Validation protective potential of CagA, as the the infection was observed, possibly of a volunteer model of cholera with frozen strain was selected as CagA negative related to the precautions undertaken bacteria as the challenge. Infect Immun for safety reasons. As virulence factors to minimise this risk. The size of the 1998;66:1968–72. often represent important vaccine anti- experiment however was far too small 2 Tacket CO, Binion SB, Bostwick E, et al. Efficacy of bovine milk immunoglobulin gens, this is certainly a limitation of the to ascertain that the strain can be safely concentrate in preventing illness after Shigella model which could be resolved only eradicated and is not transmitted, with flexneri challenge. Am J Trop Med Hyg after extensive safety data are obtained type II errors for these observations 1992;47:276–83. 3 Miller FG, Grady C. The ethical challenge of with the challenge model. When the being possible. Therefore, the challenge infection-inducing challenge experiments. Clin goal is ultimately to develop a therapeu- model can not be considered safe, but as Infect Dis 2001;33:1028–33. tic vaccine however, testing could be progress is made towards the control of 4 Marwick C. Volunteers in typhoid infection study will aid future vaccine development. JAMA conducted by vaccinating adults natu- H pylori infection, the threshold for 1998;279:1423–4. rally infected with H pylori.15 As far as acceptability of this model, as for other 5 Suerbaum S, Michetti P. Helicobacter pylori vaccine development against H pylori is infections, is likely to shift. infection. N Engl J Med 2002;347:1175–86. 6 Rowland M, Kumar D, Daly L, et al. Low rates of considered, it is not yet clearly estab- After the experiment had started, an Helicobacter pylori reinfection in children. lished that an infection inducing chal- expert panel assembled on 30 November Gastroenterology 1999;117:336–41. lenge model is likely to facilitate or 2000 by the National Institute for 7 Graham DY, Alpert LC, Smith JL, et al. Iatrogenic Campylobacter pylori infection is a cause of speed up vaccine development. Health, Division of Microbiology and epidemic achlorhydria. Am J Gastroenterol In exploring the ethical justifica- Infectious Diseases, concluded that the 1988;83:974–80. tion of particular infection inducing development of the H pylori infection 8 Graham DY, Opekun AR, Osato MS, et al. Challenge model for Helicobacter pylori challenge experiments, the nature of challenge model should not be pursued infection in human volunteers. Gut 17 the infection has to be considered. at this time. In brief, the reasons for 2004;53:1235–43. Legitimate challenge experiments may this conclusion were that (1) there was 9 Tacket CO, Kotloff KL, Losonsky G, et al. include infection likely to produce mild no compelling evidence for the need for Volunteer studies investigating the safety and efficacy of live oral El Tor Vibrio cholerae O1 http://gut.bmj.com/ symptoms not interfering with the sub- a Helicobacter vaccine in the USA, (2) no vaccine strain CVD 111. Am J Trop Med Hyg ject’s daily activities and those that are clear vaccination strategy was defined 1997;56:533–7. self limiting or that can be adequately for the USA, (3) volunteers were at risk 10 Czinn SJ, Cai A, Nedrud JG. Protection of germ- 3 free mice from infection by Helicobacter felis after eradicated with certainty. From the of unsuccessful eradication with pos- active oral or passive IgA immunization. Vaccine reports of the two investigators that sible late consequences, (4) animals 1993;11:637–42. infected themselves with H pylori and models had not been fully investigated 11 Chen M, Lee A, Hazell S. Immunisation against gastric helicobacter infection in a mouse/ from iatrogenic infection reports, it was for vaccine efficacy, and (5) more con- Helicobacter felis model. Lancet

not expected that acute H pylori would ventional field trials were available to 1992;339:1120–1. on September 30, 2021 by guest. Protected copyright. cause more than mild dyspeptic symp- measure vaccine efficacy. These conclu- 12 Corthe´sy-Theulaz I, Porta N, Glauser M, et al. Oral immunization with Helicobacter pylori toms. To further minimise this risk, the sions however may underestimate the urease B subunit as a treatment against strain was obtained from a patient with full extent of the H pylori problem Helicobacter infection in mice. Gastroenterology no alarming symptom, histological worldwide and do not address the 1995;109:115–21. 13 Keller WC, Michetti P. Vaccination against changes, or complications of infection. scientific rationale for the use of a Helicobacter pylori—an old companion of man. The primary concern however was that challenge model for Helicobacter vaccine 2001;1:795–802. experimental H pylori infection may development. The H pylori challenge 14 Ruggiero P, Peppoloni S, Rappuoli R, et al. The become chronic, and that a treatment experiment reported by Graham quest for a vaccine against Helicobacter pylori: et al how to move from mouse to man? Microbes Infect regimen with 100% effectiveness to was conducted under the best available 2003;5:749–56. eradicate the infection was not avail- conditions and resulted in minimal 15 Michetti P, Kreiss C, Kotloff KL, et al. Oral able.16 Careful attention was paid by the harm to volunteers. The results shed immunization with urease and Escherichia coli heat-labile enterotoxin is safe and immunogenic investigators to selection of the chal- light on interesting aspects of the in Helicobacter pylori-infected adults. lenge strain. It was derived from a pathogenesis of the bacteria, and pro- Gastroenterology 1999;116:804–12. patient that was eradicated without vide validation information for the 16 Perri F, Qasim A, Marras L, et al. Treatment of Helicobacter pylori infection. Helicobacter difficulty, and the strain tested for its development of animal models. How- 2003;8(suppl 1):53–60. lack of resistance to the antibiotics ever, there are many more questions to 17 Blaser MJ, Lang D. NIAID Workshop: commonly used to treat the infection. be answered before this model can be Development of a Standardized Human Challenge Model for Helicobacter pylori, 1999. However, it was CagA negative and the validated and proven useful for vaccine http://www.niaid.nih.gov/dmid/meetings/ lack of this virulence factor had been development. The scientific community HpChallengeMdl.htm.

www.gutjnl.com 1222 COMMENTARIES

Mucosal resection low vertical axis which extends laterally ...... along the interior luminal wall. Such LSTs can be further subdivided into Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from granular and non-granular types.7 Endoscopic mucosal resection: an In this issue of Gut, Hurlstone and colleagues18 report their experience with evolving therapeutic strategy for non- colonoscopic resection of 82 LSTs of the colorectum (see page 1334). Firstly, polypoid colorectal neoplasia identification of these lesions was made with the help of dye spraying, under- R Kiesslich, M F Neurath lining the high impact of targeted chromoendoscopy during screening ...... colonoscopy. Using this approach, a total of 26 non-granular and 56 granular Endoscopic management for laterally spreading tumours of the LSTs were identified. The non-granular colorectum is a safe and effective treatment and may be an type of LSTs was more often associated alternative to surgery in selected patients with invasive disease and was fre- quently localised to the right colon. The theory of the adenoma-carcinoma olorectal cancer develops in Recently, it was shown in a prospective sequence is well established for polypoid approximately 5–6% of the adult randomised trial that magnifying endo- adenomas.19 Some authors have advo- population and is one of the scopy is superior to standard videoendo- C cated a new theory of cancer develop- leading causes of cancer death in scopy because of the possibility of ment with special attention to the Europe and the USA.12Screening colon- analysing stained colorectal lesions at 12 superficial forms of cancer that exhibit oscopy is the widely accepted gold higher magnifications. Differentiation non-polypoid growth. They postulated a standard for early diagnosis of colorectal between non-neoplastic and neoplastic higher risk of malignant invasion com- cancer and should be offered to patients lesions was possible however when 20 21 3–5 pared with polypoid neoplasms. older than 50 years. However, colon- using high magnification chromoendo- Hurlstone and colleagues15 observed oscopy is increasingly in competition scopy to differentiate between non- areas of high grade dysplasia in 25% of with computed tomographic virtual invasive and invasive neoplasias but flat adenomas whereas high grade 6 10 endoscopy. Therefore, it is essential to sensitivity decreased to 50%. dysplasia was observed in only 12% of understand whether total colonoscopy Initially, knowledge about non-poly- protuberant adenomas, consistent with rather than virtual endoscopy enables poid colorectal lesions was based on the idea that such flat lesions have a 16 the detection of all relevant premalig- observations in Japanese patients. high risk for malignancy. However, nant and malignant lesions as only early Similar observations have since been previous studies suggested that among diagnosis of neoplasias offers the possi- made in Western populations using non-polypoid lesions, only depressed bility for curative endoscopic or surgical magnifying chromoendoscopy. In fact, lesions have a higher risk of malignant resection. introduction of newly developed magni- infiltration. Up to a size of 2 cm, the rate http://gut.bmj.com/ Polypoid lesions are easy to detect by fying and high resolution endoscopes, as of malignant infiltration is 40% for endoscopy. In contrast, non-polypoid well as the renaissance of chromoendo- depressed lesions compared with only 7 lesions are often overlooked. The scopy, have led to a significant increase 1.3% for polypoid lesions.7 Surprisingly, endoscopist must be aware of flat and in the detection of non-polypoid lesions the rate of lymph node metastasis is depressed lesions because the primarily in the Western world. In recent pro- similar for polypoid (11%) and non- visible mucosal changes of such lesions spective chromoendoscopic studies, polypoid lesions (11%). Even for are often very discrete. In this context, 7–31% of diagnosed lesions were flat or depressed lesions with the tendency to on September 30, 2021 by guest. Protected copyright. the endoscopist should look for slight depressed.813–15Furthermore, Saitoh and grow towards the muscularis propria, colour changes, interruption of the colleagues8 observed that 62% of flat the rate of lymph node metastasis has capillary network pattern, slight defor- and depressed lesions could be judged been reported as only 6–7%.22 mation of the colonic wall, sponta- correctly by the help of intravital stain- Early colorectal carcinoma is defined neously bleeding spots, shape changes ing. Thus Western endoscopists have by its depth of infiltration (the carci- of the lesion with insufflation or defla- to recognise these newly discovered noma is invading the submucosal layer). tion of air, and interruption of the lesions. For this reason, it is essential The degree of submucosal invasion is innominate grooves.7 to discuss the malignant potential, classified into three types based on An important diagnostic tool for screening, and therapeutic strategies of depth of invasion: when less than one detection of non-polypoid colorectal such non-polypoid lesions. third of the submucosa is invaded the lesions consists of chromoendoscopy.8 Macroscopic appearance of colorectal stage is called sm1 and if more than two In contrast with ulcerative colitis where lesions can be classified according to the thirds are invaded the stage is sm3. pan-chromoendoscopy is favourable,9 Japanese classification of colorectal can- Stage sm2 is intermediate. Sm1 early targeted staining with indigo carmine cer (see fig 1).17 However, Kudo et al colorectal cancers without lymphatic or or methylene blue in a selective fashion have recommended a more practical venous invasion have almost no risk of is sufficient for non-polypoid colorectal classification.7 Here, early non-polypoid lymph node metastasis.7 Thus they can lesions.10 Chromoendoscopy helps to colorectal neoplasias can be classified as be removed endoscopically. unmask such lesions in the colon and slightly elevated (small flat adenoma Severe dysplastic epithelium confined to delineate its borders. By the help of ,1 cm), laterally spreading tumours to the mucosa is called severe dysplasia magnifying endoscope, the stained sur- (large flat adenoma .1 cm) or or intramucosal carcinoma, although face and crypt architecture can be depressed (see fig 2). Specifically, later- there is some controversy as to the analysed, and differentiation between ally spreading tumours (LST) of the terminology for this particular lesion. non-neoplastic and neoplastic lesions colorectum are defined as lesions of Japanese pathologists tend to use becomes possible with high accuracy.11 more than 10 mm in diameter with a the term ‘‘intramucosal or mucosal

www.gutjnl.com COMMENTARIES 1223

718 Rate of resection. Peacemeal resection har- submucosal bours the risk of local recurrence, as 18 Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from Chromoendoscopy Type Macroscopic appearance invasion shown by Hurlstone and colleagues, and an intense follow up programme should therefore be added (see fig 2). Ip Pedunculated Complete resection should always be intended. However, chromoendoscopy Not necessary Protruded after resection helps to identify remnant type Ips Subpedunculated Low risk of islands of neoplastic tissue which can be 18 23 submucosal removed by argon beam coagulation. Is Sessile 18 invasion Hurlstone and colleagues reported a local recurrence rate of 17% and recur- Flat type IIb Flat rent disease in 10%. In patients with local recurrence, eight of 10 underwent IIa Flat-elevated further endoscopic resection without Superficial residual disease present at 24 months elevated of follow up. These data are comparable Highly type Flat-elevated IIa + IIc with previous Japanese experience.24 25 recommended with depression High risk of In the current study, Hurlstone and submucosal colleagues18 thus showed that neither IIc Depressed Slightly invasion the diagnosis nor the treatment of non- type depressed polypoid lesions is a Japanese phenom- IIc + IIa enon or expertise. Western endoscopists should be aware of these lesions and need to include mucosal resection for Figure 1 Diagnostic classification17 and malignant potential of early colorectal cancer. flat and depressed colorectal lesions into their therapeutic strategies. carcinoma’’ whereas Western patholo- Therefore, this group of lesions should In conclusion, introduction of magni- gists tend to avoid this term to prevent always be treated. Those confined to the fying endoscopy in combination with unnecessary major surgery. However, mucosa or that only slightly invade the intravital staining has led to a signifi- these lesions are ideally suited for submucosa (without vessel infiltration) cant improvement in the diagnosis of endoscopic resection.22 can be completely removed and cured by non-polypoid neoplasms. Flat and Selection of suitable lesions for endo- endoscopic resection. If histology shows depressed lesions are now increasingly scopic resection is of crucial clinical that the cancer has massively invaded diagnosed in the Western world during importance. Hurlstone et al have devel- the submucosa or permeates the vessels, colonoscopy using this new techni- oped essential inclusion and exclusion additional surgery is required. que. In contrast, the diagnostic yield 18 criteria. According to these criteria, 56 Signs of advanced carcinoma are of alternative approaches such as http://gut.bmj.com/ of 82 identified LSTs underwent endo- depressed lesions .10 cm and unstruc- computed tomography virtual endoscopic mucosal resection and high tured crypt architecture after staining, scopy decreases for small lesions.626 ’’cure’’ rates were achieved. The results as well as lacking the lifting sign after Differentiated knowledge about non- show impressively that endoscopic man- saline injection. These lesions should polypoid lesions is essential to identify agement for LST is possible and justi- undergo surgery primarily. suitable lesions for endoscopic resection. fied. However, the question remains Flat lesions up to a size of 5 mm Endoscopic mucosal resection is the how the endoscopic or surgical manage- without suspected staining pattern or treatment of choice for defined non- on September 30, 2021 by guest. Protected copyright. ment of non-polypoid lesions should be depression can be removed by biopsy. polypoid lesions and offers a sufficient performed? Depressed lesions are often Lesions up to 2 cm should be resected and curative therapy. Because of the malignant and have been shown to en bloc whereas lesions larger than possible higher malignant potency of rapidly invade the submucosal layer. 2 cm can be treated by peacemeal non-polypoid lesions, an intense endoscopic follow up programme should be initiated and offers the possibility of Type Macroscopic appearance Size (cm) Recommended therapy repeat endoscopic treatment for local recurrence. Slightly elevated type <0.5 Biopsy Gut 2004;53:1222–1224. (flat adenoma) 0.51.0 Mucosal resection doi: 10.1136/gut.2004.043281 ...... Authors’ affiliations 1.02.0 Mucosal resection, en bloc R Kiesslich, M F Neurath, I Med Clinic, Laterally spreading type >2.0 Mucosal resection, peace meal Johannes Gutenberg University Mainz, Germany

Correspondence to: Dr R Kiesslich, I Med Klinik und Poliklinik, Johannes Gutenberg Universita¨t <1.0 Mucosal resection, en bloc Mainz, Langenbeckstr 1 55101 Mainz, Depressed type >1.0 Surgery Germany; [email protected]

Prerequisite for endoscopic therapy: neoplastic straining pattern, positive lifting sign after REFERENCES saline injection, maximum diameter <1/3 of the luminal circumference. 1 Grady WM. Genetic testing for high-risk colon cancer patients. Gastroenterology Figure 2 Classification of non-polypoid colorectal lesions7 and recommended therapy. 2003;124:1574–94.

www.gutjnl.com 1224 COMMENTARIES

2 Burt RW. Colon cancer screening. the diagnosis of neoplasia in flat and depressed tumours: a prospective analysis of endoscopic Gastroenterology 2000;119:837–53. lesions of the colorectum: a prospective analysis. mucosal resection. Gut 2004;53:1334–9. 3 Gatta G, Ciccolallo L, Capocaccia R, et al. Gut 2004;53:284–90. 19 Muto T, Bussey HJR, Morson BC. The evaluation Differences in colorectal cancer survival between 11 Kudo S, Tamura S, Nakajima T, et al. Diagnosis of cancer of the colon and rectum. Cancer Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from European and US populations: the importance of of colorectal tumorous lesions by magnifying 1975;36:2251. sub-site and morphology. Eur J Cancer endoscopy. Gastrointest Endosc 1996;44:8–14. 20 Kuramoto S, Oohara T. Flat early cancers of the 2003;39:2214–22. 12 Konishi K, Kaneko K, Kurahashi T, et al. A large intestine. Cancer 1989;15:950–5. 4 Weir HK, Thun MJ, Hankey BF, et al. Annual comparison of magnifying and nonmagnifying 21 Minamoto T, Sawaguchi K, Ohta T, et al. report to the nation on the status of cancer, 1975– colonoscopy for diagnosis of colorectal polyps: A Superficial-type adenomas and adenocarcinomas 2000, featuring the uses of surveillance data for prospective study. Gastrointest Endosc of the colon and rectum: a comparative cancer prevention and control. J Natl Cancer Inst 2003;57:48–53. morphological study. Gastroenterology 2003;95:1276–99. 13 Rembacken BJ, Fujii T, Cairns A, et al. Flat and 1994;106:1436–43. 5 Winawer S, Fletcher R, Rex D, et al. Colorectal depressed colonic neoplasms: a prospective study 22 Ajioka Y, Watanabe H, Kazama S, et al. Early cancer screening and surveillance: clinical of 1000 colonoscopies in the UK. Lancet colorectal cancer with special reference to the guidelines and rationale—Update based on new 2000;8:1211–14. superficial nonpolypoid type from a evidence. Gastroenterology 2003;124:544–60. 14 Kiesslich R, von Bergh M, Hahn M, et al. histopathologic point of view. World J Surg 6 Pickhardt PJ, Choi JR, Hwang I, et al. Computed Chromoendoscopy with indigocarmine improves 2000;24:1075–80. tomographic virtual colonoscopy to screen for the detection of adenomatous and 23 Brooker JC, Saunders BP, Shah SG, et al. colorectal neoplasia in asymptomatic adults. nonadenomatous lesions in the colon. Endoscopy Treatment with argon plasma coagulation N Engl J Med 2003;349:2191–200. 2001;33:1001–6. reduces recurrence after piecemeal resection of 7 Kudo S, Kashida H, Tamura T, et al. Colonoscopic 15 Hurlstone DP, Cross SS, Adam I, et al. A large sessile colonic polyps: a randomized trial diagnosis and management of nonpolypoid prospective clinicopathological and endoscopic and recommendations. Gastrointest Endosc colorectal cancer. World J Surg evaluation of flat and depressed colorectal lesions 2002;55:371–5. 2000;24:1081–90. in the United Kingdom. Am J Gastroenterol 24 Tanaka S, Haruma K, Oka S, et al. Clinicopatho- 8 Saitoh Y, Waxman I, West AB, et al. Prevalence 2003;98:2543–9. logic features and endoscopic treatment of super- and distinctive biologic features of flat colorectal 16 Matsui T, Yao T, Iwashita A. Natural history of ficially spreading colorectal neoplasms larger adenomas in a North American population. colorectal cancer. World J Surg than 20 mm. Gastrointest Endosc 2001;54:62–6. Gastroenterology 2001;120:1657–65. 2000;24:1022–8. 25 Tamura S, Nakajo K, Yokoyama Y, et al. 9 Kiesslich RJ, Fritsch M, Holtmann M, et al. 17 Japanese Research Society for Cancer of the Evaluation of endoscopic mucosal resection for Methylene blue-aided chromoendoscopy for the Colon and Rectum. General rules for clinical and laterally spreading rectal tumors. Endoscopy detection of intraepithelial neoplasia and colon pathological studies on cancer of the colon, 2004;36:306–12. cancer in ulcerative colitis. Gastroenterology rectum and anus. Part II. Histological 26 Laghi A, Iannaccone R, Carbone I, et al. Detection 2003;124:880–8. classification. Jpn J Surg 1983;23:574. of colorectal lesions with virtual computed 10 Hurlstone DP, Cross SS, Adam I, et al. Efficacy of 18 Hurlstone DP, Sanders DS, Cross SS, et al. tomographic colonography. Am J Surg high magnification chromoscopic colonoscopy for Colonoscopic resection of lateral spreading 2002;183:124–31.

Hepatocellular carcinoma without Doppler ultrasonography12 13 ...... may improve early diagnosis further. However, no adequate comparisons between all competing techniques are Diagnostic biopsy for hepatocellular available to date. Most studies have used two or a maximum of three carcinoma in cirrhosis: useful, neces- modalities and have come to different conclusions.14 The role of positron emis- sary, dangerous, or academic sport? sion tomography in different modes is http://gut.bmj.com/ also not clear but it does not seem to be J Scho¨lmerich, D Schacherer very helpful.15 Well performed large studies on modalities for early diagnosis ...... of HCC are therefore welcome. In this issue of Gut, a study from At present, in experienced hands, ultrasound guided biopsy using Italy16 reports on 4375 patients with either cut needles or fine needle aspiration, seems to be newly detected cirrhosis and no focal technically appealing and most effective for diagnosis of lesions at initial US followed over a on September 30, 2021 by guest. Protected copyright. hepatocellular carcinoma period of nine years (see page 1356). Of 4581 patients agreeing to partici- epatocellular carcinoma (HCC) is diagnosisseemstobeareasonablegoal pate, 206 had nodular lesions at initial the eighth most frequent tumour in improving the poor prognosis of this US. Patients were followed every Hentity worldwide (one million per cancer.2–4 Interestingly, it has recently 4–6 months using a fetoprotein (AFP) year). The incidence in the USA and been shown by comparison of time and abdominal US. If a lesion was elsewhere has increased, which is prob- periodsthatbesidesbeingpartofthe diagnosed and AFP levels were above ably due to the rise in the incidence of more recent cohort, tumour staging was 400 ng/ml, HCC was assumed and hepatitis C virus infection. A number of the only independent predictor of survi- treated accordingly. Lesions associated cohort studies have shown that surveil- val.1 Thus it seems obvious that detection with an AFP level below 400 ng/ml lance using abdominal ultrasound (US) of early carcinoma in liver cirrhosis would underwent US guided fine needle biopsy improves management of HCC develop- be a reasonable goal. Progress made in (FNB) performed by an experienced ing in cirrhotic patients as it allows early recent years regarding imaging, in parti- gastroenterologist using fine cutting detection and application of poten- cular spiral computed tomography and needles and non-cutting spinal needles. tially curative treatments.1 Thus far it using the increased vascularity of HCC, Choice between needles was left to the is not entirely clear whether surveillance allowing for better scanning protocols, individual examiner. When histology or increases survival as no randomised con- has increased the detection rate signifi- cytology indicated HCC, the patient was trolledtrialshavebeenconductedto cantly. In addition, magnetic resonance treated. If biopsy revealed a negative answer this question. However, as even- imaging has become the diagnostic pro- result, computed tomography and, in tually curative treatments such as resec- cedure of choice in some institutions. one centre, contrast enhanced magnetic tion, ablation, and in particular liver However, most experts still advocate liver resonance imaging were performed. A transplantation have emerged as effective biopsy.5–11 total of 688 new focal lesions of means of treating HCC as long as it is not In addition, improved US techniques 6–42 mm in diameter were detected. too large and too widespread, early such as contrast enhancement with and In 294 patients (43%), lesions were

www.gutjnl.com COMMENTARIES 1225

~5000 newly diagnosed patients nodules using imaging are taken for with cirrhosis granted. Only 274 nodules in 4375 patients were biopsied because their size Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from was (20 mm in diameter. 4581 agreed to participate

COMPLICATIONS 206 excluded (initially 4375 followed for With regard to the second question, data detected nodular lesions) 13108 months are also contradictory. Overall, complications of US guided diagnostic and 688 new focal lesions therapeutic procedures on focal liver lesions are rare. In a large series from the same authors, 16 648 guided biop- 294 ≤20 mm 117 >30 mm 277 2130 mm sies and 3035 therapeutic procedures in 13 222 patients were analysed. Overall mortality was 0.06% (none after diag- 274 US guided FNP 20 AFP >400 ng/ml nostic procedures).26 However, the bleeding risk is greater in HCC as these 16 Figure 1 Overview of the study. AFP, a fetoprotein; US, ultrasound; FNP, fine needle biopsy. lesions are usually hypervascular. In a series of 159 HCC patients undergoing (20 mm; 258 of these (87.6%) were by any other technique. However, a fine needle aspiration, four (2.5%) bled diagnosed as having HCC. In those with recent study following 12 cases of significantly, one of whom died.27 Thus very small nodules ((10 mm), HCC lesions with an initial fine needle bleeding has to be considered as a was diagnosed in 69% of cases in aspiration diagnosis of regenerating possible complication of biopsy in this contrast with 91% in those with nodules nodules revealed subsequently that 10 tumour entity, in particular with sub- of 11–20 mm. Only 20 of 294 patients of these patients had a HCC. Thus a capsular localisation of the tumour. had an elevated AFP .400 ng/ml. The negative biopsy result is probably not Another ‘‘complication’’ is the risk of remaining 274 underwent biopsy (fig 1). helpful in suspected HCC, as in any mismanagement after interpretative In 245 patients, a correct diagnosis was other suspected tumour. If the initial errors. It is sometimes very difficult to made which was confirmed by follow biopsy is negative, repeat biopsy leads to distinguish well differentiated HCC up: 210 patients had HCC, four had a diagnostic gain in one third of patients, from benign regenerative nodules, and non-Hodgkin lymphoma, and 31 were in particular if the first biopsy provided criteria used by pathologists need to be found to be benign, as judged by follow a non-diagnostic sample (necrosis) or a very clear.28 With fine needle aspiration up. Twenty nine of 274 cases were false false negative result due to well differ- in particular, diagnosis may be difficult; negatives. Thus in a screening popula- entiated HCC.24 It is not entirely clear if aspiration of necrotic material seems to tion, more than 66% of small nodules there is a major difference between be another limiting factor for correct arising in cirrhotic livers may prove to cutting and aspiration needles—early interpretation.10 be HCC and 90% of those can be studies did not find a difference Others consider needle tract seeding http://gut.bmj.com/ 16 identified using US guided FNB. when an experienced pathologist was of malignant cells an important risk This topic has been discussed pre- involved.25 The positive predictive value factor. This risk was found to be 5.1% 17–22 viously in this and other journals of imaging modalities has recently been (three patients) in a series of 59 15–20 years ago. Very similar results shown to be rather high.11 However, this patients.29 It was smaller in a series of regarding the sensitivity of AFP and is not the case in all studies. A recent 122 patients (1.6% (two patients)).30 No good sensitivity of US guided biopsy study from Japan demonstrated that recurrence was observed after local 21 have been reported. A larger series many nodules detected by ultrasonogra- excision in both series. However, in conducted 10 years ago again revealed phy could not even be found by com- other series the risk of seeding was on September 30, 2021 by guest. Protected copyright. 23 similar results. puted tomography.8 Another study used much lower.31 These series were how- Why then do we need new studies digital subtraction angiography and ever the result of questionnaires, with and why do we still not know how to magnetic resonance imaging and found all the uncertainties inherent in this proceed? This may be due to the fact that imaging alone was sufficient to approach. that three questions still remain un- diagnose HCC in 58% of well differen- In our view, there are risks present answered: tiated and 87% of moderately and poorly but these are negligible and, in particu- (1) How good is the technique and differentiated nodules in 139 patients lar, the risk of needle seeding should not which is the preferred technical with chronic liver disease, where 207 preclude biopsy if the result influences modality to perform a biopsy? nodules had been found by periodic treatment choice. The risk of seeding surveillance.9 Imaging alone allowed for has to be weighted against a 2.5% (2) How dangerous is such a procedure diagnosis of HCC in 60% of cases: 55% of unnecessary surgery rate when patients and does it interfere with later patients with nodules of 10 mm in are not biopsied.11 An exception may be treatment? diameter or less required biopsy. With patients with advanced liver cirrhosis (3) Is a biopsy necessary at all, and does these conflicting results and the data of who are awaiting transplantation, inde- it change the outcome? Caturelli and colleagues,16 it appears pendent of a possible tumour. Here that at present, in experienced hands, tumour seeding could be disastrous EFFICACY OF US GUIDED BIOPSY US guided biopsy using either cut and a ‘‘wait and see strategy’’ may be When considering the historical review needles or fine needle aspiration seems reasonable. However, a biopsy may be given above and data from the paper to be technically appealing and most helpful because in the case of a tumour, published in this issue of Gut,16 it effective. However, the approach chosen one would have to do an adequate appears at first glance that US guided by the authors does not necessitate ‘‘bridging therapy’’ until transplantation biopsy has a very high efficacy in biopsy in all patients as those patients (for example, chemoembolisation) or defining HCC. A rate of 10% false with AFP levels .400 ng/ml and those consider early transplantation (for negatives is probably difficult to beat with typical features of HCC in larger example, by a living donor).

www.gutjnl.com 1226 COMMENTARIES

NEED FOR BIOPSY surgeons, hepatologists, and interven- 17 Schwerk WB, Schmitz-Moormann P. Ultrasonically guided fine-needle biopsies in The question of whether biopsy is tionalists, and will be dependent on the neoplastic liver disease: cytohistologic diagnoses necessary is also difficult to answer with treatment options available for the and echo pattern of lesions. Cancer Gut: first published as 10.1136/gut.2004.042135 on 11 August 2004. Downloaded from the data available. Examining again the individual patient. 1981;48:1469–77. 16 18 Sautereau D, Vire O, Cazes PY, et al. Value of paper of Caturelli and colleagues it is Gut 2004;53:1224–1226. sonographically guided fine needle aspiration obvious that in many patients an early doi: 10.1136/gut.2004.040816 biopsy in evaluating the liver with sonographic diagnosis was made which may allow abnormalities. Gastroenterology ...... 1987;93:715–18. for different treatment options, such as 19 Brady PG, Goldschmid S, Chappel G, et al. A percutaneous ethanol injection, resec- Authors’ affiliations comparison of biopsy techniques in suspected tion, or liver transplantation. A recent J Scho¨lmerich, D Schacherer, Department of focal liver disease. Gastrointest Endosc study demonstrated that percutaneous Internal Medicine I, University Regensburg, 1987;33:289–92. Regensburg, Germany 20 Limberg B,Ho¨pker WW, Kommerell B. Histologic ethanol injection had a similar survival differential diagnosis of focal liver lesions by rate as surgery and was actually ultrasonically guided fine needle biopsy. Gut approximately 60% at five years for Correspondence to: Professor J Scho¨lmerich, 1987;28:237–41. 21 Sbolli G, Fornari F, Civardi G, et al. Role of patients with small HCC.32 However, Department of Internal Medicine I, University Regensburg, D-93042 Regensburg, Germany; ultrasound guided fine needle aspiration biopsy the European Association for the Study [email protected] in the diagnosis of hepatocellular carcinoma. Gut of Liver currently recommends a ‘‘wait 1990;31:1303–5. 33 22 Buscarini L, Fornari F, Bolondi L, et al. and see policy’’ for small lesions as Ultrasound-guided fine-needle biopsy of focal only half of these are considered to be REFERENCES liver lesions: techniques, diagnostic accuracy and HCC. The current study contradicts this 1 Sangiovanni A, Del Ninno E, Fasani P, et al. complications. A retrospective study on 2091 34 Increased survival of cirrhotic patients with a biopsies. J Hepatol 1990;11:344–8. assumption. However, the question of hepatocellular carcinoma detected during 23 Fornari F, Filice C, Rapaccini GL, et al. Small whether obtaining a liver biopsy affects surveillance. Gastroenterology ((3 cm) hepatic lesions. Results of clinical outcome cannot yet be answered 2004;126:1005–14. sonographically guided fine-needle biopsy in 385 2 Befeler AS, Di Bisceglie AM. Hepatocellular patients. Dig Dis Sci 1994;39:2267–75. with the current information. Early carcinoma: diagnosis and treatment. 24 Caturelli E, Biasini E, Bartolucci F, et al. detection of HCC seems to increase the Gastroenterology 2002;122:1609–19. Diagnosis of hepatocellular carcinoma chance of treatment, at least in Hong 3 Llovet JM, Bruix J. Systematic review of complicating liver cirrhosis: utility of repeat ultrasound-guided biopsy after unsuccessful first Kong,35 and also seems to improve randomized trials for unresectable hepatocellular carcinoma: chemoembolization improves sampling. Cardiovasc Intervent Radiol prognosis, as recently described in this survival. Hepatology 2003;37:429–42. 2002;25:295–9. journal.36 It is doubtful if random- 4 Dick EA, Taylor-Robinson SD, Thomas HC, et al. 25 Atterbury CE, Enriquez RE, Desuto-Nagy GI, ised controlled studies will ever be Ablative therapy for liver tumours. Gut et al. Comparison of the histologic and cytologic 2002;50:733–9. diagnosis of liver biopsies in hepatic cancer. conducted to answer this question 5 Lencioni R, Cioni D, Crocetti L, et al. Magnetic Gastroenterology 1979;76:1352–7. definitively. resonance imaging of liver tumors. J Hepatol 26 Giorgio A, Tarantino L, de Stefano G, et al. In the past, outcome for most 2004;40:162–71. Complications after interventional sonography of 6 Stoker J, Romijn MG, de Man RA, et al. focal liver lesions: a 22-year single-center patients, regardless of stage, was Prospective comparative study of spiral computer experience. J Ultrasound Med believed to be poor. However, as trans- tomography and magnetic resonance imaging for 2003;22:193–205. plantation results in a prolonged disease detection of hepatocellular carcinoma. Gut 27 Bret PM, Labadie M, Bretagnolle M, et al. 2002;51:105–7. Hepatocellular carcinoma: diagnosis by free survival or even cure, as shown by 7 Brancatelli G, Baron RL, Peterson MS, et al. percutaneous fine needle biopsy. Gastrointest http://gut.bmj.com/ some studies (for example, Mazzaferro Helical CT screening for hepatocellular carcinoma Radiol 1988;13:253–5. and colleagues37), biopsy with needle in patients with cirrhosis: frequency and causes of 28 Roncalli M. Hepatocellular nodules in cirrhosis: false-positive interpretation. AJR Am J Roentgenol focus on diagnostic criteria on liver biopsy. A seeding would actually deny the patient 2003;180:1007–14. Western experience. Liver Transpl treatment and therefore change the 8 Tanaka Y, Sasaki Y, Katayama K, et al. 2004;10:S9–15. clinical outcome significantly. Thus in Probability of hepatocellular carcinoma of small 29 Takamori R, Wong LL, Dang C, et al. Needle- hepatocellular nodules undetectable by computed tract implantation from hepatocellular cancer: is our opinion, patients waiting for a tomography during arterial portography. needle biopsy of the liver always necessary? Liver transplant should only be biopsied if Hepatology 2000;31:890–8. Transpl, 2000;6:67–72. 30 Durand F, Regimbeau JM, Belghiti J, et al. AFP levels are less than 20 ng/ml, if the 9 Horigome H, Nomura T, Saso K, et al. Limitations on September 30, 2021 by guest. Protected copyright. of imaging diagnosis for small hepatocellular Assessment of the benefits and risks of lesion cannot be characterised as HCC carcinoma: comparison with histological finding. percutaneous biopsy before surgical resection of using other modalities (all of them J Gastroenterol Hepatol 1999;14:559–65. hepatocellular carcinoma. J Hepatol should be performed), and if the pre- 10 Souto E, Gores GJ. When should a liver mass 2001;35:254–8. suspected of being a hepatocellular carcinoma be 31 Smith EH. Complications of percutaneous sence of a HCC favourably alters the biopsied? Liver Transpl 2000;6:73–5. abdominal fine-needle biopsy. Radiology patient’s candidacy for liver transplan- 11 Torzilli G, Minagawa M, Takayama T, et al. 1991;178:253–8. tation. If resection appears to be the best Accurate preoperative evaluation of liver mass 32 Yamamoto J, Okada S, Shimada K, et al. Treatment strategy for small hepatocellular option, biopsy may or may not be lesions without fine-needle biopsy. Hepatology 1999;30:889–93. carcinoma: comparison of long-term results performed. When palliative treatment 12 Fracanzani AL, Burdick L, Borzio M, et al. after percutaneous ethanol injection therapy and is planned, biopsy makes sense in order Contrast-enhanced doppler ultrasonography in surgical resection. Hepatology 2001;34:707–13. to avoid unnecessary treatment. the diagnosis of hepatocellular carcinoma and 33 Bruix J, Sherman M, Llovet JM, et al. Clinical premalignant lesions in patients with cirrhosis. management of hepatocellular carcinoma. In summary, the study of Caturelli Hepatology 2001;34:1109–12. Conclusions of the Barcelona-2000 EASL and colleagues16 provides long sought 13 Schlottmann K, Klebl F, Zorger N, et al. Contrast- conference. J Hepatol 2001;35:421–30. after and important information on the enhanced ultrasound allows for interventions of 34 Caturelli E, Ghittoni G, Roselli P, et al. Fine needle hepatic lesions which are invisible on biopsy of focal liver lesions: the hepatologist’s possible efficacy of US guided biopsy of conventional B-mode. Z Gastroenterol point of view. Liver Transpl 2004;10:S26–9. HCC in experienced hands and will 2004;42:303–10. 35 Yuen MF, Cheng CC, Lauder IJ, et al. Early probably define the gold standard of 14 Colagrande S, La Villa G, Bartolucci M, et al. detection of hepatocellular carcinoma increases Spiral computed tomography versus ultrasound in the chance of treatment: Hong Kong experience. quality of this procedure. However, it the follow-up of cirrhotic patients previously Hepatology 2000;31:330–5. does not answer the question of treated for hepatocellular carcinoma: a 36 Bolondi L, Sofia S, Siringo S, et al. Surveillance whether biopsy should be performed prospective study. J Hepatol 2003;39:93–8. programme of cirrhotic patients for early 15 Ho CL, Yu SCH, Yeung DWC. 11C-Acetate PET diagnosis and treatment of hepatocellular and, if so, in which patients when HCC imaging in hepatocellular carcinoma and other carcinoma: a cost effectiveness analysis. Gut is suspected on the basis of a cirrhotic liver masses. J Nucl Med 2003;44:213–21. 2001;48:251–9. liver. This remains to some extent a 16 Caturelli E, Solmi L, Anti M, et al. Ultrasound 37 Mazzaferro V, Regalia E, Doci R, et al. Liver guided fine needle biopsy of early hepatocellular transplantation for the treatment of small bedside decision, which probably should carcinoma complicating liver cirrhosis: a hepatocellular carcinomas in patients with be made after discussion between multicentre study. Gut 2004;53:1356–62. cirrhosis. N Engl J Med 1996;334:693–9.