Advertisement

Performance of radiographic imaging after incomplete colonoscopy for nonmalignant causes in clinical practice

Published:February 04, 2020DOI:https://doi.org/10.1016/j.gie.2020.01.043

      Background and Aims

      CT colonography (CTC) or barium enema are commonly ordered to complete colorectal imaging after an incomplete colonoscopy. We evaluated the sensitivity of radiographic studies performed for this purpose in clinical practice outside clinical trials.

      Methods

      Adult patients referred to an expert endoscopist for incomplete colonoscopy because of a redundant colon or a difficult sigmoid and who underwent previous radiographic imaging between July 2001 and July 2019 were identified. None of the patients had a malignant obstruction as the cause of incomplete colonoscopy. Data on polyp size, location, and pathology were obtained from colonoscopy and radiology reports. Polyps identified on imaging and colonoscopy were matched based on polyp size and location.

      Results

      Among 769 patients referred for incomplete colonoscopy, we identified 65 with a radiographic examination performed within 36 months of colonoscopy at our center. Per-patient sensitivity for CTC was suboptimal (70%) and was very low for barium enema (26.7%). Per-polyp sensitivity for both CTC and barium enema was poor (23.8% and 7.6%). Quality of the examination did not seem to affect procedure sensitivity.

      Conclusions

      Radiographic imaging after incomplete colonoscopy for reasons other than malignant obstruction had poor sensitivity for polyps. Patients with incomplete colonoscopies should be considered for repeat colonoscopy by an expert.

      Abbreviations:

      CCE (colon capsule endoscopy), CTC (CT colonography)

      Introduction

      Incomplete colonoscopy, defined as failure to pass the colonoscope tip fully into the cecal base, is associated with a risk of interval cancer,
      • Baxter N.N.
      • Goldwasser M.A.
      • Paszat L.F.
      • et al.
      Association of colonoscopy and death from colorectal cancer.
      ,
      • Britton E.J.
      • Sidhu S.
      • Geraghty J.
      • et al.
      The 5-year outcome of patients having incomplete colonoscopy.
      and in clinical practice, is often followed by radiographic imaging to search for neoplasia proximal to the extent reached. Compared with colonoscopy, radiographic tests have a higher polyp miss rate,
      • Weinberg D.S.
      • Pickhardt P.J.
      • Bruining D.H.
      • et al.
      Computed tomography colonography vs colonoscopy for colorectal cancer surveillance after surgery.
      ,
      • Winawer S.J.
      • Stewart E.T.
      • Zauber A.G.
      • et al.
      A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. National Polyp Study Work Group.
      particularly for flat and sessile lesions.
      • IJspeert J.E.
      • Tutein Nolthenius C.J.
      • Kuipers E.J.
      • et al.
      CT-colonography vs. colonoscopy for detection of high-risk sessile serrated polyps.
      Moreover, imaging studies lack therapeutic capacity, and positive radiographic findings require another attempt at colonoscopy.
      Our center has the largest reported experience in colonoscopies performed in patients with previous incomplete colonoscopy with a cecal intubation rate in this population of 97.3%.
      • Bick B.L.
      • Vemulapalli K.C.
      • Rex D.K.
      Regional center for complex colonoscopy: yield of neoplasia in patients with prior incomplete colonoscopy.
      We have described techniques used to complete colonoscopy in patients with previous incomplete colonoscopy
      • Vemulapalli K.C.
      • Rex D.K.
      Water immersion simplifies cecal intubation in patients with redundant colons and previous incomplete colonoscopies.
      • Rex D.K.
      Achieving cecal intubation in the very difficult colon.
      • Rex D.K.
      • Chen S.C.
      • Overhiser A.J.
      Colonoscopy technique in consecutive patients referred for prior incomplete colonoscopy.
      • Rex D.K.
      • Goodwine B.W.
      Method of colonoscopy in 42 consecutive patients presenting after prior incomplete colonoscopy.
      and described a high yield of neoplasia identified by repeat colonoscopy in patients with and without radiographic imaging after the previous colonoscopy.
      • Bick B.L.
      • Vemulapalli K.C.
      • Rex D.K.
      Regional center for complex colonoscopy: yield of neoplasia in patients with prior incomplete colonoscopy.
      Others have reported high success rates of repeat colonoscopy after incomplete colonoscopy.
      • Ponte A.
      • Pinho R.
      • Rodrigues A.
      • et al.
      High efficacy of repeating colonoscopy by an advanced endoscopist after an incomplete colonoscopy.
      • Hermans C.
      • Zee D.V.
      • Gilissen L.
      Double-balloon endoscopy after incomplete colonoscopy and its comparison with computed tomography colonography.
      • Sulz M.C.
      • Frei R.
      • Semadeni G.M.
      • et al.
      The role of single-balloon colonoscopy for patients with previous incomplete standard colonoscopy: Is it worth doing it?.
      • Gawron A.J.
      • Veerappan A.
      • Keswani R.N.
      High success rate of repeat colonoscopy with standard endoscopes in patients referred for prior incomplete colonoscopy.
      • Becx M.C.
      • Al-Toma A.
      Double-balloon endoscopy: an effective rescue procedure after incomplete conventional colonoscopy.
      • Kobayashi K.
      • Mukae M.
      • Ogawa T.
      • et al.
      Clinical usefulness of single-balloon endoscopy in patients with previously incomplete colonoscopy.
      • Yamada A.
      • Watabe H.
      • Takano N.
      • et al.
      Utility of single and double balloon endoscopy in patients with difficult colonoscopy: a randomized controlled trial.
      • Dzeletovic I.
      • Harrison M.E.
      • Pasha S.F.
      • et al.
      Comparison of single- versus double-balloon assisted-colonoscopy for colon examination after previous incomplete standard colonoscopy.
      • Suzuki T.
      • Matsushima M.
      • Tsukune Y.
      • et al.
      Double-balloon endoscopy versus magnet-imaging enhanced colonoscopy for difficult colonoscopies, a randomized study.
      • Keswani R.N.
      Single-balloon colonoscopy versus repeat standard colonoscopy for previous incomplete colonoscopy: a randomized, controlled trial.
      • Schembre D.B.
      • Ross A.S.
      • Gluck M.N.
      • et al.
      Spiral overtube-assisted colonoscopy after incomplete colonoscopy in the redundant colon.
      • Hotta K.
      • Katsuki S.
      • Ohata K.
      • et al.
      A multicenter, prospective trial of total colonoscopy using a short double-balloon endoscope in patients with previous incomplete colonoscopy.
      • Teshima C.W.
      • Aktas H.
      • Haringsma J.
      • et al.
      Single-balloon-assisted colonoscopy in patients with previously failed colonoscopy.
      • Moreels T.G.
      • Macken E.J.
      • Roth B.
      • et al.
      Cecal intubation rate with the double-balloon endoscope after incomplete conventional colonoscopy: a study in 45 patients.
      • Monkemuller K.
      • Knippig C.
      • Rickes S.
      • et al.
      Usefulness of the double-balloon enteroscope in colonoscopies performed in patients with previously failed colonoscopy.
      • Pasha S.F.
      • Harrison M.E.
      • Das A.
      • et al.
      Utility of double-balloon colonoscopy for completion of colon examination after incomplete colonoscopy with conventional colonoscope.
      • Gay G.
      • Delvaux M.
      Double-balloon colonoscopy after failed conventional colonoscopy: a pilot series with a new instrument.
      • Kaltenbach T.
      • Soetikno R.
      • Friedland S.
      Use of a double balloon enteroscope facilitates caecal intubation after incomplete colonoscopy with a standard colonoscope.
      • Akahoshi K.
      • Kubokawa M.
      • Matsumoto M.
      • et al.
      Double-balloon endoscopy in the diagnosis and management of GI tract diseases: methodology, indications, safety, and clinical impact.
      • May A.
      • Nachbar L.
      • Ell C.
      Push-and-pull enteroscopy using a single-balloon technique for difficult colonoscopy.
      • Yung D.E.
      • Koulaouzidis A.
      • Fraser C.
      • et al.
      Double-balloon colonoscopy for failed conventional colonoscopy: the Edinburgh experience and systematic review of existing data.
      • Nemoto D.
      • Isohata N.
      • Utano K.
      • et al.
      Double-balloon colonoscopy carried out by a trainee after incomplete conventional colonoscopy.
      • Gomez V.
      • Patel M.K.
      • Stark M.E.
      • et al.
      Double-balloon enteroscopy in patients with previous incomplete colonoscopy: should we perform them more often?.
      • Sato K.
      • Shigiyama F.
      • Ito S.
      • et al.
      Colonoscopy using a small-caliber colonoscope with passive-bending after incomplete colonoscopy due to sharp angulation or pain.
      • Lee Y.T.
      • Hui A.J.
      • Wong V.W.
      • et al.
      Improved colonoscopy success rate with a distally attached mucosectomy cap.
      • Coppola F.
      • Gaia S.
      • Cosimato M.
      • et al.
      Enteroscope without overtube for cecal intubation after an incomplete colonoscopy.
      • Morini S.
      • Zullo A.
      • Hassan C.
      • et al.
      Endoscopic management of failed colonoscopy in clinical practice: to change endoscopist, instrument, or both?.
      • Shida T.
      • Takano S.
      • Kaiho M.
      • et al.
      Transparent hood attached to a gastroscope: a simple rescue technique for patients with difficult or incomplete colonoscopy.
      • Moreels T.G.
      • Pelckmans P.A.
      Double-balloon endoscope for failed conventional colonoscopy.
      • Fritscher-Ravens A.
      • Fox S.
      • Swain C.P.
      • et al.
      CathCam guide wire-directed colonoscopy: first pilot study in patients with a previous incomplete colonoscopy.
      • Horiuchi A.
      • Nakayama Y.
      • Kajiyama M.
      • et al.
      Usefulness of a small-caliber, variable-stiffness colonoscope as a backup in patients with difficult or incomplete colonoscopy.
      In this report, we describe the sensitivity of radiographic studies performed after a previous incomplete colonoscopy and before referral to our center for repeat colonoscopy. We used the results of colonoscopy at our center to define the sensitivity and specificity of the radiographic studies. Our goal was to determine if the performance of these radiographic tests after incomplete colonoscopy is sufficient to be relied on in clinical practice.

      Methods

      Study population and study overview

      We retrospectively reviewed a prospectively created database of all adult patients with previous incomplete colonoscopy referred to the senior author (D.K.R.) from July 2001 to July 2019. This database included patient demographics, procedure indication(s), extent of previous colonoscopy, and reason(s) for incomplete colonoscopy. Those who underwent barium enema and/or CT colonography (CTC) before referral to our institution and for whom we had the results available in our database were identified. Exclusions included if the radiographic imaging report was missing, if the time between the radiographic imaging and repeat colonoscopy was more than 36 months, if there was a missing pathology report from the colonoscopy at our center, or if the colonoscopy attempt at our center was unsuccessful in intubating the cecum. The study was approved by the Institutional Review Board at Indiana University.

      Colonoscopy

      Data on polyp number, polyp size (according to endoscopist estimate), polyp histopathology (as reported in the routine pathology report), and polyp location in the colon (according to endoscopist estimate) were identified from the colonoscopy report at our center. Polyps were grouped based on size as ≤5 mm (diminutive), 6 to 9 mm (small), 10 to 19 mm (medium), and ≥20 mm (large). Conventional adenomas were defined as those interpreted by the pathology report as tubular, tubulovillous, or villous adenomas. Serrated class lesions were defined as those interpreted as hyperplastic polyp, sessile serrated polyp, sessile serrated adenoma, serrated adenoma, or traditional serrated adenoma.
      In patients who had a lesion detected by CTC or barium enema, and no corresponding lesion was detected at our colonoscopy, we examined the area in question multiple times during colonoscopy to determine whether a lesion was actually present.

      Radiographic imaging

      Data regarding the quality of imaging studies, number of polyps, size of polyps, and location of reported polyps in the colon were obtained from the radiology report. Studies were considered adequate (excellent/satisfactory) when the radiologist indicated in the report that the entire colon was examined sufficiently. Studies were considered inadequate if the radiologist indicated in the report that any segment of the colon was examined inadequately or suboptimally.

      Polyp matching

      Polyps were considered a match if the size measured by the radiographic imaging and its ±50% range overlapped with the ±50% range of the size measured by the colonoscopy, and the location estimates by the 2 methods were in the same or adjacent segments.
      • Liedenbaum M.H.
      • de Vries A.H.
      • Halligan S.
      • et al.
      CT colonography polyp matching: differences between experienced readers.
      For location, there were 5 colonic segments: cecum and ascending colon, transverse colon and hepatic flexure, descending colon and splenic flexure, sigmoid colon, and rectum. Thus, if a polyp measured 10 mm by radiographic imaging, the ±50% range would be 5 to 15 mm. If a polyp in the same or an adjacent segment by conventional colonoscopy was 6 mm (±50% range, 3–9 mm), then the 2 polyps were a potential match because the ±50% ranges overlapped.

      Definition of diagnostic yield

      A polyp on radiographic imaging was considered true positive when it was also confirmed at colonoscopy, a polyp that was in the same or an adjacent colonic segment, and a polyp for which the size correlated within 50% of the diameter. A false-positive polyp was defined as a polyp reported on imaging that did not match any polyp detected at colonoscopy. A polyp found on colonoscopy was considered false negative (missed polyp) when it could not be matched with any findings reported on radiographic imaging.
      Analysis was performed on a per-patient and per-polyp basis. For example, in a patient with 4 reported polyps during colonoscopy and one true-positive polyp based on the imaging report, the sensitivities of radiologic imaging per patient and per polyp would be 100% and 25% (1/4), respectively.

      Statistical analysis

      We report descriptive measures for the study population. Performance characteristics for the tests are reported as percentages with 95% confidence intervals. We calculated sensitivity, specificity, positive predictive value, and negative predictive value by using 2 × 2 cross-tabulation of the endoscopy and imaging results. All analyses were performed using SAS version 9.4 (SAS Institute, Charlotte, NC, USA).

      Results

      Among 769 patients referred for incomplete colonoscopy between July 2001 and July 2019, the overall cecal intubation rate was 97.4%, and we identified 77 patients with a previous radiology report. Sixty-five patients among the 77 had a radiographic study performed <36 months before the colonoscopy at our center. None of the 65 patients had a missing pathology report at our center. Colonoscopy was complete to the cecum at our center in all 65 patients. Baseline characteristics of these 65 patients who underwent either CTC or barium enema before their colonoscopy are shown in Table 1. The mean age of the patients was 67 years, and 28% were male. Median time between the radiology examination and colonoscopy was 3 months, and only 3 radiographic examinations were performed >12 months before the colonoscopy at our center. The polyp detection rate and conventional adenoma detection rate on optical colonoscopy were 55 of 65 (84.6%) and 44 of 65 (67.7%), respectively.
      Table 1Baseline characteristics of 65 patients with radiographic examination data
      CT colonography (n = 45)Barium enema (n = 20)All (n = 65)
      Age (years), mean (SD)65.8 (11.2)68.6 (10.3)66.7 (9.9)
      Male, n (%)11 (24.4)7 (35)18 (27.7)
      Previous surgery, n (%)35 (77.8)18 (90)53 (81.5)
      Referral doctor specialty, n (%)
       Gastroenterology26 (57.8)9 (45)35 (53.8)
       Surgery12 (26.7)6 (30)18 (27.7)
       Internal medicine3 (6.7)1 (5)4 (6.2)
       Family practice4 (8.9)4 (20)8 (12.3)
      Median months between radiographic examination and colonoscopy (min-max)2.9 (0.1-20.1)5 (0.5-30)3.1 (0.1-30)
      SD, Standard deviation.
      The primary indications for colonoscopy at our center were abnormal imaging (n = 30), polyp surveillance (n = 8), cancer surveillance (n = 2), screening (n = 3), bleeding or anemia (n = 9), and nonbleeding symptoms (n = 6). The extent reached during the previous incomplete colonoscopy was ileocecal valve (n = 2), ascending colon (n = 7), hepatic flexure (n = 4), transverse colon (n = 8), splenic flexure (n = 2), descending colon (n = 4), and sigmoid colon (n = 27). Extent reached during the previous colonoscopy was unclear in 11 of the cases. The reasons for the previous incomplete colonoscopy included looping and redundancy (n = 21), angulated narrow sigmoid (n = 21), and both redundancy and narrow sigmoid (n = 10). In 13 cases, the reasons for previous colonoscopy were unclear (but likely redundancy).
      Table 2 shows the performance characteristics of CTC and barium enema on a per-patient basis. CTC sensitivity was 70% by patient and 62.5% for patients with adenomas. Barium enema had lower sensitivity for patients with polyps (26.7%) and patients with adenomas (8.3%).
      Table 2Performance characteristics of CT colonoscopy and barium enema (per-patient basis)
      CT colonoscopyBarium enemaOverall
      All polyps
      Sensitivity
      Number of patients with true-positive radiographic findings/number of patients with polyps or adenomas (%, 95% confidence interval).
      28/40 (70, 53.5-83.4)4/15 (26.7, 7.8-55.1)32/55 (58.2, 44.1-71.4)
      Specificity
      Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas (%, 95% confidence interval).
      2/5 (40, 5.3-85.3)3/5 (60, 14.7-94.7)5/10 (50, 18.7-81.3)
      Positive predictive value
      Number of patients with true-positive radiographic findings/number of patients who tested positive for polyps or adenomas (%, 95% confidence interval).
      28/31 (90.3, 74.3-97.9)4/6 (66.7, 22.3-95.7)32/37 (86.5, 71.2-95.5)
      Negative predictive value
      Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas at radiology (%, 95% confidence interval).
      2/14 (14.3, 1.8-42.8)3/14 (21.4, 4.7-51.0)5/28 (17.9, 6.1-36.9)
      Only for conventional adenomas
      Sensitivity
      Number of patients with true-positive radiographic findings/number of patients with polyps or adenomas (%, 95% confidence interval).
      20/32 (62.5, 43.7-78.9)1/12 (8.3, 0.2-38.5)21/44 (47.7, 32.5-63.3)
      Specificity
      Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas (%, 95% confidence interval).
      4/13 (30.8, 9.1-61.4)4/8 (50, 15.7-84.3)8/21 (38.1, 18.1-61.6)
      Positive predictive value
      Number of patients with true-positive radiographic findings/number of patients who tested positive for polyps or adenomas (%, 95% confidence interval).
      20/29 (69, 52.1-85.8)1/5 (20, 0.5-71.6)21/34 (61.8, 43.6-77.8)
      Negative predictive value
      Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas at radiology (%, 95% confidence interval).
      4/16 (25, 7.3-52.4)4/15 (26.7, 7.8-55.1)8/31 (25.8, 11.9-44.6)
      Number of patients with true-positive radiographic findings/number of patients with polyps or adenomas (%, 95% confidence interval).
      Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas (%, 95% confidence interval).
      Number of patients with true-positive radiographic findings/number of patients who tested positive for polyps or adenomas (%, 95% confidence interval).
      § Number of patients with true-negative radiographic findings/number of patients without polyps or adenomas at radiology (%, 95% confidence interval).
      Table 3 shows the performance characteristics of CTC and barium enema on a per-polyp basis stratified by quality of examination and size, location in the colon, and polyp pathology. There were 281 confirmed lesions by colonoscopy. Only 54 of the 281 lesions were seen on radiology (overall sensitivity for polyps, 19.2%). Twenty-six lesions were found by radiographic study and not at colonoscopy and were considered false positive. Both CTC and barium enema were better at recognizing polyps with increasing size but both of them missed all 5 polyps that were ≥20 mm in size. These 5 lesions were a fungating cecal cancer (in this case CTC detected a sigmoid polyp but no cecal lesion), 1 sessile adenoma measuring 20 mm in the cecum, 2 sessile adenomas measuring 24 and 30 mm in the ascending colon, and a 20-mm pedunculated adenoma in the sigmoid colon. Both CTC and barium enema had poor sensitivity in the right colon segment (19.6% and 2.8%, respectively) compared with the left colon segment (57.1% and 22.2%). CTC performed better with adenomas (25.4% sensitivity for adenomas vs 15% for serrated lesions), and barium enema performed better with serrated class lesions (20% sensitivity for serrated lesions vs 3.6% for adenomas). There were 14 lesions that were not confirmed by pathology as adenomas or serrated lesions, and 5 polyps where no biopsy was taken. We performed two separate analyses (not shown) excluding all 19 lesions in one analysis and then excluding only the 9 lesions reported as normal tissue, and in both analyses the per-patient and per-polyp sensitivities of CTC and barium enema declined.
      Table 3Performance characteristics of CT colonoscopy and barium enema (per-polyp basis)
      CT colonoscopy sensitivityBarium enema sensitivityOverall (sensitivity and positive predictive value)
      Overall
      Number of true-positive radiographic findings/number of polyps or adenomas seen on colonoscopy or number of positive radiographic findings (sensitivity or positive predictive value as %, 95% confidence interval).
      48/202 (23.8, 18.1-30.2)6/79 (7.6, 2.8-15.8)54/281 (19.2, 14.8-24.3)

      54/80 (67.5, 56.1-77.6)
      By quality of examination
      Limited
      Number of true-positive radiographic findings/number of polyps or adenomas seen on colonoscopy or number of positive radiographic findings (sensitivity or positive predictive value as %, 95% confidence interval).
      17/79 (21.5)2/40 (5)19/119 (16), 19/33 (57.6)
      Excellent/ satisfactory31/123 (25.2)4/39 (10.3)35/162 (21.6), 35/47 (74.5)
      By size
      1-5 mm28/142 (19.7)4/59 (6.8)32/201 (15.9), 32/37 (86.5)
      6-9 mm14/41 (34.1)1/11 (9.1)15/52 (28.8), 15/23 (65.2)
      10-19 mm6/15 (40)1/8 (12.5)7/23 (30.4), 7/12 (58.3)
      ≥20 mm0/40/10/5
      By location
      Cecum-ascending colon18/92 (19.6)1/36 (2.8)19/128 (14.8), 19/26 (73.1)
      Hepatic flexure-transverse colon17/45 (37.8)1/17 (5.9)18/62 (29), 18/25 (72)
      Splenic flexure-descending colon4/7 (57.1)4/18 (22.2)8/25 (32), 8/10 (80)
      Sigmoid7/31 (22.6)0/87/39 (17.9), 7/13 (53.8)
      Rectum2/27 (7.4)2/27 (7.4), 2/5 (40)
      By pathology
      Inflammatory polyp (n = 1), condyloma (n = 1), juvenile polyp (n = 1), normal mucosa or lymphoid follicle (n = 9), material insufficient for diagnosis or no biopsy taken (n=7).
      Adenomas32/126 (25.4)2/56 (3.6)34/182 (18.7), N/A
      Serrated lesions9/60 (15)4/20 (20)13/80 (16.3), N/A
      N/A, Not applicable.
      Number of true-positive radiographic findings/number of polyps or adenomas seen on colonoscopy or number of positive radiographic findings (sensitivity or positive predictive value as %, 95% confidence interval).
      Inflammatory polyp (n = 1), condyloma (n = 1), juvenile polyp (n = 1), normal mucosa or lymphoid follicle (n = 9), material insufficient for diagnosis or no biopsy taken (n=7).
      Eighteen CTC and 6 barium enema reports had language indicating the study was technically inadequate (Table 4). There were no substantial differences in CTC or barium enema performance between adequate and inadequate studies (Table 3). Overall sensitivity of the 65 radiographic studies on a per-patient basis was 57% for studies with adequate technical quality and 60% for technically inadequate studies, with specificities of 50% for both. Neither CTC nor barium enema had better per-patient sensitivity in technically adequate studies.
      Table 4Reasons for low-quality radiology examination
      Barium enema
      Overlapping of loops4
      Inability to retain barium1
      Poor bowel preparation1
      CT colonography
      Inadequate distension of 1 or more segments10
      Inadequate distension of 1 or more segments and poor stool tagging4
      Poor stool tagging2
      Poor stool tagging and poor patient positioning2
      The primary reason for an inadequate barium enema examination was overlapping of colon loops. For CTC, the cause was inadequate distension (Table 4).

      Discussion

      When colonoscopy is incomplete to the cecum, it is common for clinicians to order a radiographic study to complete the colorectal imaging. To our knowledge, this is the largest study to report the yield of radiographic imaging in patients with previous incomplete colonoscopy (for reasons other than malignant obstruction) in clinical practice, who then underwent a successful repeat colonoscopy that tested the accuracy of the radiographic study. We found that radiographic imaging in patients with previous incomplete colonoscopy by either CTC or barium enema had poor sensitivity for identifying polyps found on repeat colonoscopy, and there were substantial numbers of false-positive lesions. Sensitivity was low for lesions of all sizes, histology, and location in the colorectum. These results call into question the utility of performing radiographic imaging after an incomplete colonoscopy for reasons other than malignant obstruction and suggest repeat colonoscopy as a better diagnostic measure.
      There are several limitations of our study. First, it is retrospective. However, it represents a prospectively collected series of consecutive cases in which radiographic studies were obtained after an incomplete colonoscopy and patients were then referred to our center. Also, the radiographic studies were performed in a range of hospitals in our area, and therefore may better represent performance in clinical practice compared with radiographic performance in centers of excellence or single centers. Second, the repeat colonoscopies were performed by an expert colonoscopist with a proven high adenoma detection rate, which could affect the generalizability of the results. However, high-level performance of colonoscopy more accurately defines the absolute sensitivity of radiographic colonic imaging. Third, we included radiographic studies performed up to 36 months before the colonoscopy at our center, which might be considered too long an interval by some readers. However, according to the CT Colonography Reporting and Data System recommendations, patients with lesions of 6 to 9 mm on CTC can undergo repeat CTC in 36 months, indicating that CTC findings are considered stable over this interval.
      • Zalis M.E.
      • Barish M.A.
      • Choi J.R.
      • et al.
      CT colonography reporting and data system: a consensus proposal.
      Further, the median interval between the radiographic study and colonoscopy at our center was 3 months, and only 3 patients had intervals >12 months. In a separate analysis (not shown), excluding these 3 patients had no effect on the results or conclusions. Finally, the overall number of patients in the study is small. Admittedly, series of this type take time to accumulate, and achieving large numbers of patients is challenging.
      • Ridolfi T.J.
      • Valente M.A.
      • Church J.M.
      Achieving a complete colonic evaluation in patients with incomplete colonoscopy is worth the effort.
      ,
      • Copel L.
      • Sosna J.
      • Kruskal J.B.
      • et al.
      CT colonography in 546 patients with incomplete colonoscopy.
      Despite this, the number of polyps detected at repeat colonoscopy was substantial, and the very low sensitivity described for both CTC and barium enema makes it unlikely that larger numbers of patients would lead to a result that would be interpreted as CTC and barium enema showing adequate sensitivity. In addition, many of the lesions undetected by radiographic studies were ≤5 mm in size. It is not even recommended that these lesions be reported for CTC.
      • Zalis M.E.
      • Barish M.A.
      • Choi J.R.
      • et al.
      CT colonography reporting and data system: a consensus proposal.
      Despite that, we saw no substantial tendency for the sensitivity of radiographic studies to improve with increasing lesion size. Finally, there was only 1 cancer detected in the 65 study patients by repeat colonoscopy at our center. Because the goal of radiographic examination after incomplete colonoscopy may sometimes be simply exclusion of cancer, and we cannot describe cancer sensitivity because of low numbers of patients with cancer, we also cannot evaluate a role for radiographic imaging in ruling out cancer in patients with incomplete colonoscopies. However, we expect that in most instances of incomplete colonoscopy, exclusion of precancerous lesions as well as malignant lesions is the diagnostic goal of radiographic imaging.
      CTC has largely replaced barium enema in many centers because of higher sensitivity.
      • Johnson C.D.
      • MacCarty R.L.
      • Welch T.J.
      • et al.
      Comparison of the relative sensitivity of CT colonography and double-contrast barium enema for screen detection of colorectal polyps.
      However, CTC is still not available in all community hospitals, so that barium enema may be still be used in cases with incomplete colonoscopy. Recently, clinical trials have found that colon capsule endoscopy (CCE) outperforms CTC in patients with previous incomplete colonoscopy and for screening colonoscopy.
      • Cash B.D.
      • Fleisher M.R.
      • Fern S.
      • et al.
      A multicenter, prospective, randomized study comparing the diagnostic yield of colon capsule endoscopy versus computed tomographic colonography in a screening population. Results of the Topaz Study [abstract].
      ,
      • Spada C.
      • Hassan C.
      • Barbaro B.
      • et al.
      Colon capsule versus CT colonography in patients with incomplete colonoscopy: a prospective, comparative trial.
      Among 98 patients with previous incomplete colonoscopy who underwent both CCE and CTC, CCE was found to have a 2-fold increase in diagnostic yield for identifying polyps ≥6 mm compared with CTC (95% confidence interval, 1.34-2.98), without increase in the false-positive results.
      • Spada C.
      • Hassan C.
      • Barbaro B.
      • et al.
      Colon capsule versus CT colonography in patients with incomplete colonoscopy: a prospective, comparative trial.
      In a population of 286 patients who were randomized to undergo either CCE or CTC for colorectal cancer screening (145 CCE, 141 CTC), CCE was found to be superior to CTC (24.1% vs 9.4%) for identifying polyps ≥6 mm.
      • Cash B.D.
      • Fleisher M.R.
      • Fern S.
      • et al.
      A multicenter, prospective, randomized study comparing the diagnostic yield of colon capsule endoscopy versus computed tomographic colonography in a screening population. Results of the Topaz Study [abstract].
      These studies suggest that CCE is a better test than CTC to complete colorectal imaging after incomplete colonoscopy. However, the high rate of completing colonoscopy in multiple series after a previous incomplete colonoscopy
      • Bick B.L.
      • Vemulapalli K.C.
      • Rex D.K.
      Regional center for complex colonoscopy: yield of neoplasia in patients with prior incomplete colonoscopy.
      ,
      • Ponte A.
      • Pinho R.
      • Rodrigues A.
      • et al.
      High efficacy of repeating colonoscopy by an advanced endoscopist after an incomplete colonoscopy.
      ,
      • Gawron A.J.
      • Veerappan A.
      • Keswani R.N.
      High success rate of repeat colonoscopy with standard endoscopes in patients referred for prior incomplete colonoscopy.
      ,
      • Schembre D.B.
      • Ross A.S.
      • Gluck M.N.
      • et al.
      Spiral overtube-assisted colonoscopy after incomplete colonoscopy in the redundant colon.
      ,
      • Brahmania M.
      • Park J.
      • Svarta S.
      • et al.
      Incomplete colonoscopy: maximizing completion rates of gastroenterologists.
      • Tan M.
      • Lahiff C.
      • Bassett P.
      • et al.
      Efficacy of balloon overtube-assisted colonoscopy in patients with incomplete or previous difficult colonoscopies: a meta-analysis.
      • Moreels T.G.
      • Macken E.J.
      • Pelckmans P.A.
      Renewed attention for overtube-assisted colonoscopy to prevent incomplete endoscopic examination of the colon.
      and the superior sensitivity of colonoscopy over both radiographic examinations
      • Weinberg D.S.
      • Pickhardt P.J.
      • Bruining D.H.
      • et al.
      Computed tomography colonography vs colonoscopy for colorectal cancer surveillance after surgery.
      ,
      • IJspeert J.E.
      • Tutein Nolthenius C.J.
      • Kuipers E.J.
      • et al.
      CT-colonography vs. colonoscopy for detection of high-risk sessile serrated polyps.
      ,
      • Kao K.T.
      • Tam M.
      • Sekhon H.
      • et al.
      Should barium enema be the next step following an incomplete colonoscopy?.
      ,
      • Gawron A.J.
      • Veerappan A.
      • McCarthy S.T.
      • et al.
      Impact of an incomplete colonoscopy referral program on recommendations after incomplete colonoscopy.
      and CCE
      • Morgan D.R.
      • Malik P.R.
      • Romeo D.P.
      • et al.
      Initial US evaluation of second-generation capsule colonoscopy for detecting colon polyps.
      ,
      • Rex D.K.
      • Adler S.N.
      • Aisenberg J.
      • et al.
      Accuracy of capsule colonoscopy in detecting colorectal polyps in a screening population.
      support repeating colonoscopy as the initial diagnostic strategy in most patients with previous incomplete colonoscopy.
      In conclusion, radiographic imaging by either CTC or barium enema has limited effectiveness for detection of precancerous lesions in patients with previous incomplete colonoscopy where malignant obstruction was not the cause of the previous failure. Radiographic imaging in patients with previous incomplete colonoscopy may still have a valuable role in excluding colon cancer proximal to the extent of the incomplete colonoscopy. Radiographic imaging could be reserved for situations where exclusion of cancer is the only clinical goal. When exclusion of proximal precancerous lesions is the preeminent goal, patients with incomplete colonoscopy should undergo repeat colonoscopy using techniques established for this indication or be directly referred to an expert endoscopist.

      Acknowledgments

      This work was supported by a grant from Scott Schurz and his children of Bloomington, IN, to the Indiana University foundation in the name of Douglas K. Rex.

      References

        • Baxter N.N.
        • Goldwasser M.A.
        • Paszat L.F.
        • et al.
        Association of colonoscopy and death from colorectal cancer.
        Ann Intern Med. 2009; 150: 1-8
        • Britton E.J.
        • Sidhu S.
        • Geraghty J.
        • et al.
        The 5-year outcome of patients having incomplete colonoscopy.
        Colorectal Dis. 2015; 17: 298-303
        • Weinberg D.S.
        • Pickhardt P.J.
        • Bruining D.H.
        • et al.
        Computed tomography colonography vs colonoscopy for colorectal cancer surveillance after surgery.
        Gastroenterology. 2018; 154: 927-934 e4
        • Winawer S.J.
        • Stewart E.T.
        • Zauber A.G.
        • et al.
        A comparison of colonoscopy and double-contrast barium enema for surveillance after polypectomy. National Polyp Study Work Group.
        N Engl J Med. 2000; 342: 1766-1772
        • IJspeert J.E.
        • Tutein Nolthenius C.J.
        • Kuipers E.J.
        • et al.
        CT-colonography vs. colonoscopy for detection of high-risk sessile serrated polyps.
        Am J Gastroenterol. 2016; 111: 516-522
        • Bick B.L.
        • Vemulapalli K.C.
        • Rex D.K.
        Regional center for complex colonoscopy: yield of neoplasia in patients with prior incomplete colonoscopy.
        Gastrointest Endosc. 2016; 83: 1239-1244
        • Vemulapalli K.C.
        • Rex D.K.
        Water immersion simplifies cecal intubation in patients with redundant colons and previous incomplete colonoscopies.
        Gastrointest Endosc. 2012; 76: 812-817
        • Rex D.K.
        Achieving cecal intubation in the very difficult colon.
        Gastrointest Endosc. 2008; 67: 938-944
        • Rex D.K.
        • Chen S.C.
        • Overhiser A.J.
        Colonoscopy technique in consecutive patients referred for prior incomplete colonoscopy.
        Clin Gastroenterol Hepatol. 2007; 5: 879-883
        • Rex D.K.
        • Goodwine B.W.
        Method of colonoscopy in 42 consecutive patients presenting after prior incomplete colonoscopy.
        Am J Gastroenterol. 2002; 97: 1148-1151
        • Ponte A.
        • Pinho R.
        • Rodrigues A.
        • et al.
        High efficacy of repeating colonoscopy by an advanced endoscopist after an incomplete colonoscopy.
        GE Port J Gastroenterol. 2018; 25: 285-290
        • Hermans C.
        • Zee D.V.
        • Gilissen L.
        Double-balloon endoscopy after incomplete colonoscopy and its comparison with computed tomography colonography.
        Clin Endosc. 2018; 51: 66-71
        • Sulz M.C.
        • Frei R.
        • Semadeni G.M.
        • et al.
        The role of single-balloon colonoscopy for patients with previous incomplete standard colonoscopy: Is it worth doing it?.
        Surg Endosc. 2016; 30: 1876-1882
        • Gawron A.J.
        • Veerappan A.
        • Keswani R.N.
        High success rate of repeat colonoscopy with standard endoscopes in patients referred for prior incomplete colonoscopy.
        BMC Gastroenterol. 2014; 14: 56
        • Becx M.C.
        • Al-Toma A.
        Double-balloon endoscopy: an effective rescue procedure after incomplete conventional colonoscopy.
        Eur J Gastroenterol Hepatol. 2014; 26: 519-522
        • Kobayashi K.
        • Mukae M.
        • Ogawa T.
        • et al.
        Clinical usefulness of single-balloon endoscopy in patients with previously incomplete colonoscopy.
        World J Gastrointest Endosc. 2013; 5: 117-121
        • Yamada A.
        • Watabe H.
        • Takano N.
        • et al.
        Utility of single and double balloon endoscopy in patients with difficult colonoscopy: a randomized controlled trial.
        World J Gastroenterol. 2013; 19: 4732-4736
        • Dzeletovic I.
        • Harrison M.E.
        • Pasha S.F.
        • et al.
        Comparison of single- versus double-balloon assisted-colonoscopy for colon examination after previous incomplete standard colonoscopy.
        Dig Dis Sci. 2012; 57: 2680-2686
        • Suzuki T.
        • Matsushima M.
        • Tsukune Y.
        • et al.
        Double-balloon endoscopy versus magnet-imaging enhanced colonoscopy for difficult colonoscopies, a randomized study.
        Endoscopy. 2012; 44: 38-42
        • Keswani R.N.
        Single-balloon colonoscopy versus repeat standard colonoscopy for previous incomplete colonoscopy: a randomized, controlled trial.
        Gastrointest Endosc. 2011; 73: 507-512
        • Schembre D.B.
        • Ross A.S.
        • Gluck M.N.
        • et al.
        Spiral overtube-assisted colonoscopy after incomplete colonoscopy in the redundant colon.
        Gastrointest Endosc. 2011; 73: 515-519
        • Hotta K.
        • Katsuki S.
        • Ohata K.
        • et al.
        A multicenter, prospective trial of total colonoscopy using a short double-balloon endoscope in patients with previous incomplete colonoscopy.
        Gastrointest Endosc. 2012; 75: 813-818
        • Teshima C.W.
        • Aktas H.
        • Haringsma J.
        • et al.
        Single-balloon-assisted colonoscopy in patients with previously failed colonoscopy.
        Gastrointest Endosc. 2010; 71: 1319-1323
        • Moreels T.G.
        • Macken E.J.
        • Roth B.
        • et al.
        Cecal intubation rate with the double-balloon endoscope after incomplete conventional colonoscopy: a study in 45 patients.
        J Gastroenterol Hepatol. 2010; 25: 80-83
        • Monkemuller K.
        • Knippig C.
        • Rickes S.
        • et al.
        Usefulness of the double-balloon enteroscope in colonoscopies performed in patients with previously failed colonoscopy.
        Scand J Gastroenterol. 2007; 42: 277-278
        • Pasha S.F.
        • Harrison M.E.
        • Das A.
        • et al.
        Utility of double-balloon colonoscopy for completion of colon examination after incomplete colonoscopy with conventional colonoscope.
        Gastrointest Endosc. 2007; 65: 848-853
        • Gay G.
        • Delvaux M.
        Double-balloon colonoscopy after failed conventional colonoscopy: a pilot series with a new instrument.
        Endoscopy. 2007; 39: 788-792
        • Kaltenbach T.
        • Soetikno R.
        • Friedland S.
        Use of a double balloon enteroscope facilitates caecal intubation after incomplete colonoscopy with a standard colonoscope.
        Dig Liver Dis. 2006; 38: 921-925
        • Akahoshi K.
        • Kubokawa M.
        • Matsumoto M.
        • et al.
        Double-balloon endoscopy in the diagnosis and management of GI tract diseases: methodology, indications, safety, and clinical impact.
        World J Gastroenterol. 2006; 12: 7654-7659
        • May A.
        • Nachbar L.
        • Ell C.
        Push-and-pull enteroscopy using a single-balloon technique for difficult colonoscopy.
        Endoscopy. 2006; 38: 395-398
        • Yung D.E.
        • Koulaouzidis A.
        • Fraser C.
        • et al.
        Double-balloon colonoscopy for failed conventional colonoscopy: the Edinburgh experience and systematic review of existing data.
        Gastrointest Endosc. 2016; 84: 878-881
        • Nemoto D.
        • Isohata N.
        • Utano K.
        • et al.
        Double-balloon colonoscopy carried out by a trainee after incomplete conventional colonoscopy.
        Dig Endosc. 2014; 26: 392-395
        • Gomez V.
        • Patel M.K.
        • Stark M.E.
        • et al.
        Double-balloon enteroscopy in patients with previous incomplete colonoscopy: should we perform them more often?.
        J Clin Gastroenterol. 2012; 46: 887-888
        • Sato K.
        • Shigiyama F.
        • Ito S.
        • et al.
        Colonoscopy using a small-caliber colonoscope with passive-bending after incomplete colonoscopy due to sharp angulation or pain.
        Surg Endosc. 2013; 27: 4171-4176
        • Lee Y.T.
        • Hui A.J.
        • Wong V.W.
        • et al.
        Improved colonoscopy success rate with a distally attached mucosectomy cap.
        Endoscopy. 2006; 38: 739-742
        • Coppola F.
        • Gaia S.
        • Cosimato M.
        • et al.
        Enteroscope without overtube for cecal intubation after an incomplete colonoscopy.
        Dig Liver Dis. 2011; 43: 475-477
        • Morini S.
        • Zullo A.
        • Hassan C.
        • et al.
        Endoscopic management of failed colonoscopy in clinical practice: to change endoscopist, instrument, or both?.
        Int J Colorectal Dis. 2011; 26: 103-108
        • Shida T.
        • Takano S.
        • Kaiho M.
        • et al.
        Transparent hood attached to a gastroscope: a simple rescue technique for patients with difficult or incomplete colonoscopy.
        Endoscopy. 2008; 40: E139
        • Moreels T.G.
        • Pelckmans P.A.
        Double-balloon endoscope for failed conventional colonoscopy.
        Clin Gastroenterol Hepatol. 2008; 6 (author reply 259): 259
        • Fritscher-Ravens A.
        • Fox S.
        • Swain C.P.
        • et al.
        CathCam guide wire-directed colonoscopy: first pilot study in patients with a previous incomplete colonoscopy.
        Endoscopy. 2006; 38: 209-213
        • Horiuchi A.
        • Nakayama Y.
        • Kajiyama M.
        • et al.
        Usefulness of a small-caliber, variable-stiffness colonoscope as a backup in patients with difficult or incomplete colonoscopy.
        Am J Gastroenterol. 2004; 99: 1936-1940
        • Liedenbaum M.H.
        • de Vries A.H.
        • Halligan S.
        • et al.
        CT colonography polyp matching: differences between experienced readers.
        Eur Radiol. 2009; 19: 1723-1730
        • Zalis M.E.
        • Barish M.A.
        • Choi J.R.
        • et al.
        CT colonography reporting and data system: a consensus proposal.
        Radiology. 2005; 236: 3-9
        • Ridolfi T.J.
        • Valente M.A.
        • Church J.M.
        Achieving a complete colonic evaluation in patients with incomplete colonoscopy is worth the effort.
        Dis Colon Rectum. 2014; 57: 383-387
        • Copel L.
        • Sosna J.
        • Kruskal J.B.
        • et al.
        CT colonography in 546 patients with incomplete colonoscopy.
        Radiology. 2007; 244: 471-478
        • Johnson C.D.
        • MacCarty R.L.
        • Welch T.J.
        • et al.
        Comparison of the relative sensitivity of CT colonography and double-contrast barium enema for screen detection of colorectal polyps.
        Clin Gastroenterol Hepatol. 2004; 2: 314-321
        • Cash B.D.
        • Fleisher M.R.
        • Fern S.
        • et al.
        A multicenter, prospective, randomized study comparing the diagnostic yield of colon capsule endoscopy versus computed tomographic colonography in a screening population. Results of the Topaz Study [abstract].
        Gastrointest Endosc. 2019; 89: AB87-AB88
        • Spada C.
        • Hassan C.
        • Barbaro B.
        • et al.
        Colon capsule versus CT colonography in patients with incomplete colonoscopy: a prospective, comparative trial.
        Gut. 2015; 64: 272-281
        • Brahmania M.
        • Park J.
        • Svarta S.
        • et al.
        Incomplete colonoscopy: maximizing completion rates of gastroenterologists.
        Can J Gastroenterol. 2012; 26: 589-592
        • Tan M.
        • Lahiff C.
        • Bassett P.
        • et al.
        Efficacy of balloon overtube-assisted colonoscopy in patients with incomplete or previous difficult colonoscopies: a meta-analysis.
        Clin Gastroenterol Hepatol. 2017; 15: 1628-1630
        • Moreels T.G.
        • Macken E.J.
        • Pelckmans P.A.
        Renewed attention for overtube-assisted colonoscopy to prevent incomplete endoscopic examination of the colon.
        Dis Colon Rectum. 2013; 56: 1013-1018
        • Kao K.T.
        • Tam M.
        • Sekhon H.
        • et al.
        Should barium enema be the next step following an incomplete colonoscopy?.
        Int J Colorectal Dis. 2010; 25: 1353-1357
        • Gawron A.J.
        • Veerappan A.
        • McCarthy S.T.
        • et al.
        Impact of an incomplete colonoscopy referral program on recommendations after incomplete colonoscopy.
        Dig Dis Sci. 2013; 58: 1849-1855
        • Morgan D.R.
        • Malik P.R.
        • Romeo D.P.
        • et al.
        Initial US evaluation of second-generation capsule colonoscopy for detecting colon polyps.
        BMJ Open Gastroenterol. 2016; 3e000089
        • Rex D.K.
        • Adler S.N.
        • Aisenberg J.
        • et al.
        Accuracy of capsule colonoscopy in detecting colorectal polyps in a screening population.
        Gastroenterology. 2015; 148: 948-957 e2