Clinical impact of endoscopic ultrasonography on the management of malignancies

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2004;2:1069 –1073

Clinical Impact of Endoscopic Ultrasonography on the Management of Malignancies JANAK N. SHAH, NUZHAT A. AHMAD, MICHELLE C. BEILSTEIN, GREGORY G. GINSBERG, and MICHAEL L. KOCHMAN Division of Gastroenterology, Department of Medicine, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

Background & Aims: There are limited data evaluating the clinical impact of endoscopic ultrasonography (EUS). We assessed the impact of EUS on management of known or suspected malignancies. Methods: Physicians requesting EUS were contacted before the examination and asked: “How would you manage this patient if EUS were not available?” Pre-EUS responses were recorded on standardized forms, and compared with management recommendations as determined by the same requesting physician after EUS. Endosonographers performing the study were blinded to the pre-EUS responses. Results: Responses were obtained from requesting physicians before and after EUS in 90 patients. Requesting physicians were composed of surgeons (33%), non–EUS-performing gastroenterologists (58%), oncologists (3%), internists (4%), and a pulmonologist (1%). After EUS, referring clinicians altered management plans in 46 of 90 patients (51%). By anatomic site, management changes occurred in 12 of 22 patients (56%) undergoing esophageal EUS, 9 of 15 (60%) undergoing gastric EUS, 21 of 43 (49%) undergoing pancreatic EUS, and 4 of 10 (40%) undergoing rectal EUS. There were no significant differences in the frequency of post-EUS management changes with respect to examination site (P ⴝ .76). EUS fine-needle aspiration (FNA) altered management in 9 of 20 (45%) patients. Management changes involved less complex or decreased risk-associated approaches in the majority (70%), and included 14 of 50 (28%) patients in whom surgical procedures were no longer planned. Conclusions: Based on EUS examination findings, clinicians requesting EUS alter patient management in one half of cases, and more often pursue a less-complicated approach. EUS substantially impacts clinical care, and should be used in appropriate settings to guide patient management.

here are limited data evaluating the clinical use of endoscopic ultrasonography (EUS) and its effect on patient management, especially as determined by the physician who is directly caring for the patient. Prior investigations examining this topic have been limited by retrospective designs or study methods that used the endosonographer’s clinical decision pattern rather than

T

the requesting physician’s management approach.1– 8 Results from these studies may not truly reflect the manner in which the clinicians that are directing patient care are using EUS findings in routine practice. Some investigators have proposed that the contribution of EUS to patient management has not been investigated thoroughly and that further studies are needed.9,10 Therefore, we decided to examine prospectively the impact of EUS in the evaluation of patients with known or suspected malignancies, with specific emphasis on the direct effect of EUS on clinical management, and from the perspective of the non–EUS-performing requesting physician.

Materials and Methods A prospective study was performed over a 6-month period (March 2002–August 2002) to evaluate the effect of EUS on patient management. All patients undergoing EUS for the evaluation of known or suspected malignancies were eligible for the study. The physicians requesting EUS for these patients were contacted before the EUS examination by telephone or direct personal communication by 1 of 2 study authors (J.N.S., M.C.B.), and were asked: “How would you manage this patient if EUS were not available?” Their pre-EUS responses were recorded on a standardized form. Endosonographers (N.A.A., G.G.G., M.L.K.) were blinded to the requesting physician responses. After the examination the referring clinicians were recontacted within 1 week of the procedure, informed of the EUS findings, and asked: (1) “What management plan will you recommend to this patient given the EUS findings?” and if the management strategy differed compared with the pre-EUS response, (2) “Is the recommended change in the management plan directly the result of the EUS findings?” Post-EUS responses were recorded on the standardized form. None of the requesting physicians performed EUS. Exclusion criteria included: (1) EUS referral by endosonographers or study authors, (2) inability to contact the requesting physician before EUS, and (3) prior communication between Abbreviations used in this paper: EUS, endoscopic ultrasonography; FNA, fine-needle aspiration. © 2004 by the American Gastroenterological Association 1542-3565/04/$30.00 PII: 10.1053/S1542-3565(04)00444-6

1070

SHAH ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 2, No. 12

Table 1. Patient, Requesting Physician, and EUS Examination Site Characteristics

Patient characteristics Mean age (range) Male sex Requesting physician specialties Gastroenterology Surgery Oncology Internal medicine Pulmonary Anatomic sites of EUS examinations Esophagus or mediastinum Stomach Pancreas Colorectal

Study group (n ⫽ 90)

Excluded group (n ⫽ 399)

P value

61 y (30–84) 47 (52%)

62 y (19–95) 212 (53%)

.45 .88

52 (58%) 30 (33%) 3 (3%) 4 (4%) 1 (1%)

248 (62%) 115 (29%) 23 (6%) 10 (3%) 3 (1%)

.44 .40 .44 .30 .56

22 (24%)

99 (25%)

.94

15 (17%) 43 (48%) 10 (11%)

68 (17%) 192 (48%) 40 (10%)

.93 .95 .76

the endosonographer and requesting physician regarding the management strategy. All other EUS examinations were included in a prospective and consecutive manner. Specific data that were collected included: patient demographics, medical specialties of referring physicians, EUS indications and examination results, pre- and post-EUS management plans as per the referring clinician, number of EUS– fine-needle aspiration (FNA) procedures leading to patient management changes, and number of patients in whom surgical procedures were no longer planned as a result of EUS findings. Management changes post-EUS were characterized as more complex or increased-risk associated, or less complex or decreased-risk associated. The order of increasing complexity was defined as: (1) clinical follow-up (including nonprocedural palliative measures), (2) radiologic imaging, (3) diagnostic endoscopy, (4) therapeutic endoscopy (ie, endoscopic mucosal resection) or interventional radiologic procedure/radiologyguided biopsy examination, (5) endoscopic retrograde cholangiopancreatography, (6) chemotherapy or radiation therapy, (7) surgical procedure, and (8) neoadjuvant therapy (preoperative chemotherapy and/or radiation therapy followed by surgery). Categoric variables were summarized by counts and proportions, and continuous variables were summarized by means, SDs, medians, and ranges. Pearson’s ␹2 test or Fisher exact test when cell counts had expected values of less than 5 were used to test for differences between the patients included in the study vs. those excluded. The Student t test was used for age comparisons. Pearson’s ␹2 test also was used to compare the frequency of post-EUS management changes with respect to examination site. All analyses were performed using SAS version 8.2 (SAS Institute, Cary, NC). All EUS examinations were performed using radial scanning echoendoscopes (GF-UM20 or GF-UM160; Olympus America, Inc., Melville, NY) by 1 of 3 experienced endosonographers (N.A.A., G.G.G., M.L.K.). EUS-FNA procedures were

performed by using curved linear array echoendoscopes (GFUC140P; Olympus America, Inc.; or FG-32 UA; Pentax, Tokyo, Japan). The University of Pennsylvania Institutional Review Board approved this study.

Results Of 489 patients undergoing EUS over the 6-month study period, 90 were included in the study group. The remainder were excluded for the following reasons: EUS referral by study authors (42%), prior communication between endosonographers and referring clinicians regarding management strategy (35%), and inability to contact referring clinicians before EUS (24%). Responses on patient management were obtained from requesting physicians before and after EUS for 90 examinations. Patient demographics, requesting physician specialties, and anatomic sites of EUS examinations for both study and excluded groups are summarized in Table 1. Indications for esophageal EUS included staging of known cancers (n ⫽ 12), evaluation of suspected cancers (n ⫽ 4), evaluation of mediastinal masses detected by cross-sectional imaging (n ⫽ 3), and evaluation of submucosal lesions (n ⫽ 3). Requesting physicians altered management in 56% of patients after esophageal EUS (Table 2). This involved a less-complex management strategy in more than one half (58%) of patients, and included 4 of 14 patients (29%) in whom surgery was no longer planned. Of the 4 patients undergoing EUS FNA, all had changes in management after the procedure. Gastric EUS procedures were performed to evaluate known or suspected cancers (n ⫽ 5) and submucosal masses (n ⫽ 10). Requesting physicians altered patient management in 60% of patients after gastric EUS (Table 3). A less-complex management strategy was involved in the Table 2. Esophageal/Mediastinal EUS Management plan changes post-EUS More complex Surgery to neoadjuvant therapy and surgery (n ⫽ 1) Chemotherapy to surgery (n ⫽ 1) Clinical follow-up (palliative measures) to neoadjuvant therapy and surgery (n ⫽ 1) Diagnostic endoscopy to endoscopic mucosal resection (n ⫽ 1) Imaging follow-up to chemotherapy (n ⫽ 1) Less complex Surgery to radiation and/or chemotherapy (n ⫽ 3) Surgery to imaging and/or endoscopic follow-up (n ⫽ 1) Imaging and/or endoscopic follow-up to clinical follow-up (n ⫽ 3) Number of EUS FNA leading to management change

12/22 (56%) 5/12 (42%)

7/12 (58%)

4/4 (100%)

December 2004

CLINICAL IMPACT OF EUS

Table 3. Gastric EUS Management plan changes post-EUS More complex Imaging follow-up to surgery (n ⫽ 1) Endoscopic follow-up to endoscopic mucosal resection (n ⫽ 1) Less complex Surgery to endoscopic mucosal resection (n ⫽ 1) Surgery to clinical follow-up (n ⫽ 1) Imaging and/or endoscopic follow-up to clinical follow-up (n ⫽ 5) Number of EUS FNA leading to management change

9/15 (60%) 2/9 (22%)

7/9 (78%)

0/1 (0%)

majority (78%), and included 2 of 5 patients in whom surgery was no longer considered. One of these cases involved a gastric polypoid lesion. Previous forceps biopsy examination revealed adenoma, but the referring gastroenterologist was concerned about an invasive lesion and, without EUS, was recommending a localized surgical resection. After EUS confirmed a mucosal-based lesion, the referring gastroenterologist changed the management plan to an endoscopic mucosal resection. The other case involved a submucosal mass. An interval increase in size raised the suspicion of a malignant stromal tumor. Without EUS the referring clinician was recommending surgical exploration. EUS showed a normal gastric wall layer pattern; the endoscopically seen abnormality corresponded to an area of extrinsic compression from the liver. Only 1 patient underwent EUS FNA of a gastric or perigastric site. In this case, FNA confirmed a gastrointestinal stromal tumor, but did not alter the plan to proceed with surgery. Indications for pancreatic EUS included the evaluation of solid pancreatic masses (n ⫽ 19), cystic lesions (n ⫽ 6), and suspected pancreatic masses (n ⫽ 18). Requesting physicians changed management strategies in 49% of patients after pancreatic EUS procedures (Table 4). This most often involved a less-complex approach (71%), and included 7 of 22 patients (32%) in whom surgery was no longer recommended. Four of 13 patients (31%) undergoing EUS FNA of the pancreas or peripancreatic lymph nodes had alterations of management strategy after the procedure. In 2 cases, referring clinicians altered patient management after the pancreatic EUS procedure, but these were not attributed directly to the EUS results. In 1 case, EUS identified a resectable pancreatic mass, but the patient was a high-risk surgical candidate and declined other therapies. In the second case, EUS imaging of the pancreas was suboptimal owing to anatomic deformity from duodenal stenosis, and a management change occurred as a result of subsequent magnetic resonance imaging findings that revealed unresectable pancreatic cancer.

1071

Rectal EUS procedures were performed to stage known cancers (n ⫽ 4), and to evaluate large rectal polyps (n ⫽ 3), perirectal masses (n ⫽ 2), and a submucosal lesion (n ⫽ 1). Requesting physicians altered management plans in 40% of patients after rectal EUS (Table 5). The change in management was less complex or decreased-risk associated in one half of patients. This included 1 of 9 patients (11%) in whom surgery was no longer recommended. In this particular case, EUS FNA diagnosed metastatic gastric adenocarcinoma to the perirectal region (manifested as a perirectal mass on cross-sectional imaging that was inaccessible to radiology-guided biopsy examination). If EUS were not available, the referring physician was planning for a surgical biopsy examination for diagnosis. Overall, requesting physicians altered management plans in 46 of 90 (51%) patients after EUS procedures. There were no differences in the proportion of post-EUS management changes with respect to examination site (P ⫽ .76 by ␹2 test). In the majority (70%) of patients, the change involved a less-complex or decreased-risk associated approach. Surgical procedures were no longer planned in 14 of 50 patients (28%) as a result of EUS. Nine of 20 patients (45%) had management changes after EUS-FNA procedures.

Discussion Although the clinical use of EUS appears to be gaining popularity, its true use in guiding patient care has not been well characterized previously. The interest in examining the clinical impact of EUS on patient management is not new, and others have attempted to study this topic. Prior studies have suggested that EUS and EUS FNA influence management decisions in 33%–

Table 4. Pancreatic EUS Management plan changes post-EUS More complex ERCP to surgery (n ⫽ 1) Radiology-guided biopsy to radiation and/or chemotherapy (n ⫽ 1) Imaging to surgery (n ⫽ 2) Imaging to ERCP (n ⫽ 2) Less complex Surgery to clinical follow-up (n ⫽ 3) Surgery to radiation and/or chemotherapy (n ⫽ 4) ERCP to clinical follow-up (n ⫽ 2) ERCP to imaging follow-up (n ⫽ 3) Radiology-guided biopsy examination to imaging follow-up (n ⫽ 2) Imaging to clinical follow-up (n ⫽ 1) Number of EUS FNA leading to management change

21/43 (49%) 6/21 (29%)

15/21 (71%)

ERCP, endoscopic retrograde cholangiopancreatography.

4/13 (31%)

1072

SHAH ET AL

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 2, No. 12

Table 5. Rectal EUS Management plan changes post-EUS More complex Surgery alone to neoadjuvant therapy and surgery (n ⫽ 2) Less complex Surgery to chemotherapy (n ⫽ 1) Neoadjuvant therapy and surgery to surgery alone (n ⫽ 1) Number of EUS FNA leading to management change

4/10 (40%) 2/4 (50%)

2/4 (50%)

1/2 (50%)

76% and 44%– 87%, respectively.1– 8,11,12 However, interpretation of these investigations has been limited because of various pitfalls. In several studies, the pre-EUS management strategy (eg, what would be performed if EUS were not available?) was determined by the endosonographer, and not by the clinician directly caring for the patient.1,3,7,8 This raises a potential bias and may not reflect actual patient care decisions. The largest study to date by Nickl et al.1 involved 428 subjects, and suggested that EUS changes treatment plans in 74%. In this study the endosonographers, and not the requesting physician, established the pre-EUS management plans. Jafri et al.11 attempted to overcome this limitation by recording clinical impressions and management strategy from the referring clinicians, and found management alterations in 48%. However, referring physicians were comprised of study authors in nearly one quarter of cases. Other investigations of the clinical impact of EUS may be less generalizable because of their retrospective designs.2,4 – 6 A few investigators have designed their studies to account for these potential limitations, but focused on particular indications at specific anatomic sites. Chang et al.12 found EUS FNA of pancreatic lesions influenced clinical management in 68% of patients, and their results were based on the clinical decisions of referring physicians. In a blinded assessment, Harewood et al.13 detected a change in treatment plan in 31% of patients undergoing EUS for staging of rectal cancer. We designed our study to reflect the actual manner in which clinicians requesting EUS were using the results, and included all the commonly imaged anatomic sites of examination. For this study, no referring physicians were endosonographers or study authors. Our results indicate that clinicians requesting EUS and EUS FNA alter patient care in one half of cases as a direct result of examination findings. This effect on clinical management is nearly equal for various anatomic sites, including esophagus (56%), stomach (60%), pancreas (49%), and rectum (40%). When management changes occur postEUS, they are most often (70%) more simplified or

less-risk associated. Perhaps most importantly, surgical procedures are no longer considered in over one quarter of patients as a result of EUS findings. Although cost savings were not evaluated in our study, the proportion of patients with more simplified management plans, including those no longer requiring surgery, may likely have been associated with overall cost savings. This has been described with EUS for esophageal cancer management.14 Our study may be limited because of the exclusion of a large number of patients undergoing EUS over the study period. However, most (77%) were excluded because of referral by study authors or because of prior communication regarding clinical management between referring clinicians and endosonographers. We purposely identified these criteria for exclusion to minimize potential bias from using the endosonographer’s clinical decision patterns. Moreover, the excluded group had similar patient and examination characteristics as compared with the study group. Thus, if these patients were to be included, we would anticipate similar study results. We conclude that EUS provides useful information to referring physicians, impacts clinical decision making, and should be used in appropriate settings to guide patient care.

References 1. Nickl NJ, Bhutani MS, Catalano M, Hoffman B, Hawes R, Chak A, Roubein LD, Kimmey M, Johnson M, Affronti J, Canto M, Sivak M, Boyce HW, Lightdale CJ, Stevens P, Schmitt C. Clinical implications of endoscopic ultrasound: the American Endosonography Club study. Gastrointest Endosc 1996; 44:371–377. 2. Kaffes AJ, Mishra A, Simpson SB, Jones DB. Upper gastrointestinal ultrasound and its impact on patient management: 1990 – 2000. Intern Med J 2002;32:372–378. 3. Ainsworth AP, Mortensen MB, Durup J, Wamberg PA. Clinical impact of endoscopic ultrasonography at a county hospital. Endoscopy 2002;34:447– 450. 4. Erickson RA, Garza AA. Impact of endoscopic ultrasound on the management and outcome of pancreatic carcinoma. Am J Gastroenterol 2000;95:2248 –2254. 5. Fritscher-Ravens A, Brand L, Knofel T, Bobrowski C, Topalidis T, Thonke F, de Werth A, Soehendra N. Comparison of endoscopic ultrasound-guided fine needle aspiration for focal pancreatic lesions in patients with normal parenchyma and chronic pancreatitis. Am J Gastroenterol 2002;97:2768 –2775. 6. Larsen SS, Krasnik M, Vilmann P, Jacobsen GK, Pederson JH, Faurschou P, Folke K. Endoscopic ultrasound guided biopsy of mediastinal lesions has a major impact on patient management. Thorax 2002;57:98 –103. 7. Catalano MF, Nayar R, Gress F, Scheiman J, Wassef W, Rosenblatt ML, Kochman ML. EUS-guided fine needle aspiration in mediastinal lymphadenopathy of unknown etiology. Gastrointest Endosc 2002;55:863– 869. 8. Catalano MF, Rosenblatt ML, Chak A, Sivak MV, Scheiman J, Gress F. Endoscopic ultrasound-guided fine needle aspiration in the diagnosis of mediastinal masses of unknown origin. Am J Gastroenterol 2002;97:2559 –2565.

December 2004

9. Caletti G, Fusaroli P. Endoscopic ultrasonography (EUS) and EUSfine-needle procedures: fascinating technique but little clinical impact? Endoscopy 2001;33:537–540. 10. Fusaroli P, Vallar R, Togliani T, Khodadadian E, Caletti G. Scientific publications in endoscopic ultrasonography: a 20-year global survey of the literature. Endoscopy 2002;34:451– 456. 11. Jafri IH, Saltzman JR, Colby JM, Krims PE. Evaluation of the clinical impact of endoscopic ultrasonography in gastrointestinal disease. Gastrointest Endosc 1996;44:367–370. 12. Chang KJ, Nguyen P, Erickson RA, Durbin TE, Katz KD. The clinical utility of endoscopic ultrasound-guided fine-needle aspiration in the diagnosis and staging of pancreatic carcinoma. Gastrointest Endosc 1997;45:387–393. 13. Harewood GC, Wiersema MJ, Nelson H, Maccarty RL, Olson JE, Clain JE, Ahlquist DA, Jondal ML. A prospective, blinded assess-

CLINICAL IMPACT OF EUS

1073

ment of the impact of preoperative staging on the management of rectal cancer. Gastroenterology 2002;123:24 –32. 14. Shumaker DA, de Garmo P, Faigel DO. Potential impact of preoperative EUS on esophageal cancer management and cost. Gastrointest Endosc 2002;56:391–396.

Address requests for reprints to: Michael L. Kochman, MD, FACP, Division of Gastroenterology, Hospital of the University of Pennsylvania, 3400 Spruce Street, 3rd Floor Ravdin Building, Philadelphia, Pennsylvania 19104. e-mail: [email protected]; fax: (215) 349-5915. Presented in part at the annual meeting of the American Society for Gastrointestinal Endoscopy, May 18 –21, 2003, Orlando, Florida (Gastrointest Endosc 2003;57:AB249).