297 / 412
via the SAS macro MAGREE. Kendall’s coefficients do not treat
all misclassifications equally. For instance, Kendall’s coefficients
considers the consequences of misclassifying a perfect (rat-
ing
5
5) object as bad (rating
5
1) as more serious than misclas-
sifying it as very good (rating
5
4). For all statistical analysis,
data was analyzed using the SAS System software (SAS Insti-
tute, Inc., Cary, NC).
RESULTS
Seventy-six subjects underwent evaluation with 81
office biopsies with subsequent direct microlaryngoscopy
under general anesthesia. The age range of the subjects
was 21 to 84 years, with a median age of 62 and a male
to female ratio of 5:1. There were 76 laryngeal biopsies
and five oropharyngeal biopsies. The oropharynx sub-
sites included two for tonsil, two for tongue base, and
one for soft palate. There were no complications from
any of the office or operative procedures performed.
The results of office biopsy and their subsequent
direct microlaryngoscopy (DML) and biopsy/excision are
summarized in Table I.
When considered as separate groups, Kendall’s coeffi-
cient for each group was all 0.5, indicating “moderate
correlation” only. None of these approached statistical sig-
nificance (
P
5
0.5). When groups 1 and 2 (i.e., lesions of
uncertain significance and premalignant/malignancy) were
considered together, the coefficient was 0.64 (
P
5
0.029),
indicating “substantial correlation.” For malignant/prema-
lignant lesions, the office biopsy analysis was as follows:
sensitivity
5
60%, specificity
5
87%, positive predictive val-
ue
5
78%, and negative predictive value
5
74%.
DISCUSSION
Medical decision making, patient counseling, and
surgical planning benefit from understanding the nature
of a lesion based on patient demographics, clinical his-
tory, the physical examination, and cytologic or patho-
logic diagnosis. The nature of a laryngeal lesion will
affect prioritizing the different surgical goals of the fol-
lowing: 1) confidence in a pathologic diagnosis, 2) control
of disease, and 3) voice preservation or improvement.
See Table II for proposed office biopsy candidacy.
While office biopsy increases in popularity, there
needs to be further clarification regarding its utility. The
results of this study highlight the concern of office biopsy
being used as a substitute for traditional DML. The small
tissue sample, limited depth past the basement membrane,
and ability to sample only portions of a suspicious lesion
are known disadvantages to this technique, especially with
leukoplakia and lesions that have some degree of dyspla-
sia. Serrated or other forceps, however, may favourably
influence the sensitivity and specificity of the findings, and
appropriate selection of forceps is a major consideration in
the management of early laryngeal malignancy by any
method. Therefore, just as in the OR under direct laryngo-
scopic conditions for which the surgeon would choose the
forceps for biopsy carefully, the same considerations must
be applied when performing transnasal biopsy.
Although it may be acceptable for some screening
tests to have a high specificity and lower sensitivity, this
is not appropriate for this diagnostic test. Sensitivity for
malignancy/premalignancy was only 60%, indicating that
it is inadequate as a diagnostic test: clinical suspicion
alone in this setting would seem at least equivalent. Only
15% of invasive SCC was identified at office biopsy, and it
is evident that any clinically suspicious neoplasm must
TABLE I.
Summary of Results.
Office Biopsy
Pathology
Office
Biopsy
N
5
Accuracy
Compared
to Final
Pathology
Pathology of
Missed
Diagnosis
(False Negatives)
Rate of
False
Negatives%
SCC
4 100.0% N/A
0.0%
Severe
dysplasia/
CIS
23 17.4% SCC
56.5%
Mild–Moderate
dysplasia
8.7%
Inflammation only
8.7%
Polyps or nodule
4.3%
Keratosis
4.3%
Mild to
moderate
dysplasia
12 25.0% SCC
25.0%
Severe dysplasia/CIS 33.3%
Polyps or nodule
8.3%
Keratosis
8.3%
Keratosis
without
dysplasia
7 14.3% SCC
28.6%
Severe dysplasia/CIS 14.3%
Polyps or nodule
28.6%
Inflammation only
14.3%
Inflammation
only
8 12.5% SCC
12.5%
Severe dysplasia/CIS 25.0%
Keratosis
12.5%
Polyps or nodule
25.0%
Papilloma
12.5%
Polyp/nodule 10 100.0% N/A
0.0%
Papilloma
11 81.8% Inflammation only
9.1%
Other
9.1%
Other benign 6
0.0% SCC
16.7%
Keratosis
33.3%
Inflammation only
33.3%
Papilloma
16.7%
Inadequate 4
N/A SCC
50.0%
Polyp or nodule
50.0%
CIS
5
carcinoma in situ; SCC
5
squamous cell carcinoma.
TABLE II.
Candidates for Office Biopsy.
Inclusion
Exclusion
Anatomic limitations for DML
Anticoagulation
Voice less critical
Anterior commissure location
Following a known benign
diagnosis
Submucosal lesion
High risk for general anesthesia Lesion associated with
obstruction
DML
5
direct microlaryngoscopy.
Laryngoscope 125: April 2015
Richards et al.: Office-Based Biopsy for Laryngopharyngeal Lesions
72