Keywords
Ultrasound tongue - oral tongue cancer - squamous cell cancer
Introduction
Intraoral ultrasonography (USG) is a useful modality to evaluate the extend of tumor
involvement of tongue cancer and to ascertain its tumor thickness (TT).[1],[2],[3],[4] There exist several studies proving the statistical correlation between ultrasound
TT and histopathological gross TT [2],[5],[6],[7] postsurgical resection, making ultrasound an essential modality in ascertaining
the TT as well as to ascertain the regional lymph nodal metastasis.[8],[9],[10]
Several studies have proved that the TT is an important factor to predict occult nodal
metastasis. In literature review by Pentenero et al.,[11] showed the TT cut off ranged from 2 mm to 10 mm, as prognostic factor. Other important
prognostic predictive parameters include age, sex, tobacco intake, TNM (tumor, node,
and metastasis) status, histopathological parameters like tumor border, infiltration,
perineural, or vascular invasions.[11],[12],[13]
When considering OTSCC, the risk of occult metastasis is 50% in clinically N0 neck.[14],[15] Thus in cases of early N0-OTSCC, it is prudent on the surgical team to decide on
to whether to undertake an elective neck dissection or not.
The studies have used magnetic resonance imaging (MRI),[16] ultrasound [7] in preoperative assessment of TT with histopathology being the gold standard. Sonographic
evaluation has the advantage of being rapid, less costly over MRI, although both modalities
are noninvasive.
Materials and Methods
The study was carried out in the Department of Radiodiagnosis and Interventional Radiology
from January 2012 to December 2013. 24 patients were included in the study. The ethics
committee approved this prospective cross-sectional study. An informed consent was
taken from all the patients undergoing study.
TT measured by macroscopic technique and intraoral ultrasound compared to the gold
standard microscopic thickness statistically using: 1) Pearson correlation coefficient;
2) Interclass correlation; 3) Bland-Altman plot —95% confidence interval (CI). Chi-square
test was used to find the association of pathological T stage and TT with pathological
nodal status.
Our inclusion criteria included patient population of any age or sex with, 1) Biopsy
proven T1N0 or T2N0 primary squamous cell carcinoma of tongue; 2) Tumors located on
lateral tongue in anterior two-third.
Our exclusion criteria included: 1) Tumor of tongue crossing the midline or involving
the tip of tongue; 2) Tumor of lateral surface of anterior two-third of tongue infiltrating
into surrounding structures; 3) Irradiated tumor of anterior two-third of tongue;
4) Recurrent tumor of anterior two-third of tongue; 5) Tumor of other subsites of
oral cavity.
TT measurement was done as a part of preoperative assessment by intraoral ultrasound.
All patients underwent resection of primary lesion and ipsilateral elective neck dissection,
i.e. supraomohyoid neck dissection (Level I to IV). Measurement of TT was obtained
intraoperatively from fresh glossectomy specimen and postoperatively from histopathological
paraffin section examination.
Ultrasound measurement was done using 17 or 9 MHz conventional linear probe with sterile
cover. Tongue protruded, held gently with gauze and probe placed directly on tumor
surface such that deformation of tumor was not caused [Figure 1]. TT measured from tumor surface to deepest point of invasion was used in protruding
lesions; and for ulcerative lesions an imaginary line was drawn over the ulcerated
area joining the normal mucosa on both ends and the deepest point of invasion was
measured [Figure 2] and [Figure 3]a, [Figure 3]b. Neck ultrasound screening was done using 17 MHz linear conventional probe in all
patients to assess suspicious nodes. There were no suspicious nodes that were detected
in our sample size that included T1/T2, N0 early OTSCC.
Figure 1: TT measurement from the lesion surface to the greatest depth of lesion
Figure 2 (A and B): (A) TT measurement in exophytic lesion; (B) showing TT measurements in ulcerative
lesion
Figure 3 (A and B): (A) Exophytic lesion in ultrasound; (B) ulcerative lesion in ultrasound
After resection, specimen was placed in saline, and immediately shifted to the Department
of Pathology for sectioning. To avoid shrinkage and distortion error, the specimen
was not treated with formalin. Specimen measurements, tumor measurements (except for
thickness), and clear margins (except deep clear margins) were noted before sectioning.
Specimen was later cut into approximately 2–3 mm thick transverse slices. Tumor tissue
was visualized as whitish hard tumor mass from surrounding reddish uninvolved tongue
muscle. Cut sections were examined for TT. Tissue section in which mucosa adjacent
to the tumor was observed and which was considered to have greatest infiltration into
the underlying tissue was noted in both ulcerated and exophytic lesion, disregarding
any superficial keratin or inflammatory infiltrate that existed. Macroscopic TT measurements
were obtained using hand held lens and Vernier caliper. This macroscopic TT was verified
and confirmed by two other pathologist.
Microscopically lesion staged as per sixth edition of American Joint Committee on
Cancer system (AJCC 2010). Histological measurements obtained using an ocular micrometer.
Maximum thickness recorded from imaginary line reconstructing the intact mucosa to
deepest point of invasion into underlying tissue, in both ulcerated and exophytic
lesion, disregarding any superficial keratin or inflammatory infiltrates [Figure 4].
Figure 4: Microscopic section of tongue with infiltrating carcinoma
Slides were examined for lymphovascular invasion. Lymph node sections stained by hematoxylin-eosin
(H and E). Serial sections of the same were stained using immunohistochemical marker
PAN CK and observed under ocular microscope to record metastatic status by pathologist.
Results
Twenty four patients with biopsy proven OTSCC were included in our study. Age group
ranged from 22 years to 76 years (mean age 55) with 16 males (66.7%) and 8 females
[Table 1]. Eighteen cases (75%) were pT1 and 6 cases (25%) were pT2 [Table 2].
Table 1: Age group distribution
Table 2: Case distribution of tumor stage T1 and T2
Ultrasound TT vs. microscopic thickness [Table 3] was within 1 mm in 37.5% (9/24) of cases, within 2 mm in 29.16% (7/24), and greater
than 2 mm in 33.34% (8/24) of cases. Very thin lesions (as thin as 1 mm included in
the study) could be detected with ultrasound. Pearson correlation for ultrasound and
microscopic thickness was r = 0.692 (P < 0.001) and interclass correlation for ultrasound and microscopic thickness, (interclass
correlation coefficient) ICC = 0.821 [Table 4].
Table 3: Graphical representation of USG vs microscopic thickness distribution
Table 4: Bland-Altman and interclass correlation of USG vs microscopic thickness distribution
Macroscopic vs. microscopic thickness [Table 5] was within 1 mm in 37.5% (9/24) of cases, within 2 mm in 25% (6/24), greater than
2 mm in 16.6% (4/24) of cases, no difference or exactly same in 20.8% (5/24) of cases.
Macroscopic vs. microscopic thickness was within 1 mm in 37.5% (9/24) of cases, within
2 mm in 25% (6/24), greater than 2 mm in 16.6% (4/24) of cases, no difference or exactly
same in 20.8% (5/24) of cases. The Pearson correlation for macroscopic and microscopic
thickness r = 0.834 (P < 0.001) and interclass correlation for macroscopic and microscopic thickness, ICC
= 0.898 [Table 6].
Table 5: Graphical representation of macroscopic vs. microscopic thickness distribution
Table 6: Bland-Altman and interclass correlation of macroscopic vs. microscopic thickness
distribution
There is significant correlation between macroscopic vs. microscopic measures of TT
(correlation coefficient 0.834, P < 0.001) as well as between USG and microscopic thickness (correlation coefficient
0.692, P < 0.001). Among the two techniques, macroscopic measurement showed better agreement
to microscopic thickness. Average difference between microscopic thickness and US
thickness (Bias) is -0.14637 with 95% limits of agreement.
TT of 4 mm and above is an indicator of cervical lymph node metastasis. Overall rate
of occult lymph node metastasis was 16.6% [Table 7].
Table 7
Occult nodal metastasis distribution in relation to TT
|
Pathological parameter
|
Total (24)
|
Positive nodes
|
Negative nodes
|
P
|
|
n
|
%
|
n
|
%
|
n
|
%
|
|
pT1
|
18
|
75
|
3
|
16.7
|
15
|
83.3
|
1
|
|
pT2
|
6
|
25
|
1
|
16.7
|
5
|
83.3
|
|
|
TT <4 mm
|
6
|
25
|
0
|
0
|
6
|
100
|
0.533
|
|
TT ≥4 mm
|
18
|
75
|
4
|
22.2
|
14
|
77.8
|
|
Discussion
Oral tongue squamous cell carcinoma (OTSCC) is the most prevalent malignant neoplasm
of oral cavity with worst survival rates.[17] Various preoperative techniques have been described for TT evaluation of OTSCC by
ultrasound, computed tomography (CT) and MRI with variable results and the gold standard
being histopathological evaluation. There are studies that point towards the superiority
of ultrasound in TT assessment over CT and MRI.[18],[19] Intraoperative assessment technique by frozen section have also been described.
In our study, there was significant correlation between microscopic vs. macroscopic
and USG vs. microscopic measurements. Between the two, macroscopic measurements showed
better agreement to microscopic thickness than to ultrasound.
Cervical nodal metastasis is the most important prognosticator of survival in OTSCC.
The risk of occult cervical nodal metastasis in early OTSCC varies from 6% to 46%.[10] The regional rate of recurrence in clinically untreated N0 neck varies from 30%
to 47%[20],[21] for early OTSCC (T1 & T2). Controversies exist onto whether to treat early N0-OTSCC
with elective neck node dissection. The TT cut off for deciding to treat the neck
varies greatly in literature from 1.5 mm to 10 mm, being most often 4 mm.[22],[23]
In our study, the rate of occult metastasis was 16.6%. The incidence of occult nodal
metastasis was zero when TT cut off was less than 4 mm.
Our study had few limitations:
Discrepancy arising due to tumor shrinkage during the fixation process was not addressed
in the study. The effect reported is variable ranging in literature from 4.1% to 30%.[24] In our study, however, we tried to minimize the variability due to tumor shrinkage
by fixation being done in less than 24 hExophytic lesions, while being measured with
USG probe over the lesion, may cause some pressure effect over the tumor, thus resulting
in variations in TT and depth of involvement. However, we tried to reduce this effect
by minimizing the pressure of application, by placing the probe over the tumor as
gently as possible and with the same radiologist doing all the TT measurementAlmost
all studies depicting USG TT was done intraoperatively after anesthesia. However,
we tried in this study to do an USG TT as a part of preoperative assessment and to
confirm its statistical validity.
Conclusion
Ultrasound TT evaluation is a reliable and cost-effective tool to measure the TT preoperatively,
which will be of help in deciding the management of N0 neck in early OTSCCTT of 4
mm and above was predictor of occult cervical lymph node metastasis.
