We thank Dr. Eisen for his letter. The aim of our review was not to criticize currently available adjuncts for detection of oral mucosal disease. Rather, we sought to delineate, using strict inclusion criteria and standard systematic review methodology, the status of our existing knowledge to further progress in early detection of oral mucosal disease. Ultimately, the goal is to improve detection and facilitate diagnosis of oral cancer to improve cure rates and quality of life and to reduce the cost of care of this devastating disease.
In our systematic review, we used histopathology as the gold-standard basis for comparison against the performance of the adjunctive techniques, including the OralCDx brush test. In the abstract, we stated that "OralCDx is useful in assessment of dysplastic changes in clinically suspicious lesions; however, there are insufficient data meeting the inclusion criteria to assess usefulness in innocuous mucosal lesions."
In the pivotal 1999 clinical study published in JADA by Sciubba,1 the Class II lesions (denoted clinically innocuous) did not undergo simultaneous brush and scalpel biopsy, and as such, we could not include the data in our analysis. We did not state, as alleged, that "the sensitivity of OralCDx for testing ‘nonsuspicious’ lesions has not yet been established," only that we could not assess the test performance based on matched histology. This was presented in the discussion.
We are in agreement with Dr. Eisen’s statement that "two biopsies of any type, performed at the same time but on different areas of a lesion, or performed on the same lesion but at different times, are expected to yield different results," and that disparities in accuracy data from the Sciubba’s1 study compared with those reported by Svirsky and colleagues2 and Poate and colleagues3 can be explained by such methodological flaws.
As for the National Cancer Institute (NCI) publication,4 with all due respect, four lines devoted to the assessment of two studies and without there being any distinction between Class I (clinically suspicious) and Class II (apparently innocuous) lesions is hardly equivalent to the rigorous process we executed in our systematic review. Readers are encouraged to review the online tables and figures that document the nature of the systematic review and methods used in our July article (Supplemental Data, "http://jada.ada.org/cgi/content/full/139/7/896/DC1").
Furthermore, the improved sensitivity and specificity of the brush-biopsy technique over prior oral exfoliative cytology techniques cited in the NCI publication4 is taken from the article by Scheifele and colleagues,5 in which the inclusion criteria were "a lesion with the clinical diagnosis oral leukoplakia, oral lichen planus, or obvious oral squamous cell carcinoma." Thus we believe the NCI publication4 is consistent with our JADA article in indicating the value of OralCDx on "clinically suspicious lesions," which were the Class I lesions of the Sciubba1 article included in our review.
We disagree with the comment that "how suspicious an oral lesion appears matters little," as there are clinical features that should kindle a diagnostic pathway commensurate with degree of suspicion. A mixed red and white lesion on the lateral tongue carries a far higher risk of developing malignant transformation than does a small sesame seed–sized homogeneous leukoplakia on the buccal mucosa. Identification of premalignant and malignant lesions is a desirable goal.
In this time of escalating medical costs, there remain several unknowns that apply to all the adjunctive techniques we included in our review. What is the natural history of the lesions or dysplasias detected (most are likely low grade)? What is the possibility of lesion regression, and how are the different lesions identified? How will these adjuncts affect cost of care and oral cancer mortality rates? Our article highlights the current state of knowledge and encourages use based on evidence reported in the scientific literature to date. We also seek to promote continuing research and publication of results in the peer-reviewed scientific literature so that evidence supports our clinical practices and improves the care we provide.
We encourage Dr. Eisen to publish the descriptive results of the "close to 500,000 oral lesions" that have been analyzed to date by OralCDx Laboratories to demonstrate the population prevalence of OralCDx brush biopsy results among current users. Furthermore, analysis of the "tens of thousands of patients" who had follow-up scalpel biopsies would be of significant interest to the JADA readership and all practicing clinicians. Evidence-based practice is anchored by available and appropriately evaluated evidence.
TOP
REFERENCES
