- Email: [email protected]
Endoscopic Measurement of Colorectal Polyps: How Do We Measure Up?
Gastroenterology 2016;-:1–3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
SELECTED SUMMARIES Philip S. Schoenfeld, Section Editor John Y. Kao, Section Editor STAFF OF CONTRIBUTORS Joseph Anderson, White River Junction, VT Johanna L. Chan, Houston, TX Matthew A. Ciorba, St. Louis, MO Massimo Colombo, Milan, Italy Gregory A. Cote, Charleston, SC Evan S. Dellon, Chapel Hill, NC Alex Ford, Leeds, United Kingdom Lauren B. Gerson, San Francisco, CA David S. Goldberg, Philadelphia, PA Samir Gupta, San Diego, CA
Reena Khanna, London, Ontario, Canada W. Ray Kim, Rochester, MN Paul Y. Kwo, Indianapolis, IN Uma Mahadevan, San Francisco, CA Baha Moshiree, Miami, FL Swati G. Patel, Ann Arbor, MI Laurent Peyrin-Biroulet, Vandoeuvre-lès-Nancy, France Jesus Rivera-Nieves, San Diego, CA
Endoscopic Measurement of Colorectal Polyps: How Do We Measure Up? Anderson BW, Smyrk TC, Anderson KS, et al. Endoscopic overestimation of colorectal polyp size. Gastrointest Endosc 2015;83:201–2088. Accurate measurement of polyp size is crucial to screening or surveillance for colorectal cancer (CRC). Adenoma size has been shown to be an important predictor of long-term CRC risk (N Engl J Med 1992;326:658–662). Accordingly, adenomas that are measured as 1 cm may warrant shorter surveillance intervals than smaller polyps, as recommended by the United States Multicenter Task Force guidelines (Gastroenterology 2012;143:844–857). Owing to technical issues, such as polyp fragmentation, the endoscopist’s assessment is often the only measurement of polyp size in clinical practice. Since the nascent stages of endoscopy, there have been concerns regarding the endoscopic measurement of polyps (Dis Colon Rectum 1982;25:669–672). Many of the earlier studies examining the accuracy of endoscopists in assessing polyp size have suffered from low numbers of polyps in their analyses (Gastrointest Endosc 1997;46:492–496; Gastrointest Endosc 1997;46:497–502). Polyp size measurement, which may be subjective, can have a large impact on CRC risk stratification and recommended surveillance intervals. One study reported that nearly half of all exams with polypectomies resulted in earlier surveillance owing to inaccurate size measurements (Dis Colon Rectum 2013;56:315–321). In the present study, the authors examine and compare the measurement of polyp size by endoscopists and pathologists for polyps removed in 2012 at the Mayo Clinic in Rochester (Gastrointest Endosc 2015;83:201–8). The authors also examined endoscopist- and polyp-related factors that may be associated with inaccurate endoscopic polyp size estimation. In their analysis, the authors included only those polyps which were removed in toto and had both endoscopic and pathologic size estimation. In adults with multiple polyps, only the largest lesion meeting inclusion criteria was included. As part of the center’s usual routine, polyps were placed in formalin immediately after resection and sent
Sameer Saini, Ann Arbor, MI Ekihiro Seki, Los Angeles, CA Amit Singal, Dallas, TX Ryan W. Stidham, Ann Arbor, MI Akbar Waljee, Ann Arbor, MI Sachin Wani, Aurora, CO Alastair J. M. Watson, Norwich, United Kingdom Yana Zavros, Cincinnati, OH
to pathology for processing. Objective pathologic size, measured with a ruler to the nearest millimeter, was defined as the “gold standard.” Polyp sizes were dichotomized into those that were <1 versus 1 cm. Polyp sizes were also categorized into those that were overcalled, or had a larger endoscopic estimation than the pathology-based size and those that were undercalled, or had a smaller endoscopic estimation. Polyps were also broken down by gender, morphology, histology, and location. Fifty staff gastroenterologists and 12 gastroenterology fellows performed all of the 9146 colonoscopies during the study period for the year 2012. Of 6067 polyps resected during the study period, 1528 lesions (25.2%; median size, 0.4 cm) from 1422 patients met inclusion criteria. With regard to the endoscopists, the median number of colonoscopies performed during 2012 was 146, the median polyp detection rate was 44%, and the median endoscopic experience was 15 years. An analysis of the endoscopically estimated sizes revealed clustering, especially around 1 and 2 cm. The estimated sizes in the clusters were largely represented by the overcalled measurements versus the undercalled or accurate measurements. Conversely, the observed curve for the pathology-based measurements was smooth without clustering. Furthermore, 72% of the 99 polyps endoscopically measured as 1 cm and 86% of the 43 polyps endoscopically measured as 2 cm were found to be smaller as measured by pathology. Conversely, a smaller proportion (47%) of the size measurements for 186 polyps estimated to be 0.5 cm were found to be overcalled. Overall, endoscopic overcalled measurements were twice as common as undercalled measurements based on a cutoff polyp size of 1 cm. Univariate analyses revealed that polyps that were hyperplastic, flat, or proximally located were more likely to have overcalled measurements. Overestimation was also more common in women but not influenced by patient age. After multivariable regression analyses, nonpedunculated polyp morphology was the only significant risk factor for endoscopic overestimation (P ¼ .015) with female gender having borderline significance (P ¼ .055). No endoscopist factors influenced overcall rates including procedure numbers, adenoma detection rate, experience, or presence of fellow.
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
2
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
Selected Summaries
Gastroenterology Vol.
Underestimation was the most prevalent in sessile serrated adenomas/polyps. Of the sessile serrated adenomas/ polyps with an endoscopic estimated size of <1 cm, 9.8% had been undercalled, a proportion that was significantly higher than hyperplastic polyps or adenomas (P ¼ .0002). Age, gender, site, and configuration did not affect risk for undercalled measurements. Comment. There are some important findings in the current study that have implications for practicing endoscopists. The investigators observed a clustering of endoscopic size measurements around 1 and 2 cm. When compared with the pathology measurement, which is the gold standard, most of the size estimates in the clusters were overcalled. Furthermore, the ratio of overcalled to undercalled polyp size measurements was higher for the 1and 2-cm categories than for the 0.5-cm group. Thus, the observed clustering demonstrated a tendency for the participating endoscopists to round up to a larger size estimate, especially for the larger polyps. The overestimation of the polyps’ sizes likely resulted in an earlier follow up colonoscopy, especially for those polyps that were estimated to be 1 cm. Nearly one-half (46%) of all polyps endoscopically measured to be 1 cm were actually <1 cm on pathology measurement. These individuals might have received more aggressive surveillance, 3 years versus the appropriate interval of 5–10 years. Conversely, a smaller proportion of patients had undercalled polyp measurements and might have received inappropriately less aggressive surveillance. In the current study, the authors estimated that 10% of all examinations would have resulted in inappropriate surveillance if size alone was the only factor in determining if the adenoma was advanced. One of the study’s strong points is the inclusion of common endoscopy-related factors that may be associated with endoscopic over estimation of polyp size. Although only nonpedunculated morphology, flat or sessile, was associated with size overestimation after multivariable analysis, other clinically important factors were examined. Two factors, hyperplastic histology and flat morphology, were identified in the univariate analyses, but these findings failed to attain significance in the multivariable analysis. Inaccurate measurement seems plausible in hyperplastic polyps because these lesions are typically flat and often are reported to lack a clear border. Interestingly, the factor associated with undercalled polyp size measurements was sessile serrated polyp histology, again likely owing to flat configuration and unclear margins. Thus, endoscopists may want to be especially careful when estimating the size of flat polyps in the right side of the colon, which may represent sessile serrated adenoma/ polyps. Female gender was found to have a borderline significant association with endoscopic overestimation of polyp size. The authors propose an explanation that, because women may have colons with smaller diameters, the endoscopic appearance may be altered (Surg Radiol Anat 1992;14:251–257). It has also been shown that larger adenomas in women tend to be proximal and
-,
No.
-
flat as compared to men (Am J Gastroenterol 2011;106:2018–2022). Thus, many of the factors in the analysis may have been interrelated. Limitations of the study include the exclusion of almost 75% of all of the polyps resected in 2012. Excluding adults with larger, proximal polyps may have also had an impact on the overcall rate. Including these polyps in the analysis would have been interesting because many sessile serrated and hyperplastic polyps can present as flat lesions. Furthermore, histology, proximal location, and morphology were factors that were associated with inaccurate polyp size measurement in this analysis. Another limitation is the retrospective design. A prospective design would have allowed for standardization or careful recording of various polypectomy techniques. Optical lens distortion, use of diathermy, or lack thereof, saline, cap assistance, and water immersion could all influence visual size estimation. Furthermore, postpolypectomy processes also influence polyp size. An older study observed that formalin fixing may shrink resected polyps (Gastrointest Endosc 1997;46:497–502). This also raises the possibility of pathology measurement bias. Measurement of the polyp in the endoscopy suite before fixation might have provided useful data to examine this issue. In addition, the assessment of polyp morphology, the only significant factor associated with size overestimation, can be subjective. Previous studies have demonstrated a higher percentage of flat lesions (Gastroenterology 2007;133:42–47). A prospective design might have allowed for uniform assessment of polyp morphology as well as a method for capturing the size of polyps that fragmented. For example, photodocumentation of polyps next to the snare may have allowed for confirmation of size and morphology for all polyps. This study has several strengths, including a large cohort, which provided sufficient power to analyze the issue of polyp measurement. There were 50 endoscopists who performed the examinations with extensive experience, increasing the generalizability of the results. In addition, the assessment of endoscopic size measurement was conducted in a population undergoing colonoscopy with newer endoscopic technology. Endoscopes have evolved in terms of optics and definition. These changes may affect estimation by the even the most experienced trained eye. Finally, important factors such as polyp morphology and endoscopist factors were taken into account. In summary, the current study present provocative data, that may have implications for CRC screening and surveillance with colonoscopy. The authors state that accuracy of endoscopic polyp size estimation as a standard quality metric may become an added incentive for improved outcomes. Unfortunately, the size estimation for most polyps in current practice cannot be based on pathologic measurement. In the current study, more than one-half (56%) were excluded owing to fragmentation. Proposed new technological tools, designed to improve endoscopic estimation include the use of disposable graduated biopsy forceps (World J Gastroenterol 2015;21:623–628), calibrated hoods (Gastroenterol Res Pract 2014;2014:714294; Int J
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
-
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
2016
Selected Summaries
Colorectal Dis 2015;30:933–938), and a ruler snare (Gastrointest Endosc 2015 Sep 14 [Epub ahead of print]). In the field of surgery, an augmented reality system is described in laparoscopic surgery (IEEE Trans Biomed Eng 2014;61:2609–2620). Although these technologies may offer assistance in polyp size estimation, a new study similar to the current with a prospective design may be the next step. For now, endoscopists may want to check their size estimations with their pathology reports. Q1
IMAD AHMAD Division of Gastroenterology
3
JOEL B. LEVINE Hepatology and Colon Cancer Prevention Program University of Connecticut School of Medicine Farmington, Connecticut JOSEPH C. ANDERSON Department of Veterans Affairs Medical Center White River Junction, Vermont and The Geisel School of Medicine at Dartmouth Hanover, New Hampshire The contents of this work do not represent the views of the Department of Veterans Affairs or the United States Government.
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
SELECTED SUMMARIES Philip S. Schoenfeld, Section Editor John Y. Kao, Section Editor STAFF OF CONTRIBUTORS Joseph Anderson, White River Junction, VT Johanna L. Chan, Houston, TX Matthew A. Ciorba, St. Louis, MO Massimo Colombo, Milan, Italy Gregory A. Cote, Charleston, SC Evan S. Dellon, Chapel Hill, NC Alex Ford, Leeds, United Kingdom Lauren B. Gerson, San Francisco, CA David S. Goldberg, Philadelphia, PA Samir Gupta, San Diego, CA
Reena Khanna, London, Ontario, Canada W. Ray Kim, Rochester, MN Paul Y. Kwo, Indianapolis, IN Uma Mahadevan, San Francisco, CA Baha Moshiree, Miami, FL Swati G. Patel, Ann Arbor, MI Laurent Peyrin-Biroulet, Vandoeuvre-lès-Nancy, France Jesus Rivera-Nieves, San Diego, CA
Endoscopic Measurement of Colorectal Polyps: How Do We Measure Up? Anderson BW, Smyrk TC, Anderson KS, et al. Endoscopic overestimation of colorectal polyp size. Gastrointest Endosc 2015;83:201–2088. Accurate measurement of polyp size is crucial to screening or surveillance for colorectal cancer (CRC). Adenoma size has been shown to be an important predictor of long-term CRC risk (N Engl J Med 1992;326:658–662). Accordingly, adenomas that are measured as 1 cm may warrant shorter surveillance intervals than smaller polyps, as recommended by the United States Multicenter Task Force guidelines (Gastroenterology 2012;143:844–857). Owing to technical issues, such as polyp fragmentation, the endoscopist’s assessment is often the only measurement of polyp size in clinical practice. Since the nascent stages of endoscopy, there have been concerns regarding the endoscopic measurement of polyps (Dis Colon Rectum 1982;25:669–672). Many of the earlier studies examining the accuracy of endoscopists in assessing polyp size have suffered from low numbers of polyps in their analyses (Gastrointest Endosc 1997;46:492–496; Gastrointest Endosc 1997;46:497–502). Polyp size measurement, which may be subjective, can have a large impact on CRC risk stratification and recommended surveillance intervals. One study reported that nearly half of all exams with polypectomies resulted in earlier surveillance owing to inaccurate size measurements (Dis Colon Rectum 2013;56:315–321). In the present study, the authors examine and compare the measurement of polyp size by endoscopists and pathologists for polyps removed in 2012 at the Mayo Clinic in Rochester (Gastrointest Endosc 2015;83:201–8). The authors also examined endoscopist- and polyp-related factors that may be associated with inaccurate endoscopic polyp size estimation. In their analysis, the authors included only those polyps which were removed in toto and had both endoscopic and pathologic size estimation. In adults with multiple polyps, only the largest lesion meeting inclusion criteria was included. As part of the center’s usual routine, polyps were placed in formalin immediately after resection and sent
Sameer Saini, Ann Arbor, MI Ekihiro Seki, Los Angeles, CA Amit Singal, Dallas, TX Ryan W. Stidham, Ann Arbor, MI Akbar Waljee, Ann Arbor, MI Sachin Wani, Aurora, CO Alastair J. M. Watson, Norwich, United Kingdom Yana Zavros, Cincinnati, OH
to pathology for processing. Objective pathologic size, measured with a ruler to the nearest millimeter, was defined as the “gold standard.” Polyp sizes were dichotomized into those that were <1 versus 1 cm. Polyp sizes were also categorized into those that were overcalled, or had a larger endoscopic estimation than the pathology-based size and those that were undercalled, or had a smaller endoscopic estimation. Polyps were also broken down by gender, morphology, histology, and location. Fifty staff gastroenterologists and 12 gastroenterology fellows performed all of the 9146 colonoscopies during the study period for the year 2012. Of 6067 polyps resected during the study period, 1528 lesions (25.2%; median size, 0.4 cm) from 1422 patients met inclusion criteria. With regard to the endoscopists, the median number of colonoscopies performed during 2012 was 146, the median polyp detection rate was 44%, and the median endoscopic experience was 15 years. An analysis of the endoscopically estimated sizes revealed clustering, especially around 1 and 2 cm. The estimated sizes in the clusters were largely represented by the overcalled measurements versus the undercalled or accurate measurements. Conversely, the observed curve for the pathology-based measurements was smooth without clustering. Furthermore, 72% of the 99 polyps endoscopically measured as 1 cm and 86% of the 43 polyps endoscopically measured as 2 cm were found to be smaller as measured by pathology. Conversely, a smaller proportion (47%) of the size measurements for 186 polyps estimated to be 0.5 cm were found to be overcalled. Overall, endoscopic overcalled measurements were twice as common as undercalled measurements based on a cutoff polyp size of 1 cm. Univariate analyses revealed that polyps that were hyperplastic, flat, or proximally located were more likely to have overcalled measurements. Overestimation was also more common in women but not influenced by patient age. After multivariable regression analyses, nonpedunculated polyp morphology was the only significant risk factor for endoscopic overestimation (P ¼ .015) with female gender having borderline significance (P ¼ .055). No endoscopist factors influenced overcall rates including procedure numbers, adenoma detection rate, experience, or presence of fellow.
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
2
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
Selected Summaries
Gastroenterology Vol.
Underestimation was the most prevalent in sessile serrated adenomas/polyps. Of the sessile serrated adenomas/ polyps with an endoscopic estimated size of <1 cm, 9.8% had been undercalled, a proportion that was significantly higher than hyperplastic polyps or adenomas (P ¼ .0002). Age, gender, site, and configuration did not affect risk for undercalled measurements. Comment. There are some important findings in the current study that have implications for practicing endoscopists. The investigators observed a clustering of endoscopic size measurements around 1 and 2 cm. When compared with the pathology measurement, which is the gold standard, most of the size estimates in the clusters were overcalled. Furthermore, the ratio of overcalled to undercalled polyp size measurements was higher for the 1and 2-cm categories than for the 0.5-cm group. Thus, the observed clustering demonstrated a tendency for the participating endoscopists to round up to a larger size estimate, especially for the larger polyps. The overestimation of the polyps’ sizes likely resulted in an earlier follow up colonoscopy, especially for those polyps that were estimated to be 1 cm. Nearly one-half (46%) of all polyps endoscopically measured to be 1 cm were actually <1 cm on pathology measurement. These individuals might have received more aggressive surveillance, 3 years versus the appropriate interval of 5–10 years. Conversely, a smaller proportion of patients had undercalled polyp measurements and might have received inappropriately less aggressive surveillance. In the current study, the authors estimated that 10% of all examinations would have resulted in inappropriate surveillance if size alone was the only factor in determining if the adenoma was advanced. One of the study’s strong points is the inclusion of common endoscopy-related factors that may be associated with endoscopic over estimation of polyp size. Although only nonpedunculated morphology, flat or sessile, was associated with size overestimation after multivariable analysis, other clinically important factors were examined. Two factors, hyperplastic histology and flat morphology, were identified in the univariate analyses, but these findings failed to attain significance in the multivariable analysis. Inaccurate measurement seems plausible in hyperplastic polyps because these lesions are typically flat and often are reported to lack a clear border. Interestingly, the factor associated with undercalled polyp size measurements was sessile serrated polyp histology, again likely owing to flat configuration and unclear margins. Thus, endoscopists may want to be especially careful when estimating the size of flat polyps in the right side of the colon, which may represent sessile serrated adenoma/ polyps. Female gender was found to have a borderline significant association with endoscopic overestimation of polyp size. The authors propose an explanation that, because women may have colons with smaller diameters, the endoscopic appearance may be altered (Surg Radiol Anat 1992;14:251–257). It has also been shown that larger adenomas in women tend to be proximal and
-,
No.
-
flat as compared to men (Am J Gastroenterol 2011;106:2018–2022). Thus, many of the factors in the analysis may have been interrelated. Limitations of the study include the exclusion of almost 75% of all of the polyps resected in 2012. Excluding adults with larger, proximal polyps may have also had an impact on the overcall rate. Including these polyps in the analysis would have been interesting because many sessile serrated and hyperplastic polyps can present as flat lesions. Furthermore, histology, proximal location, and morphology were factors that were associated with inaccurate polyp size measurement in this analysis. Another limitation is the retrospective design. A prospective design would have allowed for standardization or careful recording of various polypectomy techniques. Optical lens distortion, use of diathermy, or lack thereof, saline, cap assistance, and water immersion could all influence visual size estimation. Furthermore, postpolypectomy processes also influence polyp size. An older study observed that formalin fixing may shrink resected polyps (Gastrointest Endosc 1997;46:497–502). This also raises the possibility of pathology measurement bias. Measurement of the polyp in the endoscopy suite before fixation might have provided useful data to examine this issue. In addition, the assessment of polyp morphology, the only significant factor associated with size overestimation, can be subjective. Previous studies have demonstrated a higher percentage of flat lesions (Gastroenterology 2007;133:42–47). A prospective design might have allowed for uniform assessment of polyp morphology as well as a method for capturing the size of polyps that fragmented. For example, photodocumentation of polyps next to the snare may have allowed for confirmation of size and morphology for all polyps. This study has several strengths, including a large cohort, which provided sufficient power to analyze the issue of polyp measurement. There were 50 endoscopists who performed the examinations with extensive experience, increasing the generalizability of the results. In addition, the assessment of endoscopic size measurement was conducted in a population undergoing colonoscopy with newer endoscopic technology. Endoscopes have evolved in terms of optics and definition. These changes may affect estimation by the even the most experienced trained eye. Finally, important factors such as polyp morphology and endoscopist factors were taken into account. In summary, the current study present provocative data, that may have implications for CRC screening and surveillance with colonoscopy. The authors state that accuracy of endoscopic polyp size estimation as a standard quality metric may become an added incentive for improved outcomes. Unfortunately, the size estimation for most polyps in current practice cannot be based on pathologic measurement. In the current study, more than one-half (56%) were excluded owing to fragmentation. Proposed new technological tools, designed to improve endoscopic estimation include the use of disposable graduated biopsy forceps (World J Gastroenterol 2015;21:623–628), calibrated hoods (Gastroenterol Res Pract 2014;2014:714294; Int J
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
-
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
2016
Selected Summaries
Colorectal Dis 2015;30:933–938), and a ruler snare (Gastrointest Endosc 2015 Sep 14 [Epub ahead of print]). In the field of surgery, an augmented reality system is described in laparoscopic surgery (IEEE Trans Biomed Eng 2014;61:2609–2620). Although these technologies may offer assistance in polyp size estimation, a new study similar to the current with a prospective design may be the next step. For now, endoscopists may want to check their size estimations with their pathology reports. Q1
IMAD AHMAD Division of Gastroenterology
3
JOEL B. LEVINE Hepatology and Colon Cancer Prevention Program University of Connecticut School of Medicine Farmington, Connecticut JOSEPH C. ANDERSON Department of Veterans Affairs Medical Center White River Junction, Vermont and The Geisel School of Medicine at Dartmouth Hanover, New Hampshire The contents of this work do not represent the views of the Department of Veterans Affairs or the United States Government.
DIS 5.4.0 DTD YGAST60247_proof 25 January 2016 5:28 pm ce
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276