Critical appraisal:Shi XY, Bhagwandeen B, Leong AS 2008 2

From Cancer Guidelines Wiki

Critical Appraisal

Article being appraised

Shi XY, Bhagwandeen B, Leong AS. CDX2 and villin are useful markers of intestinal metaplasia in the diagnosis of Barrett esophagus. Am J Clin Pathol 2008 Apr;129(4):571-7 Abstract available at http://www.ncbi.nlm.nih.gov/pubmed/18343784.


Applicable clinical question

Key Facts

Study Design

diagnostic accuracy study

Number of Patients enrolled:

Not applicable

Number of Patients evaluated:

Not applicable

Number of samples:

Not applicable


Includes an economic evaluation

no

Evidence ratings

Level of evidence

III-1


Risk of bias assessment: diagnostic accuracy study

Patient Selection
Prior tests and any referral filters
108 esophageal biopsy and resection specimens (clinically diagnosed as BE) and 20 control samples of morphologically normal gastric mucosa (no evidence of inflammation or IM) were obtained. Control samples were selected from patients unassociated with GERD.
Condition that defined entry into study
Biopsy and resection specimens (endoscopy performed from 1999 to 2006) were obtained from the database of the Division of Anatomical Pathology, Hunter Area Pathology Service, Newcastle, Australia. Clinical data and pathological slides were reviewed in all cases.
Setting
Biopsy and resection specimens were obtained from the database of the Division of Anatomical Pathology, Hunter Area Pathology Service, Newcastle, Australia.
Was a diagnostic case-control design avoided?
Yes
Consecutive or random sample?
No
Did the study avoid inappropriate exclusions?
Yes
Reasons
94 of the 108 BE sections were confirmed to contain goblet cells using the reference standard. The remaining 14 cases that were clinically diagnosed as short-segment BE but failed to show goblet cells were not excluded. A single section was excluded as a result of insufficient tissue for staining.
If comparing more than one index test was the design fully paired or paired randomly?
Not applicable
If a paired randomised design was used, was allocation to groups concealed and was the generation of allocation sequence adequate?
Not applicable
What is the risk that the selection of participants introduced bias?
Low
Comments
Acceptable numbers of BE and control samples were used. It was not stated whether each biopsy came from a unique patient or whether several biopsies were taken from a single patient.
Index test 1
Describe index test and how it was conducted and interpreted
Index test was immunohistochemical staining of biopsy sections for CDX2, villin, HepPar-1 and cytokeratin-7 (CK7). All immunohistochemical stains were separately evaluated by two pathologists. The staining was interpreted in a binary fashion where any staining was regarded as positive and no staining regarded as negative.
Were the index test results interpreted without knowledge of the results of the reference standard?
Unclear
If a threshold was used, was it pre-specified?
Yes
If two tests are being compared, have they been assessed independently / blind to each other?
Not applicable
What is the risk that the conduct or interpretation of the index test introduced bias?
Low
Comments
Two independent pathologists were used to evaluate results of the index test.
Index test 2
Describe index test and how it was conducted and interpreted, if applicable
Not applicable
Were the index test results interpreted without knowledge of the results of the reference standard?
Not applicable
If a threshold was used, was it pre-specified?
Not applicable
What is the risk that the conduct or interpretation of the index test introduced bias?
Not applicable
Comments
Not applicable
Reference Standard
Describe the reference standard and how it was conducted and interpreted
Reference standard was Alcian blue-periodic acid-Schiff staining to confirm the presence of definite goblet cells. The test was interpreted in a binary fashion where any staining was regarded as positive and no staining regarded as negative.
Is the reference standard likely to correctly classify the target condition?
Yes
Were the reference standard results interpreted without knowedge of the results of the index test/s?
Unclear
Was the reference test standard independent of the index test?
(i.e. the index test did not form part of the reference standard)
Yes
What is the risk that the reference standard, its conduct or interpretation introduced bias?
Low
Comments
The reference standard confirms the presence of goblet cells, required by definition for the diagnosis of Barrett’s oesophagus.
Flow and timing
Describe any patients who did not receive the index test(s) and/or reference standard or who were excluded from the 2x2 table
All samples underwent both the index test and reference standard.
Describe the time interval and any interventions between index test(s) and reference standard
Both the index test and reference standard were performed on biopsy sections taken from a pathology database. As such the time interval between performing the reference standard and index test is negligible.
If a predictive test (the reference standard is a later event that the test aims to predict) were any subsequent interventions between test and later event blind to test result?
Not applicable
Was there an appropriate interval between index test(s) and reference standard?
Yes
Did either all participants or a random sample of participants receive a reference standard test?
Yes
Did all patients receive the same reference standard irrespective of index test result?
Yes
Were all test results including unclear results reported?
Yes
Were all patients included in the analysis?
Yes
What is the risk that the patient flow introduced bias?
Low
Comments
Samples selected from a database over a period of 7 years based on clinical data specimen pathology.
Size of effect
2 Reason for decision: CDX2 and Villin staining highlighted goblet cells in 100% of the cases where they existed, with specificity of 66% and 45% respectively (staining also present in cardia type mucosa). Granular cytoplasmic HepPar-1 expression was present only in columnar mucosa with goblet cells (100% specificity) however staining was weak and variable with relatively poor sensitivity (57%). 96% (90/94) of the BE samples exhibited CK7 staining (42% specificity), however the staining pattern was highly variable. CDX2 and Villin are sensitive markers for the detection of early IM and are suitable supplements for the histologic diagnosis of BE. HepPar-1 has limited sensitivity and CK7 is not specific.
Relevance of evidence
2 Additional comments: Provides evidence that CDX2 and Villin are senstitive markers for BE and can supplement the histologic identification of Barrett’s oesophagus. However these results do not overcome the sampling issue associated diagnosis of BE (BE may be present, but if not correctly biopsied may be missed). Does not provide evidence that the test improves outcome for the patient.
Result of appraisal

Jutta's tick icon.png Included




Completed by

Melissa Thomas


Jutta's exclamation mark icon.png This appraisal is pending.


Article
Shi XY, Bhagwandeen B, Leong AS. CDX2 and villin are useful markers of intestinal metaplasia in the diagnosis of Barrett esophagus. Am J Clin Pathol 2008 Apr;129(4):571-7 Abstract available at http://www.ncbi.nlm.nih.gov/pubmed/18343784.
Assigned to
User:angelique.levert
Topic area
Guidelines:Barrett's
Clinical question
Form
Form:Critical appraisal


Section below only relevant for Cancer Council Project Officer

Edit appraisal assignment